Chemical Communication Among Bacteria

Chemical
Communication
among Bacteria
This page intentionally left blank
Chemical
Communication
among Bacteria
Edited by
Stephen C. Winans
Department of Microbiology
Cornell University
Ithaca, New York
and
Bonnie L. Bassler
Howard Hughes Medical Institute
Chevy Chase, Maryland and
Department of Molecular Biology
Princeton University
Princeton, New Jersey
Washington, DC
Address editorial correspondence to ASM Press, 1752 N St. NW, Washington, DC
20036-2904, USA
Send orders to ASM Press, P.O. Box 605, Herndon,VA 20172, USA
Phone: (800) 546-2416 or (703) 661-1593
Fax: (703) 661-1501
E-mail: books@asmusa.org
Online: estore.asm.org
Copyright © 2008 ASM Press

American Society for Microbiology
1752 N Street NW
Washington, DC 20036-2904
Library of Congress Cataloging-in-Publication Data
Chemical communication among bacteria / edited by Stephen C.Winans and

Bonnie L. Bassler.
p. ; cm.
Includes bibliographical references and index.
ISBN 978-1-55581-404-5 (alk. paper)
1. Quorum sensing (Microbiology) 2. Cellular signal transduction.
3. Bacteria—Physiology. I. Winans, Stephen Carlyle. II. Bassler, Bonnie L.
III. American Society for Microbiology.
[DNLM: 1. Bacteria—chemistry. 2. Bacterial Physiology. 3. Cell Communication.
4. Intercellular Signaling Peptides and Proteins. QW 52 C5166 2008]
QR96.5.C54 2008
571.7′4—dc22 2007052056
10 9 8 7 6 5 4 3 2 1
All Rights Reserved

Printed in the United States of America
Cover: A Staphylococcus aureus-Pseudomonas aeruginosa co-culture biofilm. Aggregates of S.

aureus (colored red from SYTO 62 staining) are surrounded by a monolayer of P.
aeruginosa (green-GFP tagged) in this 24-h-old biofilm. How and when interspecies
signaling occurs to form organized mixed species communities represent an emerging
area. Photo courtesy of Dinding An and Matthew R. Parsek.
CONTENTS
Contributors ix
Preface xv
I. CELL-CELL SIGNALING DURING

DEVELOPMENT AND DNA EXCHANGE 1
1. Intercompartmental Signal Transduction during
Sporulation in Bacillus subtilis
David Z. Rudner and Thierry Doan
3
2. Extracellular Peptide Signaling and Quorum Responses
in Development, Self-Recognition, and Horizontal Gene
Transfer in Bacillus subtilis
Jennifer M. Auchtung and Alan D. Grossman
13
3. New Insights into Pheromone Control and Response
in Enterococcus faecalis pCF10
Heather A. H. Haemig and Gary M. Dunny
31
4. C-Signal Control of Aggregation and Sporulation
Dale Kaiser
51
5. The Dif Chemosensory System Is Required for S Motility,
Biofilm Formation, Chemotaxis, and Development
in Myxococcus xanthus
Lawrence J. Shimkets
65
v
vi ■ CONTENTS
6. Heterocyst Development and Pattern Formation

M. Ramona Aldea, Krithika Kumar, and James W. Golden
75
7. Diverse Cell-Cell Signaling Molecules Control Formation of
Aerial Hyphae and Secondary Metabolism in Streptomycetes
Joanne M.Willey and Justin R. Nodwell
91
8. Metabolites as Intercellular Signals for Regulation of
Community-Level Traits
Russell D. Monds and George A. O’Toole
105
II. CELL-CELL SIGNALING IN MUTUALISTIC

AND PATHOGENIC ASSOCIATIONS WITH
HUMANS, ANIMALS, AND PLANTS 131
9. LuxR-Type Proteins in Pseudomonas aeruginosa Quorum
Sensing: Distinct Mechanisms with Global Implications
Martin Schuster and E. P. Greenberg
133
10. Quorum Sensing in Vibrio cholerae Pathogenesis
Fiona R. Stirling, Zhi Liu, and Jun Zhu
145
11. Signal Integration and Virulence Gene Regulation in
Staphylococcus aureus
Edward Geisinger and Richard P. Novick
161
12. Quorum Sensing in the Soft-Rot Erwinias
Sarah J. Coulthurst, Rita E. Monson, and George P. C. Salmond
185
13. Role of Quorum-Sensing Regulation in Pathogenesis of
Pantoea stewartii subsp. stewartii
Susanne B. von Bodman, Aurelien L. Carlier, and Ann M. Stevens
201
14. Cell-to-Cell Communication in Rhizobia: Quorum Sensing
and Plant Signaling
J. Allan Downie and Juan E. González
213
15. Quorum Signaling and Symbiosis in the Marine Luminous
Bacterium Vibrio fischeri
E.V. Stabb, A. Schaefer, J. L. Bose, and E. G. Ruby
233
CONTENTS ■ vii
16. Acylated Homoserine Lactone Signaling in Marine

Bacterial Systems
Elisha M. Cicirelli, Holly Williamson, Karen Tait, and Clay Fuqua
251
III. PRODUCTION, DETECTION, AND QUENCHING

OF CHEMICAL SIGNALS 273
17. Acyl-Homoserine Lactone Biosynthesis:
Structure and Mechanism
Mair E. A. Churchill and Jake P. Herman
275
18. Cell-Cell Signaling within Crown Gall Tumors
Stephen C.Winans
291
19. A New Look at Secondary Metabolites
Michael G. Surette and Julian Davies
307
20. Signal Integration in the Vibrio harveyi and Vibrio cholerae
Quorum-Sensing Circuits
Brian Hammer and Bonnie L. Bassler
323
21. Signal Trafficking with Bacterial Outer Membrane Vesicles
Lauren Mashburn-Warren and Marvin Whiteley
333
22. Cooperative Regulation of Competence Development in
Streptococcus pneumoniae: Cell-to-Cell Signaling via a Peptide
Pheromone and an Alternative Sigma Factor
Marco R. Oggioni and Donald A. Morrison
345
23. The A Factor Regulatory Cascade That Triggers Secondary
Metabolism and Morphological Differentiation in Streptomyces
Sueharu Horinouchi
363
24. Quorum Quenching: Impact and Mechanisms
Lian-Hui Wang,Yi-Hu Dong, and Lian-Hui Zhang
379
25. Quorum-Sensing Inhibition
Staffan Kjelleberg, Diane McDougald,Thomas Bovbjerg Rasmussen,
and Michael Givskov
393
viii ■ CONTENTS
IV. EUKARYOTIC QUORUM SENSING

AND INTERACTIONS WITH
QUORUM-SENSING BACTERIA 417
26. Interdomain Cross Talk
Carla Cugini, Roberto Kolter, and Deborah A. Hogan
419
27. Intercellular Signaling by Rhomboids in Eukaryotes
and Prokaryotes
Matthew Freeman and Philip Rather
431
28. Quorum Sensing in Fungi
Claire C.Tseng and Gerald R. Fink
443
29. Quorum Sensing in Rotifers
Julia Kubanek and Terry W. Snell
453
30. “Quorum Sensing” in Honeybees: Pheromone Regulation
of Division of Labor
Yves Le Conte, Zachary Huang, and Gene E. Robinson
463
Index 469
CONTRIBUTORS
M. Ramona Aldea
Department of Biology,Texas A&M University, College Station,TX 77843
Jennifer M. Auchtung
Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139
Bonnie L. Bassler
Howard Hughes Medical Institute, Chevy Chase, MD, and Department of Molecular
Biology, Princeton University, Princeton, NJ 08544-1014
J. L. Bose
Department of Microbiology, University of Georgia, Athens, GA 30602
Aurelien L. Carlier
Department of Plant Science, University of Connecticut, Storrs, CT 06269-4163
Mair E. A. Churchill
Department of Pharmacology and Program in Biomolecular Structure,The University of
Colorado at Denver and Health Sciences Center, Aurora, CO 80045
Elisha M. Cicirelli
Department of Biology, Indiana University, Bloomington, IN 47405
Sarah J. Coulthurst
Department of Biochemistry, University of Cambridge, Cambridge CB2 1QW,
United Kingdom
Carla Cugini
Department of Microbiology and Immunology, Dartmouth Medical School,
Hanover, NH 03755
Julian Davies
Department of Microbiology and Immunology, University of British Columbia,
Vancouver, British Columbia V6T 1Z3, Canada
ix
x ■ CONTRIBUTORS
Thierry Doan
Department of Microbiology and Molecular Genetics, Harvard Medical School,
Boston, MA 02115
Yi-Hu Dong
Institute of Molecular and Cell Biology, Singapore 138673
J. Allan Downie
John Innes Centre, Norwich NR4 7UH, United Kindgom
Gary M. Dunny
Department of Microbiology, University of Minnesota, Minneapolis, MN 55455
Gerald R. Fink
Whitehead Institute for Biomedical Research, Cambridge, MA 02142
Matthew Freeman
MRC Laboratory of Molecular Biology, Cambridge CB2 2QH, United Kingdom
Clay Fuqua
Department of Biology, Indiana University, Bloomington, IN 47405
Edward Geisinger
Molecular Pathogenesis Program,The Helen L. and Martin S. Kimmel Center for Biology
and Medicine at the Skirball Institute for Biomolecular Medicine, New York University
School of Medicine, New York, NY 10016
Michael Givskov
BioScience and Technology,Technical University of Denmark, Lyngby,
Copenhagen, Denmark
James W. Golden
Juan E. González
Department of Molecular & Cell Biology, University of Texas at Dallas, Richardson,TX
E. P. Greenberg
Department of Microbiology, University of Washington, Seattle,WA 98195
Alan D. Grossman
Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139
Heather A. H. Haemig
Department of Microbiology, University of Minnesota, Minneapolis, MN 55455
Brian Hammer
Department of Molecular Biology, Princeton University, Princeton, NJ 08544-1014
Jake Herman
Department of Pharmacology,The University of Colorado at Denver and Health Sciences
Center, Aurora, CO 80045
Deborah A. Hogan
Hanover, NH 03755
CONTRIBUTORS ■ xi
Sueharu Horinouchi
Department of Biotechnology, Graduate School of Agriculture and Life Sciences,
The University of Tokyo, Bunkyo-ku,Tokyo 113-8657, Japan
Zachary Huang
Department of Entomology, Michigan State University, East Lansing, MI 48824
Dale Kaiser
Departments of Biochemistry and Developmental Biology,
Stanford University School of Medicine, Stanford, CA 94305
Staffan Kjelleberg
Centre for Marine Biofouling and Bio-Innovation,The University of New South Wales,
New South Wales, Australia
Roberto Kolter
Boston, MA 02115
Julia Kubanek
School of Biology and School of Chemistry & Biochemistry,
Georgia Institute of Technology, Atlanta, GA 30332
Krithika Kumar
Yves Le Conte
INRA, UMR406 INRA/UAPV Ecologie des Invertébrés, Laboratoire Biologie et
Protection de l’Abeille, Site Agroparc, Domaine Saint-Paul,
84914 Avignon Cedex 9, France
Zhi Liu
Department of Microbiology, University of Pennsylvania School of Medicine,
Philadelphia, PA 19104-6076
Lauren Mashburn-Warren
Section of Molecular Genetics and Microbiology,The University of Texas at Austin,
Austin,TX 78712
Diane McDougald
Centre for Marine Biofouling and Bio-Innovation,The University of New South Wales,
New South Wales, Australia
Russell D. Monds
Hanover, NH 03755
Rita E. Monson
United Kingdom
Donald A. Morrison
Laboratory for Molecular Biology, Department of Biological Sciences,
University of Illinois at Chicago, Chicago, IL 60607
xii ■ CONTRIBUTORS
Justin R. Nodwell
Department of Biochemistry, Health Sciences Centre, McMaster University, Hamilton,
Ontario L8N 3Z5, Canada
Richard Novick
Molecular Pathogenesis Program,The Helen L. and Martin S. Kimmel Center for Biology
and Medicine at the Skirball Institute for Biomolecular Medicine, New York University
School of Medicine, New York, NY 10016
Marco R. Oggioni
Laboratorio di Microbiologia Molecolare e Biotecnologia, Dipartimento di Biologia
Molecolare, Universitá di Siena, 53100 Siena, Italy
George A. O’Toole
Hanover, NH 03755
Thomas Bovbjerg Rasmussen

Chr. Hansen A/S, Bøge Allé 10-12, 2970 Hørsholm, Denmark
Philip Rather
Department of Microbiology and Immunology, Emory University School of Medicine,
Atlanta, GA 30322
Gene E. Robinson
Department of Entomology, Neuroscience Program, and Institute for Genomic Biology,
University of Illinois at Urbana-Champaign, Urbana, IL 61801
E. G. Ruby
Department of Medical Microbiology and Immunology, University of Wisconsin—
Madison, Madison,WI 53706
David Z. Rudner
Boston, MA 02115
George P. C. Salmond
United Kingdom
A. Schaefer
Department of Microbiology, University of Washington, Seattle,WA 98195
Martin Schuster
Department of Microbiology, Oregon State University, Corvallis, OR 97331
Terry W. Snell
School of Biology, Georgia Institute of Technology, Atlanta, GA 30332
E.V. Stabb
Ann M. Stevens
Department of Biological Sciences,Virginia Tech, Blacksburg,VA 24061
CONTRIBUTORS ■ xiii
Fiona R. Stirling
Michaela G. Surette
Department of Microbiology and Infectious Diseases and Department of Biochemistry
and Molecular Biology, University of Calgary, Calgary, Alberta T2N 4N1, Canada
Karen Tait
Plymouth Marine Laboratory, Prospect Place, Plymouth PL1 3DH, United Kingdom
Claire C. Tseng
Whitehead Institute for Biomedical Research, Cambridge, MA 02142
Susanne B. von Bodman

Departments of Plant Science and Molecular and Cell Biology, University of
Connecticut, Storrs, CT 06269-4163
Lian-Hui Wang
Marvin Whiteley
Section of Molecular Genetics and Microbiology,The University of Texas at Austin,
Austin,TX 78712
Joanne M. Willey
Department of Biology, Hofstra University, Hempstead, NY 11549
Holly Williamson
Plymouth Marine Laboratory, Prospect Place, Plymouth PL1 3DH, United Kingdom
Stephen C. Winans
Department of Microbiology, Cornell University, Ithaca, NY 14853
Lian-Hui Zhang
Jun Zhu
PREFACE
lthough a few groups of bacteria have long been known to communicate

A via diffusible chemical signals, we are only now learning that this process
is enormously widespread.We are now in a position to begin to appreciate the
importance of cell-cell communication in areas as fundamental as bacterial
physiology, ecology, evolution, and pathogenesis. Approximately one decade
ago, ASM Press published the first comprehensive review of the topic of bac-
terial cell-cell communication, Cell-Cell Signaling in Bacteria, with chapters
contributed by leaders in the then-nascent field.We hope readers of this new
volume will agree that an enormous amount of information on major aspects
of signaling has surfaced since that first book was published and that a fresh
view of the topic is now appropriate and important for a diverse audience of
researchers, educators, and clinicians.
The past decade has witnessed new insights about the chemical composi-
tion, synthesis, and turnover of a variety of bacterial signal molecules. First, the
enzymes that synthesize signal molecules are far better understood than they
were 10 years ago.At the close of the 20th century, no signal synthase had been
studied at the structural level. Currently, the structures of three bacterial signal
synthases have been solved, two of which produce AHLs (chapter 16) and one
of which synthesizes AI-2 (chapter 19). In other developments, the Streptomyces
coelicolor 15-residue SapB peptide, required for aerial fruiting body formation,
is now known to be synthesized by a nonribosomal peptide synthase (chapter
6).We recently learned that at least one class of extremely hydrophobic signal
travels as a component of vesicles derived from the cell outer membrane (chap-
ter 20). This signal, designated PQS (Pseudomonas quorum signal) also has
antimicrobial properties against gram-positive bacteria. Many new types of sig-
nal molecules with a variety of novel structures are under study, including
polyamines, rhamnolipids, and metabolites such as indole and amino acids
(chapters 3, 7, and 17). During the past decade, a variety of enzymes capable of
degrading bacterial communication signals have been described, as well as nat-
ural and synthetic small molecules that agonize or antagonize signaling (chap-
ters 10, 24, and 25). Future studies may help us understand whether the
xv
xvi ■ PREFACE
substrates for which these enzymes were selected are signaling molecules or
whether the destruction of signal molecules is incidental to the activity for
which they were selected.
There has been an explosion of new information on signal receptors and
mechanisms of signal transduction.Where 10 years ago there was no structural
information about quorum sensing receptors, there now exists structural infor-
mation for seven of these receptors (chapters 21, 13, 19, and 23) (2, 7). It is
striking that at least five of these receptors (PrgX, CprB,TraR, LasR, and SdiA)
fully or partially engulf their respective ligands, which contribute to the
hydrophobic cores of these proteins. In the case of the cytoplasmic TraR, LasR,
and SdiA receptors, ligand binding is required for protein folding and resist-
ance to proteolysis, while PrgX and CprB function as apo-proteins, so their
folding must occur in the absence of ligand. Quorum sensing structural stud-
ies have provided other surprises. For example, it was found that AI-2 bound
to LuxP includes a boron atom and, perhaps equally surprising, that AI-2
bound to the homologous Lsr receptor lacks boron (chapter 19). Also surpris-
ing is that the LuxPQ structures provide a new mechanism for two-compo-
nent signal transduction across the bacterial membrane that differs dramatically
from that proposed for signal relay in chemotaxis systems. A cocrystal contain-
ing TraR and its antiactivator TraM provides insight into the mechanism for
how TraM allosterically prevents TraR from binding DNA (3).Thus, studies of
the molecular biology of cell-cell signaling are providing unexpected insights
into other areas of molecular biology.
New discoveries about signal transduction pathways and the expression of
target genes have also been made. For example, a decade ago we could not have
guessed that at the heart of the Vibrio harveyi and Vibrio cholerae quorum sens-
ing cascades would lie several redundant small RNAs.We could not have pre-
dicted that two autoinducers and two AI synthases in Vibrio fischeri would
influence the activity of LuxR. Large sets of new target genes have been iden-
tified using global high throughput techniques such as proteomics and DNA
microarrays (6) (chapter 8).
The repertoire of phenotypes affected by cell-cell communication has
grown considerably. For example, it has long been appreciated that oligopep-
tides stimulate sporulation and competence for transformation in Bacillus sub-
tilis, but only recently has it been reported that peptides also stimulate the
conjugation of the integrative and conjugative element ICEBs1 (chapter 2).
Expression of these genes is also induced by DNA damage, similar to the cor-
responding genes of the STX element of V. cholerae (1). A decade ago it was
clear that communication is required for biofilm formation in Pseudomonas
aeruginosa (4), but recently it has been discovered that communication has the
opposite effect on biofilms in V. cholerae (chapter 9).The surprising finding that
increases in population densities activate this trait in P. aeruginosa and inhibit the
same trait in V. cholerae most likely defines the persistent versus acute diseases,
respectively, caused by these pathogens.
Studies of P. aeruginosa have been especially fast-paced in this past decade. It
was already known that this organism has two AHL signals, as well as a
quinilone signal called PQS, and that there were two AHL synthases and two
AHL receptors.We now know that a third AHL receptor exists that detects one
of the known AHLs.The P. aeruginosa LasR protein has been extensively stud-
PREFACE ■ xvii
ied in vitro and binds both to cannonical las box binding sites as well as to
completely different sites (chapter 8). Binding to some sites is cooperative,
while at other sites, the protein binds noncooperatively. Microarrays and ran-
dom fusions have shown that hundreds of genes are controlled by one or more
of these systems, including many genes that encode exported proteins. The
crystal structure of the LasR N-terminal domain, complexed with the cognate
AHL, was recently determined (2) and revealed interesting structural similari-
ties to TraR of A. tumefaciens (chapter 13).
What discoveries might we anticipate in the coming decade? We will look
for the elucidation of new classes of signaling molecules, such as the one
described late in 2007 (5). We expect that the next 10 years will also witness
advances in structural and biochemical studies of signal synthases and recep-
tors, including structural determination of peptide signals complexed with
their receptors. There will likely be surprises about the mechanisms of diffu-
sion of different signals, perhaps aided by nanofabrication technologies.
Genetic approaches and transcriptional profiling will likely lead to new under-
standing of network design principles that provide noise reduction, signal inte-
gration, and signal amplification.We will learn how bacterial regulatory circuits
are wired to provide ordered temporal and spatial expression of large sets of
target genes.We also will expect further collaborations between biologists and
molecular modelers. We can be reasonably confident about progress in all of
these areas. On the other hand, the most exciting discoveries will generally be
the ones that no one can even begin to anticipate.To paraphrase a former U.S.
Secretary of Defense, it is hard to predict the unknown, but much harder still
to predict the unknown unknowns. If the rate of progress of the past decade is
to continue at the present pace for another 10 years, we will be able to glimpse
a body of knowledge about which we could not even have dreamt when the
first ASM volume on this topic went to press just 10 years ago.
REFERENCES
1. Beaber, J. W., B. Hochhut, and M. K. Waldor. 2004. SOS response promotes hori-
zontal dissemination of antibiotic resistance genes. Nature 427:72–74.
2. Bottomley, M. J., E. Muraglia, R. Bazzo, and A. Carfi. 2007. Structure of the P.
aeruginosa LasR ligand-binding domain bound to its autoinducer. Direct submission to
protein Data Bank, submission number 2uv0.
3. Chen, G., P. D. Jeffrey, C. Fuqua, Y. Shi, and L. Chen. 2007. Structural basis for
antiactivation in bacterial quorum sensing. Proc. Natl. Acad. Sci. USA 104:16474–16479.
4. Davies, D. G., M. R. Parsek, J. P. Pearson, B. H. Iglewski, J. W. Costerton, and
E. P. Greenberg. 1998.The involvement of cell-to-cell signals in the development of a
bacterial biofilm. Science 280:295–298.
5. Higgins, D. A., M. E. Pomianek, C. M. Kraml, R. K.Taylor, M. F. Semmelhack,
and B. L. Bassler. 2007.The major Vibrio cholerae autoinducer and its role in virulence
factor production. Nature 450:883–886.
6. Walters, M., M. P. Sircili, and V. Sperandio. 2006. AI-3 synthesis is not dependent
on luxS in Escherichia coli. J. Bacteriol. 188:5668–5681.
7. Yao, Y., M. Martinez-Yamout, T. Dickerson, A. Brogan, P. Wright, and H.
Dyson. 2006. Structure of the Escherichia coli quorum sensing protein SdiA: activation
of the folding switch by acyl homoserine lactones. J. Mol. Biol. 355:262–273.
STEPHEN C.WINANS
BONNIE L. BASSLER
December 2007
CELL-CELL SIGNALING
DURING DEVELOPMENT
AND DNA EXCHANGE
I
INTERCOMPARTMENTAL SIGNAL
TRANSDUCTION DURING
SPORULATION IN BACILLUS SUBTILIS
David Z. Rudner and Thierry Doan
1
Many of the chapters in this book describe cell- introduction to sporulation in B.subtilis,we will
cell signaling among bacteria within their com- describe each pathway in turn, discuss the cur-
munities. Here we consider a much more rent state of understanding, and suggest testable
intimate conversation that occurs between two hypotheses. Spore formation in B. subtilis has
siblings: the two cells that comprise the spo- served as an important model for cell-cell sig-
rangium during spore formation in Bacillus sub- naling in bacteria. As you will see,the conversa-
tilis. The conversation involves three separate tion between the forespore and mother cell
signal transduction pathways that together provides insight into how and why cells com-
coordinate the transcriptional programs of the municate and highlights the diversity of ways in
two cells, called the mother cell and the fore- which organisms transduce information across
spore.These two cells follow different programs their membranes.
of developmental gene expression, in which
several classes of coordinately regulated genes INTRODUCTION TO SPORULATION
are sequentially activated. Cell-cell signaling B. subtilis enters the sporulation pathway in
ensures that the developmental programs in the response to nutrient deprivation (15, 20, 45,
mother cell and forespore are maintained in 61). The sporulating cell proceeds through a
register with each other such that earlier events series of well-defined morphological stages that
are initiated or in some cases completed before culminate in the production of a dormant cell
later events can begin. Our understanding of type known as the spore (or more properly,
the molecular mechanisms underlying each the endospore).The first landmark morpholog-
of these three pathways is at a different stage of ical event during sporulation is the formation
maturity. But it is already clear that each path- of an asymmetrically positioned septum that
way has something different to teach us about divides the developing cell (or sporangium)
how cells send and interpret signals.After a brief into two daughter cells of unequal size: a
forespore (the smaller cell) and a mother cell
(Color Plate 1). Initially, the two cells lie side-
David Z.Rudner and Thierry Doan Department of Microbiol-
ogy and Molecular Genetics, Harvard Medical School, by-side, but later in development the mother
Boston, Massachusetts 02115. cell engulfs the forespore. Engulfment is a
Chemical Communication among Bacteria, Edited by S.C.Winans and B.L. Bassler
© 2008 ASM Press,Washington, DC
3
4 ■ RUDNER AND DOAN
phagocytic-like process in which the forespore spore plays a key role in coordinating the two
is pinched off within the mother cell, creating developmental programs and that keeping these
a cell within a cell (Color Plate 1). At this late programs in register is critical for morphogene-
stage, the mother cell packages the forespore sis and successful sporulation.
in a protective proteinaceous coat while the It is now clear that there are three signal
forespore prepares for dormancy. Later in transduction pathways between the mother cell
development, the mother cell lyses, liberating and forespore (constituting a true conversation)
the fully mature spore. that ensure that gene expression in one com-
A cascade of developmental transcription partment is linked to gene expression in the
factors are activated in a stage- and cell-type- other throughout the sporulation process
specific manner during this developmental (Color Plate 1). Here we examine each signal-
process (37). Upon asymmetric division, the ing pathway separately and in chronological
RNA polymerase sigma factor F is activated in order: (i) F in the forespore initiates a pathway
the forespore followed by E in the mother cell. that triggers activation of E in the mother
After engulfment is complete (or near com- cell;(ii) E in the mother cell is then responsible
plete), G replaces F in the forespore and, in for the activation of G in the forespore; and
turn, K replaces E in the mother cell (34).The (iii) G in the forespore sets in motion events
stage- and compartment-specific activities of that activate K in the mother cell (Color Plate
these four transcription factors were originally 1).These three pathways highlight the creative
demonstrated using promoter fusions to lacZ and diverse ways evolution has solved the
and heroic immunoelectron microscopy (12, challenge of transducing information across
38, 71). Today, the stage- and compartment- a lipid bilayer.
specific activities of these four sigma factors
can be visualized with ease by fluorescence The Forespore Starts the Conversation
microscopy using strains in which sigma- by Sending a Signal to the Mother Cell
dependent promoters direct the expression of In response to starvation, two stationary-phase
the genes encoding the yellow and cyan fluores- transcription factors, Spo0A and H, are acti-
cent proteins (yfp and cfp) (Color Plate 1). For vated in the predivisional cell (45). Spo0A and
many years it was unclear whether the mother H direct the transcription of early-acting
cell and forespore followed independent devel- sporulation genes, including those involved in
opmental programs like two separate free- repositioning the division machinery to an
running clocks. Elegant genetic analysis has asymmetric site and the synthesis of the first
revealed that, in fact, these two cells communi- two compartment-specific transcription fac-
cate with each other throughout the sporula- tors, F and E. Both F and E are synthesized
tion process. Signaling between the forespore prior to septation in the predivisional cell but
and mother cell was first demonstrated in the are held inactive until the formation of the
early 1990s when it was discovered that gene polar septum. Shortly after the septum is com-
expression in the mother cell under the control plete, F becomes active in the forespore com-
of K was blocked by a mutation in the fores- partment. F is the first cell type-specific
pore transcription factor G (7).The identifica- transcription factor, and its activity confers
tion of mutants that restored mother cell gene forespore identity on the small cell and com-
expression in the absence of G argued in favor mits it to the sporulation program (13, 45).The
of a bona fide signal transduction pathway. activation of F is thought to be directly cou-
Finally, the discovery that these bypass muta- pled to the completion of septation through the
tions in a wild-type background uncouple action of a membrane phosphatase that plays a
mother cell and forespore gene expression and key role in shifting the division site from the
cause a defect in spore formation demonstrated mid-cell to the pole (20, 53).The exact mecha-
that the conversation between mother and fore- nism by which septation triggers F activation
1. SIGNAL TRANSDUCTION DURING SPORULATION IN B. SUBTILIS ■ 5
remains unclear and a source of some contro- E is derived from an inactive proprotein
versy (5, 23, 24, 31, 66). Focus and scrutiny on precursor called pro-E (33,63) (Color Plate 2).
this key morphological event in committing This precursor has an amino-terminal exten-
the cell to the sporulation pathway are well jus- sion of 27 amino acids that must be proteolyti-
tified as it defines how one cell type (the predi- cally removed in order for the transcription
visional cell) becomes two different cell types. factor to associate efficiently with RNA poly-
We refer the interested reader to several recent merase and direct gene expression. The
reviews about this important control point in prodomain both holds E inactive and localizes
this developmental pathway (15, 20, 53). the proprotein to the membrane (Color Plate
The first signal transduction pathway 2) (17, 21, 27).The pathway that begins with F
between the forespore and mother cell has as in the forespore and results in pro-E pro-
input F activity in the forespore and as output cessing in the mother cell is thought to be
the activation of E in the mother cell. E is mediated by a secreted protein ligand and a
the first mother-cell-specific transcription putative membrane protease. The signaling
factor in the large compartment and, as such, molecule (SpoIIR, hereafter referred to as R) is
confers mother cell identity upon it.Activation encoded by a gene under the control of F
of E acts as a switch that sets in motion the (22, 28, 36). R is translocated into the space
mother cell developmental program and at the between the forespore and mother cell mem-
same time halts the program of the predivi- branes (51) where it activates SpoIIGA (here-
sional cell (45). after referred to as GA), resulting in the
In the absence of E, the large compartment conversion of pro-E to its mature and active
continues to follow the program of the predivi- form (Color Plate 2). GA is a polytopic mem-
sional cell, and a second asymmetric division brane protein with homology in its cytoplasmic
occurs at the distal pole, creating a sporangium domain to aspartyl proteases (60).This putative
with two forespore compartments (45). These membrane-tethered protease localizes to the
disporic sporangia are terminally arrested in polar septum by an unknown mechanism,
sporulation. Thus, E-dependent gene expres- where it is perfectly positioned to receive an
sion is required to prevent septation at the activating signal from the forespore (Color
unused polar site. In fact, E activation appears Plate 2) (16).The putative catalytic residues in
to occur minutes after F activation. If the acti- the GA protease have not been shown to be
vation of E is delayed 10 to 15 min after septa- critical for pro-E processing nor has this cleav-
tion, then the efficiency of sporulation is age reaction been reconstituted in vitro. How-
reduced, with a concomitant increase in dis- ever, cells engineered to artificially synthesize
poric sporangia (30, 72). Indeed, simultane- GA, pro-E, and R during vegetative growth
ously visualizing F and E activity using a can process pro-E to its mature and active
F-responsive promoter fusion to cfp and a E- form (60).This result indicates that GA and R
responsive promoter fusion to yfp in the same are the only sporulation-specific proteins
strain, it is hard to find a cell with CFP fluores- required to activate E and suggests that GA is
cence (F activity) that does not also have YFP indeed the pro-E processing enzyme.
fluorescence (E activity) (Color Plate 1). On To ensure that E is activated in a timely
the other hand, if E appears prematurely, that fashion after polar division, both pro-E and its
is, in the predivisional cell, then formation putative processing enzyme GA are made
of the first polar septum is inhibited (14, 17). predivisionally under the control of the Spo0A
These findings indicate that the cell-cell signal- transcription factor (45). Thus, both proteins
ing pathway serves as a timing device to ensure are present in the mother cell and forespore
that E is activated during a critical temporal compartments. Since R can activate pro-E
window following the completion of the polar processing in cis and in trans (60), why is E
septum (14, 46). activity restricted to the mother cell? The
answer appears to be that pro-E is specifically gene transcription is regulated, but it appears
degraded in the forespore compartment to require E-dependent gene expression (61).
(Color Plate 2) (17). The protease (or proteases) G is also controlled posttranslationally. Prema-
responsible for clearing pro-E has not yet been ture expression of the G gene (by fusing it to
identified. In addition, there is a second mecha- a strong F-dependent promoter that is tran-
nism that contributes to compartmentalized E scribed 30 to 45 min earlier) has no effect on
activity.After polar division, Spo0A acts prefer- the timing of G activity (61).
entially in the mother cell,resulting in increased G only becomes active when engulfment is
accumulation of pro-E and GA in this com- complete (or nearly complete),and it is thought
partment (17, 18). that G activation is linked to this morphologi-
The mechanism by which information is cal process (this linkage is referred to as mor-
transduced across the lipid bilayer by R and GA phological coupling) (53). In support of this
is not yet known.We favor a model in which the idea, all mutants that block engulfment prevent
cytosolic aspartyl protease domain of GA is G activation but do not block transcription of
active only as a dimer and the R signaling pro- the G gene (35, 41, 44, 58). It remains unclear
tein promotes dimerization by binding GA in how the completion of engulfment is sensed
the intermembrane space (Color Plate 2).Signal and how this triggers G activity. The surveil-
transduction by ligand-induced dimerization lance of engulfment may be achieved by a puta-
has been well characterized for receptor tyro- tive complex of eight proteins (encoded by the
sine kinases (55,56) and seems like a compelling spoIIIA operon) that are synthesized in the
mechanism here.Alternatively,R binding to the mother cell under E control and localize to
extracellular face of GA could cause a confor- the mother cell membrane that surrounds the
mational change in the protease that activates forespore (1, 9, 29).The spoIIIA locus (referred
the catalytic center or allows the protease to to as IIIA) is not required for the engulfment
bind its substrate. Dissection of the molecular process but is essential for G activation (29).
mechanisms of R-mediated GA activation and Perhaps this putative membrane complex mon-
GA-mediated pro-E processing promises to itors the engulfment process and, once it is
reveal general principles of how information complete,transduces a signal to the forespore to
can be transduced across a lipid bilayer. activate G (Color Plate 3).With this view of G
regulation, the “cell-cell” signaling pathway
The Mother Cell Responds between the mother cell and forespore would
to the Forespore be indirect and therefore considered nontradi-
The second signal transduction pathway, the tional. The role of E would be to promote
activation of G in the forespore under the transcription of the engulfment proteins (45)
control of E in the mother cell, has been and the IIIA surveillance complex (Color Plate
the most refractory to genetic and molecular 3), and not to transcribe a signaling molecule
dissection and therefore is the least well under- like R in the first pathway. In this scenario, IIIA
stood of the three signaling pathways described transduces an”engulfment signal” and cell-cell
here. In part, this is because G activity is signaling is only a genetic formalism.
regulated at multiple levels and appears to be While this model is certainly reasonable, we
coupled to the morphological process of hypothesize that G activity requires both the
engulfment (53, 61).The gene encoding G is completion of engulfment and a more tradi-
under the control of the forespore transcription tional signal sent by the mother cell (Color
factor F. However, transcription of the G Plate 3). In the parlance of systems biology,
gene is delayed by approximately 30 to 45 min these two signals would constitute an “AND
relative to other genes in the F regulon, until gate” and the regulation of G would act as a
the time when engulfment is complete (or coincidence sensor. In this model, a key func-
nearly complete) (44). It is not clear how G tion of the IIIA complex is to transduce the
mother cell signal (and could, in fact, encode the mother cell signal or engulfment status
the signal itself).Although no signaling protein to the forespore. No other proteins have been
under E control has been identified in the acti- identified except SpoIIIJ, which probably
vation of G, it is likely that we have not yet does not play a direct role in signal transduction
found (or characterized) all the players in this (42). Proteins that function at earlier steps in
pathway. In support of this idea, we still do not sporulation could have a second role in G
know what holds G inactive prior to receiving signaling. These factors would likely not be
a signal from the mother cell (see below). identified in genetic screens because mutations
Importantly, there are no data to our knowl- would block the sporulation pathway at an
edge that can distinguish between a role for the earlier step. For example, any of the engulfment
IIIA in monitoring engulfment and a role in a proteins could participate in both the mechan-
classical cell-cell signaling pathway. ics of engulfment and in signal transduction.
What might be the engulfment signal? It has Currently, the best candidate of this type is
been proposed previously that the completion the forespore protein SpoIIQ (referred to as
of engulfment is sensed as the separation of the IIQ). IIQ is required at a late stage for normal
forespore from the external environment (61). engulfment and for activation of G (35, 62).
Once engulfment is complete, the forespore Recently, it has been shown that IIQ localizes
becomes a free protoplast within the mother to the septal membrane on the forespore
cell and is isolated from the outside world. At side (51) where it interacts directly with
this time, the intermembrane space between one of the membrane proteins in the IIIA com-
mother cell and forespore could change in its plex (IIIAH) (Color Plate 3) (1, 9). In fact,
redox potential, electrochemical gradient, or IIQ is required to anchor IIIAH and presum-
metabolic state. Any of these changes could ably the rest of the putative IIIA complex in
trigger G activation. In this scenario, IIIA, the mother cell membrane that surrounds
which is thought to reside on the mother cell the forespore (1, 9). Based on the physical inter-
side of this double membrane (Color Plate 3), action between IIIAH on the mother cell
would probably have no role in surveillance or side of the septum and IIQ on the forespore
in transducing the engulfment signal. side, it was proposed that IIQ might partici-
Alternatively, it is possible that the activity of pate in triggering G activation (1). This
the IIIA complex in transducing a classical cell- exciting and provocative hypothesis awaits fur-
cell signal to the forespore is sensitive to the ther investigation.
engulfment status of the cell. For example, Another outstanding question in the regula-
before the completion of engulfment (or if tion of G is what holds the transcription factor
engulfment is perturbed) the IIIA proteins inactive prior to completion of engulfment.
might be unable to fold or assemble into an It was previously hypothesized that the SpoI-
active signaling complex. Thus, only when IAB protein, which serves as an anti-F factor,
engulfment is complete would IIIA become might also inhibit G (29). However, more
competent to receive and transduce the mother recent evidence suggests that SpoIIAB, which
cell signal. In this model, IIIA would serve both is indeed capable of holding G inactive, does
to transduce a classical E-dependent signal not perform this function in the forespore (57).
from the mother cell and to sense engulfment. Nonetheless, these data reinforce the possibility
At first glance,the distinction between a classical that the understanding of the activation of
signaling pathway and morphological coupling G might require the identification and charac-
may seem artificial; however, we believe consid- terization of proteins required at earlier stages
ering them separately has both mnemonic and in sporulation, and spur the development of
predictive value. clever strategies for temporally controlled inac-
It remains unclear whether additional pro- tivation of proteins after their early roles are
teins participate with IIIA in transducing complete (18).
The Forespore Talks Back cholesterol metabolism and uptake (2).It is now
to the Mother clear that members of this family play key roles
The third and final signal transduction path- in signaling pathways in many bacteria and have
way, the activation of K in the mother cell been implicated in transducing a variety of sig-
under the control of G in the forespore, is the nals (40).The B processing enzyme is regulated
most well understood of the three.The require- by two other integral membrane proteins,
ment for G-dependent gene expression for the SpoIVFA (referred to as A) and BofA (7, 8, 50).
activation of K was first recognized when it In the absence of either of these proteins, B is
emerged that the cotA gene, responsible for the capable of processing pro-K without a fore-
characteristic brown pigment of spores, was spore signal.All three proteins are synthesized in
itself controlled by K (7,39).In G mutants the the mother cell under the control of E and
spores failed to become pigmented, indicating reside in a multimeric complex in the mother
that activation of K in the mother cell required cell membrane that surrounds the forespore,
G-dependent gene expression in the fore- perfectly positioned to receive a signal emanat-
spore. If K activation is uncoupled from its ing from the forespore compartment (Color
dependence on G,expression of K-controlled Plate 4) (48, 54).The A protein is required for
genes commences approximately 30 min early, the proper localization of B and BofA and is
resulting in a reduction in sporulation effi- necessary for their interaction (Color Plate 4)
ciency (7). Thus, K is synthesized in the (54). In turn, A is held in the forespore mem-
mother cell before the activation of G but brane through a network of interactions along
remains inactive until the forespore signals that and across the sporulation septum (9, 25).
it is time for late mother cell gene expression. Indeed, A requires both IIIAH on the mother
K, like E, is derived from an inactive pro- cell side and IIQ on the forespore side for its
protein precursor called pro-K (32) (Color localization.Our current view of the regulation
Plate 4).This precursor has an amino-terminal of pro-K processing is that A anchors the com-
extension of 20 amino acids with no detectable plex in the mother cell membrane that sur-
sequence similarity to the prodomain of E. rounds the forespore and acts as a platform
However, like pro-E, the prodomain on K bringing BofA and B together, wherein BofA
serves as a covalently attached antisigma factor, inhibits the processing activity of B until a sig-
which prevents interaction with core RNA nal has been received from the forespore (54).
polymerase (26, 69). Also, like the prodomain The signaling protein SpoIVB (referred to as
on E, the amino-terminal prodomain of K IVB) is produced in the forespore under G
localizes the proprotein to the membrane (69). control (6, 19) and is translocated across the
The polytopic membrane protein, SpoIVFB membrane into the space between the forespore
(hereafter referred to as B), is absolutely and mother cell membranes (3). IVB is a serine
required for the proteolytic activation of pro- protease (64) and is capable of cleaving and
K and is likely to be the processing enzyme releasing itself from the membrane (10). IVB
(Color Plate 4) (8, 49). Expression of B and then cleaves the extracellular domain of the reg-
pro-K in Escherichia coli is sufficient to trigger ulatory protein A at multiple sites (3, 11).These
processing (70). B is a founding member of a cleavages are critical for efficient K activation.
family of membrane-embedded metallopro- A mutant A protein that cannot be cleaved by
teases whose catalytic centers reside adjacent to IVB delays pro-K processing by over 2 h but
or within the lipid bilayer (52, 67). The other does not completely block it (3).When IVB is
founding member of this family of proteases is unable to cleave A, a second signaling protease,
the Site-2 protease (47, 68), which is required CtpB, is able to compensate (3). CtpB, like IVB,
for proteolytic activation of the sterol response is a serine protease with similar overall domain
element binding protein, a transcription factor structure (3, 43). In its absence, pro-K process-
required for the activation of genes involved in ing is delayed by ~30 min, suggesting that the
normal function of CtpB is to fine-tune the membrane-embedded metalloprotease B. We

timing of K activation. CtpB, like IVB, is capa- hypothesize that cleavage of the regulatory pro-
ble of cleaving the extracellular domain of A (3). tein A in the intermembrane space causes a con-
Importantly,an A mutant that cannot be cleaved formational change in the B-A-BofA signaling
by either of the two signaling proteases is complex that allows pro-K access to the caged
blocked in pro-K processing, and sporulation interior of the membrane metalloprotease B
efficiency is significantly impaired (3). CtpB is (3).We anticipate that further elucidation of the
synthesized in both the mother cell and the molecular mechanism of B activation and the
forespore compartments (43, 59, 65). However, consequences of the branched structure of
synthesis in the forespore is both necessary and the pathway will reveal general principles of
sufficient for CtpB activity (Color Plate 4) (4). how these membrane-embedded proteases are
Thus, there are two forespore signaling pro- regulated and employed for signal transduction.
teases that both target the same regulatory pro-
tein. We originally hypothesized that IVB SUMMARY
cleaves and activates CtpB (3). CtpB is indeed a We have described a conversation between two
substrate for IVB both in vitro and in vivo;how- siblings during the developmental process of
ever, this cleavage does not appear to be critical sporulation.These signal transduction pathways
for CtpB activity (4). Since IVB is absolutely ensure that gene expression in the mother cell is
required to trigger pro-K processing but CtpB coordinated with gene expression in the fore-
becomes essential for signaling only when IVB spore and vice versa. As we have seen, if these
is present but cannot cleave A, these results sug- development programs are uncoupled, sporula-
gest that IVB has yet another target (other than tion efficiency is impaired. Both signaling path-
A and CtpB) in this signaling pathway. ways that emanate from the forespore result
It is clear that once the time is right,the fore- in the proteolytic activation of membrane-
spore takes no chances in the activation of K. tethered transcription factors,yet the molecular
The primary forespore signal IVB triggers pro- mechanisms that govern information transduc-
K processing by cleaving A at one of two possi- tion are completely different in the two path-
ble sites.If IVB is unable to cleave A,then CtpB ways. In the case of the signaling pathway
compensates. It remains unknown whether emanating from the mother cell, it remains
these are fail-safe mechanisms for K activation unclear whether this conforms to a traditional
or whether the exact timing of K activation cell-cell signaling pathway or acts indirectly via
is critical, and if so, why. Under laboratory a mechanism that couples gene expression to
conditions, subtle differences in the temporal morphogenesis.We hypothesize that the signal-
activation of K have very modest impacts on ing components in this pathway serve to moni-
sporulation efficiency. Nonetheless, in the wild tor the completion of engulfment and to
(and in the course of multiple cycles of sporula- transduce a more classical signal from the
tion and germination), small defects in the mother cell. All three of these signaling path-
timing of late mother cell gene expression ways have served as powerful models for study-
might have more pronounced consequences for ing cell-cell signaling in bacteria.Their further
the organism. elucidation will continue to provide insights
In summary, G activity in the forespore into how information is transduced across a
sends a signal to the mother cell via a two-step lipid bilayer.
proteolytic cleavage pathway that results in acti-
vation of K. In this signaling pathway informa-
ACKNOWLEDGMENTS
tion is transduced across the lipid bilayer by the
We acknowledge Amy Hitchcock Camp, Patrick
action of IVB and CtpB on one side of the Stragier, and members of the Rudner laboratory for
membrane that triggers cleavage of pro-K valuable discussions; Nathalie Campo for microscopy;
within or adjacent to the membrane by the and Kirsten Benjamin for extensive editorial advice.
The work in the Rudner laboratory is supported by 12. Driks, A., and R. Losick. 1991. Compartmen-
National Institutes of Health grant GM073831-01A1, talized expression of a gene under the control of
the Giovanni Armenise-Harvard Foundation, and the sporulation transcription factor sigma E in Bacillus
Hellman Family Faculty Fund. D. Z. R. is a Damon subtilis. Proc. Natl.Acad. Sci. USA 88:9934–9938.
Runyon Scholar supported by the Damon Runyon 13. Dworkin, J., and R. Losick. 2005.Developmen-
Cancer Research Foundation (DRS-44-05). tal commitment in a bacterium.Cell 121:401–409.
14. Eichenberger, P., P. Fawcett, and R. Losick.
REFERENCES 2001.A three-protein inihibitor of polar septation
1. Blaylock, B., X. Jiang,A. Rubio, C. P. Moran, during sporulation in Bacillus subtilis.Mol.Microbiol.
Jr., and K. Pogliano. 2004. Zipper-like inter- 42:1147–1162.
action between proteins in adjacent daughter 15. Errington, J. 2003. Regulation of endospore for-
cells mediates protein localization. Genes Dev. mation in Bacillus subtilis. Nat. Rev. Microbiol.
18:2916–2928. 1:117–126.
2. Brown, M. S., and J. L. Goldstein. 1997. The 16. Fawcett, P., A. Melnikov, and P. Youngman.
SREBP pathway: regulation of cholesterol metab- 1998. The Bacillus SpoIIGA protein is targeted
olism by proteolysis of a membrane-bound tran- to sites of spore septum formation in a SpoIIE-
scription factor. Cell 89:331–340. independent manner. Mol. Microbiol. 28:931–943.
3. Campo, N., and D. Z. Rudner. 2006. A 17. Fujita, M., and R. Losick. 2002. An investiga-
branched pathway governing the activation of a tion into the compartmentalization of the sporula-
developmental transcription factor by regulated tion transcription factor sigmaE in Bacillus subtilis.
intramembrane proteolysis. Mol. Cell 23:25–35. Mol. Microbiol. 43:27–38.
4. Campo, N., and D. Z. Rudner. 2007. SpoIVB 18. Fujita, M., and R. Losick. 2003. The master
and CtpB are both forespore signals in the activa- regulator for entry into sporulation in Bacillus
tion of the sporulation transcription factor sigmaK subtilis becomes a cell-specific transcription
in Bacillus subtilis. J. Bacteriol. 189:6021–6027. factor after asymmetric division. Genes Dev.
5. Clarkson, J., I. D. Campbell, and M. D. 17:1166–1174.
Yudkin. 2004. Efficient regulation of sigmaF, 19. Gomez, M., S. Cutting, and P. Stragier. 1995.
the first sporulation-specific sigma factor in Transcription of spoIVB is the only role of sigma G
B. subtilis. J. Mol. Biol. 342:1187–1195. that is essential for pro-sigma K processing during
6. Cutting, S., A. Driks, R. Schmidt, B. Kunkel, spore formation in Bacillus subtilis. J. Bacteriol.
and R. Losick. 1991. Forespore-specific tran- 177:4825–4827.
scription of a gene in the signal transduction path- 20. Hilbert, D.W., and P. J. Piggot. 2004.Compart-
way that governs Pro-sigma K processing in mentalization of gene expression during Bacillus
Bacillus subtilis.Genes Dev. 5:456–466. subtilis spore formation. Microbiol. Mol. Biol. Rev.
7. Cutting, S.,V. Oke,A. Driks, R. Losick, S. Lu, 68:234–262.
and L. Kroos. 1990. A forespore checkpoint for 21. Hofmeister, A. 1998. Activation of the propro-
mother cell gene expression during development tein transcription factor pro-sigmaE is associated
in B. subtilis. Cell 62:239–250. with its progression through three patterns of sub-
8. Cutting, S., S. Roels, and R. Losick. 1991. cellular localization during sporulation in Bacillus
Sporulation operon spoIVF and the characteriza- subtilis. J. Bacteriol. 180:2426–2433.
tion of mutations that uncouple mother-cell from 22. Hofmeister, A. E., A. Londono-Vallejo, E.
forespore gene expression in Bacillus subtilis. J. Mol. Harry, P. Stragier, and R. Losick. 1995. Extra-
Biol. 221:1237–1256. cellular signal protein triggering the proteolytic
9. Doan, T., K. A. Marquis, and D. Z. Rudner. activation of a developmental transcription factor
2005. Subcellular localization of a sporulation in B. subtilis. Cell 83:219–226.
membrane protein is achieved through a network 23. Iber, D., J. Clarkson, M. D.Yudkin, and I. D.
of interactions along and across the septum. Mol. Campbell. 2006.The mechanism of cell differen-
Microbiol. 55:1767–1781. tiation in Bacillus subtilis. Nature 441:371–374.
10. Doan, T., and D. Z. Rudner. 2007. Perturba- 24. Igoshin, O. A., C. W. Price, and M. A.
tions to engulfment trigger a degradative response Savageau. 2006. Signalling network with a
that prevents cell-cell signaling during sporulation bistable hysteretic switch controls developmental
in Bacillus subtilis. Mol. Microbiol. 64:500–511. activation of the sigma transcription factor in
11. Dong, T. C., and S. M. Cutting. 2003.SpoIVB- Bacillus subtilis. Mol. Microbiol. 61:165–184.
mediated cleavage of SpoIVFA could provide 25. Jiang, X., A. Rubio, S. Chiba, and K.
the intercellular signal to activate processing of Pogliano. 2005. Engulfment-regulated proteoly-
Pro-sigmaK in Bacillus subtilis. Mol. Microbiol. sis of SpoIIQ: evidence that dual checkpoints con-
49:1425–1434. trol sigma activity. Mol. Microbiol. 58:102–115.
26. Johnson, B. D., and A. J. Dombroski. 1997. 39. Lu, S., R. Halberg, and L. Kroos. 1990. Pro-
The role of the pro sequence of Bacillus subtilis sig- cessing of the mother-cell sigma factor, sigma K,
maK in controlling activity in transcription initia- may depend on events occurring in the forespore
tion. J. Biol. Chem. 272:31029–31035. during Bacillus subtilis development. Proc. Natl.
27. Ju, J., Luo, T., and W. G. Haldenwang. 1997. Acad. Sci. USA 87:9722–9726.
Bacillus subtilis pro-sigmaE fusion protein localizes 40. Makinoshima, H., and M. S. Glickman.
to the forespore septum and fails to be processed 2006. Site-2 proteases in prokaryotes: regulated
when synthesized in the forespore. J. Bacteriol. intramembrane proteolysis expands to microbial
179:4888–4893. pathogenesis. Microbes Infect. 8:1882–1888.
28. Karow, M. L., P. Glaser, and P. J. Piggot. 1995. 41. Margolis, P. S.,A. Driks, and R. Losick. 1993.
Identification of a gene, spoIIR, that links the acti- Sporulation gene spoIIB from Bacillus subtilis. J.
vation of sigma E to the transcriptional activity of Bacteriol. 175:528–540.
sigma F during sporulation in Bacillus subtilis. Proc. 42. Murakami, T., K. Haga, M. Takeuchi, and T.
Natl.Acad. Sci. USA 92:2012–2016. Sato. 2002. Analysis of the Bacillus subtilis spoIIIJ
29. Kellner, E. M., A. Decatur, and C. P. Moran, gene and its paralogue gene, yqjG. J. Bacteriol.
Jr. 1996. Two-stage regulation of an anti-sigma 184:1998–2004.
factor determines developmental fate during 43. Pan, Q., R. Losick, and D. Z. Rudner. 2003.
bacterial endospore formation. Mol. Microbiol. A second PDZ-containing serine protease con-
21:913–924. tributes to activation of the sporulation transcrip-
30. Khvorova, A.,V. K. Chary, D. W. Hilbert, and tion factor sigmaK in Bacillus subtilis. J. Bacteriol.
P. J. Piggot. 2000.The chromosomal location of 185:6051–6056.
the Bacillus subtilis sporulation gene spoIIR is 44. Partridge, S. R., and J. Errington. 1993. The
important for its function. J. Bacteriol. 182: importance of morphological events and intercel-
4425–4429. lular interactions in the regulation of prespore-
31. King, N., O. Dreesen, P. Stragier, K. Pogliano, specific gene expression during sporulation in
and R. Losick. 1999. Septation, dephosphoryla- Bacillus subtilis. Mol. Microbiol. 8:945–955.
tion, and the activation of sigmaF during sporula- 45. Piggot, P. J., and R. Losick. 2002. Sporulation
tion in Bacillus subtilis. Genes Dev. 13:1156–1167. genes and intercompartmental regulation,
32. Kroos, L., B. Kunkel, and R. Losick. 1989. p. 483–517. In A. L. Sonenshein, J.A. Hoch, and R.
Switch protein alters specificity of RNA poly- Losick (ed.),Bacillus subtilis and Its Closest Relatives:
merase containing a compartment-specific sigma from Genes to Cells. ASM Press,Washington, DC.
factor. Science 243:526–529. 46. Pogliano, J., N. Osborne, M. D. Sharp, A.
33. LaBell, T. L., J. E. Trempy, and W. G. Abanes-De Mello, A. Perez,Y. L. Sun, and K.
Haldenwang. 1987. Sporulation-specific sigma Pogliano. 1999. A vital stain for studying mem-
factor sigma 29 of Bacillus subtilis is synthesized brane dynamics in bacteria: a novel mechanism
from a precursor protein, P31. Proc. Natl.Acad. Sci. controlling septation during Bacillus subtilis sporu-
USA 84:1784–1788. lation. Mol. Microbiol. 31:1149–1159.
34. Li, Z., and P. J. Piggot. 2001. Development 47. Rawson, R. B., N. G. Zelenski, D. Nijhawan,
of a two-part transcription probe to determine J.Ye, J. Sakai, M. T. Hasan, T.Y. Chang, M. S.
the completeness of temporal and spatial com- Brown, and J. L. Goldstein. 1997.Complemen-
partmentalization of gene expression during bac- tation cloning of S2P, a gene encoding a putative
terial development. Proc. Natl. Acad. Sci. USA metalloprotease required for intramembrane
98:12538–12543. cleavage of SREBPs. Mol. Cell 1:47–57.
35. Londono-Vallejo, J. A., C. Frehel, and P. 48. Resnekov, O., S. Alper, and R. Losick. 1996.
Stragier. 1997. SpoIIQ, a forespore-expressed Subcellular localization of proteins governing the
gene required for engulfment in Bacillus subtilis. proteolytic activation of a developmental tran-
Mol. Microbiol. 24:29–39. scription factor in Bacillus subtilis. Genes Cells
36. Londono-Vallejo, J. A., and P. Stragier. 1995. 1:529–542.
Cell-cell signaling pathway activating a develop- 49. Resnekov, O., and R. Losick. 1998. Negative
mental transcription factor in Bacillus subtilis. Genes regulation of the proteolytic activation of a devel-
Dev. 9:503–508. opmental transcription factor in Bacillus subtilis.
37. Losick, R., and P. Stragier. 1992.Crisscross reg- Proc. Natl.Acad. Sci. USA 95:3162–3167.
ulation of cell-type-specific gene expression dur- 50. Ricca, E., S. Cutting, and R. Losick. 1992.
ing development in B.subtilis.Nature 355:601–604. Characterization of bofA, a gene involved in inter-
38. Losick, R., P. Youngman, and P. J. Piggot. compartmental regulation of pro-sigma K process-
1986. Genetics of endospore formation in Bacillus ing during sporulation in Bacillus subtilis.J.Bacteriol.
subtilis.Annu. Rev. Genet. 20:625–669. 174:3177–3184.
51. Rubio, A., and K. Pogliano. 2004. Septal 63. Trempy, J. E., J. Morrison-Plummer, and
localization of forespore membrane proteins W. G. Haldenwang. 1985. Synthesis of sigma 29,
during engulfment in Bacillus subtilis. EMBO J. an RNA polymerase specificity determinant, is a
23:1636–1646. developmentally regulated event in Bacillus subtilis.
52. Rudner, D. Z., P. Fawcett, and R. Losick. J. Bacteriol. 161:340–346.
1999. A family of membrane-embedded metallo- 64. Wakeley, P. R., R. Dorazi, N. T. Hoa, J. R.
proteases involved in regulated proteolysis of Bowyer, and S. M. Cutting. 2000. Proteolysis of
membrane-associated transcription factors. Proc. SpolVB is a critical determinant in signalling of
Natl.Acad. Sci. USA 96:14765–14770. pro-sigmaK processing in Bacillus subtilis. Mol.
53. Rudner, D. Z., and R. Losick. 2001. Morpho- Microbiol. 36:1336–1348.
logical coupling in development: lessons from 65. Wang, S. T., B. Setlow, E. M. Conlon, J. L.
prokaryotes. Dev. Cell 1:733–742. Lyon, D. Imamura, T. Sato, P. Setlow, R.
54. Rudner, D. Z., and R. Losick. 2002.A sporula- Losick, and P. Eichenberger. 2006.The fores-
tion membrane protein tethers the prosigmaK pore line of gene expression in Bacillus subtilis. J.
processing enzyme to its inhibitor and dictates its Mol. Biol. 358:16–37.
subcellular localization. Genes Dev. 16:1007–1018. 66. Wu, L. J., A. Feucht, and J. Errington. 1998.
55. Schlessinger, J. 2000. Cell signaling by receptor Prespore-specific gene expression in Bacillus subtilis
tyrosine kinases. Cell 103:211–225. is driven by sequestration of SpoIIE phosphatase to
56. Schlessinger, J. 2002. Ligand-induced, receptor- the prespore side of the asymmetric septum. Genes
mediated dimerization and activation of EGF Dev. 12:1371–1380.
receptor. Cell 110:669–672. 67. Yu, Y. T., and L. Kroos. 2000. Evidence that
57. Serrano, M., A. Neves, C. M. Soares, C. P. SpoIVFB is a novel type of membrane metallopro-
Moran, Jr., and A. O. Henriques. 2004. Role of tease governing intercompartmental communica-
the anti-sigma factor SpoIIAB in regulation of tion during Bacillus subtilis sporulation. J. Bacteriol.
sigmaG during Bacillus subtilis sporulation. J. Bacte- 182:3305–3309.
riol. 186:4000–4013. 68. Zelenski, N. G., R. B. Rawson, M. S. Brown,
58. Smith, K., M. E. Bayer, and P. Youngman. and J. L. Goldstein. 1999. Membrane topology
1993. Physical and functional characterization of of S2P, a protein required for intramembranous
the Bacillus subtilis spoIIM gene. J. Bacteriol. cleavage of sterol regulatory element-binding pro-
175:3607–3617. teins. J. Biol. Chem. 274:21973–21980.
59. Steil, L., M. Serrano,A. O. Henriques, and U. 69. Zhang, B.,A. Hofmeister, and L. Kroos. 1998.
Volker. 2005. Genome-wide analysis of tempo- The prosequence of pro-sigmaK promotes mem-
rally regulated and compartment-specific gene brane association and inhibits RNA polymerase
expression in sporulating cells of Bacillus subtilis. core binding. J. Bacteriol. 180:2434–2441.
Microbiology 151:399–420. 70. Zhou, R., and L. Kroos. 2004. BofA protein
60. Stragier, P., C. Bonamy, and C. Karmazyn- inhibits intramembrane proteolysis of pro-sigmaK
Campelli. 1988. Processing of a sporulation in an intercompartmental signaling pathway dur-
sigma factor in Bacillus subtilis: how morpho- ing Bacillus subtilis sporulation. Proc. Natl.Acad. Sci.
logical structure could control gene expression. USA 101:6385–6390.
Cell 52:697–704. 71. Zuber, P., and R. Losick. 1983. Use of a lacZ
61. Stragier, P., and R. Losick. 1996. Molecular fusion to study the role of the spoO genes of Bacil-
genetics of sporulation in Bacillus subtilis. Annu. lus subtilis in developmental regulation. Cell
Rev. Genet. 30:297–241. 35:275–283.
62. Sun,Y. L., M. D. Sharp, and K. Pogliano. 2000. 72. Zupancic, M. L., H. Tran, and A. E.
A dispensable role for forespore-specific gene Hofmeister. 2001. Chromosomal organization
expression in engulfment of the forespore governs the timing of cell type-specific gene
during sporulation of Bacillus subtilis. J. Bacteriol. expression required for spore formation in Bacillus
182:2919–2927. subtilis. Mol. Microbiol. 39:1471–1481.
EXTRACELLULAR PEPTIDE SIGNALING
AND QUORUM RESPONSES IN
DEVELOPMENT, SELF-RECOGNITION,
AND HORIZONTAL GENE TRANSFER
IN BACILLUS SUBTILIS
Jennifer M.Auchtung and Alan D. Grossman
2
Microbes generally have mechanisms for sens- sporulation, the ComA-mediated general quo-
ing and responding to environmental condi- rum response,and horizontal gene transfer.This
tions, including aspects of nutrient availability signaling provides mechanisms for the cell to
and the presence or absence of other cells.Many monitor population density (or limited diffu-
microbes are particularly adept at monitoring sion indicative of colonization) and to distin-
the presence or absence of neighbors and guish whether neighboring cells are similar or
determining whether the neighbors are similar different. All the responses described above are
to themselves. Bacillus subtilis generally uses stimulated by high population densities. Both
secreted peptides to monitor the presence of mechanisms of horizontal gene transfer are also
other cells.This extracellular signaling controls a controlled by recognition of self. Competence
variety of processes, including sporulation, development is stimulated in the presence of
biofilm development, swarming, the ComA- genetically similar cells whereas transfer of
mediated quorum response,and at least two dif- ICEBs1 is inhibited in the presence of other
ferent mechanisms for horizontal gene transfer. cells that contain ICEBs1. Some strains of B.
These mechanisms of horizontal gene transfer subtilis also use AI-2 signaling to regulate
are (i) competence development and transfor- biofilm development and swarming in response
mation and (ii) excision and mating of the inte- to population density (69). However, the focus
grative and conjugative element ICEBs1 (a of this chapter is cell-cell signaling mediated by
conjugative transposon). peptides.
This chapter focuses on the mechanisms of
extracellular peptide signaling that B. subtilis CELL DENSITY PHENOMENA
utilizes to control gene expression involved in IN B. SUBTILIS
In B.subtilis,both competence development and
sporulation have long been known to be more
Jennifer M. Auchtung and Alan D. Grossman Department of
Biology, Massachusetts Institute of Technology, Cambridge, efficient at higher population densities (2, 3, 37,
Massachusetts 02139. 38, 52, 56, 121, 123). Such population density
13
14 ■ AUCHTUNG AND GROSSMAN
effects often indicate that there is signaling Population-density-dependent control of

between cells. However, since nutritional con- competence likely serves to coordinate expres-
ditions change as cells grow to high density, sion of the DNA uptake machinery with the
especially in complex medium,it can be difficult likely presence of conspecific exogenous DNA,
to distinguish nutrient effects from effects due to which could arise from cell lysis.The secreted
cell-cell signaling. For both competence devel- peptides that stimulate competence develop-
opment and sporulation in B. subtilis, it is clear ment are species-specific, thereby causing com-
that population-density-dependent regulation petence to develop when a cell is crowded by
involves specific cell-cell signaling peptides. similar cells (4, 5). It is interesting to note that
Changes in gene expression at the end of of the four well-studied organisms that are
exponential growth are highly regulated. One naturally transformable, the two that take up
of the most significant changes is the initiation DNA independently of sequence, B. subtilis and
of sporulation. Spore formation allows the cell Streptococcus pneumoniae, use cell-cell signaling
to survive a variety of environmental stresses. to regulate competence development. In con-
However, it comes at a cost. Sporulating cells trast, the two organisms that have sequence-
undergo asymmetric septation generating two specific DNA uptake, Neisseria gonorrhoeae and
cell types: a forespore and a mother cell. The Haemophilus influenzae, do not use cell-cell sig-
forespore develops inside the mother cell, naling to regulate competence development
which then lyses to release the mature spore. (reviewed in reference 107).
This is in contrast to cell division in nonsporu- Cell-density control of conjugation of
lating cells where two daughter cells result from ICEBs1 has two aspects. ICEBs1 excision and
every division. Many of the regulatory changes gene expression are stimulated at high density
are at the end of exponential growth and, as (and nutrient deprivation) but inhibited by a
cells get crowded,appear to function to increase secreted peptide if neighboring cells also contain
the chances of growth and reduce the chances ICEBs1 (7), another form of self-recognition.
of sporulation. If nutritional conditions do not
support sufficient growth, then sporulation can QUORUM SENSING AND
efficiently start. CONTROL OF SPORULATION
Nutrient depletion is the major condition Nutrient deprivation is a primary signal that
needed to initiate sporulation. However, it is causes cells to initiate sporulation. However,
not the only condition. Sporulation is most sporulation is inefficient in cells at low popula-
efficient upon nutrient depletion when cells are tion density, even upon starvation, indicating
crowded. One role for population-density- that nutrient depletion is not sufficient for effi-
dependent control of sporulation may be to cient sporulation (38, 52, 68, 121, 123). This
serve as an early warning signal of starvation by population-density-dependent regulation of
indicating crowding and possible competition sporulation is mediated by extracellular signal-
for scarce nutrients. If cells are starving in a ing molecules that affect the earliest steps in
crowded colony, it may be advantageous to sporulation initiation.
sporulate rather than compete for additional
nutrients. If cells are dispersed (i.e., at low cell Cell-Cell Signaling
density), then the chances of finding additional and the Phosphorelay
nutrients might be higher and sporulation The spo0A gene product is the key transcription
might be less desirable. In addition, spores factor required for the initiation of sporulation
appear to form at specific locations within mul- (reviewed in references 37 and 51),and its activ-
ticellular communities (12, 122). Population- ity is regulated, in part, by population-density
density-dependent control of sporulation may signals (38). Spo0A is a member of the response
indicate the presence of cells at a concentration regulator family of proteins and receives phos-
sufficient to form these structures. phate from a phosphorelay consisting of multi-
2. PEPTIDE SIGNALING AND QUORUM RESPONSES IN B. SUBTILIS ■ 15
also directly activates genes involved in produc-

tion of the spore-killing factor (30, 36, 76) and
sporulation (37, 51, 76).
The pleiotropic functions of Spo0A~P are
dependent on its concentration (22, 33, 105).
Certain promoters, such as those for abrB, sinI,
and skf, are regulated by low concentrations of
Spo0A~P. Other promoters (e.g., spoIIA,
spoIIE, spoIIG) require a high concentration of
FIGURE 1 Regulation of the transcription factor Spo0A~P. In general, those promoters that
Spo0A by the phosphorelay and extracellular peptide require a high concentration of Spo0A~P are
signaling. The response regulator and transcription involved in sporulation and those that require a
factor Spo0A is activated by a phosphorelay that
transfers phosphate from histidine protein kinases,
lower concentration of Spo0A~P are involved
KinA to KinE, to the response regulator Spo0F, then in other postexponential phase processes. This
to Spo0B, and finally to Spo0A. Spo0A~P directly regulation likely allows cells to try other, less
activates (→) transcription of some genes and represses drastic, adaptive approaches to starvation before
(— ) transcription of others. The phosphatases RapA,
committing to sporulation.
RapB, RapE, and RapH promote dephosphorylation
of Spo0F~P, thereby inhibiting the activation of
Spo0A.The activity of RapA, RapB, RapE, and RapH Rap-Phr Signaling and
is inhibited by the PhrA, PhrC (CSF), PhrE, and Control of Sporulation
PhrH pentapeptides, respectively. ComA~P binds to The best-characterized signaling peptides that
sites upstream from rapA, C, E, and F to activate stimulate sporulation are the pentapeptides
transcription.
PhrA, PhrC (a.k.a., CSF, for competence and
sporulation-stimulating factor), and PhrE. Each
ple kinases and Spo0F and Spo0B (Fig. 1) (14). peptide inhibits the activity of a phosphatase
Accumulation of Spo0A~P is negatively regu- (RapA, RapB, and RapE, respectively) of the
lated by the activities of multiple phosphatases phosphorelay component Spo0F~P, thereby
(55,89,92,95),several of which target Spo0F~P indirectly increasing the level of Spo0A~P
and are inhibited by cell-cell signaling mole- (Fig. 1).
cules (Fig. 1). The PhrA pentapeptide stimulates sporula-
In addition to its role in regulating sporulation by antagonizing the activity of the Spo0F
tion, Spo0A~P regulates many other processes, phophatase, RapA (93, 95). The role that the
including biofilm development, the production PhrA peptide plays in population-density sig-
of antibiotics, competence development, and naling is not clear, as it is not known if the PhrA
the activity of the mobile genetic element peptide accumulates in culture supernatants as
ICEBs1. cells grow to high density, a prerequisite for its
Many genes and processes are affected by ability to serve as a cell-density signal. There-
Spo0A~P indirectly. Spo0A~P activates biofilm fore,it has been proposed that the PhrA peptide
development by inhibiting expression of the serves as a timing mechanism for the control of
transition state regulator,AbrB, which represses the initiation of sporulation (93). However, the
the transcription of genes required for biofilm sporulation defect of phrA mutant cells is res-
development (12,46,47) and by activating tran- cued by growth in the presence of wild-type
scription of SinI, an antagonist of SinR, the cells (97), indicating that the PhrA peptide can
master regulator of biofilm formation (9, 21, serve as a cell-cell signaling molecule. In addi-
34, 57). The effects of Spo0A~P on repressing tion, transcription of rapA and phrA, which are
transcription of abrB also control antibiotic encoded together in an operon, is activated by
production (73, 74, 101, 112), competence ComA~P (78), a response regulator activated at
development, and ICEBs1 (below). Spo0A~P high population density (see below), indicating
that PhrA peptide signaling conveys informa- initially sequenced, encodes 11 Rap proteins
tion about population density indirectly (RapA–K) (64). Eight of the rap genes are
through its regulation by ComA. immediately upstream of and cotranscribed
The PhrC pentapeptide stimulates sporula- with recognizable phr genes (phrA, phrC, phrE,
tion by inhibiting RapB, a Spo0F~P phos- phrF, phrG, phrI, phrH, and phrK) (48, 64). The
phatase,and another member of the Rap family physiological roles of most of these Rap pro-
of proteins. The PhrC pentapeptide prevents teins and Phr peptides have been investigated
RapB from dephosphorylating Spo0F~P in (Fig. 2). RapC, RapF, RapG, RapH, RapK, and
vitro (93). Similarly, the PhrE pentapeptide their cognate peptides all play a role in regulat-
antagonizes the activity of the Spo0F~P phos- ing the activity of ComA (discussed below).
phatase, RapE (55). RapG/PhrG also regulates a second response
regulator, DegU. RapI and PhrI are encoded by
Production, Export, and the mobile genetic element ICEBs1 and con-
Import of Phr Pentapeptides trol its mobility (discussed below). The targets
The mature forms of the Phr pentapeptides are of RapD and RapJ have not been identified.
generated by processing of precursor peptides Rap protein activity is mediated by tetratri-
that are ~40 amino acids long. Each precursor copeptide repeat domains, protein-protein
peptide contains a putative signal recognition interaction domains that compose approxi-
sequence and cleavage site in its N terminus for mately two-thirds of each Rap protein (25, 53,
the B. subtilis type I secretion machinery. How- 94). Six Rap proteins have been characterized
ever, it is unclear whether the secretion
machinery plays a role in export and processing
of precursor Phr peptides (113). Furthermore,
even if secretion and processing were mediated
by the type I secretion machinery, each peptide
would have to undergo at least one additional
processing step to generate the mature form of
the peptide.
The Phr peptides are brought back into the
cell through the oligopeptide permease (Opp;
a.k.a., Spo0K) (68, 93), which is required for
efficient sporulation (96, 103). Opp is a rela-
tively sequence-independent peptide importer
that is a member of the ATPase-binding cassette
family of transporters (reviewed in references
49 and 115). A second oligopeptide permease,
App,may import Phr peptides in some strains of
B. subtilis.This permease is nonfunctional in the
strain of B. subtilis initially sequenced due to a
frame-shift mutation in its peptide-binding
protein, AppA. However, restoration of the
appropriate frame of appA allows App to largely FIGURE 2 Multiple rap and phr genes in B. subtilis.
substitute for Opp (63). The 11 rap and 8 phr genes in the B. subtilis genome are
indicated.All eight phr genes are downstream from and
cotranscribed with a cognate rap. Six of the eight phr
OTHER Rap PROTEINS genes are known to be transcribed from at least one
AND Phr PEPTIDES promoter that is sigma-H-dependent and internal to
There are families of rap and phr genes in Bacil- the cognate rap.The rapA, C, E, and F operons are acti-
lus species (94, 100). B. subtilis 168, the strain vated by ComA~P.
in molecular detail. RapA, RapB, and RapE REGULATION OF ComA BY

bind to Spo0F~P and stimulate dephosphoryla- POPULATION DENSITY
tion (53, 55, 95). In contrast, RapC, RapF, and At least two population-density signals, the
RapG have been shown to interfere with bind- ComX pheromone and the PhrC pentapep-
ing of response regulators to DNA (11, 25, 87). tide, stimulate the activity of ComA (72, 108,
For both types of regulators, Phr peptides are 109) (Fig. 3A). These peptides accumulate in
thought to competitively inhibit binding of and can be isolated from growth medium from
Rap proteins to response regulators (25, 53). cells grown to high density. Once these signals
achieve a sufficient concentration, reflective of
the number of producing cells in the popula-
A GENERAL QUORUM RESPONSE
CONTROLLED BY THE tion, they stimulate ComA activity. ComA then
TRANSCRIPTION FACTOR ComA activates expression of the genes in its regulon.
Several Rap-Phr pairs (RapC/PhrC, RapF/
PhrF,RapG/PhrG,RapH/PhrH,RapK/PhrK) ComX
regulate the activity of the transcription factor ComX is an isoprenylated oligopeptide that
ComA, a response regulator that is activated by can vary in size from 5 to 10 amino acids,
phosphorylation (Fig. 3A). ComA receives depending on the strain of B. subtilis (4, 5, 72,
phosphate from the histidine kinase, ComP, 116, 117). In addition to varying in amino acid
which is activated by an extracellular signaling length and sequence, the number of isoprenyl
peptide, ComX (discussed below). ComA can groups added to the mature ComX peptide also
also receive phosphate from small molecule varies among different strains, although iso-
donors (60). prenylation is required for function in all strains
ComA mediates an adaptive response to (5, 72, 116, 117). These differences in the
high population density. ComA activates ComX peptide contribute to strain specificity
expression of the large srfA operon (81, 82, 102) in ComX signaling (see below).
that is required for production of the lipopep- ComX is produced as a precursor polypep-
tide antibiotic surfactin (79, 80) and contains tide, and ComQ is required for modifica-
comS, encoding a protein required for compe- tion and production of the mature ComX
tence development (discussed below). Surfactin pheromone (72). Expression of comX and comQ
is a lipopeptide antibiotic that is required for is sufficient for Escherichia coli to produce
swarming motility and fruiting body formation mature ComX pheromone (5, 116). Further-
(12,58,59).ComA also stimulates expression of more, mutations in the isoprenoid-binding
other genes involved in antibiotic production, domain of ComQ inhibit production of mature
such as the B. subtilis spore-killing factor and ComX pheromone (8). It is not known how
sublancin, a lantibiotic that is active against ComX is exported.
gram-positive bacteria (24, 36, 91). ComA acti- ComX interacts extracellularly with its
vates transcription of the degradative enzymes receptor, the histidine kinase, ComP (5, 72, 99,
pectate lyase and lipase, as well as a regulator of 109, 117). This likely stimulates ComP
degradative enzyme synthesis, degQ (24, 77). In autophosphorylation; phosphate is transferred
addition, ComA stimulates transcription of to ComA, resulting in its activation for DNA
rapA, rapC, rapE, rapF, and several genes encod- binding (83, 99, 102, 124, 125). ComX and
ing products involved in membrane and extra- ComP appear to function exclusively to regu-
cellular processes (24, 54, 55, 78, 108, 111). late the activity of ComA; all genes whose
Cell-density signaling controls expression of expression decreases as a result of deletion of
many of these genes in other organisms as well, comQ, comX, or comP are also affected by dele-
demonstrating the generality of this control tion of comA (24).
mechanism for the production of secreted comQ, comX, comP, and comA are together
products. and apparently in a single operon (72, 125),
although there might be internal promoters domain) (4, 5, 116, 117). Diversity in comQ and
(Fig. 3A). This arrangement is similar to that comX results in production of different forms of
observed in the signaling cassettes of several the mature ComX peptide, and variations in
other gram-positive bacteria (13, 61, 85, 114). comP allow response to different ComX pep-
The comQXPA locus is conserved in several B. tides (5, 116, 117).
subtilis strains and closely related Bacillus
species, although an extensive amount of diver- Strain Specificity in ComX Signaling
sity exists in the sequence of comQ, comX, and The different forms of the mature ComX pep-
the 5′ end of comP (the signal reception tide have been classified into pherotypes based
FIGURE 3 Complex regulation, including autoregulatory loops, involving the transcription factor ComA and
multiple extracellular signaling peptides. Stimulatory effects are indicated with arrows (→). Inhibitory effects are
indicated by lines with bars (— ). (A) ComA is activated by extracellular peptide signaling.The cell is represented by
the contents of the large rectangular box with components that are intracellular, extracellular, or in the membrane.
Numbers are for descriptive purposes and do not necessarily indicate sequential events. (a) The ComX pheromone
is encoded together with a protein required for its production, ComQ, and the two-component signal transduction
system, ComP-ComA, that regulates competence development and a quorum response. In addition to the promoter
upstream from comQ, there are likely to be minor promoters (not shown) internal to the operon. (b) After transcrip-
tion and translation, a precursor to ComX pheromone (pre-ComX) is modified by ComQ and exported (c) by an
unknown mechanism. (d) ComX interacts on the cell surface with the membrane receptor-histidine kinase ComP
and stimulates auto-kinase activity of ComP.(e) Phosphate is transferred from ComP~P to ComA.(f) ComA~P acti-
vates expression of several genes, including comS, the only ComA-target that is required for competence develop-
ment. The activity of ComA~P is antagonized by the indicated Rap proteins in the absence of sufficient
concentrations of the cognate Phr pentapeptides. It is currently unclear whether RapG, RapH, and RapK affect
ComA directly or indirectly. (g) PhrC, PhrF, PhrG, PhrH, and PhrK pentapeptides are encoded in operons with their
cognate Rap proteins. (h) The pre-Phr peptides are exported and processed through an unknown mechanism. (i)
Mature Phr pentapeptides are imported into the cell through the oligopeptide permease (Opp; a.k.a., Spo0K). (j)
PhrC, PhrF, PhrG, PhrH, and PhrK stimulate ComA-dependent gene expression by antagonizing the activities of
their cognate Rap proteins. (Continued on following page)
FIGURE 3 Continued (B) Complex regulation of ComA and Spo0A by multiple signals, including several extra-
cellular peptides,allows for signal integration in the control of gene expression and many autoregulatory loops.Tran-
scription of the rap/phr genes is regulated by a variety of different proteins. The activities of these proteins are
controlled by a variety of physiological signals, some of which are described in more detail in the text. Spo0A is reg-
ulated indirectly by the Rap proteins via the effects of the Raps on Spo0F. This diagram is an oversimplification of
the regulatory circuits, and regulation of ComK by the Spo0A and ComA pathways is shown in Fig. 4.
on the ability of these peptides to affect ComP- rapC and phrC are encoded together in
ComA signaling in other strains (5). Some an operon, and transcription is activated by
forms of the mature ComX peptide stimulate ComA (66).This regulation establishes multiple
ComP-ComA signaling of strains that produce autoregulatory loops: ComA limits its own
different peptides, while other forms of mature activity by activating transcription of rapC;phrC
ComX peptides antagonize ComP-ComA sig- stimulates its own transcription by indirectly
naling of noncognate strains (5). This cross- increasing the activity of ComA.The autoregu-
regulation is similar to that observed for the latory loop established by PhrC and ComA may
AgrD signaling peptides of Staphylococcus be important for autoactivation of phrC tran-
species (reviewed in reference 85). It is thought scription once a sufficient concentration of the
that specificity in ComX-ComP-ComA sig- PhrC pentapeptide has accumulated.
naling may improve fitness of these strains by phrC is also expressed from a second pro-
limiting competence development in the pres- moter that is dependent on the alternative
ence of divergent strains and thereby providing sigma factor, sigma-H (18, 66).This regulation
a mechanism for sexual isolation (4). causes phrC transcription to increase as cells
transition into postexponential phase (66).
PhrC Sigma-H also plays a role in posttranscriptional
The PhrC pentapeptide stimulates ComA- control of synthesis of the PhrC pentapeptide,
dependent gene expression by inhibiting the likely by controlling transcription of a protein
activity of RapC, a negative regulator of ComA or proteins involved in pre-PhrC processing
(25, 68, 108). RapC binds to ComA and pre- (66). Therefore, regulation of transcription of
vents it from binding to DNA; the PhrC pep- phrC and production of the mature PhrC pen-
tide is a competitive inhibitor of the interaction tapeptide ensures that the levels of this peptide
between RapC and ComA (25). increase at high cell density and may be impor-
tant for the multiple roles of the PhrC pen- mechanisms have been observed in other
tapeptide. organisms (70, 98).
In addition to stimulating ComA-depend- RapG, RapH, and RapK have all been
ent gene expression, the PhrC pentapeptide has shown to inhibit ComA-dependent gene
two additional activities: stimulating sporula- expression. This inhibition is relieved by the
tion and inhibiting ComA-dependent gene actions of their cognate Phr peptides (6,48).It is
expression (68, 108). PhrC pentapeptide stimu- not clear whether these proteins affect the
lates ComA-dependent gene expression when activity of ComA directly or affect the activity
present at relatively low concentrations; maxi- of another regulator that influences ComA-
mal stimulation occurs when CSF peptide is dependent gene expression.All three Rap pro-
present at a concentration of 5 to 10 nM (108). teins also inhibit the expression of genes not
PhrC pentapeptide inhibits ComA-dependent regulated by ComA. RapG has been shown to
gene expression and stimulates sporulation inhibit the activity of the response regulator
when present at high concentrations (20 DegU by binding to this protein and prevent-
nM), although the effects of PhrC pentapeptide ing it from binding to DNA (87). RapG and
on sporulation are observed when the levels of RapH have also been reported to interact with
other extracellular signaling molecules are the C terminus of ComA in yeast two-hybrid
reduced (68). As described above, PhrC pen- experiments (48).
tapeptide stimulates sporulation by inhibiting
the activity of another Rap protein, RapB (93). Regulation of ComA by Multiple
It is not known how the PhrC pentapeptide Peptides Allows for Complex
inhibits ComA-dependent gene expression, Integration of Signals
although it is known that this is not mediated At least six peptides regulate the activity of
through RapC (68). ComA. Although this may appear ridiculous
and perhaps largely redundant, these multiple
Multiple Rap Proteins and peptides probably serve to integrate a variety of
Phr Peptides Regulate ComA complex signals into the decision to activate the
As mentioned above, several additional Rap ComA regulon (Fig. 3B).
proteins inhibit ComA-dependent gene comX appears to be transcribed constitu-
expression. RapF binds to ComA in vitro tively during normal exponential growth, and
and prevents it from binding to DNA (11). the production of ComX pheromone occurs
RapF also inhibits ComA-dependent gene consistently throughout growth, and provides
expression in vivo, although large effects of cells with information about population den-
RapF on ComA-dependent gene expression sity. Recent work has shown that comX tran-
are seen only in the absence of its inhibitory scription is inhibited by superoxide stress (90).
peptide (in either phrF or opp mutants). Fur- ComA activates transcription of both the
thermore, ComA also activates transcription rapC phrC and rapF phrF operons (24, 54, 66).
of the rapF-phrF operon; activation occurs at RNA polymerase containing sigma-H tran-
very low levels of ComA~P (24, 54), and could scribes phrC, phrF, phrG, and phrK and plays a
be important under conditions in which role in the production of the mature PhrC pep-
ComA might be activated by small-molecule tide (18, 66, 75). Sigma-H is regulated by a vari-
phosphate donors (60).This circuit may func- ety of signals, including inhibition of its activity
tion to keep ComA-dependent gene expres- by certain carbon sources and low pH (reviewed
sion low at very low cell densities and thereby in reference 15).Transcription of the rapC phrC
establish a sequential progression of cell- operon is also repressed by CodY (66), a protein
cell signaling with RapC-PhrC signaling. that senses the levels of branched chain amino
Sequential activation of cell-cell signaling acids and GTP (reviewed in reference 110).
Transcription of the rapG phrG and rapH

phrH operons is repressed by RghR (48). It is
currently not known what signals regulate
RghR.Transcription of the rapH phrH operon
is also activated by ComK (10, 88), the compe-
tence transcription factor (discussed below),
thereby imposing a different type of autoregu-
lation in the decision to become competent.
The ComK-dependent activation of rapH may
be important to inhibit expression of comS and
allow cells to exit from competence.
Transcription of the rapK phrK operon is
thought to be indirectly activated by the master
regulator of sporulation, Spo0A (31).The activ-
ity of Spo0A is regulated by DNA replication
and damage, CodY, sigma-H, and peptide sig-
naling (reviewed in reference 15).The peptide
signals that control the activity of Spo0A are
described above.
In summary, the six peptides that regulate
the activity of ComA are themselves regulated
by several different proteins that respond to a
variety of different signals.This complex regu-
latory network seems to allow cells to coordi-
nate the ComA response with appropriate
environmental conditions and with the activi- FIGURE 4 Complex regulation of the transcription
ties of other intracellular signaling networks. factor ComK. ComK is the master transcriptional reg-
ulator of competence development. Both its stability
(A) and transcription (B) are regulated by extracellular
REGULATION OF COMPETENCE peptide signaling. For simplicity in the figure, genes
DEVELOPMENT IN B. SUBTILIS (e.g., comK) and proteins (e.g., ComK) are not distin-
One aspect of the ComA response is to tran- guished. (A) Control of the stability of ComK. (a) At
scribe a gene, comS, that leads to competence low cell density,MecA binds to ComK and targets it for
development (27, 28, 40, 42, 84). Under the degradation by the ClpCP protease complex. (b) As cell
appropriate conditions (described above), the density increases, ComA is activated by quorum sens-
ing, and activates transcription of comS. (c) ComS dis-
ComA response is activated in all cells in rupts the complex between ComK and MecA,
the population. However, competence devel- liberating ComK. (d) ComK is free to activate tran-
opment occurs in a subset of cells in the popu- scription of genes that are required for competence. (e)
lation: 10 to 20% of cells in certain laboratory ComS and MecA are degraded by ClpCP. (B) Control
strains and 1% or less of cells in some wild iso- of transcription of comK. ComK activates its own
expression. Its activity is controlled by quorum
lates (17, 39, 71). Therefore, several additional sensing via ComA and ComS (see above). Several other
layers of regulation influence whether a cell proteins also regulate comK transcription, including
becomes competent. Spo0A and DegU, whose activities are also regulated
Competence requires the formation of a by cell-cell signaling. The regulatory network that
multicomponent DNA uptake apparatus that controls comK transcription is described more fully
in the text.
assembles at the poles of competent cells as well
as the activities of several recombination and
DNA repair proteins (reviewed in references 20
and 29). The master regulator of competence ComK to its own promoter (45).RapG inhibits
development, ComK, activates transcription of binding of DegU to the comK promoter in vitro
the operons encoding these proteins (10, 44, and inhibits the expression of other DegU-
88). Regulation of ComK occurs at two levels: controlled genes in vivo (87). Although the
transcription and stability (Fig. 4). Stabilization effects of the RapG protein and PhrG peptide
of ComK is mediated by ComS, which is action comK transcription in vivo have not been
vated by the signaling peptides that activate the evaluated, this may be one mechanism through
ComA response (Fig. 4A). Cell-cell signaling which peptide signaling controls transcription
also plays a role in regulating transcription of of comK.
comK (Fig. 4B). Transcription of comK is also repressed by
AbrB (41, 43, 120). As described above, tran-
Regulation of ComK Stability scription of AbrB is inhibited by Spo0A, a tran-
At low population density, ComK is unstable sciption factor whose activity is inhibited by
due to the activities of MecA and the ClpCP several intercellular signaling molecules (PhrA,
protease complex (Fig. 4A) (118, 119). MecA CSF, and PhrE).
binds to ComK and prevents it from binding to In summary, the regulation of transcription
DNA; MecA binding also serves to target of comK is complex and highly integrated with
ComK for degradation by ClpCP (118, 119). other signaling pathways in the cell.This com-
Interestingly, MecA was the first adaptor pro- plex regulation serves at least two purposes: it
tein for Clp proteases to be identified; other prevents expression of multiple developmental
adaptors have since been identified (reviewed processes in a single cell, and it limits compe-
in reference 1). When cells reach high popu- tence development to conditions in which it is
lation density, peptide signaling activates most likely to be beneficial to the cell.
ComA, and comS is expressed. ComS binds to
MecA and disrupts the interaction between GENERAL ASPECTS OF
ComK and MecA, thereby preventing degra- HORIZONTAL GENE TRANSFER
dation of ComK and allowing it to activate Horizontal gene transfer plays an important
transcription of competence genes (118, 119). role in bacterial evolution (26, 32, 35, 86).
Horizontal gene transfer can provide the cell
Regulation of comK Transcription with benefits, such as acquisition of a gene or
ComK activates its own transcription (Fig. 4B) genes that allows increased fitness. However,
(44, 120).This autoregulatory loop is the source horizontal gene transfer can also be detrimen-
of bistability in competence development, i.e., tal, such as when the incoming DNA encodes a
the segregation of the population of cells into product that is toxic to the cell.Horizontal gene
competent and noncompetent cells (71, 106). transfer is often regulated by cell-cell signaling,
In addition, several other proteins regulate including the pheromone-responsive conjugal
transcription of comK (Fig. 4B).The activities of plasmids of Entercoccus faecalis, the autoinducer-
these proteins coordinate comK transcription sensing conjugal plasmids of Agrobacterium
with nutritional and population-density sig- tumefaciens, and competence development in
nals. Rok represses comK transcription; ComK Streptococcus and Bacillus species (reviewed in
also represses rok transcription, thereby estab- references 19, 23, 65, 67, and 126).
lishing a second autoregulatory loop that The intercellular signaling peptides that acti-
controls comK transcription (50). CodY, a tran- vate competence limit its development to con-
scription factor that is inactive when the levels ditions in which there is most likely to be high
of branched chain amino acids and GTP are concentrations of DNA from related B. subtilis
low, represses transcription of comK (104). or Streptococcus cells. For the conjugal plasmids
Binding of DegU stimulates comK transcrip- of A. tumefaciens, cell-cell signaling promotes
tion, most likely by promoting binding of conjugal transfer of the plasmid to other A.
tumefaciens cells. Cell-cell signaling also regu- CELL-CELL SIGNALING AND

lates conjugal transfer of ICEBs1 and E. faecalis RECOGNITION OF SELF IN THE
pheromone-responsive plasmids. In these cases, CONTROL OF TRANSFER OF ICEBs1
chromosomal encoded signaling peptides stim- ICEBs1 is an integrative and conjugative ele-
ulate transfer of the conjugal element, whereas ment (Fig. 5A) that was identified in the chro-
element-encoded peptides inhibit transfer into mosome of B. subtilis 168 and is found in some
cells that already contain a copy of the element. other related B. subtilis strains (7, 16). Like other
These regulatory strategies appear to reduce integrative and conjugative elements (a.k.a.,
the potential risks of horizontal gene transfer conjugative transposons), ICEBs1 spends much
and increase the potential benefits. of its time integrated into the chromosome of
FIGURE 5 The integrative and conjugative element ICEBs1 and its regulation by peptide signaling. (A) The
genetic map of ICEBs1. Open reading frame and direction of transcription are indicated by thick black arrows.The
ends of the element are marked by 60-bp direct repeats (black rectangles at the ends). Genes whose functions are
known experimentally include rapI and phrI, toward the right end of the element; int, encoding integrase that is
needed for site-specific integration and excision; immA, encoding an antirepressor; immR, encoding a transcriptional
repressor; and xis, encoding excisionase.The known promoters are indicated by lines with arrows above the genes.
ImmR represses transcription from the promoter immediately upstream of xis, and both activates and represses its
own promoter, which also drives transcription of immA and int.The ydc and ydd genes in the central part of the ele-
ment are coregulated with xis and encode most of the machinery needed for conjugal transfer. (B) Regulation of
transcription of ICEBs1 by peptide signaling and the SOS response. Expression of the ICEBs1 excisionase (xis) and
conjugation genes is repressed by ImmR. ImmA appears to antagonize the activity of ImmR, thereby stimulating
expression of the excisionase and conjugation genes and causing excision and transfer.The activity of ImmA is stim-
ulated by RapI. However, rapI and phrI are not significantly expressed at low cell densities and in the presence of
abundant nutrients due to repression by AbrB.When cells are at high cell density and starved, Spo0A~P accumulates
and inhibits abrB,leading to increased transcription of rapI and phrI.phrI is also transcribed by RNA polymerase con-
taining sigma-H, an alternative sigma factor whose activity increases as cells enter stationary phase.The activity of
RapI is inhibited by the PhrI pentapeptide, thereby inhibiting excision and transfer of ICEBs1.The concentration of
the PhrI pentapeptide (encoded in ICEBs1) reflects the concentration of surrounding cells that contain ICEBs1.
ImmA is also activated by RecA under conditions that induce the SOS response.The RecA and RapI pathways are
independent of each other.
its host where it is passively propagated through ImmR, apparently by stimulating the activity
host chromosomal replication and cell division of the antirepressor ImmA (Fig. 5B). Prelimi-
(7). However, under certain conditions ICEBs1 nary evidence indicates that RapI may stabi-
excises from the host chromosome and directs lize ImmA, causing degradation of ImmR
its own transfer to recipient cells (7). Con- (J. M. A. uchtung, C. A. Lee, A. D. Grossman,
ditions that promote excision and transfer are unpublished data).
those that ensure efficient propagation of
ICEBs1.These conditions include activation of Inhibition of ICEBs1 Excision and
ICEBs1 excision and transfer by the host cell’s Transfer by a Self-Encoded Peptide
DNA damage response, which allows the ele- The activity of RapI is inhibited by the PhrI
ment a chance to escape its distressed host, pentapeptide (7). Because PhrI is encoded in
and activation of excision and transfer by the ICEBs1, its concentration in the environment
presence of high concentrations of recipient reflects the concentration of cells that contain
cells that lack a copy of ICEBs1 (Fig. 5B).This ICEBs1. ICEBs1 uses this self-recognition
second condition, the presence of potential mechanism to limit excision and transfer when
recipient cells that lack a copy of ICEBs1, is the majority of the cells in the population
sensed by cell-cell signaling through a combi- already contain ICEBs1. Peptide signaling
nation of element-encoded and host-encoded inhibits excision and transfer of ICEBs1 at an
signaling peptides. early step, before expression of the genes
required for excision and mating pore forma-
Regulation of ICEBs1 Excision tion. This regulation minimizes the potential
and Transfer by Host-Encoded risks to the element by allowing it to remain
Signaling Peptides integrated into the chromosome of the host
Transcription of the majority of genes in when suitable recipient cells are unavailable.
ICEBs1 is under the control of the ICEBs1-
encoded immunity repressor, ImmR. ImmR Potential for Regulation of
represses transcription of a putative operon Other Mobile Genetic Elements
containing the excisionase (xis) and several by Cell-Cell Signaling
genes that are required for mating and likely Several Bacillus mobile genetic elements con-
form the conjugal pore apparatus (ydcO-yddJ). tain rap and phr genes (7). rapE and phrE, which
Host-encoded signaling peptides indirectly are found in a defective prophage known as the
inhibit the activity of ImmR, thereby causing skin element, regulate the activity of Spo0A as
derepression of the xis-yddJ operon and subse- described above. Rap60, encoded by pTA1060,
quent excision and mating. can also antagonize the activity of Spo0A, pre-
Host-encoded signaling peptides regulate sumably by affecting the phosphorylation of
ICEBs1 by stimulating the activity of Spo0F, and the activity of Rap60 is antagonized
Spo0A, the master regulator of sporulation. As by Phr60 (62). However, it is not known if that
described above,Spo0A activity is stimulated by is the primary physiological function of this
three signaling peptides, PhrA, PhrC, and PhrE, rap/phr pair. In addition, the biological func-
which function in conjunction with a variety tions of the rap/phr pairs from most mobile ele-
of nutritional signals to regulate the activity ments are not known.We suspect that some of
of Spo0A. these rap/phr pairs may regulate the mobility of
Spo0A indirectly inhibits the activity their respective elements.
of ImmR by stimulating transcription of the
ICEBs1-encoded regulatory protein, rapI, SUMMARY
whose transcription is repressed by AbrB Cell-cell signaling is one mechanism that cells
(7). RapI then antagonizes the activity of use to survey their surroundings and respond
appropriately. Cells can use cell-cell signaling 5. Ansaldi, M., D. Marolt, T. Stebe, I. Mandic-
molecules to sense the high population density Mulec, and D. Dubnau. 2002. Specific activa-
tion of the Bacillus quorum-sensing systems by
and limited diffusion that are indicative of colo- isoprenylated pheromone variants. Mol. Microbiol.
nization, as well as the crowding that may indi- 44:1561–1573.
cate there will be a competition for scarce 6. Auchtung, J. M., C. A. Lee, and A. D.
resources. In addition, cell-cell signaling mole- Grossman. 2006. Modulation of the ComA-
cules are often used in the identification of self. dependent quorum response in Bacillus subtilis by
multiple Rap proteins and Phr peptides. J. Bacteriol.
Self-recognition during B. subtilis competence 188:5273–5285.
development likely serves to limit competence 7. Auchtung, J. M., C. A. Lee, R. E. Monson, A.
development to conditions when DNA from P. Lehman, and A. D. Grossman. 2005. Regu-
closely related bacteria will be present.Cell-cell lation of a Bacillus subtilis mobile genetic element
signaling influences the activity of ICEBs1 in by intercellular signaling and the global DNA
damage response. Proc. Natl. Acad. Sci. USA
two ways. First, signaling via host-encoded reg- 102:12554–12559.
ulators indicates a high population density and 8. Bacon Schneider, K.,T. M. Palmer, and A. D.
the proximity of potential mating partners.Sec- Grossman. 2002. Characterization of comQ
ond, the signaling pentapeptide PhrI is pro- and comX, two genes required for production of
duced by ICEBs1-containing cells and is used ComX pheromone in Bacillus subtilis. J. Bacteriol.
184:410–419.
to inhibit transfer to potential partners that 9. Bai, U., I. Mandic-Mulec, and I. Smith.
already contain the element. 1993. SinI modulates the activity of SinR, a
B. subtilis competence and sporulation are developmental switch protein of Bacillus subtilis,
also regulated by two regulators, ComK and by protein-protein interaction. Genes Dev. 7:
Spo0A,that are parts of multiple autoregulatory 139–148.
10. Berka, R. M., J. Hahn, M. Albano, I.
loops.They involve both positive and negative Draskovic, M. Persuh, X. Cui, A. Sloma, W.
feedback regulation and help establish and Widner, and D. Dubnau. 2002. Microarray
maintain stable subpopulations of cells that analysis of the Bacillus subtilis K-state: genome-
exhibit specific patterns of gene expression and wide expression changes dependent on ComK.
follow specific developmental fates. Mol. Microbiol. 43:1331–1345.
11. Bongiorni, C., S. Ishikawa, S. Stephenson, N.
Ogasawara, and M. Perego. 2005. Synergistic
ACKNOWLEDGMENTS regulation of competence development in Bacillus
Work in the Grossman laboratory is supported by Pub- subtilis by two Rap-Phr systems. J. Bacteriol.
lic Health Service grants GM50895 and GM41934. 187:4353–4361.
We thank members of the laboratory for useful dis- 12. Branda, S. S., J. E. Gonzalez-Pastor, S.
cussion. We apologize for not being able to cite all Ben-Yehuda, R. Losick, and R. Kolter. 2001.
appropriate references. Fruiting body formation by Bacillus subtilis. Proc.
Natl.Acad. Sci. USA 98:11621–11626.
13. Brurberg, M. B., I. F. Nes, and V. G. Eijsink.
REFERENCES 1997. Pheromone-induced production of antimi-
1. Ades, S. E. 2004. Proteolysis: adaptor, adaptor, crobial peptides in Lactobacillus. Mol. Microbiol.
catch me a catch. Curr. Biol. 14:R924–6. 26:347–60.
2. Akrigg, A., and S. R. Ayad. 1970. Studies on the 14. Burbulys, D., K. A. Trach, and J. A. Hoch.
competence-inducing factor of Bacillus subtilis. 1991. Initiation of sporulation in B. subtilis is con-
Biochem. J. 117:397–403. trolled by a multicomponent phosphorelay. Cell
3. Akrigg,A., S. R.Ayad, and G. R. Barker. 1967. 64:545–52.
The nature of a competence-inducing factor in 15. Burkholder,W. F., and A. D. Grossman. 2001.
Bacillus subtilis. Biochem. Biophys. Res. Comm. 28: Initiation of sporulation in B.subtilis,p.151–166.In
1062–1067. Y.V. Brun and L. J. Shimkets (ed.), Prokaryotic Devel-
4. Ansaldi, M., and D. Dubnau. 2004. Diversifying opment.ASM Press,Washington, DC.
selection at the Bacillus quorum-sensing locus and 16. Burrus, V., G. Pavlovic, B. Decaris, and G.
determinants of modification specificity during Guedon. 2002.The ICESt1 element of Streptococ-
synthesis of the ComX pheromone. J. Bacteriol. cus thermophilus belongs to a large family of integra-
186:15–21. tive and conjugative elements that exchange
modules and change their specificity of integra- lus subtilis and Its Closest Relatives. ASM Press,
tion. Plasmid 48:77–97. Washington, DC.
17. Cahn, F. H., and M. S. Fox. 1968. Fractionation 30. Ellermeier, C. D., E. C. Hobbs, J. E.
of transformable bacteria from ocompetent cul- Gonzalez-Pastor, and R. Losick. 2006.A three-
tures of Bacillus subtilis on renografin gradients. J. protein signaling pathway governing immunity
Bacteriol. 95:867–875. to a bacterial cannibalism toxin. Cell 124:549–559.
18. Carter, H. L., III, K. M.Tatti, and C. P. Moran, 31. Fawcett, P., P. Eichenberger, R. Losick, and P.
Jr. 1990. Cloning of a promoter used by sigma H Youngman. 2000. The transcriptional profile of
RNA polymerase in Bacillus subtilis. Gene 96: early to middle sporulation in Bacillus subtilis. Proc.
101–105. Natl.Acad. Sci. USA 97:8063–8068.
19. Chandler, J. R., and G. M. Dunny. 2004. Ente- 32. Frost, L. S., R. Leplae,A. O. Summers, and A.
rococcal peptide sex pheromones: synthesis and Toussaint. 2005. Mobile genetic elements: the
control of biological activity. Peptides 25:1377– agents of open source evolution. Nat. Rev. Micro-
1388. biol. 3:722–732.
20. Chen, I., P. J. Christie, and D. Dubnau. 2005. 33. Fujita, M., J. E. Gonzalez-Pastor, and R.
The ins and outs of DNA transfer in bacteria. Sci- Losick. 2005. High- and low-threshold genes in
ence 310:1456–1460. the Spo0A regulon of Bacillus subtilis. J. Bacteriol.
21. Chu, F., D. B. Kearns, S. S. Branda, R. Kolter, 187:1357–1368.
and R. Losick. 2006.Targets of the master regula- 34. Gaur, N. K., K. Cabane, and I. Smith. 1988.
tor of biofilm formation in Bacillus subtilis. Mol. Structure and expression of the Bacillus subtilis sin
Microbiol. 59:1216–1228. operon. J. Bacteriol. 170:1046–1053.
22. Chung, J. D., G. Stephanopoulos, K. Ireton, 35. Gogarten, J. P., and J. P.Townsend. 2005. Hor-
and A. D. Grossman. 1994. Gene expression in izontal gene transfer, genome innovation and evo-
single cells of Bacillus subtilis: evidence that a lution. Nat. Rev. Microbiol. 3:679–687.
threshold mechanism controls the initiation of 36. Gonzalez-Pastor, J. E., E. C. Hobbs, and R.
sporulation. J. Bacteriol. 176:1977–1984. Losick. 2003. Cannibalism by sporulating bacte-
23. Claverys, J. P., and L. S. Havarstein. 2002. ria. Science 301:510–513.
Extracellular-peptide control of competence for 37. Grossman, A. D. 1995. Genetic networks con-
genetic transformation in Streptococcus pneumoniae. trolling the initiation of sporulation and the devel-
Front. Biosci. 7:d1798–d1814. opment of genetic competence in Bacillus subtilis.
24. Comella, N., and A. D. Grossman. 2005. Con- Annu. Rev. Genet. 29:477–508.
servation of genes and processes controlled by the 38. Grossman, A. D., and R. Losick. 1988. Extra-
quorum response in bacteria: characterization of cellular control of spore formation in Bacillus sub-
genes controlled by the quorum sensing transcrip- tilis. Proc. Natl.Acad. Sci. USA 85:4369–4373.
tion factor ComA in Bacillus subtilis. Mol. Microbiol. 39. Hadden, C., and E. W. Nester. 1968. Purifica-
57:1159–1174. tion of competent cells in the Bacillus subtilis trans-
25. Core, L., and M. Perego. 2003.TPR-mediated formation system. J. Bacteriol. 95:876–885.
interaction of RapC with ComA inhibits response 40. Hahn, J., and D. Dubnau. 1991. Growth stage
regulator-DNA binding for competence develop- signal transduction and the requirements for srfA
ment in Bacillus subtilis. Mol. Microbiol. 49:1509– induction in development of competence. J. Bacte-
1522. riol. 173:7275–7282.
26. Doolittle, W. F. 1999. Phylogenetic classification 41. Hahn, J., L. Kong, and D. Dubnau. 1994.The
and the universal tree. Science 284:2124–2129. regulation of competence transcription factor syn-
27. D’Souza, C., M. M. Nakano, D. L. Frisby, and thesis constitutes a critical control point in the reg-
P. Zuber. 1995. Translation of the open reading ulation of competence in Bacillus subtilis.J.Bacteriol.
frame encoded by comS, a gene of the srf operon, is 176:5753–5761.
necessary for the development of genetic compe- 42. Hamoen, L. W., H. Eshuis, J. Jongbloed, G.
tence, but not surfactin biosynthesis, in Bacillus sub- Venema, and D. van Sinderen. 1995. A small
tilis. J. Bacteriol. 177:4144–4148. gene, designated comS, located within the coding
28. D’Souza, C., M. M. Nakano, and P. Zuber. region of the fourth amino acid-activation domain
1994. Identification of comS, a gene of the srfA of srfA, is required for competence development in
operon that regulates the establishment of genetic Bacillus subtilis. Mol. Microbiol. 15:55–63.
competence in Bacillus subtilis. Proc. Natl.Acad. Sci. 43. Hamoen, L. W., D. Kausche, M. A. Marahiel,
USA 91:9397–9401. D. van Sinderen, G. Venema, and P. Serror.
29. Dubnau, D., and C. M. Lovett. 2002.Transfor- 2003.The Bacillus subtilis transition state regulator
mation and recombination, p. 453–472. In A. L. AbrB binds to the -35 promoter region of comK.
Sonenshein, J.A. Hoch, and R. Losick (ed.), Bacil- FEMS Microbiol. Lett. 218:299–304.
44. Hamoen, L. W., A. F. Van Werkhoven, J. J. 57. Kearns, D. B., F. Chu, S. S. Branda, R. Kolter,
Bijlsma, D. Dubnau, and G. Venema. 1998. and R. Losick. 2005. A master regulator for
The competence transcription factor of Bacillus biofilm formation by Bacillus subtilis. Mol. Microbiol.
subtilis recognizes short A/T-rich sequences 55:739–749.
arranged in a unique, flexible pattern along the 58. Kearns, D. B., F. Chu, R. Rudner, and R.
DNA helix. Genes Dev. 12:1539–1550. Losick. 2004.Genes governing swarming in Bacil-
45. Hamoen, L. W., A. F. Van Werkhoven, G. lus subtilis and evidence for a phase variation mech-
Venema, and D. Dubnau. 2000.The pleiotropic anism controlling surface motility. Mol. Microbiol.
response regulator DegU functions as a priming 52:357–369.
protein in competence development in Bacillus 59. Kearns, D. B., and R. Losick. 2003. Swarming
subtilis. Proc. Natl.Acad. Sci. USA 97:9246–9251. motility in undomesticated Bacillus subtilis. Mol.
46. Hamon, M.A., and B.A. Lazazzera. 2001.The Microbiol. 49:581–590.
sporulation transcription factor Spo0A is required 60. Kim, S. B., B. S. Shin, S. K. Choi, C. K. Kim,
for biofilm development in Bacillus subtilis. Mol. and S. H. Park. 2001. Involvement of acetyl
Microbiol. 42:1199–1209. phosphate in the in vivo activation of the response
47. Hamon, M.A., N. R. Stanley, R.A. Britton,A. regulator ComA in Bacillus subtilis. FEMS Micro-
D. Grossman, and B.A. Lazazzera. 2004.Iden- biol. Lett. 195:179–183.
tification of AbrB-regulated genes involved in 61. Kleerebezem, M., and L. E. Quadri. 2001.
biofilm formation by Bacillus subtilis.Mol.Microbiol. Peptide pheromone-dependent regulation of
52:847–860. antimicrobial peptide production in gram-positive
48. Hayashi, K., T. Kensuke, K. Kobayashi, N. bacteria: a case of multicellular behavior. Peptides
Ogasawara, and M. Ogura. 2006.Bacillus subtilis 22:1579–1596.
RghR (YvaN) represses rapG and rapH, which 62. Koetje, E. J., A. Hajdo-Milasinovic, R.
encode inhibitors of expression of the srfA operon. Kiewiet, S. Bron, and H. Tjalsma. 2003. A
Mol. Microbiol. 59:1714–1729. plasmid-borne Rap-Phr system of Bacillus subtilis
49. Higgins, C. F., and K. J. Linton. 2001. Struc- can mediate cell-density controlled production of
tural biology.The xyz of ABC transporters. Science extracellular proteases. Microbiology 149:19–28.
293:1782–1784. 63. Koide, A., and J. A. Hoch. 1994. Identification
50. Hoa, T. T., P. Tortosa, M. Albano, and D. of a second oligopeptide transport system in Bacil-
Dubnau. 2002. Rok (YkuW) regulates genetic lus subtilis and determination of its role in sporula-
competence in Bacillus subtilis by directly repress- tion. Mol. Microbiol. 13:417–426.
ing comK. Mol. Microbiol. 43:15–26. 64. Kunst, F., N. Ogasawara, I. Moszer, A. M.
51. Hoch, J.A. 1993. Regulation of the phosphorelay Albertini, G. Alloni, V. Azevedo, M. G.
and the initiation of sporulation in Bacillus subtilis. Bertero, P. Bessieres, A. Bolotin, S. Borchert,
Annu. Rev. Microbiol. 47:441–465. R. Borriss, L. Boursier,A. Brans, M. Braun, S.
52. Ireton, K., D. Z. Rudner, K. J. Siranosian, and C. Brignell, S. Bron, S. Brouillet, C.V. Bruschi,
A. D. Grossman. 1993. Integration of multiple B. Caldwell, V. Capuano, N. M. Carter, S. K.
developmental signals in Bacillus subtilis through Choi, J. J. Codani, I. F. Connerton, N. J. Cum-
the Spo0A transcription factor. Genes Dev. mings, R.A. Daniel, F. Denizot, K. M. Devine,
7:283–294. A. Dusterhoft, S. D. Ehrlich, P.T. Emmerson,
53. Ishikawa, S., L. Core, and M. Perego. 2002. K. D. Entian, J. Errington, C. Fabret, E.
Biochemical characterization of aspartyl phos- Ferrari, D. Foulger, C. Fritz, M. Fujita, Y.
phate phosphatase interaction with a phosphory- Fujita, S. Fuma,A. Galizzi, N. Galleron, S.-Y.
lated response regulator and its inhibition by a Ghim, P. Glaser,A. Goffeau, E. J. Golightly, G.
pentapeptide. J. Biol. Chem. 277:20483–20489. Grandi, G. Guiseppi, B. J. Guy, K. Haga, J.
54. Jarmer, H., T. S. Larsen, A. Krogh, H. H. Haiech, C. R. Harwood, A. Henaut, H.
Saxild, S. Brunak, and S. Knudsen. 2001. Hilbert, S. Holsappel, S. Hosono, M.-F. Hullo,
Sigma A recognition sites in the Bacillus subtilis M. Itaya, L. Jones, B. Joris, D. Karamata, Y.
genome. Microbiology 147:2417–2424. Kasahara, M. Klaerr-Blanchard, C. Klein, Y.
55. Jiang, M., R. Grau, and M. Perego. 2000. Dif- Kobayashi, P. Koetter, G. Koningstein, S.
ferential processing of propeptide inhibitors of Krogh, M. Kumano, K. Kurita, A. Lapidus, S.
Rap phosphatases in Bacillus subtilis. J. Bacteriol. Lardinois, J. Lauber, B. Lazarevic, S.-M. Lee,
182:303–310. A. Levine, H. Liu, S. Masuda, C. Mauel, C.
56. Joenje, H., M. Gruber, and G.Venema. 1972. Medigue, N. Medina, R. P. Mellado, M.
Stimulation of the development of competence by Mizuno, D. Moestl, S. Nakai, M. Noback, D.
culture fluids in Bacillus subtilis transformation. Noone, M. O’Reilly, K. Ogawa, A. Ogiwara,
Biochim. Biophys.Acta 262:189–199. B. Oudega, S.-H. Park,V. Parro,T. M. Pohl, D.
Portetelle, S. Porwollik, A. M. Prescott, E. sion of the Bacillus subtilis pleiotropic regulatory

Presecan, P. Pujic, B. Purnelle, et al. 1997.The gene degQ. J. Bacteriol. 173:2366–2377.
complete genome sequence of the gram-positive 78. Mueller, J. P., G. Bukusoglu, and A. L. Sonen-
bacterium Bacillus subtilis. Nature 390:249–256. shein. 1992.Transcriptional regulation of Bacillus
65. Lazazzera, B. A., and A. D. Grossman. 1998. subtilis glucose starvation-inducible genes: control
The ins and outs of peptide signaling. Trends Micro- of gsiA by the ComP-ComA signal transduction
biol. 6:288–294. system. J. Bacteriol. 174:4361–4373.
66. Lazazzera, B. A., I. G. Kurtser, R. S. 79. Nakano, M. M., R. Magnuson, A. Myers, J.
McQuade, and A. D. Grossman. 1999. An Curry, A. D. Grossman, and P. Zuber. 1991.
autoregulatory circuit affecting peptide signaling srfA is an operon required for surfactin production,
in Bacillus subtilis. J. Bacteriol. 181:5193–5200. competence development, and efficient sporula-
67. Lazazzera, B. A., T. Palmer, J. Quisel, and A. tion in Bacillus subtilis. J. Bacteriol. 173:1770–1778.
D. Grossman. 1999. Cell density control of gene 80. Nakano, M. M., M.A. Marahiel, and P. Zuber.
expression and development in Bacillus subtilis, p. 1988. Identification of a genetic locus required for
27–46.In G.M.Dunny and S.C.Winans (ed.),Cell- biosynthesis of the lipopeptide antibiotic surfactin
Cell Signaling in Bacteria. ASM Press,Washington, in Bacillus subtilis. J. Bacteriol. 170:5662–5668.
DC. 81. Nakano, M. M., L. Xia, and P. Zuber. 1991.
68. Lazazzera, B. A., J. M. Solomon, and A. D. Transcription initiation region of the srfA operon,
Grossman. 1997. An exported peptide functions which is controlled by the ComP-ComA signal
intracellularly to contribute to cell density signal- transduction system in Bacillus subtilis. J. Bacteriol.
ing in B. subtilis. Cell 89:917–925. 173:5487–5493.
69. Lombardia, E.,A. J. Rovetto,A. L.Arabolaza, 82. Nakano, M. M., and P. Zuber. 1989. Cloning
and R. R. Grau. 2006. A LuxS-dependent cell- and characterization of srfB, a regulatory gene
to-cell language regulates social behavior and involved in surfactin production and competence
development in Bacillus subtilis. J. Bacteriol. in Bacillus subtilis. J. Bacteriol. 171:5347–5353.
188:4442–4452. 83. Nakano, M. M., and P. Zuber. 1993.Mutational
70. Lupp, C., and E. G. Ruby. 2005. Vibrio fischeri analysis of the regulatory region of the srfA operon
uses two quorum-sensing systems for the regula- in Bacillus subtilis. J. Bacteriol. 175:3188–3191.
tion of early and late colonization factors. J. Bacte- 84. Nakano, M. M., and P. Zuber. 1991.The pri-
riol. 187:3620–3629. mary role of ComA in establishment of the com-
71. Maamar, H., and D. Dubnau. 2005. Bistability petent state in Bacillus subtilis is to activate
in the Bacillus subtilis K-state (competence) system expression of srfA. J. Bacteriol. 173:7269–7274.
requires a positive feedback loop. Mol. Microbiol. 85. Novick, R. P. 2003. Autoinduction and signal
56:615–624. transduction in the regulation of staphylococcal
72. Magnuson, R., J. Solomon, and A. D. virulence. Mol. Microbiol. 48:1429–1449.
Grossman. 1994. Biochemical and genetic 86. Ochman, H., J. G. Lawrence, and E.A. Grois-
characterization of a competence pheromone man. 2000. Lateral gene transfer and the nature of
from B. subtilis. Cell 77:207–216. bacterial innovation. Nature 405:299–304.
73. Marahiel, M.A., M. M. Nakano, and P. Zuber. 87. Ogura, M., K. Shimane, K. Asai, N.
1993. Regulation of peptide antibiotic production Ogasawara, and T. Tanaka. 2003. Binding of
in Bacillus. Mol. Microbiol. 7:631–636. response regulator DegU to the aprE promoter
74. Marahiel, M. A., P. Zuber, G. Czekay, and R. is inhibited by RapG, which is counteracted by
Losick. 1987. Identification of the promoter for a extracellular PhrG in Bacillus subtilis. Mol. Microbiol.
peptide antibiotic biosynthesis gene from Bacillus 49:1685–1697.
brevis and its regulation in Bacillus subtilis. J. Bacte- 88. Ogura, M., H.Yamaguchi, K. Kobayashi, N.
riol. 169:2215–2222. Ogasawara, Y. Fujita, and T. Tanaka. 2002.
75. McQuade, R. S., N. Comella, and A. D. Whole-genome analysis of genes regulated by the
Grossman. 2001. Control of a family of phos- Bacillus subtilis competence transcription factor
phatase regulatory genes (phr) by the alternate ComK. J. Bacteriol. 184:2344–2351.
sigma factor sigma-H of Bacillus subtilis. J. Bacteriol. 89. Ohlsen, K. L., J. K. Grimsley, and J. A. Hoch.
183:4905–4909. 1994.Deactivation of the sporulation transcription
76. Molle, V., M. Fujita, S. T. Jensen, P. factor Spo0A by the Spo0E protein phosphatase.
Eichenberger, J. E. Gonzalez-Pastor, J. S. Liu, Proc. Natl.Acad. Sci. USA 91:1756–1760.
and R. Losick. 2003. The Spo0A regulon of 90. Ohsawa, T., K. Tsukahara, T. Sato, and M.
Bacillus subtilis. Mol. Microbiol. 50:1683–1701. Ogura. 2006. Superoxide stress decreases expres-
77. Msadek, T., F. Kunst, A. Klier, and G. sion of srfA through inhibition of transcription of
Rapoport. 1991. DegS-DegU and ComP- the comQXP quorum-sensing locus in Bacillus sub-
ComA modulator-effector pairs control expres- tilis. J. Biochem. (Tokyo) 139:203–211.
91. Paik, S. H.,A. Chakicherla, and J. N. Hansen. permease locus and is required for sporulation
1998. Identification and characterization of the and competence. J. Bacteriol. 173:1388–1398.
structural and transporter genes for, and the 104. Serror, P., and A. L. Sonenshein. 1996. CodY
chemical and biological properties of, sublancin is required for nutritional repression of Bacillus
168, a novel lantibiotic produced by Bacillus sub- subtilis genetic competence. J. Bacteriol. 178:
tilis 168. J. Biol. Chem. 273:23134–23142. 5910–5915.
92. Perego, M. 2001.A new family of aspartyl phos- 105. Siranosian, K. J., and A. D. Grossman. 1994.
phate phosphatases targeting the sporulation Activation of spo0A transcription by sigma-H is
transcription factor Spo0A of Bacillus subtilis.Mol. necessary for sporulation but not for competence
Microbiol. 42:133–143. in Bacillus subtilis. J. Bacteriol. 176:3812–3815.
93. Perego, M. 1997.A peptide export-import con- 106. Smits,W. K., C. C. Eschevins, K.A. Susanna,
trol circuit modulating bacterial development S. Bron, O. P. Kuipers, and L. W. Hamoen.
regulates protein phosphatases of the phosphore- 2005. Stripping Bacillus: ComK auto-stimulation
lay. Proc. Natl.Acad. Sci. USA 94:8612–8617. is responsible for the bistable response in compe-
94. Perego, M., and J. A. Brannigan. 2001. Pen- tence development. Mol. Microbiol. 56:604–614.
tapeptide regulation of aspartyl-phosphate phos- 107. Solomon, J. M., and A. D. Grossman. 1996.
phatases. Peptides 22:1541–1547. Who’s competent and when: regulation of natu-
95. Perego, M., C. Hanstein, K. M. Welsh, T. ral genetic competence in bacteria. Trends Genet.
Djavakhishvili, P. Glaser, and J. A. Hoch. 12:150–155.
1994. Multiple protein-aspartate phosphatases 108. Solomon, J. M., B. A. Lazazzera, and A. D.
provide a mechanism for the integration of Grossman. 1996. Purification and characteriza-
diverse signals in the control of development in tion of an extracellular peptide factor that affects
B. subtilis. Cell 79:1047–1055. two different developmental pathways in Bacillus
96. Perego, M., C. F. Higgins, S. R. Pearce, M. P. subtilis. Genes Dev. 10:2014–2024.
Gallagher, and J. A. Hoch. 1991. The 109. Solomon, J. M., R. Magnuson,A. Srivastava,
oligopeptide transport system of Bacillus subtilis and A. D. Grossman. 1995.Convergent sensing
plays a role in the initiation of sporulation. Mol. pathways mediate response to two extracellular
Microbiol. 5:173–185. competence factors in Bacillus subtilis. Genes Dev.
97. Perego, M., and J. A. Hoch. 1996. Cell-cell 9:547–558.
communication regulates the effects of protein 110. Sonenshein, A. L. 2005. CodY, a global regula-
aspartate phosphatases on the phosphorelay con- tor of stationary phase and virulence in gram-
trolling development in Bacillus subtilis. Proc. Natl. positive bacteria.Curr.Opin.Microbiol.8:203–207.
Acad. Sci. USA 93:1549–1553. 111. Stanley, N. R., and B. A. Lazazzera. 2005.
98. Pesci, E. C., J. P. Pearson, P. C. Seed, and B. Defining the genetic differences between wild
H. Iglewski. 1997. Regulation of las and rhl and domestic strains of Bacillus subtilis that affect
quorum sensing in Pseudomonas aeruginosa. J. Bac- poly-gamma-DL-glutamic acid production and
teriol. 179:3127–3132. biofilm formation.Mol.Microbiol. 57:1143–1158.
99. Piazza, F., P.Tortosa, and D. Dubnau. 1999. 112. Stein, T. 2005. Bacillus subtilis antibiotics: struc-
Mutational analysis and membrane topology of tures, syntheses and specific functions. Mol. Micro-
ComP, a quorum-sensing histidine kinase of biol. 56:845–857.
Bacillus subtilis controlling competence develop- 113. Stephenson, S., C. Mueller, M. Jiang, and M.
ment. J. Bacteriol. 181:4540–4548. Perego. 2003. Molecular analysis of Phr peptide
100. Pottathil, M., and B. A. Lazazzera. 2003.The processing in Bacillus subtilis. J. Bacteriol.
extracellular Phr peptide-Rap phosphatase signal- 185:4861–4871.
ing circuit of Bacillus subtilis.Front.Biosci. 8:32–45. 114. Sturme, M. H., M. Kleerebezem, J.
101. Robertson, J. B., M. Gocht, M. A. Marahiel, Nakayama,A. D.Akkermans, E. E.Vaughan,
and P. Zuber. 1989. AbrB, a regulator of gene and W. M. de Vos. 2002. Cell to cell communi-
expression in Bacillus,interacts with the transcrip- cation by autoinducing peptides in gram-positive
tion initiation regions of a sporulation gene and bacteria. Antonie Leeuwenhoek 81:233–243.
an antibiotic biosynthesis gene. Proc. Natl. Acad. 115. Tam, R., and M. H. Saier, Jr. 1993. Structural,
Sci. USA 86:8457–8461. functional, and evolutionary relationships among
102. Roggiani, M., and D. Dubnau. 1993. ComA, extracellular solute-binding receptors of bacteria.
a phosphorylated response regulator protein of Microbiol. Rev. 57:320–346.
Bacillus subtilis, binds to the promoter region of 116. Tortosa, P., L. Logsdon, B. Kraigher,Y. Itoh,
srfA. J. Bacteriol. 175:3182–3187. I. Mandic-Mulec, and D. Dubnau. 2001.
103. Rudner, D. Z., J. R. LeDeaux, K. Ireton, and Specificity and genetic polymorphism of the
A. D. Grossman. 1991. The spo0K locus of Bacillus competence quorum-sensing system. J. Bac-
Bacillus subtilis is homologous to the oligopeptide teriol. 183:451–460.
117. Tran, L. S., T. Nagai, and Y. Itoh. 2000. Effects of phosphorelay perturbations on archi-
Divergent structure of the ComQXPA quorum- tecture, sporulation, and spore resistance in
sensing components: molecular basis of strain- biofilms of Bacillus subtilis. J. Bacteriol. 188:3099–
specific communication mechanism in Bacillus 3109.
subtilis. Mol. Microbiol. 37:1159–1171. 123. Waldburger, C., D. Gonzalez, and G. H.
118. Turgay, K., J. Hahn, J. Burghoorn, and D. Chambliss. 1993. Characterization of a new
Dubnau. 1998. Competence in Bacillus subtilis is sporulation factor in Bacillus subtilis. J. Bacteriol.
controlled by regulated proteolysis of a transcrip- 175:6321–6327.
tion factor. EMBO J. 17:6730–6738. 124. Weinrauch, Y., N. Guillen, and D. A.
119. Turgay, K., L. W. Hamoen, G.Venema, and Dubnau. 1989. Sequence and transcription
D. Dubnau. 1997. Biochemical characterization mapping of Bacillus subtilis competence genes
of a molecular switch involving the heat shock comB and comA, one of which is related to a
protein ClpC, which controls the activity of family of bacterial regulatory determinants.
ComK, the competence transcription factor of J. Bacteriol. 171:5362–5375.
Bacillus subtilis. Genes Dev. 11:119–128. 125. Weinrauch, Y., T. Msadek, F. Kunst, and D.
120. van Sinderen, D., and G.Venema. 1994. comK Dubnau. 1991. Sequence and properties of
acts as an autoregulatory control switch in the comQ, a new competence regulatory gene of
signal transduction route to competence in Bacil- Bacillus subtilis. J. Bacteriol. 173:5685–5693.
lus subtilis. J. Bacteriol. 176:5762–5770. 126. Winans, S. C., J. Zhu, and M. I. More. 1999.
121. Vasantha, N., and E. Freese. 1979.The role of Cell density-dependent gene expression by
manganese in growth and sporulation of Bacillus Agrobacterium tumefaciens during colonization of
subtilis. J. Gen. Microbiol. 112:329–336. crown gall tumors, p. 117–128. In G. M. Dunny
122. Veening, J. W., O. P. Kuipers, S. Brul, and S. C.Winans (ed.), Cell-Cell Signaling in Bacte-
K. J. Hellingwerf, and R. Kort. 2006. ria. ASM Press,Washington, DC.
NEW INSIGHTS INTO PHEROMONE
CONTROL AND RESPONSE IN
ENTEROCOCCUS FAECALIS pCF10
Heather A. H. Haemig and Gary M. Dunny
3
One of the best-studied forms of microbial research on the pheromone systems during the
behavior controlled by intercellular signaling is past several years,particularly the pCF10 system
pheromone-inducible conjugation in Enterococ- our laboratory studies. We focus on new
cus faecalis.Transfer of conjugative plasmids from insights into the control of pheromone activity
donor cells to recipients is induced by a peptide in donor and recipient cells, on the molecular
signal molecule, whose production is encoded mechanism of the pheromone response, and on
by the chromosome, with the machinery for the evolution of these plasmids.With regard to
detection of the pheromone being carried by the latter topic, comparative sequence analysis
the plasmid. Pheromone-induced conjugation of pheromone plasmids and functional studies
plays a significant role in the dissemination of now suggest that these sensing systems probably
antibiotic resistance and virulence genes in did not evolve originally to detect the presence
E. faecalis and can also effect transfer of genes of conjugative recipients.
into other bacteria, although the pheromone
plasmids themselves do not normally repli- OVERVIEW OF pCF10
cate in nonenterococcal hosts. Enterococcal Pheromone-inducible conjugative plasmids
pheromones are also the first signal molecules have developed an extremely specific and sensi-
of gram-positive bacteria whose complete tive method for dispersal and maintenance
molecular identities were determined (19). In throughout an enterococcal population with-
an earlier publication, Clewell (15) reviewed out causing the unnecessary expenditure of
the enterococcal pheromone systems; detailed metabolic energy by the host cell. Unlike other
descriptions of the discovery of the pheromone forms of cell-cell signaling where one cell type
systems and the first 20 years of research in this can both produce and detect a signal, there are
area can be found in references 9,16, and 20. two distinct populations of cells: donors and
This chapter emphasizes significant results of recipients. Potential recipient cells produce
mating pheromones that induce plasmid
transfer from donor cells. The majority of
Heather A. H. Haemig and Gary M. Dunny Department of
Microbiology, University of Minnesota, Minneapolis, Min- research on these plasmids has focused on
nesota 55455. pAD1 (23, 24, 41) and pCF10 (9, 18, 34).
31
32 ■ HAEMIG AND DUNNY
Considerable work also was done on pPD1 cCF10; pheromone induces pCF10 transfer,
(29, 40, 43) and pAM373 (17, 46); the latter whereas cAD1 induces pAD1. For a more
plasmid/pheromone system has features that detailed listing of various pheromones and their
distinguish it from the others. These plasmids cognate plasmids, see reference 9.
share a similar mechanism of regulation by Figure 1 illustrates the basic mechanism for
utilizing a combination of host and plasmid- peptide-induced conjugative transfer of pCF10.
encoded gene products to sense pheromone, to Recipient cells release the signaling peptide
activate the expression of plasmid transfer into the growth medium where a nearby donor
genes, and to prevent self-induction. cell detects its presence at the cell surface by a
As described in detail later in this chapter, at pCF10-encoded pheromone-binding protein,
least half of the genes encoded on the approxi- PrgZ (48). Homologs of PrgZ in non-pCF10
mately 70-kb pCF10 plasmid are involved in plasmids are referred to as TraC. PrgZ and the
pheromone-inducible conjugation.The expres- TraC proteins share sequence homology and
sion of the proteins involved in plasmid transfer function with the family of peptide-binding
is induced specifically by a small chromosomally OppA proteins found in a wide range of bacte-
encoded hydrophobic peptide of 7 to 8 amino rial species.Despite their overall similarity,PrgZ
acids in length. On the basis of genetic studies and TraC proteins have variability in their
and surveys of clinical isolates, there are proba- amino acid sequences at their C termini; this
bly at least 8 to 10 families of highly related region may contribute to specificity of peptide
pheromone-responsive plasmids; each family binding. PrgZ recruits the chromosomally
determines a mating response to a different encoded oligopeptide permease system to
peptide. Pheromones are designated with a “c” import the pheromone into the cytoplasm of
followed by part of the plasmid name, e.g., the donor cell (37).Upon entry into the cell,the
FIGURE 1 Model for pheromone-inducible conjugative transfer of pCF10. Recipient cell

produces enough cCF10 pheromone to overcome iCF10 inhibitor produced by the donor cell.
Pheromone cCF10 enters the donor cell through PrgZ and interacts with PrgX, leading to
expression of the conjugation machinery including AS for plasmid transfer.
3. PHEROMONE CONTROL AND RESPONSE IN E. FAECALIS pCF10 ■ 33
pheromone interacts with PrgX, a pheromone duced at a level such that the endogenous
receptor protein. The other sex pheromone pheromone activity from the donor cells is
plasmids contain a protein with a similar func- neutralized after receipt of the plasmid. Mea-
tion called TraA, but there are no extended surement of iCF10 levels in donor cultures has
regions of high sequence similarity between been found to be 10 to 100-fold higher than
PrgX and the TraA proteins (24, 25). Recently, the pheromone level, which is sufficient to
PrgX has been identified as the critical molecu- negate any inducing activity by the endogenous
lar switch for the expression of the conjugation cCF10 (28, 42). The details of cCF10 and
machinery in pCF10 (5, 35, 51). Pheromone iCF10 synthesis and control are discussed in the
binding releases PrgX from its repressive state to next section.
allow increased production and extension of
transcripts from the prgQ locus. Gene products PHEROMONE AND INHIBITOR
much further downstream of the prgQ pro- SYNTHESIS AND CONTROL
moter encode positive and negative posttran- As noted in the previous section, the induc-
scriptional regulatory factors and several cell tion state of the prgQ operon encoding the
surface proteins. One of the early pCF10- pCF10 conjugation genes is determined by the
encoded transcripts following induction is for pheromone to inhibitor ratio in the growth
the surface protein Aggregation Substance (AS); medium. In a cell carrying pCF10, the levels of
the AS protein encoded by the pCF10 prgB these two peptides are tightly controlled such
gene is Asc10 (32).The expression of AS on the that the transfer genes are not expressed in the
donor cell surface results in tight physical con- absence of recipients, but the donor cell can
tact between the donor cell and its recipient cell detect a shift in this balance with exquisite sen-
to facilitate plasmid transfer and is responsible sitivity.This section describes various aspects of
for a clumping phenotype observed in induced cCF10 and iCF10 synthesis, as well as the role
cell cultures. Specific outcomes of pheromone of PrgY in control of production of these
induction are discussed later this chapter. peptides and the proposed mechanism by
Upon acquisition of plasmid, a newly cre- which PrgY achieves control. Here, we
ated donor cell can still produce cCF10 since describe the synthesis of cCF10 to exemplify a
the gene for cCF10 is chromosomally encoded. process that occurs in a similar fashion for sev-
This creates a concern for donor cells as their eral pheromones. As illustrated in Fig. 2A, the
own endogenous pheromone could potentially mature cCF10 sequence is contained within a
induce unnecessary conjugation.To enable the segment of the polypeptide product of the ccfA
donor cells to remain in an uninduced state gene. This gene is predicted to code for a
ready to respond to pheromone from a recipi- secreted lipoprotein; the pheromone peptide is
ent cell, an inhibitor peptide (iCF10) is pro- a proteolytic processing product of the cleaved
duced from the prgQ operon of pCF10 (28).At signal peptide of CcfA.The actual function of
the simplest level of regulation of pCF10, the mature CcfA is unknown, and we have not
levels of iCF10 and cCF10 in the surrounding observed any significant phenotypic effects
supernatant appear to be carefully balanced to (other than a lack of pheromone production) of
block self-induction while still allowing a sensi- a mutation abolishing expression of this protein
tive response to an increase in pheromone con- (11). CcfA shows amino acid sequence similar-
centration resulting from production by a ity to the essential YidC protein of Escherichia
nearby recipient cell. In the case of the pCF10 coli, which plays a role in protein export, so
plasmid, pure cultures of plasmid-containing CcfA might function similarly even though it is
donor cells and plasmid-free recipient cells not essential for E. faecalis growth.
contain pheromone in the supernatant, sug- An intensive genetic screen for chromoso-
gesting that the inhibitor peptide must be pro- mal mutations reducing pheromone produc-
FIGURE 2 A Proteolytic processing of cCF10 and iCF10. cCF10 is processed from CcfA in sev-
eral steps. SPII, signal peptidase II; Eep, membrane protease; CP, carboxy peptidase.Arrows indicate
cleavage sites. iCF10 is processed from a peptide produced from the 5′ end of the prgQ operon that is
similar to the signal sequence of CcfA. B. Model for PrgY.The CcfA lipoprotein signal sequence is
released from full-length polypeptide by SPII and further processed by Eep to generate cCF10.The
primary functional activity of PrgY is to bind and sequester, degrade, or modify newly produced
cCF10 as it is exported from the membrane following processing by Eep. In this model it is depicted
that an interaction with Eep is mediated by the C-terminal intramembrane segments of PrgY.This
would position PrgY optimally to intercept any newly released cCF10 from Eep. “N” depicts the
amino terminus of the full-length CcfA polypeptide.The large extracellular domain of PrgY corre-
sponds to the amino-terminal portion of the protein.
tion by An and Clewell did not reveal the from bacteria to humans, and where the
identity of a structural gene for pheromone active site for proteolysis is predicted to be
synthesis, but instead identified the eep in an intramembrane domain (7). It was
(enhanced expression of pheromone) gene, demonstrated that eep mutants secreted reduced
which appears to play an important role in pro- levels of several pheromones but produced a
duction of several pheromones (1). Interest- normal amount of cAM373 (1). Virtually no
ingly, the Eep protein belongs to the RIP biochemical analysis of Eep has been reported,
(regulated intramembrane proteolysis) protein but recent genetic studies noted below shed
superfamily, whose members are widespread, some light on the specificity determinants for
recognition and processing of pheromone the pPD1 and pAD1 pheromone plasmids.
precursors by Eep. Recent analysis of cCF10 With regard to the phylogenetic distribution
expression from chimeric genes where the of these proteins, they are found in a variety
mature pheromone coding sequence was fused of bacteria, but not in any “low G-C” gram-
to the adjacent upstream fragments of the signal positive organisms closely related to E. faecalis,
peptide gene segments of other pheromones including those known to employ peptides as
( 9a) suggests that the sequence specificity intercellular signal molecules. In E. faecalis,
determinants for recognition and cleavage of these proteins are encoded by pheromone-
cCF10 precursors by Eep are within the amino responsive plasmids but not by the chromo-
acid residues in the N-terminal “upstream” some. Mutagenesis and screening for amino
region rather than in the cCF10 sequence itself. acid substitutions in PrgY that gave rise to
However, in examining the corresponding stable proteins that failed to reduce pheromone
sequences from lipoprotein genes that give rise production identified several essential residues
to all known Eep-dependent pheromones, in the putative N-terminal extracellular
there are no obvious consensus sequences for domain that are highly conserved across the
Eep recognition that we have been able to superfamily (10).We think that this is a possible
identify thus far. indication of a conserved biochemical activity
Two different pCF10-encoded gene prod- involved in control of heretofore unrecognized
ucts are required for plasmid carrying donor peptide-mediated signaling circuits in many
cells to avoid self-induction of transfer genes by diverse organisms.
their endogenously produced pheromone. One Although expression of PrgY in an E. faecalis
of these genes is prgY (the corresponding genes strain reduces the level of pheromone produced
of pAD1 and pPD1 are designated traB), pre- by that strain, and the functional domain
dicted to encode a membrane-associated pro- appears to be outside the membrane, PrgY has
tein. Null mutations of prgY in the context of no detectable effect on pheromone added
pCF10 cause constitutive expression of transfer exogenously to cells. Expression of PrgY does
genes whereas expression of a cloned prgY gene not increase the ability of an E. faecalis cell to
in an E. faecalis host lacking pCF10 reduces the bind or inactivate exogenously supplied cCF10
level of pheromone produced to a level compa- in the growth medium, nor does expression of
rable to that produced by cells carrying pCF10 PrgY decrease the mating response of a pCF10-
(10). The pPD1 traB encodes a protein with containing cell to exogenous cCF10. We have
77% identity to PrgY, and this gene partially recently used engineered chimeric cCF10- and
complements a prgY null mutation, whereas cCF10-variant-encoding genes to examine the
pAD1 TraB is only 44% identical to PrgY and amino acid specificity determinants for PrgY
there is no heterologous complementation in control ( 9a). In contrast to the sequence deter-
this case (10). On the basis of sequence analysis minants described above for Eep recognition,
of PrgY and experimental analysis (8, 10, 48), our data suggest that PrgY specifically recog-
the likely topology of the protein is as depicted nizes the mature heptapeptide and that the L4
in Fig. 2B, with a large N-terminal domain residue is of particular importance in this
outside the membrane and anchored by multi- recognition. We have used the cumulative
ple membrane-spanning segments in the C- results of genetic and physiological studies
terminal portion of the protein. It is interesting to develop a model for PrgY activity (Fig. 2B).
that proteins showing significant amino acid We speculate that PrgY either degrades or
sequence similarity to PrgY are encoded by modifies cCF10 as it is released from the
organisms from all three branches of the tree of cell membrane following Eep-mediated prote-
life (10), even though the only other members olytic processing;we cannot rule out alternative
of this superfamily for which any functional models involving nonenzymatic sequestration
data are available are the traB homologs from of cCF10, but we find such models less
attractive. The C terminus may provide a exogenous pheromone production likely is

strong anchor to lock the topology of the pro- important in returning the donor cell to
tein such that the active site is poised immedi- the uninduced state following conjugation.The
ately adjacent to the outer surface of the 22-amino-acid polypeptide encoded by the
membrane where it can capture nascent cCF10 prgQ Orf resembles a signal peptide without an
molecules as they emerge from the membrane. attached secreted protein, and the carboxy-
Our analysis indicates that neither endogenous terminal residues of this peptide correspond
cCF10 that escapes interaction with the to mature iCF10 (this is also the case for the
membrane-bound PrgY during its release from inhibitor peptides of other pheromone plas-
the membrane nor exogenously supplied mids). As illustrated in Fig. 2A, conversion of
cCF10 is subject to the negative effects of pro-iCF10 to iCF10 is similar to production of
PrgY. Although biochemical and structural cCF10 from the ccfA; we recently found that
analysis of membrane proteins can be chal- Eep is also involved in iCF10 processing ( 9a).
lenging,it will be of great importance to pursue Since iCF10 needs to be processed, exported,
such studies in the future to test this model and reimported in order to function, the loca-
for PrgY function. tion of its coding sequence at the beginning of
PrgY reduces but does not abolish the induced transcript provides a built-in
pheromone production by donor cells. It is timing device to shut down expression of the
necessary but not sufficient for the cell to avoid operon quickly once the conjugation machin-
self-induction by endogenous pheromone. To ery has been produced. Interestingly, compara-
control the residual endogenous pheromone tive studies of prgQ expression using lacZ
that escapes PrgY, pCF10 encodes production transcriptional fusions in either wild-type or
of the heptapeptide iCF10, which acts as a ccfA mutant strains suggest that even though
competitive inhibitor of cCF10 (42). It is possi- endogenous pheromone produced by wild-
ble that iCF10 could interfere with either the type donor cells is not sufficient to induce a full
import of cCF10 or with its interaction with mating response, it does increase prgQ tran-
the cytoplasmic pheromone receptor PrgX. As scription significantly; in prgY mutants, prgQ
is described in detail in the next section, our transcription is further increased,but the result-
recent studies suggest that PrgX is the most ing levels of iCF10 are apparently insufficient to
important target for competition between control expression, since the cultures constitu-
these two peptides. It is significant that iCF10 is tively express all the transfer genes (10). In a
encoded by the prgQ locus, and the Orf for pure culture of wild-type donor cells, the low
iCF10 is a 66-bp gene that is the first polypep- level of endogenous pheromone apparently
tide coding gene in the pheromone-inducible serves as a “fine-tuning” device to induce suffi-
prgQ transfer operon. The details of the tran- cient levels of iCF10 production to maintain
scriptional regulation of this operon are pre- the proper ratio of iCF10 or cCF10 in the cul-
sented below, but for purposes of the present ture medium.
discussion it is important to note that the 5′
~400 nt of prgQ (including the iCF10 coding THE INDUCTION MECHANISM
region) is transcribed at significant levels con- The simplest model for pheromone-induced
stitutively and that pheromone induction plasmid transfer in E. faecalis could have
causes an increase in transcription of this 5′ included an uncomplicated mechanism by
region, as well as readthrough of transcription which a single signal molecule (pheromone)
into downstream genes. The constitutive interacted with a donor cell to result in a mat-
expression of prgQ is probably important in ing response.However,this is most certainly not
maintaining sufficient levels of iCF10 to neu- the case.As described in the previous sections of
tralize endogenous cCF10 activity, and the this chapter, multiple regulatory steps occur at
increased expression of inhibitor resulting from the extracellular level to enable a sensitive
FIGURE 3 Detailed map of the prgX-prgQ region.PQ and PQa are indicated by flags.Transcription
from PQ results in two RNA products dependent on the induction state of the cell.Qs RNA encodes
iCF10 and QL encodes a longer RNA important for readthrough to prgB. RNA transcribed from
PQa is processed to release Qa RNA, one negative regulator, and prgX mRNA, which is translated to
produce PrgX, the other negative regulatory molecule. Two binding sites for PrgX exist on the
DNA, indicated by triple lines.The lollipop structures represent stem-loop structures formed in the
message, with those downstream of PQ named IRS1 and IRS2.
response to pheromone without undergoing Transcription of prgQ in

wasteful self-induction. In addition, work by Induced/Uninduced Cells
our laboratory has resulted in exciting new As previously described in this chapter,
information regarding the regulation of pCF10 pheromone induction of the donor cells by a
conjugation at the intracellular level. PrgX, a nearby recipient cell leads to the production of
37-kDa cytoplasmic protein encoded by transcripts from the prgQ operon (outlined in
pCF10,has long been known to be required for Fig. 3) in pCF10.The prgX gene was thought to
negative regulation of pCF10 conjugation (26), encode a cytoplasmic negative regulatory pro-
but until recently, the precise mechanism by tein of prgQ as deletion of this gene led to con-
which PrgX carries out this role has been stitutive expression of Asc10 (26). In other
largely speculated on the basis of genetic and pheromone plasmid systems, pAD1 and pPD1,
biochemical evidence. A second function, PrgX homologs (TraA) were also shown to be
positive regulation of its own expression, has negative regulators (54, 55).The location of the
been proposed for PrgX, and an additional prgX promoter was unclear until further study
negative regulator of the prgQ operon, Qa of the prgQ locus revealed that mRNA was also
RNA, has been discovered. Over the past few transcribed in the antisense direction from a
years, genetic and biochemical experiments promoter (PQa) within prgQ (2). A 1.4-kb
from our laboratory and others provide mRNA was confirmed to be an unprocessed
evidence that PrgX is the key internal switch mRNA from PQa by detection with an anti-
for pheromone-induced pCF10 transfer. sense prgQ probe and a prgX probe.Transcrip-
Recently, the acquisition of crystal structures tion from PQa through an IRS sequence at
of apo-PrgX and PrgX complexed with cCF10 the end of prgX was demonstrated by reverse
or iCF10 provides a molecular basis for this transcription-PCR experiments. The absence
switching activity to explain how a donor of prgX expression from prgQ-deletion plasmids
cell can respond quickly and specifically to demonstrated that PQa is the major promoter
slight changes in external pheromone concen- for prgX expression. In addition, a smaller (102
trations (35, 51).The next two sections review nt) and more abundant RNA was identified
and link the genetic experiments, biochemi- as a processed product of the 1.4-kb mRNA
cal experiments, and structural data from our from PQa and was named Qa RNA. Qa RNA
laboratory and others to propose a single was demonstrated to be a second negative regu-
molecular mechanism by which the two dis- lator of the prgQ operon. A plasmid expressing
tinct functions of PrgX are carried out in Qa RNA in the absence of PrgX was intro-
regulation of pCF10 transfer. duced into E. faecalis OG1RF along with a
spontaneous pCF10 mutation that shows region between prgX and prgQ retained
constitutive expression of Asc10.The presence radioactivity on the filter (4). Subsequent DNA
of wild-type Qa abolished the constitutive footprinting experiments on this region of
expression of Asc10, demonstrating that Qa pCF10 DNA confirmed two binding sites for
can block readthrough of prgQ mRNA past PrgX: a palindromic binding site at 108/86
IRS1 independently of PrgX and in a cCF10- and a half-palindromic sequence found at the
insensitive fashion (5). 26/14 region upstream of the prgQ tran-
As mentioned earlier in this chapter, PrgX scription initiation site (1). Electrophoretic
does not completely repress transcription since mobility shift assay experiments demonstrated
a 430-nt RNA (Qs), which encodes the that PrgX bound each site individually but with
inhibitor peptide, is constitutively expressed a 60-fold higher affinity for the palindromic
from prgQ in the absence of cCF10. Induction sequence, leading to its designation as the pri-
by cCF10 results in an increase in the transcrip- mary binding site and the half-palindromic
tion from prgQ, leading to increased levels of sequence as the secondary binding site of PrgX.
transcriptional readthrough past the IRS1
sequence of prgQ. Readthrough results in PrgX Is a Positive Autoregulator
increased levels of a 530-nt RNA (QL) thought Analysis of PrgX protein levels revealed that the
to interact with ribosomes to enhance transla- intergenic region of prgX/prgQ is required for
tion of Asc10 (6, 14). Interestingly, cCF10 its expression and that PrgX needs to bind both
induction does not appear to completely abol- binding sites on pCF10 to support its own
ish gene products expressed from the Qa pro- expression (4). Mutations in the secondary
moter (5). The levels of prgX mRNA were binding site that abolished PrgX binding also
transiently reduced but not eliminated even abolished expression of prgX; mutations to the
after 40 min of cCF10 induction. The longer primary binding site, such that PrgX could still
1.4-kb unprocessed transcript from PQa disap- bind both sites, did not completely abolish
peared quickly upon cCF10 induction, while expression of PrgX. In contrast to prgX, Qa
levels of Qa RNA decreased only slightly.The RNA was still produced when the secondary
effect of pheromone induction on PrgX pro- PrgX DNA-binding site was mutated,although
tein was much less pronounced. Collectively, at significantly reduced levels. A series of
the data argue against the idea that pheromone defined deletions between PQa and prgX indi-
induction simply increases transcription from cated that transcription from PQa occurred nor-
the prgQ promoter while shutting off transcrip- mally in the absence of PrgX, but a sequence in
tion at the Qa promoter.We favor an alternative the intergenic region of prgX/prgQ was
model where pheromone induction affects the observed to be critical for the maintenance of
function of the negative regulator PrgX as dis- Qa and prgX mRNA levels (2, 4). When this
cussed later in this chapter. sequence was present in the mRNA tran-
scribed from PQa in the absence of PrgX, there
PrgX Is a DNA-Binding Protein was a dramatic decrease in transcript levels. In
The appearance of target sites for PrgX binding the presence of PrgX, the processing of these
within the prgQ locus and a helix-turn-helix transcripts was altered such that prgX and Qa
DNA-binding motif in PrgX suggested the RNA levels in the cell returned to normal.This
PrgX functions as a negative regulator by suggested that while PrgX does not act on its
blocking transcription from prgQ in the unin- own promoter, its presence is required for nor-
duced state.The initial evidence for PrgX bind- mal levels of products transcribed from PQa.
ing DNA was obtained by mixing 32P-labeled Collectively, these experiments were the basis
pCF10 fragments with PrgX-expressing cell of the idea that in addition to being a negative
lysate and filtering it through nitrocellulose; regulator of prgQ, PrgX also has a positive
only those fragments containing the intergenic autoregulatory function (2, 4).
To determine the functional domains of second strain in which prgX was cloned into the
PrgX, a genetic screen was devised to isolate chromosome under a constitutive promoter
dominant-negative mutants of PrgX (3). Most was tested for its ability to complement the
of the mutations isolated mapped to the C- p043lacZdX plasmid phenotype. Expression
terminal or N-terminal portions of the protein, of functional PrgX protein in trans fully
suggesting that both portions of PrgX are restored the repression of prgQ in the absence of
important for the repression of prgQ transcrip- exogenously added peptides, induction with
tion. To test the mutations’ effects on PrgX exogenous cCF10, and iCF10 inhibition of
autoregulation, the mutant plasmids were elec- cCF10 induction (22). A direct interaction
troporated into OG1RF and checked for prgX between PrgX and cCF10 was confirmed in
expression. None of the mutant forms of PrgX affinity chromatography experiments (35).
were expressed unless pCF10 was also present PrgX was specifically retained on the column
to provide wild-type PrgX to activate the containing cCF10. PrgX did bind to the other
mutant PrgX expression in trans.The dominant peptides tested (cAD1, cPD1, and iCF10), but
negative phenotype of the mutants obtained in much less protein was retained, suggesting a
this screen suggested that PrgX is an oligomeric lower affinity for these other peptides.
protein and, in fact, a one-hybrid genetic assay
using the phage lambda cI repressor fusion sys- Negative Regulation and Positive
tem, copurification of His-PrgX and PrgX, and Autoregulation Are Two Separate
in vivo cross-linking experiments all provided Functions of PrgX
strong evidence for PrgX oligomerization in The most recent genetic analysis of PrgX
both autoregulation and negative regulation of focused on mutants that retained sufficient pos-
prgQ transcription (3, 30, 33). itive autoregulatory function to produce nor-
mal levels of protein in E. faecalis but were
PrgX Is the Intracellular Target deficient in prgQ repression (33).These mutants
of cCF10 and iCF10 provide key evidence that the two functions of
In the pAD1 and pPD1 systems the PrgX PrgX (positive autoregulation and negative reg-
homologs (TraA) were shown to bind the cor- ulation) are separate; previously, mutations
responding peptide pheromones cAD1 and affecting regulation of conjugation had also
cPD1 (25, 43). Data from our laboratory sug- disrupted autoregulation. Three phenotypic
gested that it is not the expression but the func- classes of mutants were identified, and all of the
tion of PrgX that is regulated by cCF10 as levels mutants were partially defective in DNA bind-
of PrgX in the cell were maintained after ing and/or oligomerization, greatly impacting
cCF10 induction (5). We recently reported the negative regulation of PrgX while still
genetic evidence that demonstrates PrgX is the enabling PrgX to perform its autoregulatory
intracellular target of cCF10 and iCF10 (22, function.The results from cross-linking studies
35).In a strain carrying a plasmid (p043lacZdX) and repressor fusions indicate that high-affinity
containing a prgX deletion (removal of 67% dimerization is required for pheromone sensi-
of the coding sequence from the 3′ end) and a tivity and response, but the presence of several
lacZ reporter fused to the iCF10 stop codon mutants with no dimerization deficiency sug-
showed high-level constitutive expression of - gests that dimerization alone is not sufficient
galactosidase and was unresponsive to exoge- for full repression by PrgX.When expressed in
nously added cCF10 or iCF10, indicating E. faecalis, mutants still maintained low levels of
that PrgX was necessary for a response dimerization, suggesting that dimerization may
to pheromone or inhibitor. In contrast, - be important in autoregulation in addition
galactosidase activity in a similar reporter con- to DNA binding being required. Previous
struct (carrying full-length prgX) is induced in dominant-negative mutants revealed forms of
the presence of exogenously added cCF10. A PrgX that could bind DNA but are unable to
autoregulate,implying that DNA binding alone two sets of homodimers.The homodimer con-
is not sufficient for autoregulation (3). The sists of two side-by-side monomers with the N-
genetic screens and biochemical evidence terminal domain swapped (as depicted in Fig.
demonstrating two functions for PrgX resulted 4). The monomers are arranged in an “open
in the development of a model of how PrgX state”; the domain swapping suggests that the
negatively regulates from the prgQ promoter as smallest functional unit of PrgX is a dimer.This
well as positively regulates its own expression is consistent with the observation of an 11-bp
by a single mechanism. The key features and palindromic sequence at the primary binding
supporting evidence for this model are site.PrgX/iCF10 crystal structures show iCF10
described in the next section. does bind to PrgX and the helix 16-loop-helix-
17 segment forms a triangle with the loop
THE C-TERMINAL ARM OF functioning as the tetramer interface (35).
PrgX IS THE PHEROMONE- Tetramerization of PrgX may mediate
RESPONSIVE SWITCH cooperativity of DNA binding observed by
We have just presented the key genetic and bio- PrgX in vitro (4).The binding of PrgX to the
chemical evidence supporting two roles for primary binding site and formation of a PrgX
PrgX. A model that accounts for both regula- tetramer may increase the affinity for binding to
tory functions of PrgX via a single mechanism the secondary binding site. The 70-nt region
is shown in Fig. 4. Briefly, in the uninduced between the two PrgX binding sites is pre-
state, PrgX forms a tetramer that is stabilized by dicted to form a loop in the model we have
iCF10 binding as well as by pCF10 DNA loop- proposed.A case for the formation of a loop in
ing. Here, PrgX functions as a weak repressor conjunction with PrgX tetramerization can be
of prgQ such that iCF10 is constitutively made by the fact that the distance between the
expressed. The combination of DNA looping two binding sites would place bound PrgX
and PrgX binding may cause stalling of RNA molecules on the same face of the DNA helix,
polymerase to promote folding of nascent and also by the presence of a T8A9 sequence
mRNA from PQa so that it is processed into Qa between the two binding sites (4).This AT-rich
RNA and prgX transcript. Binding of cCF10 to region could facilitate the extensive bending
PrgX disrupts the C-terminal arm of PrgX that required to form this small loop. Early DNA
mediates tetramerization, resulting in a PrgX footprinting experiments suggested that several
dimer bound loosely at each DNA-binding nucleotides are hypersensitive to DNaseI
site.As shown in Fig. 5, Qa RNA interacts with attack, supporting the idea of conformational
Qs RNA, preventing its readthrough at IRS1. distortion of the prgQ promoter (4). In addi-
Upon induction, the interaction of PrgX with tion, recent genetic analysis of constructs alter-
the DNA is weak enough to allow transcription ing the spacing and positioning of the two
to occur at prgQ at higher levels such that the binding sites on the face of the DNA supports a
levels of Qs become great enough to titrate Qa, looping model (B. K. Kozlowicz and G. M.
allowing transcription to read through past Dunny, unpublished data).The small size of this
IRS1 for transcript elongation to QL. Addi- loop and the solubility of the PrgX have made it
tional supporting evidence for this model is dis- difficult to obtain concrete evidence for DNA
cussed below. looping using traditional experiments at this
The X-ray crystal structure of PrgX point. Work is under way to develop experi-
is mainly helical with 17 alpha-helices (51). ments that will enable us to visualize DNA
Three domains are clearly identified in each looping in solution.
PrgX monomer: an N-terminal DNA- Biochemical experiments previously sug-
binding domain, a large central dimerization/ gested that the oligomerization of PrgX was
pheromone-binding domain, and a C-terminal responsive to cCF10 so we investigated whether
regulatory domain. The oligomeric state of it was also responsive to iCF10 in light of this
PrgX in the crystals is a tetramer formed from new information (35). In vivo cross-linking of
FIGURE 4 Structural consequences of inhibitor and pheromone binding on PrgX

and the DNA looping model. (Top) Two PrgX molecules interlock via domain swap-
ping of the N-terminal DNA-binding domain to form an X-shaped dimer. Dimers
of PrgX complexed with iCF10 (black crosses) form a tetramer through interactions
between the C-termini arms (black curls). PrgX tetramer increases affinity of each
dimer for the DNA and is further stabilized by DNA looping. (Bottom) cCF10 occu-
pation of the binding site causes the C-terminal arm (black line) to change structure
and rotate to weaken interaction between dimers.The DNA unloops and the affinity
of PrgX for the DNA decreases, allowing increased polymerase access to the prgQ
promoter.
PrgX and copurification of PrgX and His-PrgX oligomerization alone is not sufficient for full
were both reduced by the addition of cCF10. repression of prgQ by PrgX. To determine if
The addition of iCF10 restored cross-linking iCF10 could rescue prgX regulatory functions
observed in the absence of cCF10 and a higher of these mutants, iCF10 was added to the
concentration of iCF10 compared to cCF10 culture medium and the mutants were assayed
directly inhibited cCF10-induced reduction of for lacZ expression (35). All mutant strains
PrgX cross-linking. In another study, random responded to iCF10 by repressing lacZ expres-
mutations found in the dimerization domain of sion to various degrees. Taken together, the
prgX resulted in a derepressed phenotype, but results indicate that iCF10 acts as a corepressor
dimerization was only reduced, not abolished, molecule when bound to PrgX to stabilize the
when the mutants were checked for dimeriza- carboxy-terminal domain of PrgX that forms
tion in E. faecalis cells (3). This suggested that the tetramer interface.
FIGURE 5 Flow diagram indicating the state of prgQ RNAs in the uninduced and induced cell.
In the uninduced state, transcripts from the PQa promoter are processed into prgX mRNA and Qa
RNA. Qa is a 102-nt RNA processed from the 5′ end and is complementary to the 3′ end of Qs
RNA, an RNA constitutively produced at the prgQ promoter. It is predicted that interaction
between Qa and Qs causes a folding such that Qs elongation terminates at IRS1, preventing
readthrough to QL. Upon induction, the cellular levels of Qs greatly increase beyond those of Qa,
resulting in unpaired Qs RNA. The unpaired Qs RNA does not terminate at IRS1 and is
extended into QL and beyond, activating later steps in the induction process.
It has long been known that cCF10 was the cCF10 causes a dissociation of the tetramer to
key molecule for inducing plasmid transfer. reduce affinity of PrgX toward the secondary
Recent structural data now provide insight as to binding site, thereby increasing initiation of
how cCF10 induces the expression of the plas- transcription at prgQ by RNA polymerase.This
mid transfer machinery. Comparison of pep- was confirmed by the observation that, in the
tide/PrgX crystal structures shows that cCF10 absence of the last 26 C-terminal residues of
and iCF10 bind PrgX in the same pocket PrgX, prgQ is constitutively expressed (3).The
formed by several parallel and antiparallel C-terminal deleted PrgX can still bind iCF10
helices; however, residues interacting with each and cCF10 so the only plausible explanation is
peptide differ (35, 51). Residues that are clearly that tetramer formation is necessary for the
present in the iCF10/PrgX structure become weak repression of prgQ.
disordered and are no longer seen in the crystal Initially,PrgX was thought to be required for
structure of cCF10/PrgX. Residues 283–292 the stabilization of Qa RNA due to the obser-
(loop) and 292–305 (helix), thought to stabilize vation that levels of Qa RNA were reduced
PrgX tetramers in the presence of iCF10, pull upon deletion of prgX, yet introducing prgX
away from the interface, and residues 295–306 expressed from pCF10 restored Qa RNA back
refold into a -duplex that covers cCF10 in the to wild-type levels. Attempts to observe bind-
binding pocket. This moves the C-terminal ing or interaction between PrgX and Qa
loop in a way that is thought to reduce the RNA were unsuccessful. A genetic screen for
tetrameric conformation into two dimers, cis-acting prgQ mutants provided concrete evi-
thereby unlooping the DNA. The binding of dence that PrgX and Qa act independently of
each other as negative regulators (5). Mutations We obtained evidence suggesting the presence
to the portion of prgQ encoding Qa RNA of a sequence in the prgQ promoter region that
maintained wild-type levels of PrgX despite leads to rapid degradation of Qa transcripts in
having drastically reduced Qa RNA levels. the absence of PrgX (4). Qa RNA levels pro-
Since no mutations were found in the binding duced from various prgQ plasmid constructs
sites for PrgX in this genetic screen,we hypoth- were measured in vivo.When only the second-
esized that some of the mutations within prgQ ary binding site and the prgQ promoter were
may instead affect an interaction between Qs present, Qa RNA levels were undetectable and
RNA and Qa RNA.S1 mapping defined the 3′ only barely increased slightly when the AT-rich
end of Qa RNA as 102 nt 3′ from the Qa initi- region was present, suggesting that the prgQ
ation sequence such that Qa is complementary promoter region encodes a sequence that favors
to a portion of Qs downstream from the iCF10 degradation of Qa RNA. PrgX binding to its
coding sequence (5).We propose that Qa func- primary binding site on the DNA alters this
tions by interacting with Qs RNA to prevent degradation as the level of Qa RNA increased
transcriptional readthrough at IRS1. Further- significantly upon the addition of the primary
more, in deletion constructs where Qa was binding site and PrgX protein. Future experi-
expressed in the absence of PrgX, Qa was able ments to elucidate the mechanism of autoregu-
to block production of prgQ transcripts extend- lation are of great interest as PrgX is possibly
ing past IRS1, abolishing QL RNA and greatly the first example of a repressor that simul-
reducing Qs RNA.The presence of Qa RNA taneously controls transcripts of its target
allows PrgX to act as a weak repressor, prevent- DNA loop by acting on one promoter and
ing expression of genes required for plasmid participating in posttranscriptional events in
transfer but still allowing iCF10 to be produced the opposite direction.
from prgQ while leaving the system ready to
respond to slight changes in the peptide balance PHEROMONE-INDUCIBLE
caused by the presence of nearby recipient cells. BIOLOGICAL FUNCTIONS OF pCF10
The mechanism of positive autoregulation Two important biological activities are con-
by PrgX is the least understood by our labora- trolled by the pheromone-sensing systems of
tory; however, we can offer some speculation of pCF10 and similar plasmids: (i) conjugative
how this fits into the proposed model. The transfer ability and (ii) virulence in opportunis-
genetic screen that provided proof for the two tic infections. Cumulatively, much more
separate functions for PrgX suggests that oligo- research has been done on the sensing systems
merization of PrgX may be important for than on the functions they control, but a sub-
autoregulation (33). If oligomerization is stantial quantity of new information about
required for autoregulation, it would have been these functions has been obtained in the past
impossible to isolate mutants where oligomer- few years, and these results have also shed light
ization was completely abolished.In this screen, on how these plasmids may have evolved, as
all of the mutants maintained the autoregula- described in a subsequent section.
tory function and were able to dimerize. The
looped complex formed by oligomerization of The pCF10 Conjugation Machinery
PrgX may act as a roadblock for transcription The vast majority of conjugative DNA transfer
from the Qa promoter by stalling RNA poly- systems studied to date share three critical com-
merase molecules.The blocking or delaying of ponents, including a mating pair formation
transcription past the secondary PrgX binding (Mpf) apparatus connecting donor and recipient
site could allow the folding of the 5′ end of the cells, and through which the transferred DNA
transcript into a stable structure (Qa RNA) passes. The other critical components include
before synthesis of a sequence that functions to the DNA processing machinery, a group of pro-
accelerate 3′ to 5′ decay of the entire message. teins that recognize a specific target site called
the origin of transfer (oriT), producing a single- case of pCF10 (Fig. 6) the Mpf genes likely
stranded nick that initiates the unwinding of the comprise a “gram-positive” version of a type
DNA strand that is transferred to the recipient IV secretion system (60) and include the prgD-
cell via the Mpf apparatus. It has been shown pcfD region (Fig. 6, Module B), and the DNA
that pheromone induction of Asc10 expression processing machinery is apparently encoded
contributes to efficient transfer in liquid matings by the pcfE-pcfI region (Fig. 6, Module C).The
via binding to enterococcal binding substance, a single functional oriT target of the pCF10
multicomponent receptor in the cell wall that pheromone-inducible conjugation system is
includes lipoteichoic acid as a major component contained within a short, noncoding region of
(21, 56); however, aggregation itself is not suffi- the plasmid that is between pcfE and pcfF. The
cient for transfer and Asc10 is not an integral two essential pCF10 DNA processing proteins
component of the intercellular conduit for include PcfG, the relaxase enzyme that catalyzes
DNA transfer, the Mpf apparatus (45). In the the nicking reaction, and PcfF, an accessory pro-
FIGURE 6 Map of pCF10. Each predicted Orf of pCF10 derived from its complete sequence is indicated by a
filled arrow (27). Not shown is an approximately 16-kb segment consisting of the tetracycline-resistance element
Tn925, which is not involved in pheromone-inducible conjugation.The evidence for independent evolution of the
three modules indicated on the map is described in the text and in reference 27.
tein that binds to a double-stranded target in the ers (13) have shown that Asc10 and other AS
oriT and appears to function both in the recruit- proteins increase the severity of experimental
ment of PcfG via protein-protein interactions endocarditis; the primary effects are increased
and in melting the DNA helix to facilitate bind- weight and bacterial counts of endocardiac
ing and nicking of the single-stranded substrate vegetations. Other AS-associated properties
of PcfG (12).Both the DNA processing proteins reported include enhanced attachment to tis-
and the oriT of pCF10 are highly similar in sues, enhanced internalization by epithelial
sequence and organization to the correspon- cells, and increased resistance to phagocytic
ding components of the Lactococcus lactis con- killing following ingestion by polymorphonu-
jugative element pRS01 (27, 52), and they also clear leufcougtes (47) and macrophages (53).
are more distantly related to processing systems Although the AS proteins share some sequence
from gram-negative bacteria such as those of the similarity to surface adhesin proteins from
IncP family (12). In the case of pRS01 the ltrB other gram-positive bacteria (57), the mole-
relaxase gene is interrupted by the self-splicing, cular basis for the various functional activities
mobile group II intron Ll.ltrB, and the pcfG of these proteins, including bacterial aggrega-
gene shows sufficient homology to allow effi- tion,remained elusive for many years due to the
cient insertion of the intron into pcfG in vivo. In inherent difficulties in genetic and biochemical
spite of the high degree of similarity between studies of a large secreted protein. AS proteins
the lactococcal and enterococcal processing sys- contain RGD (arginine, glycine, aspartate)
tems, each system shows high specificity for its motifs commonly found in integrin-binding
cognate oriT; our recent data indicate that the proteins, and RGD peptides partially inhibited
specificity is determined both by the interaction AS-mediated binding to canine kidney cells in
of the relaxases with their single-stranded DNA vitro (36), but no other evidence for a signifi-
targets and by interactions between the relaxase cant role of these sequences in virulence has
proteins and LtrF/PcfF (12). Remarkably, the been reported.
other characterized pheromone plasmids have Over the past several years we have devel-
completely unrelated DNA processing systems. oped improved genetic and biochemical tools
With regard to the Mpf machinery, the respec- to carry out detailed structure/function studies
tive genes of pCF10 are more similar to those of of Asc10 (38, 57, 58).Among the more signifi-
a streptococcal chromosomal conjugative ele- cant results of these studies is the identification
ment than to the other well-studied pheromone of an N-terminal lipoteichoic acid-binding
plasmids (27). domain and a second, central domain of Asc10
required for bacterial aggregation and also
Virulence Traits Associated involved in interactions leading to invasion of
with pCF10 epithelial cells (44). On the other hand, the
Pheromone-responsive plasmids are common RGD motifs play no role in these functions,
in E. faecalis clinical isolates, and we have since changing both RGDs to RADs (arginine,
demonstrated that pCF10 carriage confers a alanine,aspartate) had no effects on aggregation
strong selective advantage in animal models of or invasion (59). A comprehensive analysis is
endocarditis and subdermal abscesses (13, 28, currently in progress of the behavior of a series
31, 36, 44, 47, 49, 53). Although it has been of allelic variants of prgB (in the context of
clearly shown that the cytolysin (hemolysin) of pCF10) containing specific in-frame deletions,
certain pheromone plasmids such as pAD1 insertions, and point mutations in an experi-
clearly contributes to enterococcal virulence, mental endocarditis model. The preliminary
the only pCF10-encoded protein known to results suggest the possibility that both the
increase virulence in experimental systems is aggregation domains and at least one of the
Asc10, encoded by the prgB gene (28). Multiple RGD motifs play important roles in vivo,
studies from our group (28, 38, 39, 50) and oth- and that these different domains have distinct
functions in the interaction between the host above, two other modules are found down-
and the pathogen. stream on pCF10, but they are not conserved
For Asc10 to play a significant role in infec- on other pheromone-inducible plasmids; these
tions, it must be expressed in vivo. It was genes encode the pCF10 mating channel and
observed previously (28) that a host component DNA processing functions.
of plasma caused induction of Asc10 expression Sequence analysis suggests that as the various
via the pheromone-sensing system, but the pheromone plasmids evolved, they may have
identity of the inducing factor was not clear. shared a common pheromone-sensing module
Recently, it was shown that in vivo induction but added on different mating channel and
requires synthesis of endogenous pheromone DNA processing modules. In the case of
by the plasmid-containing bacterial cell and pCF10, it appears that cCF10 coevolved with
that the host component (probably lipid/ PrgX, PrgY, PrgZ, and iCF10, suggesting that
albumin complexes) acts by reducing the effec- the pheromone-inducible aggregation module
tive concentration of iCF10 in the growth has been in an enterococcal host for a long
medium such that the cCF10 to iCF10 ratio is time. It is interesting to consider that the acqui-
shifted in favor of pheromone (11). These sition of the pheromone-sensing module
results, along with discovery that pheromone- preceded the mating channel and DNA pro-
inducible aggregation probably evolved before cessing, yet all three modules respond to
pheromone-inducible conjugation (27), sug- cCF10. Despite the fact that all three modules
gest that the ability to express Asc10 during are required for efficient plasmid transfer, sev-
growth within a eukaryotic host may have been eral of these modules are more related to genes
a major driving force in the evolution of the in other bacterial species than similar modules
system, as described elsewhere in this chapter. in other pheromone-sensing plasmids. For
example, the mating channel formation of
EVOLUTION OF THE PHEROMONE- pCF10 is more related to group B streptococci
SENSING MACHINERY than that of pAD1 or pPD1 (27). Genes in
From molecular analysis of the pheromone- pCF10 that are involved in conjugative DNA
inducible conjugation system of E. faecalis it processing are unrelated to the corresponding
seems as if the various components and regula- genes of pAD1 but are highly similar to the lac-
tory devices were patched together at random tococcal pRSO1 conjugative element (52).
rather than designed with a simple mechanism In the design of mechanical devices and
to enable the donor cells to detect recipients computer software, the simplest design that
(34). With five of the pheromone-inducible effectively carries out the intended fuction is
plasmid sequences available for analysis,we now considered to be the best. As we continue to
have new insights into the evolution of these acquire more information regarding the regula-
plasmids and a greater appreciation for the evo- tion of pheromone-inducible conjugative
lution of these complex systems.There is a great plasmids, it is becoming clear that the
deal of sequence conservation in the contigu- pheromone-responsive plasmids of E. faecalis
ous 15-kb segment of pCF10 and the corre- did not arise by such a design process. Instead,
sponding regions of pAD1 and pPD1 that these systems were assembled over time in
encode for plasmid replication functions, plas- response to multiple selective pressures faced by
mid sensing, and production of aggregation E. faecalis.Thus, a system far more complex and
substance.These genes most likely evolved from multifunctional than initially thought has come
a common ancestor to build a module for the to exist. We propose that the pCF10 system
plasmid replication and pheromone-inducible described in this chapter has been appropriated
aggregation. However, this module alone is not by the plasmid for the purpose of increasing
responsible for all the machinery necessary for the sensitivity and adaptability to multiple
conjugative plasmid transfer. As described situations.Thus, we embrace the growing com-
plexity of this system as an insight into the 9a. Chandler, J. R., and G. M. Dunny. Character-
evolution of plasmids and microbial behavior. ization of the sequence specificity determinants
required for processing and control of sex
ACKNOWLEDGMENTS pheromone by the intramembrane protease
Eep and the plasmid-encoded protein PrgY.
We thank all previous and current members of the
J. Bacteriol., in press.
Dunny laboratory for their contributions to our under-
10. Chandler, J. R., A. R. Flynn, E. M. Bryan,
standing of the pCF10 system, and Tim Leonard for
and G. M. Dunny. 2005. Specific control of
help with the figures.
endogenous cCF10 pheromone by a conserved
This research was supported by grant GM49530
domain of the pCF10-encoded regulatory pro-
from the National Institutes of Health.
tein PrgY in Enterococcus faecalis. J. Bacteriol.
187:4830–4843.
REFERENCES 11. Chandler, J. R., H. Hirt, and G. M. Dunny.
1. An, F. Y., M. C. Sulavik, and D. B. Clewell. 2005. A paracrine peptide sex pheromone also
1999. Identification and characterization of a acts as an autocrine signal to induce plasmid
determinant (eep) on the Enterococcus faecalis transfer and virulence factor expression in vivo.
chromosome that is involved in production of Proc. Natl.Acad. Sci. USA 102:15617–15622.
the peptide sex pheromone cAD1. J. Bacteriol. 12. Chen, Y., J. H. Staddon, and G. M. Dunny.
181:5915–5921. 2007. Specificity determinants of conjugative
2. Bae, T., S. Clerc-Bardin, and G. M. Dunny. DNA processing in the Enterococcus faecalis plas-
2000.Analysis of expression of prgX, a key negative mid pCF10 and the Lactococcus lactis plasmid
regulator of the transfer of the Enterococcus faecalis pRS01. Mol. Microbiol. 63:1549–1565.
pheromone-inducible plasmid pCF10. J. Mol. Biol. 13. Chow, J. W., L. A. Thal, M. B. Perri, J. A.
297:861–875. Vazquez, S. M. Donabedian, D. B. Clewell,
3. Bae, T., and G. M. Dunny. 2001. Dominant- and M. J. Zervos. 1993. Plasmid-associated
negative mutants of prgX: evidence for a role hemolysin and aggregation substance production
for PrgX dimerization in negative regulation of contribute to virulence in experimental entero-
pheromone-inducible conjugation. Mol. Microbiol. coccal endocarditis. Antimicrob. Agents Chemother.
39:1307–1320. 37:2474–2477.
4. Bae,T., B. Kozlowicz, and G. M. Dunny. 2002. 14. Chung, J. W., and G. M. Dunny. 1995.Tran-
Two targets in pCF10 DNA for PrgX binding: their scriptional analysis of a region of the Enterococcus
role in production of Qa and prgX mRNA and in faecalis plasmid pCF10 involved in positive regu-
regulation of pheromone-inducible conjugation. J. lation of conjugative transfer functions. J. Bacte-
Mol. Biol. 315:995–1007. riol. 177:2118–2124.
5. Bae, T., B. K. Kozlowicz, and G. M. Dunny. 15. Clewell, D. B. 1999. Sex pheromone systems in
2004. Characterization of cis-acting prgQ mutants: enterococci, p. 47–65. In G. M. Dunny and S. C.
evidence for two distinct repression mechanisms by Winans (ed.), Cell-Cell Signaling in Bacteria.ASM
Qa RNA and PrgX protein in pheromone- Press,Washington, DC.
inducible enterococcal plasmid pCF10. Mol. Micro- 16. Clewell, D. B., and G. M. Dunny. 2002. Con-
biol. 51:271–281. jugation and genetic exchange in enterococci,
6. Bensing, B. A., and G. M. Dunny. 1997. p. 265–300. In D. B. Clewell and M. S. Gilmore
Pheromone-inducible expression of an aggrega- (ed.), The Enterococci: Pathogenesis, Molecular Biology,
tion protein in Enterococcus faecalis requires and Antibiotic Resistance. ASM Press, Washington,
interaction of a plasmid-encoded RNA with DC.
components of the ribosome. Mol. Microbiol. 17. De Boever, E. H., D. B. Clewell, and C. M.
24:295–308. Fraser. 2000. Enterococcus faecalis conjugative
7. Brown, M. S., J. Ye, R. B. Rawson, and J. plasmid pAM373: complete nucleotide sequence
L. Goldstein. 2000. Regulated intramembrane and genetic analyses of sex pheromone response.
proteolysis: a control mechanism conserved from Mol. Microbiol. 37:1327–1341.
bacteria to humans. Cell 100:391–398. 18. Dunny, G. M., M. H.Antiporta, and H. Hirt.
8. Buttaro, B. A., M. H. Antiporta, and G. M. 2001. Peptide pheromone-induced transfer
Dunny. 2000. Cell-associated pheromone peptide of plasmid pCF10 in Enterococcus faecalis: probing
(cCF10) production and pheromone inhibition in the genetic and molecular basis for speci-
Enterococcus faecalis. J. Bacteriol. 182:4926–4933. ficity of the pheromone response. Peptides
9. Chandler, J. R., and G. M. Dunny. 2004. Entero- 22:1529–1539.
coccal peptide sex pheromones: synthesis and con- 19. Dunny, G. M., B. L. Brown, and D. B.
trol of biological activity. Peptides 25:1377–1388. Clewell. 1978. Induced cell aggregation and
mating in Streptococcus faecalis: evidence for a bacte- pheromone receptor encoded by pPD1, in a pro-
rial sex pheromone. Proc. Natl. Acad. Sci. USA 75: moter region essential for the mating response in
3479–3483. Enterococcus faecalis. J. Bacteriol. 184:6343–6350.
20. Dunny, G. M., and B. A. Leonard. 1997. Cell- 30. Hu, J. C. 1995.Repressor fusions as a tool to study
cell communication in gram-positive bacteria. protein-protein interactions. Structure 3:431–433.
Annu. Rev. Microbiol. 51:527–564. 31. Isenmann, R., M. Schwarz, E. Rozdzinski, R.
21. Ehrenfeld, E. E., R. E. Kessler, and D. B. Marre, and H. G. Beger. 2000.Aggregation sub-
Clewell. 1986. Identification of pheromone- stance promotes colonic mucosal invasion of Ente-
induced surface proteins in Streptococcus faecalis and rococcus faecalis in an ex vivo model. J. Surg. Res.
evidence of a role for lipoteichoic acid in forma- 89:132–138.
tion of mating aggregates. J. Bacteriol. 168:6–12. 32. Kao, S. M., S. B. Olmsted,A. S.Viksnins, J. C.
22. Fixen, K. R., J. R. Chandler, T. Le, B. K. Gallo, and G. M. Dunny. 1991. Molecular and
Kozlowicz, D. A. Manias, and G. M. Dunny. genetic analysis of a region of plasmid pCF10 con-
2006. Analysis of the amino acid sequence taining positive control genes and structural genes
specificity determinants of the enterococcal encoding surface proteins involved in pheromone-
cCF10 sex pheromone in the interactions with the inducible conjugation in Enterococcus faecalis. J. Bac-
pheromone-sensing machinery. J. Bacteriol. teriol. 173:7650–7664.
189:1399–1406. 33. Kozlowicz, B. K., T. Bae, and G. M. Dunny.
23. Francia, M. V., W. Haas, R. Wirth, E. 2004. Enterococcus faecalis pheromone-responsive
Samberger, A. Muscholl-Silberhorn, M. S. protein PrgX: genetic separation of positive
Gilmore,Y. Ike, K. E.Weaver, F.Y. An, and D. autoregulatory functions from those involved in
B. Clewell. 2001. Completion of the nucleotide negative regulation of conjugative plasmid transfer.
sequence of the Enterococcus faecalis conjugative vir- Mol. Microbiol. 54:520–532.
ulence plasmid pAD1 and identification of a sec- 34. Kozlowicz, B. K., M. Dworkin, and G. M.
ond transfer origin. Plasmid 46:117–127. Dunny. 2006. Pheromone-inducible conjugation
24. Fujimoto, S., M. Bastos, K. Tanimoto, F. An, in Enterococcus faecalis: a model for the evolution of
K. Wu, and D. B. Clewell. 1997.The pAD1 sex biological complexity? Int. J. Med. Microbiol.
pheromone response in Enterococcus faecalis. Adv. 296:141–147.
Exp. Med. Biol. 418:1037–1040. 35. Kozlowicz, B. K., K. Shi, Z. Y. Gu, D. H.
25. Fujimoto, S., and D. B. Clewell. 1998. Regula- Ohlendorf, C. A. Earhart, and G. M. Dunny.
tion of the pAD1 sex pheromone response of Ente- 2006. Molecular basis for control of conjugation
rococcus faecalis by direct interaction between the by bacterial pheromone and inhibitor peptides.
cAD1 peptide mating signal and the negatively Mol. Microbiol. 62:958–969.
regulating, DNA-binding TraA protein. Proc. Natl. 36. Kreft, B., R. Marre, U. Schramm, and R.
Acad. Sci. USA 95:6430–6435. Wirth. 1992. Aggregation substance of Enterococ-
26. Hedberg, P. J., B. A. Leonard, R. E. Ruhfel, cus faecalis mediates adhesion to cultured renal
and G. M. Dunny. 1996. Identification and char- tubular cells. Infect. Immun. 60:25–30.
acterization of the genes of Enterococcus faecalis plas- 37. Leonard, B. A., A. Podbielski, P. J. Hedberg,
mid pCF10 involved in replication and in negative and G. M. Dunny. 1996. Enterococcus faecalis
control of pheromone-inducible conjugation. pheromone binding protein, PrgZ, recruits a chro-
Plasmid 35:46–57. mosomal oligopeptide permease system to import
27. Hirt, H., D. A. Manias, E. M. Bryan, J. R. sex pheromone cCF10 for induction of conjuga-
Klein, J. K. Marklund, J. H. Staddon, M. L. tion. Proc. Natl.Acad. Sci. USA 93:260–264.
Paustian, V. Kapur, and G. M. Dunny. 2005. 38. McCormick, J. K., H. Hirt, C. M.Waters,T. J.
Characterization of the pheromone response of Tripp, G. M. Dunny, and P. M. Schlievert.
the Enterococcus faecalis conjugative plasmid pCF10: 2001.Antibodies to a surface-exposed, N-terminal
complete sequence and comparative analysis of the domain of aggregation substance are not protective
transcriptional and phenotypic responses of in the rabbit model of Enterococcus faecalis infective
pCF10-containing cells to pheromone induction. endocarditis. Infect. Immun. 69:3305–3314.
J. Bacteriol. 187:1044–1054. 39. McCormick, J. K., T. J. Tripp, G. M. Dunny,
28. Hirt, H., P. M. Schlievert, and G. M. Dunny. and P. M. Schlievert. 2002.Formation of vegeta-
2002. In vivo induction of virulence and antibiotic tions during infective endocarditis excludes bind-
resistance transfer in Enterococcus faecalis mediated ing of bacterial-specific host antibodies to
by the sex pheromone-sensing system of pCF10. Enterococcus faecalis. J. Infect. Dis. 185:994–997.
Infect. Immun. 70:716–723. 40. Mori, M., H. Tanaka,Y. Sakagami, A. Isogai,
29. Horii, T., H. Nagasawa, and J. Nakayama. M. Fujino, C. Kitada, D. B. Clewell, and A.
2002. Functional analysis of TraA, the sex Suzuki. 1987. Isolation and structure of the sex
pheromone inhibitor, iPD1, excreted by Streptococ- J. W. Harmala, H. Hirt, and G. M. Dunny.

cus faecalis donor strains harboring plasmid pPD1. 1998.Aggregation and binding substances enhance
J. Bacteriol. 169:1747–1749. pathogenicity in rabbit models of Enterococcus fae-
41. Muscholl-Silberhorn, A. B. 2000. Pheromone- calis endocarditis. Infect. Immun. 66:218–223.
regulated expression of sex pheromone plasmid 51. Shi, K., C. K. Brown, Z. Y. Gu, B. K.
pAD1-encoded aggregation substance depends Kozlowicz, G. M. Dunny, D. H. Ohlendorf,
on at least six upstream genes and a cis-acting, and C.A. Earhart. 2005. Structure of peptide sex
orientation-dependent factor. J. Bacteriol. pheromone receptor PrgX and PrgX/pheromone
182:3816–3825. complexes and regulation of conjugation in
42. Nakayama, J., R. E. Ruhfel, G. M. Dunny, A. Enterococcus faecalis. Proc. Natl. Acad. Sci. USA
Isogai, and A. Suzuki. 1994.The prgQ gene of 102:18596–18601.
the Enterococcus faecalis tetracycline resistance plas- 52. Staddon, J. H., E. M. Bryan, D.A. Manias, and
mid pCF10 encodes a peptide inhibitor, iCF10. J. G. M. Dunny. 2004. Conserved target for group
Bacteriol. 176:7405–7408. II intron insertion in relaxase genes of conjugative
43. Nakayama, J., Y. Takanami, T. Horii, S. elements of gram-positive bacteria. J. Bacteriol.
Sakuda, and A. Suzuki. 1998. Molecular mech- 186:2393–2401.
anism of peptide-specific pheromone signaling in 53. Sussmuth, S. D., A. Muscholl-Silberhorn, R.
Enterococcus faecalis: functions of pheromone recep- Wirth, M. Susa, R. Marre, and E. Rozdzinski.
tor TraA and pheromone-binding protein TraC 2000.Aggregation substance promotes adherence,
encoded by plasmid pPD1. J. Bacteriol. phagocytosis, and intracellular survival of Ente-
180:449–456. rococcus faecalis within human macrophages and
44. Olmsted, S. B., G. M. Dunny, S. L. Erlandsen, suppresses respiratory burst. Infect. Immun.
and C. L.Wells. 1994.A plasmid-encoded surface 68:4900–4906.
protein on Enterococcus faecalis augments its inter- 54. Tanimoto, K., and D. B. Clewell. 1993. Regu-
nalization by cultured intestinal epithelial cells. J. lation of the pAD1-encoded sex pheromone
Infect. Dis. 170:1549–1556. response in Enterococcus faecalis: expression of
45. Olmsted, S. B., S. M. Kao, L. J. van Putte, J. C. the positive regulator TraE1. J. Bacteriol. 175:
Gallo, and G. M. Dunny. 1991. Role of the 1008–1018.
pheromone-inducible surface protein Asc10 in 55. Tanimoto, K., H.Tomita, and Y. Ike. 1996.The
mating aggregate formation and conjugal transfer traA gene of the Enterococcus faecalis conjugative
of the Enterococcus faecalis plasmid pCF10. J. Bacte- plasmid pPD1 encodes a negative regulator for the
riol. 173:7665–7672. pheromone response. Plasmid 36:55–61.
46. Ozawa, Y., E. H. De Boever, and D. B. 56. Trotter, K. M., and G. M. Dunny. 1990.
Clewell. 2005. Enterococcus faecalis sex pheromone Mutants of Enterococcus faecalis deficient as recipi-
plasmid pAM373: analyses of TraA and evidence ents in mating with donors carrying pheromone-
for its interaction with RpoB. Plasmid 54:57–69. inducible plasmids. Plasmid 24:57–67.
47. Rakita, R. M., N. N.Vanek, K. Jacques-Palaz, 57. Waters, C. M., and G. M. Dunny. 2001.Analysis
M. Mee, M. M. Mariscalco, G. M. Dunny, M. of functional domains of the Enterococcus faecalis
Snuggs, W. B. Van Winkle, and S. I. Simon. pheromone-induced surface protein aggregation
1999. Enterococcus faecalis bearing aggregation sub- substance. J. Bacteriol. 183:5659–5667.
stance is resistant to killing by human neutrophils 58. Waters, C. M., H. Hirt, J. K. McCormick, P.
despite phagocytosis and neutrophil activation. M. Schlievert, C. L. Wells, and G. M. Dunny.
Infect. Immun. 67:6067–6075. 2004. An amino-terminal domain of Enterococcus
48. Ruhfel, R. E., D.A. Manias, and G. M. Dunny. faecalis aggregation substance is required for aggre-
1993. Cloning and characterization of a region of gation, bacterial internalization by epithelial cells
the Enterococcus faecalis conjugative plasmid, and binding to lipoteichoic acid. Mol. Microbiol.
pCF10, encoding a sex pheromone-binding func- 52:1159–1171.
tion. J. Bacteriol. 175:5253–5259. 59. Waters, C. M., C. L.Wells, and G. M. Dunny.
49. Sartingen, S., E. Rozdzinski, A. Muscholl- 2003.The aggregation domain of aggregation sub-
Silberhorn, and R. Marre. 2000. Aggregation stance, not the RGD motifs, is critical for efficient
substance increases adherence and internalization, internalization by HT-29 enterocytes. Infect.
but not translocation, of Enterococcus faecalis Immun. 71:5682–5689.
through different intestinal epithelial cells in vitro. 60. Winans, S. C., D. L. Burns, and P. J. Christie.
Infect. Immun. 68:6044–6047. 1996. Adaptation of a conjugal transfer system for
50. Schlievert, P. M., P. J. Gahr, A. P. the export of pathogenic macromolecules. Trends
Assimacopoulos, M. M. Dinges, J. A. Stoehr, Microbiol. 4:64–68.
C-SIGNAL CONTROL OF
AGGREGATION AND SPORULATION
Dale Kaiser
4
Myxobacteria have a society of their own. thus evolved from its -proteobacterial progeni-
Groups of cells share extracellular digestive tor by duplicating genes for cell-cell signaling,
enzymes as they glide over the substrate in small-molecule sensing, and integrative control
search of food.When the population senses that of transcription (16). Among the nine -
starvation approaches, the cells change their proteobacterial genomes that had been com-
movement pattern. Instead of spreading out- pletely sequenced by the end of 2006, only the
ward to continue the hunt for food, they con- myxobacteria have the capacity to develop and
gregate inward and form focal aggregates that differentiate. It was observed that the most fre-
become fruiting bodies. Eventually cells inside quently duplicated genes in the evolution of M.
fruiting bodies differentiate into nonmotile, xanthus from its -proteobacterial ancestor
environmentally resistant myxospores. The life enhance its social capacities.
cycle of Myxococcus xanthus that forms a spheri-
cal fruiting body resting on a stubby pedestal is MOTILITY
illustrated in Fig. 1. Each of the 50 different M.xanthus cells are elongated,flexible rods with
species of myxobacteria builds fruiting bodies a 7:1 length-to-width ratio (59).They lack fla-
that have a distinctive form, and their morphol- gella and are unable to swim. They can only
ogy tracks their phylogeny (73). Dispersion of move on surfaces and do so by gliding in the
myxospores appears to be the primary function direction of their long axis (6). Fruiting body
of an elevated fruiting body (5); elevation facili- development requires a solid surface because
tates the transport by a small animal in soil of the the structure is built by cell movement. Two
spores to a place where food is available. Since molecular motors, retractile type IV pili at their
1970, our research group has been trying to leading end (S motility), and nozzles for secret-
decipher the instructions used by M. xanthus to ing a slime gel at the trailing end (A motility)
build its fruiting body. These structures are provide the adhesion and thrust necessary for
remarkably uniform in size and shape (76,77).A moving on surfaces.
comparative gene inventory reveals that M. xan-
PILUS ENGINE
Dale Kaiser Departments of Biochemistry and Develop-
mental Biology, Stanford University School of Medicine, At the leading end of the cell are lengthy (often
Stanford, California 94305. several micrometers long), very thin (6 to 8
51
52 ■ KAISER
FIGURE 1 The life cycle of M. xanthus.A swarm (a group of moving and interacting cells)
can have either of two fates,depending on their environment.The fruiting body (A) is a spher-
ical structure of approximately 105 cells that have become stress-resistant spores (B).The fruit-
ing body is small (1/10 mm high) and sticky, and its spores are tightly packed.When a fruiting
body receives nutrients, the individual spores germinate (C) and thousands of M. xanthus cells
emerge together as an “instant” swarm (D).When prey is available (micrococci in the figure),
the swarm becomes a predatory collective that surrounds the prey. Swarm cells feed by con-
tacting, lysing, and consuming the prey bacteria (E–F). Fruiting body development is advanta-
geous given the collective hunting behavior. Nutrient-poor conditions elicit a unified
starvation stress response. That response initiates a self-organized program that changes cell
movement behavior,leading to aggregation.The movement behaviors include wave formation
(G) and streaming into mounded aggregates (H),which become spherical (A).Spores differen-
tiate within mounded and spherical aggregates. We use the term “swarming” in its general
sense to denote a process “in which motile organisms actively spread on the surface of a suit-
ably moist solid medium” (69). Reprinted from the Proceedings of the National Academy of Sci-
ences USA (16) with permission of the publisher.
4. C-SIGNAL CONTROL OF AGGREGATION AND SPORULATION ■ 53
nm), type IV pilus fibers. M. xanthus, Neisseria Also, several hundred thick-walled pores are
gonorrhoeae, Pseudomonas aeruginosa, Synechocystis found at the cell end,but unlike the filaments of
strain PCC6803, and some other gram-nega- slime, the pores are found at both ends of the
tive bacteria share a common set of 10 pilus cell. Presumably the filaments of slime are
proteins and produce functionally similar pili emerging from a pore or a cluster of neighbor-
(11, 54). The pilus fibers are helical arrays of ing pores. Biochemical and genetic experi-
pilin (pilA in M. xanthus) monomers, whose ments (86) indicate that the thick-walled pores
sequence-conserved N termini assemble a are secretion nozzles for the slime, which
coiled coil down the center of the fiber that behaves like a repeat unit polysaccharide (58).
provides its tensile strength (12). Lacking a Differential interference contrast microscopy
unique tip structure and being helical, the M. reveals a single filament at one pole in wild-
xanthus pilus tip exposes the surfaces of several type cells (86).Apparently, the narrow filaments
pilin monomers that would have bound other associate laterally to create a gel having the
monomers within a fiber. These exposed sur- width of a cell. Slime secretion from the rear is
faces attach to cells ahead or,more accurately,to demonstrably linked to cell movement by the
multistranded polysaccharide “fibrils” that time-lapse movies of Lars Jelsbak (Fig. 2). It is
envelop clusters of adjacent cells with a net- evident that, as the cell moves, the slime trails
work resembling a fisherman’s net (2,3,8).After lengthen. Significantly, the slime engine is
firm attachment, the cell pulls itself forward by unipolar like the pilus engine and slime secre-
retracting its pilus and storing the dissociated tion pushes cells at the back, while pilus retrac-
pilin monomers in the cytoplasmic membrane tion pulls cells from the front.
for reuse.The retraction motor is PilT, an AAA Thus, at any instant, the cell is structured to
ATPase, located in the inner membrane, whose move in one particular direction: it has a dedi-
structure-based mechanism has been modeled cated head and a dedicated tail, as illustrated in
(12). Currently PilT is the strongest known Fig. 3. Two observations show that head-tail
molecular motor, capable of developing more polarity is inherent in every cell. First, the rate
than 100 pN of tension in the case of N. gonor- of swarm expansion of cells with both engines
rhoeae (52).That tension is the reason the pilus (AS) is 50% greater than the sum of the
tip must attach multivalently to a fibril,a bundle swarm rates of two mutants, one with S engines
of polysaccharide chains, because such attach- only and the other with only the A engines (32).
ment provides enough binding strength to The synergism shows that cells always have pili
withstand 100 pN of tension. Pili are too thin at one end and active slime nozzles at the other;
and flexible to push cells when they elongate; the engines seldom, if ever, oppose each other.
pushing against a solid merely causes the fiber Second, the polarized structure is revealed at
to bend.The pilus fiber passes through PilQ, a cell division, as shown in Fig. 4. The two new
gated channel in the outer membrane of M. poles created by the division septum gain
xanthus that acts as a lubricated bushing for engines that always complement the engines at
extension and retraction of the fiber.Tgl is an their old poles. Polarity is conserved through
outer membrane lipoprotein assembly factor cell division, clear evidence of their dedication.
for PilQ (55, 56). The mechanics of pulling
REVERSING THE ENGINES
together with the observation that M. xanthus
pili are unipolar (31) imply that pili are located Careful examination of Fig. 2 shows that this
at the leading end of the cell. built-in head-tail dedication lives for only
about 10 min.Both cells shown in Fig.2 reverse
SLIME ENGINE their direction of movement several times dur-
Electron microscopy shows several filaments of ing the hour-long movie. Many instances of
slime emerging from a cell end (82). Images of reversal of movement direction have been
complete cells show that filaments emerge from recorded in time-lapse movies (4, 45, 79). It is
one end only;the opposite ends do not extrude. evident in Fig. 2, and in the time-lapse movies
54 ■ KAISER
FIGURE 2 Cell movement is correlated with the secretion of slime from its back. Selected frames
from a time-lapse movie taken by Lars Jelsbak. (A) Frame 1 of movie; the upper cell has moved down,
leaving a slime trail above it.The bottom cell has not moved. (B) Frame 4 of movie; both cells have
moved down and left a slime trail above them.(C) Frame 20 of movie;both cells have moved up and left
a slime trail below them.(D) Frame 37 of movie;both cells have moved down and left a slime trail above
them.(E) Frame 58 of movie;the upper cell has moved down,leaving a slime trail above it.(F) The lower
cell has moved up, leaving a slime trail below it.
cited, that cells simply switch their polarity by second class of proteins is localized to a pole,
switching their head and tail like a locomotive. just one pole, and its position is impermanent.
Extremly few U turns have been recorded; cells One example of the permanent class is the noz-
reverse by polarity reversal. The movies also zle structures of the A engine that are visible in
reveal that switching reversals are smooth. To roughly equal numbers at both ends of a cell
reverse, a cell simply stops, pauses for about a (82). In the S engine,PilQ,which is localized in
minute, and then moves off in the opposite a patch at both cell poles, also belongs to the
direction (28). Sometimes after a pause, the cell permanent class (56). In addition to PilQ, there
moves off in the original direction; not all is evidence that PilG, PilM, PilN, PilO, and PilP
pauses lead to reversal. No differences between are also permanently at both poles, colocalized
the two types of pauses are yet apparent. The with PilQ (56). It is likely that the inner mem-
strictly coordinated polarity of two different brane PilB and PilT proteins are also perma-
engines suggests a permanent structure; yet it nent and bipolar because they have those
must be one that can be inverted smoothly. qualities in P. aeruginosa (9), and M. xanthus pili
Considering that both engines are multipro- closely resemble those of P. aeruginosa.
tein assemblies, how can their polarity be The transient unipolar class includes the
smoothly and coordinately inverted? Only the pilus fibers, which are found clustered at either
beginnings of an answer can be discerned. All pole but almost never (5%) at both (31). Pili
the S- and A-engine proteins that have been change poles upon reversal. In addition, the
localized can be divided into two classes. One Tgl lipoprotein and assembled PilQ, which
class of proteins is permanently localized to resists being dissociated in heated detergent, are
both poles; these proteins constitute almost transiently found at the piliated pole of the cell
complete engines, but they are inactive. The in the outer membrane (56). Tgl, which has
FIGURE 3 M.xanthus cells are polarized to glide in one direction.For the cell shown,it is polarized to glide
to the left.The A engine is a “pusher” and the S engine is a “puller.” Slime-secretion nozzles are always visible
at both ends of each cell, and yet only one end secretes slime.
six tetratricopeptide repeats (61, 62), is neces- of both old engines, new engines would be
sary for PilQ secretin assembly. Its assembly synthesized under direction of the cell’s polarity
permits the pilus to cross the outer membrane template, just as they are during growth shown
and to slide in or out through a sealed (gated) in Fig. 4.
channel; other proteins are excluded from the
channel.Tgl protein can be transferred at high REGULATING REVERSALS
efficiency by contact between ends of two cells, The regulatory network that triggers the coor-
a process known as stimulation (55). Stimu- dinate inactivation of old slime secretion and
latory transfer can facilitate the reversal process old pilus engines is sketched out in the right
by providing several molecules of Tgl protein half of Fig. 5. FrzCD and FrzE proteins encode
to the cell end for PilQ assembly at that site. a two-component chemosensory system (78).
CglB lipoprotein, which resides in the outer In 1985, Blackhart and Zusman discovered that
membrane (60, 68) and is stimulatable like mutations in FrzCD and FrzE change the
Tgl (55), most likely plays a role in reversal of frequency of gliding reversals. These proteins
the A engines that parallels the role played by are related by amino acid sequence but are
Tgl for the S engines. not identical to the chemotaxis proteins of
To reverse smoothly, given the protein local- Escherichia coli and Salmonella (4). The methyl-
ization data just described, it is proposed that accepting chemosensory protein, or MCP,
the transiently unipolar Tgl and CglB proteins FrzCD, is a cytoplasmic protein in M. xanthus,
are inactivated. In assembled secretins, the PilQ not a membrane receptor as in E. coli—evi-
subunits are not covalently bonded to each dence that it receives cytoplasmic, not extracel-
other but are held together by Tgl (most likely lular signals. FrzE, the histidine protein kinase
by the TPR repeats of Tgl) (56, 61, 62). Conse- for this two-component system, is signaled by
quently, when Tgl is inactivated, the PilQ MeFrzCD in turn. FrzE can autophosphorylate
secretin is expected to disassemble. But PilQ (1).Wild-type cells reverse once every 10 min
remains in the outer membrane as a patch of on average, but a frzE null mutant seldom
PilQ monomers, as observed by immunolocal- reverses (4).Therefore,FrzE~P is likely to be the
ization microscopy and sodium dodecyl sulfate signal to reverse polarity.
(SDS) gel electrophoresis (56).Inactivation may During an investigation of traveling waves,
be the result of proteolysis, and there are many Oleg Igoshin and George Oster simulated the
protease candidates in the genome (16). Having waves mathematically. They showed that the
inactivated a small number of key components levels of Me-FrzCD and FrzE~P oscillated out
56 ■ KAISER
FIGURE 4 When a cell divides, two new ends are created by the
division septum. Each daughter receives only one of the two engines at
its new pole, and always the correct one.The cell’s peptidoglycan and
cytoskeleton appear to be recognized as a polarized template specifying
different working engines at the two poles.
of phase with each other when producing poles are born with the permanent proteins
waves. They found in simulating the system of both engines; they lack Tgl, CglB, and
kinetics that there must be a negative feedback other proteins of the transient class. MglA is a
from FrzE~P back to the methylation of small G protein, homologous to Sar1 of Saccha-
FrzCD, which creates the “frizilator.”Although romyces cerevisiae. MglB is its putative guanine
direct biochemical evidence for the feedback is nucleotide release protein (GNRP), and dele-
lacking, the simulation evidence seems solid. tion of MglB alone produces a partially motile
Mutant phenotypes and epistasis experiments mutant (86).Together, MglA and MglB consti-
imply that the FrzE~P output from the oscilla- tute a G-protein switch. FrzE~P is proposed,
tor goes directly, or indirectly, to MglA. MglA directly or indirectly, to activate MglAB and to
mutants have been said to “hyper-reverse” (72), synthesize MglA•GTP.
but this description is misleading. In fact, mglA
mutants are observed to secrete slime from both REVERSAL AND FRUITING
poles as if they are trying to move forward and BODY DEVELOPMENT
backward at the same time (86). They are Growing cells reverse periodically, triggered by
observed to oscillate rapidly back and forth at the frizilator. These reversals seem to help the
very low amplitude, much faster than any frz swarm spread outward as a thin sheet of cells
mutant.The rapid oscillations can be accounted that provides them with greater access to nutri-
for as statistical fluctuations in the rate of slime ent.There is less competition for foodstuffs that
secretion from the two ends, rather than are diffusing upward from the substrate surface,
changes in structural polarity. The data and for O2 that is diffusing into cells from air above,
argument are given in detail elsewhere (30). and for CO2 that is diffusing out of cells into the
Then the G protein in its GTP-bound state atmosphere. Outward spreading stops when M.
recognizes the currently active pili at the cyto- xanthus senses that it has begun to starve;instead
plasmic face of the inner membrane for Tgl it moves inward to the swarm center to establish
inactivation and loss of the type IV pili. Simul- centers for fruiting body aggregation. By tradi-
taneously, MglA•GTP recognizes the pole that tion, the aggregation of cells has been consid-
currently has and should lose its slime secretion ered to arise from chemotaxis (48, 53), and this
activity. Despite uncertainty as to how those view was encouraged by the discovery of many
targets are recognized, the phenotype of mglA “chemotaxis genes” in M. xanthus. But the evi-
mutants clearly shows that engine inactivation dence suggests that long-range chemotactic
is the process that is regulated, while new attractions are not involved. Rather, M. xanthus
engine production follows the normal pattern builds its fruiting bodies by C signaling
of cell growth (Fig. 4). Because an mglA between adjacent cells. C signal-deficient
deletion mutant is unable to inactivate CglB, it mutants (csgA) were found to grow normally
secretes slime from both its ends (86).The new and neither to aggregate nor to sporulate (21,
FIGURE 5 Regulatory C-signal circuit. C signal is a 17-kDa cell surface protein. Cells must make end-to-end
contact to transmit the signal,as shown.Reception of C signal activates FruA by forming FruA~P.FruA~P drives the
oscillation of MeFrzCD and FrzE~P. FrzE~P switches MglA•GDP to MglA•GTP, which in turn inactivates old
engines. C signaling increases csgA transcription, directly or indirectly, via the proteins of the act operon. Rippling,
aggregation, sporulation, and C signal-dependent gene expression are induced by increasing levels of FruA~P.
MXAN4899 is proposed to be the branch from reception of C signal to FruA~P.
38, 67).The C signal is a 17-kDa cell-surface- important response regulator (14). At the same
bound protein that signals when a pair of cells time, expression of the C signal is increased by a
makes end-to-end contact with each other, as positive feedback loop involving the act operon,
indicated in Fig.5 (37).Side-by-side or end-by- also depicted in Fig. 5 (18). The complete C-
side contacts apparently do not exchange signal signaling circuit shown in Fig.5 was worked out
(64).Aggregation by local cell contact signaling from the properties of gene knockout mutants,
has been tested by mathematical simulation. A then tested by the Sozinova simulations. Due to
continuous three-dimensional simulation (70, the positive feedback, there is a progressive
71) reproduces all the experimentally observed increase in the cell surface level of C signal.The
stages of fruiting body formation: asymmetric five proteins of the act operon increase expres-
initial aggregates (known as traffic jams), linear sion of the csgA gene (19).They produce a 25-
streams, formation of hemispherical mounds kDa protein that is secreted to the cell surface
whose centers have low cell density, and finally, where it is cleaved to the active 17-kDa form
sporulation within the mounds. The simula- (50).At the start of development, there are few
tions also suggest how the fruiting body C-signal molecules per cell. Cells making end-
becomes spherical. to-end contact respond to signal exchange by
reversing their direction of gliding, which cre-
C SIGNALING AND DEVELOPMENT ates the traveling waves (79). One consequence
As shown in Fig. 5, when traffic jams enlarge of the feedback is that the wave pattern is tran-
into fruiting bodies, C signaling induces the sient (18). Each time C signal is exchanged
phosphorylation of Fru A, a developmentally between cells in the crests of two colliding
58 ■ KAISER
waves, the positive feedback loop increases fourth, and a fifth do likewise, creating a stream
expression of csgA and elevates the number of of five cells, all moving into the aggregation
signal molecules on both signaling cells. focus as a chain. Streams not moving toward a
Traveling waves are produced by the initial focus become part of the general circulation of
level of C signal found at the start of develop- cells outside the aggregate.
ment.A low level of FruA~P drives the frizilla- Some aggregations are found to have been
tor (Fig. 5), producing a very regular 8-min nucleated by small stationary traffic jams that
period. The circuit oscillates because FrzE~P form during growth and may involve a few
inhibits methylation of FrzCD or stimulates hundred cells (34).When growing cells that are
demethylation of MeFrzCD. Waves start as moving in opposite directions happen to meet
broad and diffuse ridges. Then, because the in a small area,they stall at the points of collision
frizilators in a pair of cells that are signaling to if they are prevented from turning by other cells
each other are brought into synchrony, they at their side or behind them. In submerged
sharpen. The precise period of the FruA~P- culture (46), cells settle on the bottom of the
driven frizilators and the synchronization of culture dish as domains of cells, giving rise to
oscillation between colliding cells causes the the facet pattern.Within each domain, cells are
waves to sharpen progressively. Due to the pos- aligned parallel to each other, and different
itive feedback, more signal is produced. Higher domains have different orientations. Traffic
levels of C signal induce higher levels of jams occur at the intersections where two
FruA~P (36, 44, 49). domains collide (46), and they can persist for
several hours. If fruiting body development
AGGREGATION starts during the life of a traffic jam, it can
As the number of C-signal molecules per cell become the nucleus for a fruiting body due to
rises, their signaling elevates the cytoplasmic its interaction with streams. Or, if a jam forms
level of FruA~P (Fig. 5) to a threshold. The early in development, at the intersections of
existence of a threshold that can stop the oscil- three ridges of high cell density in submerged
lation is implicit in the negative feedback agar culture, for example, it can also nucleate a
within the frizilator (23). Arresting the oscilla- fruiting body (34).
tion leads a developing population of cells to When a stream of cells encounters an
make the transition from traveling waves to impenetrable traffic jam, the stream is deflected
aggregates that is evident when the waves fade and turns to glide over or around the traffic jam.
away and the aggregation centers enlarge at the While the stream treats the jam as a surface
same time (34).The threshold level of FruA~P obstruction, individual cells must bend as they
stops the oscillation with all the FrzE in its non- pass over or around the jam. Bending has dis-
phosphorylated state (23).There being no more torted an elastic wall (81) that is composed of
signal to reverse, the cells continue to move in intercalated plates of peptidoglycan (80).Persis-
the direction they were moving before their last tence of the bent shape directs the cell into a
signaling event.This transition in cell behavior circular orbit. Cells that are streaming in orbit
has also been observed in the tracks of individ- can continue to C signal, and more positive
ual cells (29). It permits the cells to form feedback raises the C-signal level until this pop-
streams, and streaming supports the enlarge- ulation of cells reaches the still higher threshold
ment of aggregates. Consider a cell that hap- for differentiation into spores, which also has
pens to be moving toward an aggregation focus been observed experimentally (18, 36, 49).
and another cell that happens to come up When rod cells differentiate into spores,they
behind it and to make signaling contact. Both lose the capacity to move on their own (65).
cells respond to the signaling between them by Although spores can no longer glide, adjacent
continuing to move in the same direction, motile rod-shaped cells push them out of the
which is toward the focus.Then a third cell, a way. At the beginning of sporulation the cell
density in the center is about one-third the expression are necessary to adjust to starvation,
density in the outer region (66).We have previ- to aggregate, and finally to sporulate.
ously demonstrated by simulation that sig- Another group of transcription factors, the
naling by contact between M. xanthus cells is sigma-54 activator proteins, are important
sufficient to produce hemispherically mounded because many developmentally regulated
aggregates (71).When the traffic jam at the cen- genes in M. xanthus are expressed from 54-
ter of a mound resolves, those cells join the dependent promoters (15, 35, 63, 83). The
other streaming and signaling cells in the aggre- 54 promoters differ from promoters of the 70
gate, which then becomes a mound. Mounds family. All 54 promoters require a specialized
were observed by Sager to have two distinctive transcription factor, an enhancer-binding acti-
density domains (66) and this was experimen- vator protein, or EBP, to interact with RNA
tally confirmed (51). Consequently, when the polymerase at the promoter.The EBPs usually
spores are displaced by the streaming cells, they bind regulatory DNA sequences upstream from
accumulate in the low-density center of the the promoters, called enhancers. DNA bending
fruiting body. In sum, the movements of fruit- folds the EBP over onto the 54 -RNAP com-
ing body morphogenesis can be explained by C plex at the promoter for interaction between
signaling. the proteins. In addition, EBP-catalyzed ATP
hydrolysis is required to separate the DNA
GENE EXPRESSION strands, to open the 54-RNAP promoter
The expression of many genes is altered complex, and to initiate transcription.This acti-
during M. xanthus development. A sample of vation mechanism allows the level of expression
genes whose expression increases more than to be adjusted by signal input to the EBP.
threefold was obtained with the transposable Fifty-three genes in the M. xanthus genome
promoter probe, Tn5 lac (41). Each of 29 dis- encode EBPs. As probed by gene knockout
tinct Tn5 lac insertion strains began to express mutations, at least 20 EBPs have been shown to
-galactosidase at a particular time during be important for fruiting body development (7,
development, with expression times ranging 17, 18, 20, 22, 25, 39, 74, 75). Many EBPs are
from the onset of starvation to the onset of part of signal transduction circuits that respond
sporulation (43). Their dependence on starva- to environmental cues. EBPs have an almost
tion,A signal, and C signal was also determined universal domain organization, in which a cen-
(42, 47). Many of the upstream regions of those tral AAA-ATPase domain is responsible for
genes have been cloned and subjected to ATP hydrolysis and interaction with the sigma-
segmental deletion to reveal their upstream 54 factor, the C-terminal domain binds DNA,
activation sites at which transcription factors, and the N terminus is a sensory domain that
including sigma factors, might bind (13). In this regulates the ATPase activity of the central
way, the pilA promoter was shown to require domain, opening the promoter only when
PilR binding in its upstream region (27). FruA some condition is satisfied. The N-terminal
is a major transcription factor for developmen- sensory domains show the most variation from
tal gene expression (14, 57), as indicated in one EBP to another, but there are two large
Fig. 5. Since the FruA~P level rises as develop- groups with distinctive N-terminal sequences.
ment proceeds, it activates more and more Many have a response regulator sensory
genes. Since developmentally regulated genes domain, suggesting they are each part of a two-
have different FruA requirements, their expres- component system. Several examples known to
sion would be coupled to the C-signal level and be important for development are detailed in
to cell movement, as described above.Together, references 10, 18, 20, 74, 84, and 85.
cell movement and gene expression produce a Another 12 EBPs in M. xanthus have a
mature fruiting body that has the form charac- forkhead-associated (FHA) domain at their
teristic of the species.Temporal changes in gene N termini. Having the motifs G69-R70, S85-
60 ■ KAISER
XX-H88, and N107-G108, they are homologs consistent with retarding the elevation of the
of the prototypical FHA domain, RAD53FHA1 FruA~P level in the mutant.Thus, the normal
from yeast. The FHA domain in RAD53 has function of MXAN4899 EBP may be to adjust
been shown to recognize a phosphothreonine the level of FruA~P to track the increasing
moiety and is thought to interact with a protein number of C-signal molecules on the signal
partner in a process regulated by reversible pro- donor with precision. Because the mutant phe-
tein phosphorylation. The N-terminal FHA notype is produced by deleting only the FHA
domain in MXAN4899 suggests that it interdomain of MXAN4899, the EBP is likely to
acts with an autophosphorylated Ser/Thr pro- interact with a protein that is reversibly phos-
tein kinase. This holds particular current phorylated on a serine/threonine residue.
interest because the M. xanthus genome Kroos suggested that the simplest signaling cir-
encodes at least 99 different Ser/Thr protein cuit would be one in which a cognate Ser/
kinases (16),whose functions are just beginning Thr protein kinase provided the phosphory-
to be understood (24). lated serine or threonine, because the kinase
The FHA domain of enhancer-binding autophosphorylates (40). According to this
protein MXAN4899 has been shown to play scheme, the sigma-54 activator would be acti-
a role in fruiting body development (26). If vated by signal input to the cognate Ser/Thr
MXAN4899 is deleted, aggregation is delayed protein kinase.
and larger than normal aggregates are formed.
Notwithstanding their larger size, the aggre- ACKNOWLEDGMENTS
gates have fewer than 1% of the normal number This investigation was supported by U.S. Public Health
of spores. Moreover, the same phenotype was Service grant GM 23441 to D. K. from the National
Institute of General Medical Sciences.
observed when only the FHA domain was
deleted. All the defects in aggregation and
REFERENCES
sporulation could be traced to a particular
1. Acuna, G.,W. Shi, K.Trudeau, and D. Zusman.
segment of the C-signal transduction pathway 1995.The cheA and cheY domains of Myxococcus xan-
(Fig. 5), namely the branch from the C-signal thus FrzE function independently in vitro as an
receptor to FruA~P. All aspects of the pheno- autokinase and a phosphate acceptor, respectively.
type of either deletion mutant were accounted FEBS Lett. 358:31–33.
for by the hypothesis that mutant MXAN4899 2. Arnold, J.W., and L. J. Shimkets. 1988. Cell sur-
face properties correlated with cohesion in Myxo-
severely restricts the rise in the level of phos- coccus xanthus. J. Bacteriol. 170:5771–5777.
phorylated FruA. Instead of covering the 3. Behmlander, R. M., and M. Dworkin. 1994.
normal range, FruA~P seems to rise only a bit Biochemical and structural analyses of the extracel-
above the level necessary for traveling waves lular matrix fibrils of Myxococcus xanthus. J. Bacteriol.
to the level that starts aggregation.The effect of 176:6295–6303.
4. Blackhart, B. D., and D. Zusman. 1985. Frizzy
restriction can be read from Fig. 5, keeping genes of Myxococcus xanthus are involved in control
the above discussion of aggregation in mind. of frequency of reversal of gliding motility. Proc.
If FruA~P were to rise late and more slowly Natl.Acad. Sci. USA 82:8767–8770.
than normal, FruA~P would climb only part 5. Bonner, J. T. 1982. Evolutionary strategies and
way toward the sporulation threshold, explain- developmental constraints in the cellular slime
molds. Amer. Nat. 119:530–552.
ing the sporulation defect. Streaming and 6. Burchard, R. P. 1981.Gliding motility of prokary-
aggregation would be delayed, and cells would otes: ultrastructure, physiology, and genetics. Annu.
stream for a longer time than normal because Rev. Microbiol. 35:497–529.
FruA~P would not have risen to the spore- 7. Caberoy, N. B., R. D.Welch, J. S. Jakobsen, S. C.
inducing level that stops streaming. Prolonged Slater, and A. G. Garza. 2003. Global mutational
analysis of NtrC-like activators in Myxococcus xan-
streaming would give time to build larger than thus: identifying activator mutants defective for
normal aggregates.Moreover,the observed pat- motility and fruiting body development. J. Bacteriol.
tern of reporter gene expression (not shown) is 185:6083–6094.
8. Chang, B.Y., and M. Dworkin. 1994. Isolated 20. Guo, D., Y. Wu, and H. B. Kaplan. 2000.
fibrils rescue cohesion and development in the Identification and characterization of genes
Dsp mutant of Myxococcus xanthus. J. Bacteriol. required for early Myxococcus xanthus developmen-
176:7190–7196. tal gene expression. J. Bacteriol. 182:4564–4571.
9. Chiang, P., M. Habash, and L. L. Burrows. 21. Hagen, D. C., A. P. Bretscher, and D. Kaiser.
2005. Disparate subcellular localization patterns 1978. Synergism between morphogenetic mutants
of Pseudomonas aeruginosa Type IV pilus ATPases of Myxococcus xanthus. Dev. Biol. 64:284–296.
involved in twitching motility. J. Bacteriol. 22. Hager, E., H.Tse, and R. E. Gill. 2001. Identi-
187:829–839. fication and characterization of spdR mutations
10. Cho, K., and D. R. Zusman. 1999. AsgD, a that bypass the BsgA protease-dependent regula-
new two-component regulator required for A- tion of developmental gene expression in Myxococ-
signalling and nutrient sensing during early devel- cus xanthus. Mol. Microbiol. 39:765–780.
opment of Myxococcus xanthus. Mol. Microbiol. 23. Igoshin, O., A. Goldbetter, D. Kaiser, and G.
34:268–281. Oster. 2004.A biochemical oscillator explains the
11. Craig, L., M. Pique, and J.A.Tainer. 2004.Type developmental progression of myxobacteria. Proc.
IV pilus structure and pathogenicity. Nat. Rev. Natl.Acad. Sci. USA 101:15760–15765.
Microbiolo. 2:363–378. 24. Inouye, S., R. Jain, T. Ueki, H. Nariya, C.
12. Craig, L., N. Volkmann, A. Arvai, M. Pique, Xu, M. Hsu, B. A. Fernandez-Luque, J.
M. Yeager, E. Engleman, and J. A. Tainer. Munoz-Dorado, E. Farez-Vidal, and M.
2006. Type IV pilus structure by cryo-electron Inouye. 2000. A large family of eukaryotic-like
microscopy and crystallography: implications for protein Ser/Thr kinases of Myxococcus xanthus, a
pilus assembly and functions.Mol.Cell 23:651–662. developmental bacterium. Microb. Comp. Genom.
13. Downard, J., and L. Kroos. 1993. Transcrip- 5:103–120.
tional regulation of developmental gene expres- 25. Jakobsen, J. S., L. Jelsbak, L. Jelsbak, R.
sion in Myxococcus xanthus, p. 183–199. In Welch, C. Cummings, B. Goldman, E.
M. Dworkin and D. Kaiser (ed.), Myxobacteria II. Stark, S. C. Slater, and D. Kaiser. 2004.
ASM Press,Washington, DC. Sigma54 enhancer binding proteins and Myxococcus
14. Ellehauge, E., M. Norregaard-Madsen, and xanthus fruiting body development. J. Bacteriol.
L. Søgaard-Andersen. 1998. The FruA signal 186:4361–4368.
transduction protein provides a checkpoint for the 26. Jelsbak, L., M. Givskov, and D. Kaiser. 2005.
temporal coordination of intercellular signals in M. Enhancer-binding proteins with a forkhead-
xanthus development. Mol. Microbiol. 30:807–813. associated domain and the sigma54 regulon in
15. Garza,A. G., J. S. Pollack, B. Z. Harris,A. Lee, Myxococcus xanthus fruiting body development.
I. M. Keseler, E. F. Licking, and M. Singer. Proc. Natl.Acad. Sci. USA 102:3010–3015.
1998. SdeK is required for early fruiting body 27. Jelsbak, L., and D. Kaiser. 2005. Regulating
development in Myxococcus xanthus. J. Bacteriol. pilin expression reveals a threshold for type IV
180:4628–4637. pilus assembly in Myxococcus xanthus. J. Bacteriol.
16. Goldman, B. S., W. C. Nierman, D. Kaiser, 187:2105–2112.
S. C. Slater, A. S. Durkin, J. A. Eisen, C. M. 28. Jelsbak, L., and L. Søgaard-Andersen. 1999.
Ronning, W. B. Barbazuk, M. Blanchard, C. The cell-surface associated C-signal induces
Field, C. Halling, G. Hinkle, O. Iartchuk, H. S. behavioral changes in individual M. xanthus cells
Kim, C. Mackenzie, R. Madupu, N. Miller,A. during fruiting body morphogenesis. Proc. Natl.
Shvartsbeyn, S. A. Sullivan, M. Vaudin, R. Acad. Sci. USA 96:5031–5036.
Wiegand, and H. B. Kaplan 2006. Evolution 29. Jelsbak, L., and L. Søgaard-Andersen. 2002.
of sensory complexity recorded in a myxobac- Pattern formation by a cell-surface associated
terial genome. Proc. Natl. Acad. Sci. USA 103: morphogen in M.xanthus.Proc.Natl.Acad.Sci.USA
15200–15205. 99:2032–2037.
17. Gorski, L., and D. Kaiser. 1998.Targetted muta- 30. Kaiser, D. 2008. Reversing M. xanthus polarity,
genesis of sigma-54 activator proteins in Myxococ- p. 93–102. In D. E.Whitworth (ed.), Myxobacteria:
cus xanthus. J. Bacteriol. 180:5896–5905. Multicellularity and Differentiation. ASM Press,
18. Gronewold,T. M. A., and D. Kaiser. 2001.The Washington, DC.
act operon controls the level and time of C-signal 31. Kaiser, A. D. 1979. Social gliding is correlated
production for M. xanthus development. Mol. with the presence of pili in Myxococcus xanthus.
Microbiol. 40:744–756. Proc. Natl.Acad. Sci. USA 76:5952–5956.
19. Gronewold, T. M. A., and D. Kaiser. 2007. 32. Kaiser, A. D., and C. Crosby. 1983. Cell move-
Mutations of the act promoter in Myxococcus xan- ment and its coordination in swarms of Myxococcus
thus. J. Bacteriol. 184:1172–1179. xanthus. Cell Motil. 3:227–245.
62 ■ KAISER
33. Kaiser, D. 2003.Coupling cell movement to mul- gene expression in Myxococcus xanthus. Dev. Biol.
ticellular development in myxobacteria. Nature 117:267–276.
Rev. Microbiol. 1:45–54. 48. Lev, M. 1954. Demonstration of a diffusible fruit-
34. Kaiser, D., and R. Welch. 2004. Dynamics of ing factor in myxobacteria. Nature 173:501.
fruiting body morphogenesis. J. Bacteriol. 49. Li, S., B. U. Lee, and L. Shimkets. 1992. csgA
186:919–927. expression entrains Myxococcus xanthus develop-
35. Keseler, I. M., and D. Kaiser. 1995. An ment. Genes Dev. 6:401–410.
early A-signal-dependent gene in Myxococcus 50. Lobedanz, S., and L. Søgaard-Andersen.
xanthus has a sigma-54-like promoter. J. Bacteriol. 2003. Identification of the C-signal, a contact-
177:4638–4644. dependent morphogen coordinating multiple
36. Kim, S. K., and D. Kaiser. 1991. C-factor has developmental responses in Myxococcus xanthus.
distinct aggregation and sporulation thresholds Genes Dev. 17:2151–2161.
during Myxococcus development. J. Bacteriol. 51. Lux, R.,Y. Li,A. Lu, and W. Shi. 2005. Detailed
173:1722–1728. three-dimensional analysis of structural features
37. Kim, S. K., and D. Kaiser. 1990. Cell alignment of Myxococcus xanthus fruiting bodies using
required in differentiation of Myxococcus xanthus. confocal laser scanning microscopy. Biofilms
Science 249:926–928. 1:293–303.
38. Kim, S. K., and D. Kaiser. 1990. Purification 52. Maier, B., L. Potter, M. So, H. S. Seifert, and
and properties of Myxococcus xanthus C-factor, an M. P. Sheetz. 2002. Single pilus motor forces
intercellular signaling protein. Proc. Natl.Acad. Sci. exceed 100 pN. Proc. Natl. Acad. Sci. USA
USA 87:3635–3639. 99:16012–16017.
39. Kirby, J. R., and D. R. Zusman. 2003. 53. McVittie, A., and S. A. Zahler. 1962. Chemo-
Chemosensory regulation of developmental gene taxis in Myxococcus. Nature 194:1299–1300.
expression in Myxococcus xanthus. Proc. Natl. Acad. 54. Nudleman, E., and D. Kaiser. 2004. Pulling
Sci. USA 100:2008–2013. together with type IV pili. J. Mol. Microbiol. Biotech-
40. Kroos, L. 2005. Eukaryotic-like signaling and nol. 7:52–62.
gene regulation in a prokaryote that undergoes 55. Nudleman, E., D. Wall, and D. Kaiser. 2005.
multicellular development. Proc. Natl. Acad. Sci. Cell-to-cell transfer of bacterial outer-membrane
USA 102:2681–2682. lipoproteins. Science 309:125–127.
41. Kroos, L., and D. Kaiser. 1984. Construction of 56. Nudleman, E., D. Wall, and D. Kaiser. 2006.
Tn5 lac, a transposon that fuses lacZ expression to Polar assembly of the type IV pilus secretin in Myx-
exogenous promoters, and its introduction into ococcus xanthus. Mol. Microbiol. 60:16–29.
Myxococcus xanthus. Proc. Natl. Acad. Sci. USA 57. Ogawa, M., S. Fujitani, X. Mao, S. Inouye,
81:5816–5820. and T. Komano. 1996.FruA,a putative transcrip-
42. Kroos, L., and D. Kaiser. 1987. Expression of tion factor essential for the development of Myxo-
many developmentally regulated genes in Myxococ- coccus xanthus. Mol. Microbiol. 22:757–767.
cus depends on a sequence of cell interactions. 58. Raetz, C. R. H., and C. Whitfield. 2002.
Genes Dev. 1:840–854. Lipopolysaccharide endotoxins. Annu. Rev.
43. Kroos, L., A. Kuspa, and D. Kaiser. 1986. A Biochem. 71:635–700.
global analysis of developmentally regulated genes 59. Reichenbach, H. 1993.Biology of the myxobac-
in Myxococcus xanthus. Dev. Biol. 117:252–266. teria: ecology and taxonomy, p. 13–62. In M.
44. Kruse, T., S. Lobendanz, N. M. S. Dworkin and D. Kaiser (ed.), Myxobacteria II. ASM
Bertheleson, and L. Søgaard-Andersen. 2001. Press,Washington, DC.
C-signal: a cell surface-associated morphogen that 60. Rodriguez, A. M., and A. M. Spormann.
induces and coordinates multicellular fruiting 1999. Genetic and molecular analysis of cglB, a
body morphogenesis and sporulation in gene essential for single-cell gliding in Myxococcus
M. xanthus. Mol. Microbiol. 40:156–168. xanthus. J. Bacteriol. 181:4381–4390.
45. Kuhlwein, H., and H. Reichenbach. 1968. 61. Rodriguez-Soto, J. P., and D. Kaiser. 1997.
Schwarmentwicklung und Morphogenese bei Identification and localization of the tgl protein,
Myxobacterien—Archangium, Myxococcus, which is required for Myxococcus xanthus social
Chondrococcus, Chondromyces. Film C893. Inst. motility. J. Bacteriol. 179:4372–4381.
Wissensch. Film, Gottingen, Germany. 62. Rodriguez-Soto, J. P., and D. Kaiser. 1997.
46. Kuner, J., and D. Kaiser. 1982. Fruiting body The tgl gene: social motility and stimulation in
morphogenesis in submerged cultures of Myxococ- Myxococcus xanthus. J. Bacteriol. 179:4361–4371.
cus xanthus. J. Bacteriol. 151:458–461. 63. Romeo, J. M., and D. R. Zusman. 1991.Tran-
47. Kuspa,A., L. Kroos, and D. Kaiser. 1986.Inter- scription of the myxobacterial hemagglutinin
cellular signaling is required for developmental gene is mediated by a sigma 54-like promoter and
a cis-acting upstream regulatory region of DNA. 75. Sun, H., and W. Shi. 2001. Genetic studies of
J. Bacteriol. 173:2969–2976. mrp, a locus essential for cellular aggregation and
64. Sager, B., and D. Kaiser. 1994. Intercellular C- sporulation of Myxococcus xanthus. J. Bacteriol.
signaling and the traveling waves of Myxococcus. 183:4786–4795.
Genes Dev. 8:2793–2804. 76. Velicer, G., L. Kroos, and R. E. Lenski. 1998.
65. Sager, B., and D. Kaiser. 1993.Spatial restriction Loss of social behaviors by Myxococcus xanthus dur-
of cellular differentiation.Genes Dev. 7:1645–1653. ing evolution in an unstructured habitat. Proc. Natl.
66. Sager, B., and D. Kaiser. 1993.Two cell-density Acad. Sci. USA 95:12376–12380.
domains within the Myxococcus xanthus fruiting 77. Velicer, G. J. 2003. Social strife in the microbial
body. Proc. Natl.Acad. Sci. USA 90:3690–3694. world. Trends Microbiol. 11:330–337.
67. Shimkets, L. J., R. E. Gill, and D. Kaiser. 1983. 78. Ward, M. J., H. Lew,A.Treuner-Lange, and D.
Developmental cell interactions in Myxococcus xan- R. Zusman. 1998. Regulation of motility
thus and the spoC locus. Proc. Natl. Acad. Sci. USA behavior in Myxococcus xanthus may require an
80:1406–1410. extracytoplasmic-function sigma factor. J. Bacteriol.
68. Simunovic, V., F. C. Gherardini, and L. J. 180:5668–5675.
Shimkets. 2003.Membrane localization of motil- 79. Welch, R., and D. Kaiser. 2001. Cell behavior in
ity, signaling, and polyketide synthase proteins in traveling wave patterns of myxobacteria. Proc.Natl.
Myxococcus xanthus. J. Bacteriol. 185:5066–5075. Acad. Sci. USA 98:14907–14912.
69. Singleton, P., and D. Sainsbury. 2001. Dictio- 80. White, D. 1984. Structure and function of
nary of Microbiology and Molecular Biology. Wiley, myxobacteria cells and fruiting bodies, p. 51–67. In
Chichester, United Kingdom. E. Rosenberg (ed.), Myxobacteria, Development and
70. Sozinova, O., Y. Jiang, D. Kaiser, and M. Cell Interactions. Springer-Verlag, New York, NY.
Alber. 2006. A three-dimensional model of 81. Wolgemuth, C. 2005. Force and flexibility of
myxobacterial fruiting body formation. Proc. Natl. flailing myxobacteria. Biophys. J. 89:1643–1649.
Acad. Sci. USA 103:17255–17259. 82. Wolgemuth, C., E. Hoiczyk, D. Kaiser, and G.
71. Sozinova, O., Y. Jiang, D. Kaiser, and M. S. Oster. 2002. How myxobacteria glide. Curr. Biol.
Alber. 2005. Three-dimensional model of 12:369–377.
myxobacterial aggregation by contact-mediated 83. Wu, S. S., and D. Kaiser. 1997. Regulation of
interaction. Proc. Natl. Acad. Sci. USA expression of the pilA gene in Myxococcus xanthus.
102:11308–11312. J. Bacteriol. 179:7748–7758.
72. Spormann,A. M., and D. Kaiser. 1999.Gliding 84. Wu, S. S., J.Wu, and D. Kaiser. 1997.The Myx-
mutants of Myxococcus xanthus with high reversal ococcus xanthus pilT locus is required for social glid-
frequencies and small displacements. J. Bacteriol. ing motility although pili are still produced. Mol.
181:2593–2601. Microbiol. 23:109–121.
73. Sproer, C., H. Reichenbach, and E. Stacke- 85. Yang, C., and H. B. Kaplan. 1997. Myxococcus
brandt. 1999. Correlation between morphologi- xanthus sasS encodes a sensor histidine kinase
cal and phylogenetic classification of myxobacteria. required for early developmental gene expression.
Int. J. Syst. Bacteriol. 49:1255–1262. J. Bacteriol. 179:7759–7767.
74. Sun, H., and W. Shi. 2001.Analyses of mrp genes 86. Yu, R., and D. Kaiser. 2007. Gliding motility
during Myxococcus xanthus development. J. Bacte- and polarized slime secretion. Mol. Microbiol.
riol. 183:6733–6739. 63:454–467.
THE Dif CHEMOSENSORY SYSTEM IS
REQUIRED FOR S MOTILITY, BIOFILM
FORMATION, CHEMOTAXIS,
AND DEVELOPMENT IN
MYXOCOCCUS XANTHUS
5
The myxobacteria have a uniquely social devel- teins in each of these behaviors is described in
opmental cycle that is initiated by amino acid subsequent sections.
or energy limitation and the stringent response.
An orderly program of spatial and temporal S MOTILITY
gene expression culminates in the formation of Myxococcus xanthus cells are motile only when
a spore-filled fruiting body. Fruiting body morin contact with surfaces. Cells move about
phogenesis is the result of directed movement 3 m/min and reverse their direction of move-
rather than directed growth. How C signal ment every 7 min on average. For reference,
directs cell movement is explained in chapter 4. movement of flagellated bacteria in liquid is
This chapter focuses on the Dif chemosensory 3,000 times faster (21). M. xanthus utilizes two
pathway, which regulates movement and devel- distinct surface motility systems: A (adventur-
opment in different ways. ous) motility mediates the movement of single
The dif and dsp mutants are among the very cells and groups of cells, while S (social) motil-
few mutants to show no evidence of aggrega- ity controls only group translocation (23).The
tion during development. Mapping and com- mechanism of A motility is proposed to be
plementation studies have confirmed that the slime extrusion through polar nozzles (49).
dsp and dif genes are located in the same genetic Cells extrude this material out the lagging end
locus (33). The Dif nomenclature is used of the cell where it forms a trail (55). S motility,
throughout this chapter, although it should be which is mechanistically similar to twitching, is
noted that older references sometimes refer to mediated by attachment and retraction of type
these genes using the dsp name. The dif genes IV pili to pull a cell forward (35, 45; for review,
encode components of a chemosensory path- see reference 37). The ligand recognized by
way known to be essential for S motility, type IV pili may be contained in fibrils, long,
biofilm formation, lipid chemotaxis, and fruit- thin appendages that extend outward from the
ing body development.The role of the Dif pro- cell body and fuse to form an extracellular
matrix (ECM) (2). The fibrils have an
Lawrence J.Shimkets Department of Microbiology,University exopolysaccharide (fEPS) backbone associated
of Georgia,Athens, Georgia 30602. with specific proteins (6). The fEPS has been
65
66 ■ SHIMKETS
partially purified and is composed of galactose, the control of fibril production, which is
glucosamine, glucose, rhamnose, and xylose (6). stimulated by DifACE (52, 54) but not DifD
All mutants that are unable to synthesize fibrils (10). These results predict the presence of an
are defective in S motility (19, 36). Pilin protein alternate interacting partner or substrate for
binds to fEPS in vitro and fEPS stimulates pilus DifE (designated as DifX) that mediates fibril
retraction, suggesting that fEPS is the ligand for synthesis. DifX could be a response regulator
type IV pili (35). controlling fibril gene expression or an enzyme
The relationship between the A-motility required for fibril production or secretion.
“slime” and the S-motility fEPS is murky The difD mutants exhibit an increase in fibril
because neither material has been exhaustively synthesis (10), consistent with the idea that
purified and examined in detail. Genetic DifD competes with DifX for phosphorylation
approaches suggest that the two materials are by DifE.
different.The fEPS is produced by the 37-kb eps Type IV pili function upstream of DifACE
locus and the 1.7-kb eas locus, which have in regulating fibril synthesis (9). Mutations in
homology with genes known to be required for pilA that encode the pilus structural protein and
polysaccharide synthesis in other organisms other mutations that block pilus assembly cause
(36). Mutations in these genes greatly diminish reduced levels of fibrils (9, 19). The nature of
the binding of the textile dyes calcofluor white the pilus-dependent stimulus for fibril produc-
and trypan blue in addition to eliminating S tion is currently unknown but appears not to be
motility (36).While no mutations are known to a pilus retraction-based signal as pilT mutants
eliminate polar slime production, some muta- produce excessive fibrils in spite of being
tions that reduce A motility are located in genes unable to retract their pili.
encoding putative glycosyl transferases that are
different from those required for S motility and BIOFILM FORMATION
do not disrupt S motility (55).Specific stains for A biofilm is a collection of surface-associated
“slime” have not been described. Since most cells that are surrounded by an ECM (20).The
mutations in S-motility genes greatly diminish ECM is composed of exopolysaccharide,
the binding of the textile dyes Congo red, cal- protein, DNA, and other macromolecules
cofluor white, and trypan blue, it appears that (48). The ability to form a biofilm is due to
the bulk of the material staining with these dyes the presence of ligands that enable cells to
is required for S motility (2, 3, 19). Wild-type attach to surfaces (adhesion) and to each other
cells contain a high-affinity Congo red receptor (cohesion). ECM production is not normally
correlated with S motility and a low-affinity constitutive and is typically regulated by envi-
receptor whose function is unknown (3). It is ronmental stimuli.
possible that the low-affinity receptor is the M. xanthus cells adhere to a variety of sur-
polar slime. faces. Cells on plastic submerged under a layer
S motility requires the Dif chemosensory of buffer progress more rapidly through the
system (41, 52). The essential Dif proteins developmental cycle than on an agar surface
include DifACE (8, 52). DifA, a methyl-accept- (32). Although assays for biofilm formation in
ing chemotaxis protein (MCP); DifC, a CheW- other organisms typically use flow cells, biofilm
like coupler; and DifE, a CheA-like histidine formation in M. xanthus has not adopted this
kinase, form a ternary signaling complex (53). technique. In the remainder of this chapter
Mutations in difBDG, also members of the dif biofilm formation refers to adhesive and cohe-
gene cluster, do not eliminate S motility (10) in sive properties of cells that are manifested in an
spite of the fact that DifD is likely a cognate agglutination assay. In this assay a uniform cell
response regulator for DifE (34, 53). Part of the suspension left unperturbed for about 30 min
reason for this apparent paradox is that the main forms large cell clumps that settle out of
contribution of the Dif system to S motility is suspension (41). Agglutination occurs most
5. Dif CHEMOSENSORY SYSTEM IN M. XANTHUS ■ 67
rapidly in the absence of a carbon source and

the presence of divalent cations (41).The cohe-
sion efficiency is high; two cells will stick after
roughly three collisions in cell suspension (2).
The ability to cohere is correlated with the
production of an ECM, which is produced in
long, thin fibrils (2). The fEPS component of
the ECM is likely the same as the fEPS required
for S motility based on the observation that
most S-motility mutants, including difACE, are
defective in both fEPS production and aggluti-
nation (19, 54). ECM production parallels fibril
production and is induced by nutrient depriva-
tion and/or cell-cell contact (2, 7).Type IV pili FIGURE 1 The Dif chemosensory system mediates
and DifACE are required, but not DifBDG (9). lipid chemotaxis and ECM production. The left por-
While the nature of the signal(s) that initiate tion of the figure illustrates ECM production and
chemotaxis to the unique lipid 16:1 PE, both of which
ECM production is unknown, signal percep-
are features of self-sensing. ECM production is induced
tion presumably leads to a change in the phos- by nutrient deprivation in the presence of divalent
phorylation state of DifE (51) and the putative cations (2) and cell-cell contact (7).An unknown signal,
output regulator DifX (Fig 1). possibly generated following type IV pilus binding to
DifACE mutants lack an ECM but can be fEPS (9), activates DifACE, which in turn activates
DifX, a hypothetical interacting protein that induces
agglutinated with wild-type cells (42) or par-
ECM production. For ECM production, activation of
tially purified ECM (17, 54), suggesting that DifA is proposed to stimulate autophosphorylation of
cohesion involves the interaction of an ECM DifE. ECM production is required for directed move-
component with a cell surface component. ment to 16:1 PE due at least in part to a dependence on
Fibril-binding defective (fbd) mutants lack both the ECM-bound zinc metalloprotease FibA whose
substrate is not known. For lipid chemotaxis, activation
the cell surface receptor and the ECM (13).
of DifA is proposed to prevent autophosphorylation of
Neither ligand has been conclusively identi- DifE, which in turn leads to an increased reversal
fied.The ligands involved in adhesion to inert period.The right side of the figure depicts elements of
surfaces are also unknown. the Dif pathway involved in prey sensing.18:1 PE is not
found in M. xanthus and appears to be involved in prey
sensing. Response to 18:1 PE requires only DifE and
LIPID CHEMOTAXIS
DifD, suggesting that a second lipid-sensing system
M. xanthus does not respond chemotactically to converges at DifE. Modified from reference 14 with
a variety of molecules that are chemoeffectors permission from Blackwell Publishing.
in other organisms (21). It was suggested that
chemotaxis to soluble molecules is not possible reversal period (stimulation), the length of time
because of their rapid rate of diffusion relative between direction reversals, but not cellular
to the slow rate of surface motility. M. xanthus velocity (28). Following stimulation by attrac-
cells migrate up gradients of the phospholipid tant there is a decline in the reversal period to
phosphatidylethanolamine (PE) (28).The slow unstimulated levels even in the presence of
diffusion of PE is more suited to slow-moving, attractant (adaptation). Stimulation depends on
surface-motile bacteria than soluble chemoef- the length and saturation of the PE acyl chains
fectors (29). Directed movement of twitching rather than surface-active properties (28, 29).
Pseudomonas aeruginosa cells to PE has also been Signal transduction involves classical prokary-
reported (5, 27). otic transducing proteins (Table 1). The Dif
Several lines of evidence suggest that the M. chemosensory system contains genes for both
xanthus PE response is analogous to chemotaxis stimulation and adaptation (15), while FrzCD
of flagellar bacteria. PE causes an increase in and FrzE are essential for adaptation (28).There
68 ■ SHIMKETS
TABLE 1 Dif and Frz proteins are homologs of classical chemotaxis proteins
Domain Bacterial chemotaxis protein functiona Dif gene cluster Frz gene cluster
MCP Methyl-accepting chemotaxis DifA FrzCD
protein;Tar,Tsr,Trg,Tap,Aer
CheW Coupling protein DifC FrzA, FrzB
CheA Histidine kinase DifE FrzE (CheA-CheY)
CheY Response regulator DifD FrzE (CheA-CheY), FrzZ
(CheY-CheY), FrzS
CheR Methyltransferase None FrzF
CheB Methylesterase None FrzG
CheC Phosphatase DifG None
DifBb None
a
The classical chemotaxis proteins and functions are derived from the E. coli paradigm for chemotaxis, with the exception of CheC,
which is not found in E. coli but acts as a CheY phosphatase in B. subtilis.
b
DifB does not have a homolog among the proteins required for chemotaxis in either E. coli or B. subtilis.
must be some integration between the Dif and activities (28). Whether the diacylglycerol
Frz pathways (40), but thus far that connection derivatives are active because they stimulate the
has not emerged. same receptor or whether the PE is processed
Two naturally occurring lipids are attractants into the diacylglycerol needs to be resolved.
for M. xanthus. PE containing the fatty acid The ECM contains a zinc metalloprotease
18:1ω9c (18:1 PE) is not made by M. xanthus known as FibA that is essential for chemotaxis
(18) and is likely used to locate prey, which to 16:1 PE but not to 18:1 PE (25). FibA is
myxobacteria hydrolyze with extracellular remarkably conserved in M. xanthus strains col-
enzymes and consume. PE containing 16:1ω5c lected around the world (47), consistent with a
(16:1 PE) is a chemoattractant only during star- central role in the life cycle. While the fibA
vation,suggesting a development-specific func- mutant has a modest increase in basal reversal
tion. Fatty acid 16:1ω5c is rare in nature, period (the average length of time between
suggesting a role in self-recognition (18). M. direction reversals), both A- and S-motility
xanthus places much of the 16:1 at the sn-1 motors are intact (25).The fibA deficiency for
position of PE, rather than the traditional sn-2 basal reversal period and PE stimulation can be
position for unsaturated fatty acids, and may complemented by the addition of ECM mate-
have generated a novel chemical signal out of an rial extracted from wild-type, but not fibA, cells
essential structural lipid by deploying a rare (25). A mutation in the FibA active site prevents
fatty acid at an unusual position (18).M.xanthus 16:1 PE stimulation, suggesting that catalytic
also makes other lipid attractants.A strain with activity may be important, but the FibA sub-
reduced levels of 16:1 (11) produces other lipid strate is not known (12).
attractants that have not been identified (L. J. The Dif chemosensory pathway plays at least
Shimkets, unpublished data). three roles in motility. DifACE are required for
Whether PE is the actual attractant or S motility because they provide fEPS, the sub-
whether it is activated by hydrolysis has not strate for type IV pilus attachment and retrac-
been examined. PE forms large molecular tion. DifACE also control the behavior of cells
complexes whose structure is determined by moving with A motility. The difACE mutants
the temperature, fatty acid composition, and exhibit extraordinary long basal reversal peri-
presence of other lipids. Processing of PE into ods under starvation conditions (35 min as
smaller components could simplify the percep- opposed to the normal 7 min), which exceed
tion process. Diacyl glycerol derivatives with even that of frz mutants under comparable con-
the same fatty acid composition stimulate ditions (28), suggesting a link with the bio-
directed movement with comparable specific chemical oscillator for motility control (the
frizilator) (24) (see chapter 4 for more infor- need for stimulation and adaptation rates con-
mation on the frizilator). Finally, DifACE plays sistent with the slow rate of movement. Maxi-
a direct role in lipid signal transduction above mal stimulation occurs within 15 to 30 min,
and beyond stimulating FibA production (15). whereas adaptation takes 1 to 2 h (15, 28).The
Like ECM production, 16:1 PE chemotaxis half-lives of covalently modified proteins are at
is dependent on the presence of the DifACE least an order of magnitude less than the dura-
ternary complex. Unlike ECM production, tion of these responses. Unlike the situation
16:1 PE chemotaxis requires the response with Escherichia coli, several different chemosen-
regulator DifD. Furthermore, nonphosphory- sory pathways converge on a common mecha-
latable DifD mutants (D54N and D54A) have nism for coordinating the behavior of the two
increased basal reversal periods relative to cells motility engines.Given the unusual complexity
with a DifD D10K mutation that is predicted of the organism, it is not surprising that ele-
to mimic the phosphorylated state of the ments of both the Frz pathway (FrzCD and
protein (15).This result, among others, has led FrzE) (28) and Dif pathway (DifB and DifG)
to the prediction that 16:1 PE prevents phos- (15) are required for PE adaptation.
phorylation of DifE, leading to an increase M. xanthus has a novel methylation-
in unphosphorylated DifD (15). This resem- independent mechanism for adapting to 16:1
bles phosphorelay during chemotaxis of PE involving DifG and DifB.The dif gene clus-
enteric bacteria rather than Bacillus where the ter does not contain methyl transferase or
chemoattractant stimulates autophosphoryla- methyl esterase genes, and there is no evidence
tion of the kinase. In contrast to 16:1 PE, 18:1 that DifA is methylated. Mutations in the five
PE is a chemoattractant even in the presence of putative DifA methylation sites that do not
nutrients and does not require the ECM (26). inhibit 16:1 PE stimulation show normal adap-
Of the Dif proteins, only DifE and DifD are tation (15). DifG is homologous to CheC of
required for an 18:1 PE response, indicating Bacillus subtilis (10), which regulates adaptation
that another lipid-sensing system converges at by acting as a CheY phosphatase (46).The difG
DifE (Fig. 1). mutant is partially stimulated by 16:1 PE but
In summary, M. xanthus cells stimulated by fails to adapt (15). DifG has been shown to
PE exhibit suppression of direction changes, interact with DifD using the yeast two-hybrid
adaptation, and the use of a chemosensory sig- system (34, 53), suggesting that DifG may be a
nal transduction system, all of which are hall- DifD-P phosphatase. While the phosphatase
marks of chemotaxis by flagellated bacteria. activity of B. subtilus CheC is activated by
Nevertheless, there are fundamental differences CheD (46),there are no CheD homologs in the
that may reflect adaptation to a solid surface. M. xanthus genome (10), and it remains
Principal among these are (i) the short spatial unknown how DifG activity is regulated. One
range of the attractant, (ii) the long time frame possibility is that DifG may function more like
for stimulation and adaptation, and (iii) the CheX or FliY, both of which dephosphorylate
requirement for the ECM protein FibA for the CheY-P without activation (38, 46).The possi-
16:1 response. ble role of a Che-Y phosphatase in M. xanthus
The short diffusion range of the attractant adaptation is counterintuitive. Since the pres-
suggests that Myxococcus chemotaxis is contact ence of attractant is predicted to increase
based. Myxobacteria cells are frequently unphosphorylated DifD, it is unclear why a
observed to rub against one another (39), and phosphatase would be needed for adaptation.
this behavior could lead to exchange of lipids. The answer may be resolved by learning the
Myxobacteria also engage in trail-following function of DifB, which is a unique protein
behavior like ants (39), and lipids could mark (10).A difB mutant adapts faster than wild type
the trails. One of the most intriguing problems to 16:1 PE, suggesting that DifB slows adapta-
about Dif-mediated lipid chemotaxis is the tion (15). DifB does not appear to interact with
70 ■ SHIMKETS
other Dif chemosensory proteins (34, 53).This the ECM are required for fruiting body mor-
novel adaptation mechanism may reveal how phogenesis. Whether biofilm formation plays
sensory adaptation is slowed to a time frame a direct role in fruiting body development or
consistent with the rate of motility. whether the ECM components support
biofilm formation and fruiting body develop-
THE Dif PATHWAY AND ment in different ways remains unclear. More
DEVELOPMENT information will be required regarding
Fruiting body formation is probably not medi- components of the ECM that are essential for
ated by conventional, long-range chemotaxis as development.
in Dictyostelium. Myxococcus development is While some of the genes in the Dif pathway
most efficient if cells are in contact with each are essential for development (difACE), others
other. Three-dimensional modeling suggests are dispensable (fibA, pilA, and difD). These
that aggregation is directed by contact- results, coupled with the knowledge that only
mediated interactions (43). about 0.3% of the genome (or about 22 genes)
Since DifACE is required for S motility, is essential for development (31), led us to won-
biofilm formation, and lipid chemotaxis, the der whether there are partially redundant
essential nature of DifACE for fruiting body pathways for aggregation that might prevent
development is likely due to a specific require- single mutations from blocking development.
ment for one or more of these processes. Sur- In support of this idea, the pilA fibA combina-
prisingly, S motility is not required for fruiting tion fails to aggregate or produce viable spores
body development (12, 50). While many S- (12). PilA is the structural subunit for the type
motility mutants fail to develop, others com- IV pilus and FibA is a zinc metalloprotease, so it
plete development using the A-motility system. is unlikely that the proteins are providing com-
Comparison of the phenotypes of S-motility plementary biochemical functions. With the
mutants suggests that fibril production rather same approach, pilA difD blocked development
than S motility is the essential ingredient for whereas fibA difD did not.These results suggest
fruiting body morphogenesis. pilA mutants, that difD (the response regulator for 16:1 PE-
which cannot produce the major structural directed movement) and fibA lie on the same
protein pilin, complete fruiting body morpho- branch of the developmental pathway (12).
genesis with a delay of 12 to 24 h (12). The Given the importance of fruiting body devel-
delay is consistent with the observation that opment to the organism, and the fact that
pilA mutants produce reduced but significant myxobacteria have among the largest bacterial
levels of fibrils (9). Conversely, pilT mutants, genomes, it is not surprising that redundant
which cannot retract pili, overproduce fibrils genes and pathways exist.
(9) and develop normally (12).
A much stronger case can be made for a role MATCHING INPUTS AND OUTPUTS
for biofilm formation in fruiting body develop- Enteric bacteria utilize ternary complexes
ment. All mutants that are unable to form a composed of MCP, CheW, and CheA subunits
biofilm, as measured with an agglutination to modulate the phosphorylation state of the
assay, are defective in fruiting body morpho- diffusible response regulator CheY. This path-
genesis (13, 19). Furthermore, mutants that way has a single purpose, chemotaxis, and a sin-
agglutinate at a reduced rate have a noticeable gle output, the flagellar motor. In comparison,
defect in either the rate of fruiting body devel- the Dif chemosensory pathway has an astonish-
opment or the quality of the fruiting bodies. ing level of complexity. Signaling through the
Addition of ECM purified from wild-type cells Dif pathway can result in ECM production or a
restores fruiting body development to mutants change in reversal behavior (Fig 1). ECM pro-
that cannot make an ECM (17, 54). These duction does not influence the reversal behav-
results suggest that one or more components of ior of cells (26). Likewise, there is no evidence
that any of the lipid attractants regulate ECM

biogenesis.These results argue that the sensory
inputs are matched specifically with their
outputs even though they share a core of com-
mon machinery. One possibility is that the
phosphorylation state of DifE is responsible
for determining the output. The signal input
for ECM production is thought to stimulate
phosphorylation of DifE, whereas the 16:1
PE input is thought to prevent DifE autophos-
phorylation (15, 51). Alternatively, two unique
Dif complexes may exist,one for ECM produc-
tion and the other for motility. Cross-linking
assays with E. coli chemoreceptors suggest FIGURE 2 DifA is a dual-function MCP receptor.
that MCP receptors form trimers of dimers The transmembrane domains in the inner membrane
that may include more than one type of MCP (IM), periplasmic loop, and HAMP linker domain
(1, 44). Higher-order structures may provide a (white boxes) are required for ECM production but not
possible mechanism for the generation of lipid chemotaxis. None of the potential methylation
sites in the first methylation domain (MD1) are
unique Dif complexes. involved in ECM production. E110 is required for 16:1
The mechanism by which extracellular PE chemotaxis but not ECM production (black oval).
stimuli are perceived by an MCP and used to Three potential methylation sites in the second methy-
stimulate cells has been examined most closely lation domain (MD2) function in reversal period con-
in E. coli (4).Attractant binding to the periplas- trol for unstimulated cells (white outlines), and one of
these sites is required for ECM production (white oval).
mic exposed region of E.coli MCPs,such as Tar, The highly conserved domain (HCD) and potential
induces conformational changes in the con- methylation sites with no clear function are shown
served cytoplasmic domains that lead to the (black outline). There is no evidence that DifA is
CheW-dependent inhibition of CheA kinase methylated or that methylation is required for stimula-
activity. One hypothesis for the signal transduction or adaptation.This figure summarizes data found
in reference 15 and 51 and is modified from reference
tion in enteric bacteria postulates that CheW 14 with permission from Blackwell Publishing.
and CheA bind to adjacent sites within the
receptor-signaling array in an orientation that
favors CheA autophosphorylation (4). In this cyclase, methyl-accepting chemotaxis protein,
model, regulation of CheA activity by attrac- and phosphatase) linker domains of DifA are
tant is influenced by methylation or amidation replaced by the corresponding domains of the
to shift the alignment of CheW and CheA rela- enteric nitrate sensor NarX are defective in
tive to one another, resulting in a reduction of ECM production (51). ECM production is
kinase activity.The molecular details are lacking restored by nitrate in the presence of DifC and
since the crystal structures of activate and inac- DifE. These results argue that nitrate binding
tivate complexes have not been solved. induces the conformational changes normally
The DifA periplasmic loop is only 10 amino mediated by the unknown ECM biogenesis
acids in length (Fig. 2) (52), suggesting that signal leading to downstream signaling (51).
DifA may sense signals by a mechanism other However, nitrate does not induce a 16:1 PE-
than periplasmic binding. The mechanism of like chemotaxis response (15). If anything,
signal perception was examined by making a nitrate behaves like a repellant and causes a
variety of artificial DifA constructs. M. xanthus slight reduction in the reversal period, the
cells expressing a chimeric chemoreceptor opposite effect of 16:1 PE. NafA-expressing
(NafA) in which the transmembrane, periplas- cells are stimulated by 16:1 PE (in the presence
mic, and HAMP (histidine kinase, adenylyl of nitrate to stimulate ECM production),
72 ■ SHIMKETS
indicating that the cytoplasmic portion of DifA unknown receptor to initiate the response.
is responsible for lipid perception. These and many other mysteries make this an
The DifA cytoplasmic region closely follows interesting organism to study from the perspec-
the E. coli consensus and contains two methyla- tive of behavioral biochemistry and genetics.
tion domains (MD) separated by the highly
conserved domain (HCD) that binds CheW. ACKNOWLEDGMENTS
Mutation Q346D prevents stimulation, but as I am indebted to Dale Kaiser and Zhaomin Yang for
it also causes a defect in ECM production, the critical review of the manuscript.
stimulation defect may be due to loss of FibA. This material is based on work supported by the
Mutation of E110, in the first methylation National Science Foundation under grant no.0343874.
domain, prevents stimulation by 16:1 PE but
does not affect ECM production, suggesting REFERENCES
that this domain has a role in chemosensory 1. Ames, P., C. A. Studdert, R. H. Reiser, and J.
input (Fig. 2). There is considerable variation S. Parkinson. 2002. Collaborative signaling by
in the structure of MCPs and the manner of mixed chemoreceptor teams in Escherichia coli.Proc.
Natl.Acad. Sci. USA 99:7060–7065.
signal perception (56).The periplasmic exposed 2. Arnold, J. W., and L. J. Shimkets. 1988. Cell
domain of B. subtilis McpC is dispensable surface properties correlated with cohesion in
for carbohydrate chemotaxis, and the signal is Myxococcus xanthus. J. Bacteriol. 170:5771–5777.
assimilated in a methylation domain (30). The 3. Arnold, J.W., and L. J. Shimkets. 1988. Inhibi-
cytoplasmic portion of MCPs is constrained by tion of cell-cell interactions in Myxococcus xanthus
by congo red. J. Bacteriol. 170:5765–5770.
a coiled-coil helical bundle (22), making it 4. Baker, M. D., P. M. Wolanin, and J. B. Stock.
unlikely that a cytoplasmic domain directly 2006. Signal transduction in bacterial chemotaxis.
senses ligands. The binding of an interacting Bioessays 28:9–22.
protein partner to a cytoplasmic domain may 5. Barker,A. P.,A. I.Vasil,A. Filloux, G. Ball, P. J.
cause conformational changes that influence Wilderman, and M. L. Vasil. 2004. A novel
extracellular phospholipase C of Pseudomonas
the signaling state of the receptor. Phototaxis aeruginosa is required for phospholipid chemotaxis.
in Halobacterium salinarum is mediated by Mol. Microbiol. 53:1089–1098.
sensory module SRI containing sensory 6. Behmlander, R. M., and M. Dworkin. 1994.
rhodopsin, which interacts with the MCP Biochemical and structural analyses of the extra-
HtrI and changes orientation in response cellular matrix fibrils of Myxococcus xanthus.J.Bacte-
riol. 176:6295–6303.
to light. The methylation and highly con- 7. Behmlander, R. M., and M. Dworkin. 1991.
served domains of the soluble MCP FrzCD Extracellular fibrils and contact-mediated cell
are required for sensing the repellant isoamyl interactions in Myxococcus xanthus. J. Bacteriol.
alcohol (IAA) (16). While it is not clear if the 173:7810–7820.
C terminus of FrzCD directly senses IAA, a 8. Bellenger, K., X. Ma, W. Shi, and Z. Yang.
2002. A CheW homologue is required for Myxo-
constitutive signaling state of FrzCD is induced coccus xanthus fruiting body development, social
by deletion of the first 16 amino acids of the gliding motility, and fibril biogenesis. J. Bacteriol.
first methylation domain or the last 25 amino 184:5654–5660.
acids of the second methylation domain.These 9. Black,W. P., Q. Xu, and Z.Yang. 2006.Type IV
deletions may mimic conformational changes pili function upstream of the Dif chemotaxis path-
way in Myxococcus xanthus EPS regulation. Mol.
induced by protein-protein interactions, or Microbiol. 61:447–456.
alternately, site-specific methylation may be 10. Black,W. P., and Z.Yang. 2004. Myxococcus xan-
involved in sensory transduction (16). thus chemotaxis homologs DifD and DifG nega-
It is also possible the PE activates directed tively regulate fibril polysaccharide production. J.
movement by becoming directly incorporated Bacteriol. 186:1001–1008.
11. Bode, H. B., M. W. Ring, D. Kaiser, A. C.
into the membrane.While DifA is already sur- David, R. M. Kroppenstedt, and G. Schwar.
rounded by 16:1 PE, incorporated 18:1 PE 2006. Straight-chain fatty acids are dispensable in
could potentially alter the conformation of its the myxobacterium Myxococcus xanthus for vegeta-
tive growth and fruiting body formation. J. Bacte- 25. Kearns, D. B., P. J. Bonner, D. R. Smith, and
riol. 188:5632–5634. L. J. Shimkets. 2002. An extracellular matrix-
12. Bonner, P. J., W. P. Black, Z. Yang, and L. J. associated zinc metalloprotease is required for
Shimkets. 2006. FibA and PilA act cooperatively dilauroyl phosphatidylethanolamine chemotactic
during fruiting body formation of Myxococcus xan- excitation in Myxococcus xanthus. J. Bacteriol.
thus. Mol. Microbiol. 61:1283–1293. 184:1678–1684.
13. Bonner, P. J., and L. J. Shimkets. 2006. 26. Kearns, D. B., B. D. Campbell, and L. J.
Cohesion-defective mutants of Myxococcus xan- Shimkets. 2000. Myxococcus xanthus fibril
thus. J. Bacteriol. 188:4585–4588. appendages are essential for excitation by a phos-
14. Bonner, P. J., and L. J. Shimkets. 2006. Phos- pholipid attractant. Proc. Natl. Acad. Sci USA
pholipid directed motility of surface-motile bacte- 97:11505–11510.
ria. Mol. Microbiol. 61:1101–1109. 27. Kearns, D. B., J. Robinson, and L. J. Shimkets.
15. Bonner, P. J., Q. Xu, W. P. Black, Z. Li, Z. 2001. Pseudomonas aeruginosa exhibits directed
Yang, and L. J. Shimkets. 2005. The Dif twitching motility up phosphatidylethanolamine
chemosensory pathway is directly involved in gradients. J. Bacteriol. 183:763–767.
phosphatidylethanolamine sensory transduction in 28. Kearns, D. B., and L. J. Shimkets. 1998.
Myxococcus xanthus. Mol. Microbiol. 57:1499–1508. Chemotaxis in a gliding bacterium. Proc. Natl.
16. Bustamante,V. H., I. Martinez-Flores, H. C. Acad. Sci. USA 95:11957–11962.
Vlamakis, and D. R. Zusman. 2004.Analysis of 29. Kearns, D. B., and L. J. Shimkets. 2001. Lipid
the Frz signal transduction system of Myxococcus chemotaxis and signal transduction in Myxococcus
xanthus shows the importance of the conserved xanthus.Trends Microbiol 9:126–129.
C-terminal region of the cytoplasmic chemore- 30. Kristich, C. J., G. D. Glekas, and G.W. Ordal.
ceptor FrzCD in sensing signals. Mol. Microbiol. 2003. The conserved cytoplasmic module of the
53:1501–1513. transmembrane chemoreceptor McpC mediates
17. Chang, B.Y., and M. Dworkin. 1994. Isolated carbohydrate chemotaxis in Bacillus subtilis. Mol.
fibrils rescue cohesion and development in the Microbiol. 47:1353–1366.
Dsp mutant of Myxococcus xanthus. J. Bacteriol. 31. Kroos, L., A. Kuspa, and D. Kaiser. 1986. A
176:7190–7196. global analysis of developmentally regulated genes
18. Curtis, P. D., R. Geyer, D. C.White, and L. J. in Myxococcus xanthus. Dev. Biol. 117:252–266.
Shimkets. 2006. Novel lipids in Myxococcus xan- 32. Kuner, J. M., and D. Kaiser. 1982.Fruiting body
thus and their role in chemotaxis. Environ. Micro- morphogenesis in submerged cultures of Myxococ-
biol. 8:1935–1949. cus xanthus. J. Bacteriol. 151:458–461.
19. Dana, J. R., and L. J. Shimkets. 1993. Regula- 33. Lancero, H., J. E. Brofft, J. Downard, B. W.
tion of cohesion-dependent cell interactions in Birren, C. Nusbaum, J. Naylor,W. Shi, and L.
Myxococcus xanthus. J. Bacteriol. 175:3636–3647. J. Shimkets. 2002.Mapping of Myxococcus xanthus
20. Davies, D. G., M. R. Parsek, J. P. Pearson, social motility dsp mutations to the dif genes. J. Bac-
B. H. Iglewski, J. W. Costerton, and E. P. teriol. 184:1462–1465.
Greenberg. 1998.The involvement of cell-to-cell 34. Lancero, H. L., S. Castaneda, N. B. Caberoy,
signals in the development of a bacterial biofilm. X. Ma,A. G. Garza, and W. Shi. 2005.Analysing
Science 280:295–298. protein-protein interactions of the Myxococcus xan-
21. Dworkin, M., and D. Eide. 1983. Myxococcus thus Dif signalling pathway using the yeast two-
xanthus does not respond chemotactically to hybrid system. Microbiology 151:1535–1541.
moderate concentration gradients. J. Bacteriol. 35. Li, Y., H. Sun, X. Ma, A. Lu, R. Lux, D.
154:437–442. Zusman, and W. Shi. 2003. Extracellular poly-
22. Falke, J. J., and S. H. Kim. 2000. Structure of a saccharides mediate pilus retraction during social
conserved receptor domain that regulates kinase motility of Myxococcus xanthus. Proc. Natl.Acad. Sci.
activity: the cytoplasmic domain of bacterial taxis USA 100:5443–5448.
receptors. Curr. Opin. Struct. Biol. 10:462–469. 36. Lu, A., K. Cho, W. P. Black, X. Y. Duan, R.
23. Hodgkin, J., and D. Kaiser. 1979. Genetics of Lux, Z. Yang, H. B. Kaplan, D. R. Zusman,
gliding motility in Myxococcus xanthus (Myxobac- and W. Shi. 2005.Exopolysaccharide biosynthesis
terales): two gene systems control movement. Mol. genes required for social motility in Myxococcus
Gen. Genet. 171:177–191. xanthus. Mol. Microbiol. 55:206–220.
24. Igoshin, O.A.,A. Goldbeter, D. Kaiser, and G. 37. Nudleman, E., and D. Kaiser. 2004. Pulling
Oster. 2004. A biochemical oscillator explains together with type IV pili. J. Mol. Microbiol Biotech-
several aspects of Myxococcus xanthus behavior nol 7:52–62.
during development. Proc. Natl. Acad. Sci. USA 38. Park, S. Y., X. Chao, G. Gonzalez-Bonet,
101:15760–15765. B. D. Beel, A. M. Bilwes, and B. R. Crane.
74 ■ SHIMKETS
2004. Structure and function of an unusual xanthus at the centimeter scale. Appl. Environ.
family of protein phosphatases: the bacterial Microbiol. 72:3615–3625.
chemotaxis proteins CheC and CheX. Mol. Cell 48. Whitchurch, C. B., T. Tolker-Nielsen, P. C.
16:563–574. Ragas, and J. S. Mattick. 2002. Extracellular
39. Reichenbach, H., H. Heunert, and H. DNA required for bacterial biofilm formation.
Kuczka. 1965. Myxococcus spp. (Myxobacterales)- Science 295:1487.
Schwarmentwicklung und bildung von protocys- 49. Wolgemuth, C., E. Hoiczyk, D. Kaiser, and G.
ten. Film E779 Inst. f. Wissen., Gottingen, Oster. 2002. How myxobacteria glide. Curr Biol.
Germany. 12:369–377.
40. Shi, W., Z.Yang, H. Sun, H. Lancero, and L. 50. Wu, S. S., J. Wu, Y. L. Cheng, and D. Kaiser.
Tong. 2000. Phenotypic analyses of frz and dif 1998.The pilH gene encodes an ABC transporter
double mutants of Myxococcus xanthus. FEMS homologue required for type IV pilus biogenesis
Microbiol. Lett. 192:211–215. and social gliding motility in Myxococcus xanthus.
41. Shimkets, L. J. 1986. Correlation of energy- Mol. Microbiol. 29:1249–1261.
dependent cell cohesion with social motility in 51. Xu, Q.,W. P. Black, S. M.Ward, and Z.Yang.
Myxococcus xanthus. J. Bacteriol. 166:837–841. 2005. Nitrate-dependent activation of the Dif sig-
42. Shimkets, L. J. 1986. Role of cell cohesion in naling pathway of Myxococcus xanthus mediated by
Myxococcus xanthus fruiting body formation. J. Bac- a NarX-DifA interspecies chimera. J. Bacteriol.
teriol. 166:842–848. 187:6410–6418.
43. Sozinova, O., Y. Jiang, D. Kaiser, and M. 52. Yang, Z.,Y. Geng, D. Xu, H. B. Kaplan, and
Alber. 2005. A three-dimensional model of W. Shi. 1998. A new set of chemotaxis homo-
myxobacterial aggregation by contact-mediated logues is essential for Myxococcus xanthus social
interactions. Proc. Natl. Acad. Sc.i USA 102: motility. Mol. Microbiol. 30:1123–1130.
11308–11312. 53. Yang, Z., and Z. Li. 2005. Demonstration of
44. Studdert, C. A., and J. S. Parkinson. 2004. interactions among Myxococcus xanthus Dif
Crosslinking snapshots of bacterial chemoreceptor chemotaxis-like proteins by the yeast two-hybrid
squads. Proc. Natl.Acad. Sci. USA 101:2117–2122. system. Arch. Microbiol. 183:243–252.
45. Sun, H., D. R. Zusman, and W. Shi. 2000.Type 54. Yang, Z., X. Ma, L. Tong, H. B. Kaplan, L. J.
IV pilus of Myxococcus xanthus is a motility appara- Shimkets, and W. Shi. 2000. Myxococcus xanthus
tus controlled by the frz chemosensory system. dif genes are required for biogenesis of cell surface
Curr. Biol. 10:1143–1146. fibrils essential for social gliding motility. J. Bacte-
46. Szurmant, H., T. J. Muff, and G. W. Ordal. riol. 182:5793–5798.
2004. Bacillus subtilis CheC and FliY are members 55. Yu, R., and D. Kaiser. 2007. Gliding motility
of a novel class of CheY-P-hydrolyzing proteins in and polarized slime secretion. Mol. Microbiol.
the chemotactic signal transduction cascade. J. Biol. 63:454–467.
Chem. 279:21787–21792. 56. Zhulin, I. B. 2001.The superfamily of chemotaxis
47. Vos, M., and G. J.Velicer. 2006.Genetic popula- transducers: from physiology to genomics and
tion structure of the soil bacterium Myxococcus back. Adv. Microb. Physiol. 45:157–198.
HETEROCYST DEVELOPMENT
AND PATTERN FORMATION
6
Multicellularity has arisen several times during operates in vegetative cells that continue to
evolution, and although common to eukary- grow and divide (99). For Anabaena and Nostoc
otes, it is also found in prokaryotes (52). The strains, nitrogen-limiting conditions induce
generation of multicellularity and cell-type about 5 to 10% of cells along a filament to
diversity enables an organism to acquire spe- differentiate, resulting in a one-dimensional
cialized functions and advantages in feeding, developmental pattern of single heterocysts
dispersion, and protection (52). One example separated by an average of 10 to 20 vegetative
of prokaryotic multicellularity that clearly cells (Fig. 1B); this regulated pattern suggests
exhibits cellular differentiation and the forma- the existence of cell-to-cell communication
tion of a multicellular pattern is found in fila- within filaments (40). Filaments grown in the
mentous nitrogen-fixing cyanobacteria, e.g., presence of a combined nitrogen source consist
the genera Anabaena and Nostoc. Cyanobacteria of vegetative cells only (Fig.1A).Differentiating
comprise a diverse group of gram-negative cells pass through an intermediary prohetero-
prokaryotes that perform oxygenic photosyn- cyst stage, when differentiation is reversible
thesis. Some are also able to “fix” nitrogen if a source of combined nitrogen becomes
by reducing atmospheric dinitrogen to ammo- available. Proheterocysts become committed to
nium. Nitrogen fixation and oxygenic photo- complete heterocyst differentiation between
synthesis are incompatible processes because 9 to 14 h after nitrogen step-down (106), and
the nitrogenase enzyme is very oxygen- differentiation is completed by 18 to 24 h.
sensitive. Many filamentous cyanobacteria that Heterocysts undergo specific changes in
fix nitrogen overcome this incompatibility by physiology and morphology to generate the
undergoing cellular differentiation, which microoxic environment needed to accommo-
spatially separates the two processes; nitrogen date nitrogen fixation (101).The O2-producing
fixation takes place in terminally differentiated photosystem II, which is part of the photosyn-
cells called heterocysts and photosynthesis thetic electron transport chain, is dismantled
during heterocyst differentiation. An envelope
that consists of two layers encircles the hetero-
Department of Biology,Texas A&M University, College Sta- cyst and constrains the influx of environmental
tion,Texas 77843. oxygen. The inner layer is composed of a
75
76 ■ ALDEA ET AL.
PCC 7120, Anabaena variabilis ATCC 29413,

and Nostoc punctiforme ATCC 29133. All three
of these genomes have been sequenced as well
as those of over two dozen other cyanobacterial
species, which facilitates bioinformatics analy-
ses. Numerous genes have been identified that
are involved in heterocyst development and/or
function. Here, we will focus on those genes
involved in signaling and regulation.
Sensing Nitrogen Limitation

Ammonium is usually the preferred source of
FIGURE 1 Wild-type and mutant filaments of nitrogen for bacteria, including heterocyst-
Anabaena PCC 7120. (A) Wild-type filaments grown forming cyanobacteria. Its presence exerts an
on medium containing nitrate consist of vegetative cells inhibitory effect on assimilatory pathways of
only. (B) Wild-type filaments grown on medium with- alternative nitrogen sources and also strongly
out a source of combined nitrogen have a pattern of
single heterocysts spaced along filaments. (C) A patS
inhibits heterocyst development; nitrate has a
deletion mutant (strain AMC451) grown on medium lesser inhibitory effect (101). In Escherichia
without a source of combined nitrogen has an Mch coli,the intracellular concentration of glutamine
phenotype. (D) Green fluorescent protein fluorescence serves as the signal for nitrogen status and
of a PpatS-gfp reporter strain (AMC484) grown on 2-oxoglutarate (2-OG) serves as the signal for
medium without a source of combined nitrogen is
localized to heterocysts (lower panel).Arrowheads indi-
carbon status (49, 71). Intracellular nitrogen sta-
cate heterocysts. tus is sensed by PII (encoded by glnB or glnK),
a protein that plays a central role in interpreting
hydroxylated glycolipid, and the outer layer is and responding to changes in key carbon
composed of polysaccharide. Traces of oxygen and nitrogen metabolites (72). A UTase/UR
that cross the envelope are consumed by enzyme (encoded by glnD) controls the activity
increased respiration in heterocysts. Heterocysts of PII by uridylylation and deuridylylation in
are thought to become dependent on neigh- response to levels of glutamine (71).The uridy-
boring vegetative cells for reduced carbon, and lylated form of PII can interact with adenyl-
in return, the heterocysts provide fixed nitrogen transferase (encoded by glnE) to activate
as amino acids that probably move through the glutamine synthase (GS). Native PII protein
periplasm (32) to the vegetative cells (Fig. 2). interacts with NtrB to modify the phosphory-
Thus, these two functionally distinct cell types lation status of NtrC, and in turn, NtrC~P acts
must collaborate to support diazotrophic as a transcriptional regulator of various genes
growth and communicate to regulate the devel- involved in nitrogen metabolism including
opmental pattern for efficiency. Heterocystous glnA, which encodes GS (23, 71).
cyanobacteria provide a relatively simple model The cyanobacterial nitrogen sensory system
of prokaryotic multicellular development in diverges from that of E. coli but is similar
which two cell types must continuously interact between unicellular and filamentous cyanobac-
to support growth of the organism. teria. In the unicellular cyanobacterium Syne-
chococcus elongatus PCC 7942, PII is not
REGULATION OF HETEROCYST modified by uridylylation, as is the case in E.
DEVELOPMENT coli, but is instead phosphorylated (34). In hete-
Most of the information about heterocyst rocyst-forming filamentous cyanobacteria, PII
development to date is based on the study of is modified, but the nature of the modification
three species of heterocyst-forming filamentous is not yet clear; phosphorylation of recombi-
cyanobacteria: Anabaena (also Nostoc) sp. strain nant PII was achieved in vitro but not in vivo
6. HETEROCYST DEVELOPMENT AND PATTERN FORMATION ■ 77
HetN
PatS
Amino acids
Carbohydrates Carbohydrates
CO2 CO2
FIGURE 2 Signal and metabolite flow between heterocysts and vegetative cells.Wild-type filament
with a normal pattern of heterocysts (marked by the arrowheads) grown on medium without a source
of combined nitrogen.Vegetative cells provide carbohydrates produced by photosynthesis to hetero-
cysts,which in turn provide fixed nitrogen as amino acids to the reproductive vegetative cells.Hetero-
cyst pattern is controlled by PatS- and HetN-dependent signals and by the supply of nitrogen from
heterocysts.The transfer and movement of small molecules along filaments is thought to occur via the
periplasm (32). Horizontal arrows indicate the apparent effective range of each signal or metabolite.
(33, 41). Functional analysis of PII in N. puncti- a nonmetabolizable analog of 2-OG, DFPA
forme has been impeded by the fact that glnB (2, 2-difluoropentanedioic acid), to mimic
seems to be essential (42). In contrast to pro- nitrogen-limiting conditions and provoke
teobacteria where PII activity is modified in heterocyst development (57).
response to glutamine levels, in cyanobacteria, Another major difference in the mechanism
PII is modified by phosphorylation in response of nitrogen control present in cyanobacteria is
to 2-OG levels (33). the lack of the RpoN (sigma 54) sigma factor,
The 2-OG molecule is an intermediate in and of RpoN-dependent transcription factors,
the Krebs cycle and is considered to be the including NtrC. In cyanobacteria, nitrogen
metabolic junction between carbon and nitro- control is mediated by NtcA, a transcription
gen balance in bacteria. In cyanobacteria, the factor in the cAMP receptor protein family of
Krebs cycle is incomplete because they lack 2- DNA-binding proteins. NtcA regulates various
OG dehydrogenase (101); as a result, the main genes important for nitrogen and carbon
function of 2-OG in cyanobacteria is to serve as metabolism (46, 90). In Anabaena PCC 7120,
the primary carbon skeleton for the incorpora- ntcA mutants are not able to use nitrate as a sole
tion of ammonium.Therefore, 2-OG can serve nitrogen source and require ammonium for
as a measure of nitrogen status.Variations in the growth; moreover, they show no signs of
nitrogen supply are inversely correlated with heterocyst differentiation, which indicates an
levels of 2-OG; thus, nitrogen-limiting condi- essential role for NtcA in the initiation of dif-
tions produce an increase in 2-OG (77). The ferentiation (Fig. 3) (35, 46, 98).The ntcA gene
hypothesis that 2-OG controls heterocyst is induced after nitrogen deprivation and is pos-
development has been confirmed by use of itively autoregulated (46, 82), and is induced
78 ■ ALDEA ET AL.
FIGURE 3 Model showing the influence of cell-to-cell signals on heterocyst development and pattern forma-
tion. Nitrogen starvation results in accumulation of 2-oxoglutarate (2-OG), which leads to increased NtcA activity.
Initiation of heterocyst development is controlled mainly by NtcA and HetR,which are positively autoregulated and
mutually dependent on each other for upregulation; HetR also has autoproteolytic activity.The DNA-binding pro-
tein NrrA serves as a regulatory link between NtcA and HetR. It is directly activated by NtcA and then causes the
initial induction of hetR;further induction of hetR relies on autoregulatory positive feedback.HetR is considered the
master regulator of heterocyst development. NtcA and HetR collaborate to reduce the levels of CcbP in the differ-
entiating cell, the former by directly inhibiting transcription and the latter by proteolysis, resulting in release of free
calcium. NtcA is required for early as well as late stages of heterocyst development. HetF positively influences hete-
rocyst development by an unknown mechanism.Three factors produced by the differentiating cell are proposed to
influence heterocyst development and pattern formation by acting as cell-to-cell signals (enclosed in boxes). Fixed
nitrogen originating from the differentiating cells and mature heterocysts in the form of amino acids negatively influ-
ence differentiation of neighboring cells and establish the ultimate spacing between heterocysts. Production of PatS
is directly activated by HetR in differentiating cells,and it is thought that PatS,or a posttranslationally processed pep-
tide, is secreted and then enters neighboring cells where it inhibits HetR activity to block differentiation. HetN is
also thought to generate an inhibitory signal from mature heterocysts that inhibits differentiation of adjacent cells,
possibly by interfering with HetR activity. PatA may partially relieve the inhibitory effect of the HetN- and PatS-
dependent signals.The dashed arrow indicates putative posttranslational processing of PatS, and the large arrows rep-
resent other developmental steps that take place between the activation of HetR and completion of the
differentiation process.
most strongly in differentiating cells (79). In S. dependent genes requires PII, and when nitrate
elongatus, NtcA DNA-binding activity is is present, PII has an inhibitory effect on NtcA
enhanced by the presence of 2-OG and, more- activity (3). Additionally, NtcA regulates PII at
over, transcriptional activation by NtcA the transcriptional and posttranslational level
requires 2-OG (91, 96). The 2-OG analog (28, 81). Recently, a search for interacting part-
DFPA also stimulates DNA-binding of NtcA ners of nitrogen regulators found that the previ-
(57). Taken together, these data indicate that ously identified PipX protein, which is present
NtcA is a sensor of 2-OG that upregulates only in cyanobacteria,acts as a link between the
expression of genes for permeases and enzymes key nitrogen regulators PII and NtcA. PipX
of the nitrogen assimilatory pathways needed swaps between theses two partners in a 2-OG-
for utilizing alternative sources of nitrogen, and dependent manner (27).
triggers heterocyst differentiation in heterocys-
tous strains (46). In most cases, genes that are HetR Is a Master Regulator
directly controlled by NtcA have the signature of Heterocyst Development
binding site TGTA-(N8)-TACA centered at HetR plays a central role in heterocyst develop-
41.5 with respect to the transcriptional start ment and pattern formation (Fig. 3) (7, 14, 54,
site (46, 50). In the unicellular cyanobacterium 108). HetR is a positive regulatory factor that is
S. elongatus, full activation of some NtcA- essential for heterocyst development.The hetR
gene is expressed early in differentiating cells this regulation may result from the HetR-
after nitrogen step-down.Transcription of hetR dependent increase of NtcA levels (46, 78).
is positively autoregulated and begins to Other genes such as hetC,which is also required
increase 30 min after nitrogen deprivation, and for the early stage of heterocyst differentiation,
by 3.5 h PhetR-luxAB expression is localized to a exhibit only NtcA dependence (75). NrrA,
subset of cells, well before morphological dif- a response regulator, has recently been identi-
ferentiation is observed (7, 13). HetR has two fied as the regulatory link between NtcA and
known activities, autoproteolysis and specific HetR.The nrrA gene is directly dependent on
DNA-binding that requires formation of a NtcA and is transcribed in differentiating cells
homodimer (48, 110). Mutations that interfere within 3 h after nitrogen deprivation (Fig. 3)
with either of these activities block heterocyst (25, 74). In turn, NrrA binds specifically to the
development at an early stage. HetR is an hetR promoter region and is required for
unusual serine-type protease. Two serine increased hetR expression (24). In the absence
residues, Ser-152 and Ser-179, are required for of nrrA, HetR accumulation and, thus, hetero-
autoproteolysis and heterocyst differentiation, cyst development are delayed (25), and extra
with Ser-152 thought to be in the active site of copies of nrrA result in increased expression of
the protease (22). However, the role of Ser-152 hetR and increased heterocyst frequency (24).
has been recently questioned (84). In hetR
S179N and S152A mutants, HetR overaccu- Ca2 Has a Regulatory Role
mulates in filaments, but these mutants are het- in Heterocyst Development
erocyst defective (22). The autoproteolytic In bacteria, calcium ions play important roles in
activity may be important for regulating the various cellular processes such as pathogenesis,
accumulation of HetR in only differentiating sporulation in Bacillus, chemotaxis in E. coli, and
cells because hetR is, at least initially, transcribed heterocyst development in cyanobacteria (45,
in all cells. 87, 89, 94, 95). An early account of Ca2
Increased HetR levels or activity is sufficient involvement in heterocyst development found
to force heterocyst differentiation. Overexpres- that manipulation of the extracellular concen-
sion of hetR on a multicopy plasmid, either tration of calcium influences heterocyst fre-
from its native promoter or from the copper- quency and nitrogenase activity (87). A recent
regulated petE promoter, leads to increased het- study used the Ca2-binding luminescent pro-
erocyst frequency regardless of the presence of tein aequorin as a reporter to track intracellular
nitrate or ammonium (13).Ectopic overexpres- calcium changes in cultures during heterocyst
sion of a mutant allele of hetR (hetRR223W) is development (95). A transient increase prima-
able to bypass the main inhibitory signals of rily due to mobilization of internally stored
heterocyst pattern formation and results in a Ca2 occurs at about 1 h after nitrogen step-
conditionally lethal phenotype caused by com- down. In a hetR mutant, the Ca2 concentra-
plete differentiation of nearly all cells under tion increase is not disrupted, suggesting that
nitrogen-limiting conditions (54). the Ca2 signal acts earlier than hetR during
During heterocyst development, ntcA and heterocyst development (95). Subsequently, the
hetR exhibit a mutual dependency (Fig. 3) (75). use of obelin, a different Ca2 reporter, allowed
Activation of hetR expression in the early stage monitoring of intracellular calcium levels in
of heterocyst development precedes that of the individual cells (109). Heterocysts have a 10-
ntcA gene, suggesting that the NtcA-dependent fold higher calcium concentration than vegeta-
boost in hetR expression is mediated by preex- tive cells and the increase occurs 4 h after
isting NtcA protein (46). Expression of some removal of combined nitrogen. The free cal-
genes involved in heterocyst development,such cium levels are inversely correlated with the
as devH and the devBCA operon, is dependent expression of ccbP, which encodes a calcium-
on both ntcA and hetR (31, 44). In some cases, sequestering protein in Anabaena PCC 7120
80 ■ ALDEA ET AL.
and other heterocyst-forming cyanobacteria overexpressed, suggesting that patA acts down-
(109). In the absence of combined nitrogen, stream of hetR in the regulatory pathway con-
inactivation of ccbP results in a multiple- trolling heterocyst development (13, 62). PatA
contiguous-heterocyst (Mch) phenotype, may influence heterocyst development by
whereas overexpression has an inhibitory effect attenuating the negative effects of the main
on hetR induction and heterocyst development. inhibitory signals of heterocyst pattern forma-
Expression of ccbP is down-regulated in hetero- tion, PatS and HetN (80). This could result if
cysts, and CcbP is not present in mature hetero- PatA modifies or interacts with HetR to make
cysts (109). A regulatory pathway comprising it less sensitive to the inhibitory signals.
HetR, CcbP, and NtcA controls intracellular The patB gene was originally thought to be
free calcium during heterocyst development involved in heterocyst pattern formation
(Fig. 3) (86). HetR specifically degrades because the original mutant had a Mch pheno-
CcbP in a Ca2-dependent manner, and ccbP type (63). PatB contains an N-terminal domain
down-regulation in differentiating cells requires with two putative 4Fe-4S centers and a C-
2-OG-dependent binding of NtcA to its pro- terminal domain with a DNA-binding motif.It
moter region. The increase in free calcium in is now clear that patB is required for growth and
differentiating cells may regulate the Ca2- survival in the absence of combined nitrogen
dependent serine protease activity of HetR and (51). A PpatB-gfp reporter is expressed in hetero-
other Ca2-dependent proteases (65). It has also cysts 16 h after nitrogen step-down (51).A patB
been suggested that PII modification is inhib- frameshift mutant has a delayed Mch pheno-
ited by calcium (66), which correlates with the type in the absence of combined nitrogen,
observation that PII is unmodified in hetero- whereas a deletion mutant impairs growth and
cysts, which is a requirement for normal nitro- nitrogen fixation within 24 h of combined
gen metabolism in heterocysts (58). nitrogen step-down. The pattern formation
defect of patB mutants is apparently a result of
Other Genes Involved in the defective heterocyst function because the phe-
Regulation of Heterocyst Development notype is similar to other mutants defective for
The hetF gene is present in all heterocystous nitrogen fixation, which have a mild Mch phe-
cyanobacteria, but the predicted HetF protein notype and decreased numbers of vegetative
has no similarity to proteins of known function cells in the intervals between heterocysts (J.
(102).In N.punctiforme,hetF mutants do not ini- Golden, unpublished data).
tiate heterocyst development and HetR accu- The hetC gene encodes an ABC-type
mulates nonspecifically in all cells, while hetF exporter and is required for heterocyst differen-
overexpression produces a Mch phenotype but tiation (53).A PhetC-gfp reporter shows increased
only in the absence of combined nitrogen. expression in differentiating cells (53) and hetC
HetF is thought to have a role in constraining expression is NtcA-dependent (76). Inactiva-
the accumulation of active HetR protein and tion of the hetC gene blocks heterocyst devel-
increased hetR expression to differentiating opment at an early stage, resulting in a pattern
cells (102). of weakly autofluorescent cells that express a
PatA is a response regulator that contains a PhetR-gfp reporter (104); these cells are still able
CheY-like phosphoacceptor domain at its C to divide under certain conditions (104). It is
terminus and a newly identified domain, possible that HetC is required for normal mor-
PATAN, at its N terminus, which may be phogenesis of the heterocyst envelope and the
involved in protein-protein interactions (62, absence of HetC results in a regulatory check-
67). Inactivation of patA causes heterocysts to point that blocks further differentiation.
form almost exclusively at the ends of filaments Overexpression of the hetL gene strongly
regardless of their length (62).This phenotype is stimulates heterocyst development (64). The
maintained even when hetR or hetRR223W is predicted HetL protein is composed almost
entirely of pentapeptide repeats with a consen- rocyst development (69).The ordered sequence
sus of A(D/N)L*X, where * is a polar amino of events during heterocyst development may
acid. Pentapeptide repeat proteins may resem- be the result of a hierarchy of transcriptional
ble DNA in structure, and some members of regulators, possibly including sigma factors, that
the family bind to and inhibit DNA gyrase (97). control expression of different sets of genes at
Anabaena PCC 7120 contains 30 genes encod- particular times (15, 55). The principal sigma
ing proteins that contain this motif. The hetL factor encoded by sigA is expressed in all cells in
gene was identified because its overexpression the presence or absence of combined nitrogen
suppresses the inhibition of heterocyst dif- and it is essential for viability (12, 85). We
ferentiation caused by extra copies of PatS. showed that some combinations of double
Overexpression of hetL stimulates heterocyst mutants, such as sigD sigB2 and sigB2 sigE,
development even in an ntcA mutant back- had deficiencies in establishing diazotrophic
ground; however, the differentiation of the ntcA growth, but none of the tested alternative sigma
mutant is blocked at an intermediate stage,indi- factor genes, inactivated individually or in pairs
cating that NtcA is required for later stages of (sigB, sigC, sigD, sigB2, and sigE), were found to
differentiation (46, 64). Heterocyst develop- be essential for development (11, 55). However,
ment and diazotrophic growth appear to be our recent studies using a gfp reporter to follow
normal in a hetL knockout mutant, showing the expression of eight sigma factor genes
that HetL is not essential for normal heterocyst found that three, sigC, sigE, and sigG, were
development. Further study is needed to deter- upregulated in differentiating cells at different
mine if HetL is normally involved in heterocyst times during heterocyst development.
development or if the overexpression pheno-
type is caused indirectly, possibly through inter- LATER STAGES OF HETEROCYST
actions with other proteins that contain the DIFFERENTIATION AND
MORPHOGENESIS
pentapeptide repeat motif. Based on the char-
acteristics of the pentapeptide repeat family of Synthesis of Heterocyst Envelope
proteins, HetL might interact with a DNA- Late stages of heterocyst development are char-
binding protein or transcription factor involved acterized by structural changes that include the
in regulating heterocyst development. deposition of three cell layers: an outermost
The Anabaena PCC 7120 genome harbors fibrous layer, an envelope polysaccharide layer,
12 genes that encode putative sigma factors: 9 and an innermost glycolipid layer.The hetero-
on the chromosome, sigA (all5263), sigB2 cyst envelope is thought to limit the entry of
(alr3800, previously sigE), sigC (all1692), sigD oxygen into the heterocyst to provide it with a
(alr3810), sigE (alr4249, previously sigF), sigF microoxic environment, which is vital for the
(all3853), sigG (alr3280, previously sigma-E), function of nitrogenase (101). The outermost
sigI (all2193), and sig J (alr0277, previously layer also may be involved in associations with
sigma-37); and 3 on plasmids, sigB (all7615), specific heterotrophic epibiotic bacteria that
sigB3 (all7608, previously sigH), and sigB4 may have a mutualistic relationship with hete-
(all7179, previously sigG) (11, 12, 55, 56, 107). rocystous cyanobacteria (88). Mutants that lack
The sigma factor nomenclature for Anabaena the envelope polysaccharide or glycolipid layer
PCC 7120 has recently been modified by are unable to grow diazotrophically in air (47,
Yoshimura et al. (107), and our own phyloge- 100). A cluster of hep genes is required for the
netic grouping of cyanobacterial sigma factors deposition of the polysaccharide layer (47,
is in agreement; therefore, we have adopted the 111). The heterocyst-specific glycolipids are
suggested nomenclature changes. composed of fatty alcohols glycosidically linked
In Anabaena PCC 7120, numerous genes are to sugar residues (29).The hgl genes are required
expressed or upregulated only in differentiating for the glycolipid layer and are expressed during
cells at specific times during the course of hete- the middle stage of differentiation around the
82 ■ ALDEA ET AL.
time that cells become committed to form het- genes are expressed late during heterocyst
erocysts (5, 6, 17, 26, 30, 47, 59).The devH gene development between 18 and 24 h after nitro-
encodes a trans-acting regulatory protein gen step-down (38, 40, 43). In Anabaena PCC
required for formation of the glycolipid layer 7120,two nif operons and the hupSL operon are
(44, 83). each interrupted by DNA elements whose
developmentally regulated site-specific exci-
Metabolic Changes and sion is required to re-create the intact operons
Nitrogen Fixation before they can be correctly expressed (19, 39).
Heterocyst development culminates in the syn- All three DNA rearrangements occur late dur-
thesis of nitrogenase and the supply of fixed ing heterocyst development, between 18 and
nitrogen to vegetative cells. During nitrogen 24 h after nitrogen step-down (18, 19, 36–38).
fixation, nitrogenase reduces atmospheric
nitrogen to ammonia, which is then assimilated PATTERN FORMATION
AND MAINTENANCE
into amino acids (21). In addition to synthesiz-
ing the nitrogenase enzyme complex, hetero- The patS Gene Is Required for
cyst differentiation requires changes in the De Novo Pattern Formation
photosynthetic apparatus to stop oxygen pro- Heterocystous cyanobacteria provide an excel-
duction and increased production of enzymes lent prokaryotic model for studying pattern
for carbon metabolism to provide ATP and formation in a multicellular organism because
low-potential reductant for nitrogen fixation they form a one-dimensional developmental
(101). Compounds from photosynthetic vege- pattern composed of only two cell types, hete-
tative cells are transported into heterocysts, rocysts and vegetative cells. Regulation of hete-
probably in the form of sucrose (20), and heterocyst frequency and spacing is necessary to
rocysts deliver nitrogenous compounds to the ensure an efficient exchange of fixed nitrogen
vegetative cells, probably in the form of amino and fixed carbon between heterocysts and veg-
acids such as glutamine (32, 101). Large polar etative cells (Fig. 2). A long-standing model
cyanophycin granules, which store nitrogen proposes that heterocyst pattern formation is
as a nonribosomally synthesized branched regulated by lateral inhibition by a diffusible
polypeptide composed of multi-L-arginyl- inhibitor originating from differentiating cells
poly-L-aspartic acid, form near the intercellular that would inhibit differentiation of neighbor-
junctions of mature heterocysts. However, ing cells (69).The PatS peptide fulfills the role
interruption of the cphA gene showed that of this diffusible inhibitor. In Anabaena strain
cyanophycin production is not essential for dia- PCC 7120,the patS gene is predicted to encode
zotrophic growth (112). a 13- or 17-amino-acid peptide, depending on
The nitrogen-fixation (nif) genes are organ- the start codon chosen in vivo (105, 106).The
ized into several operons in Anabaena PCC patS ortholog in N. punctiforme contains only 13
7120. The nifHDK operon encodes the struc- codons (70).Overexpression of patS blocks het-
tural components of nitrogenase (43), and the erocyst development, whereas a patS null
nifB-fdxN-nifS-nifU operon (73) and other mutant forms heterocysts even in the presence
genes (9) are required for assembly of the nitro- of nitrate, and forms multiple contiguous hete-
genase enzyme complex. Interestingly, a closely rocysts and short vegetative-cell intervals
related cyanobacterium, A. variabilis, contains between heterocysts, resulting in about 30%
three different nitrogenase gene clusters: one heterocysts after nitrogen step-down (Fig. 1C).
expressed in heterocysts, one expressed in The last five carboxy-terminal amino acid
vegetative cells, and a third that produces a residues (RGSGR) of PatS are necessary and
vanadium-dependent nitrogenase (92, 93). Lit- sufficient for inhibiting heterocyst develop-
tle is known about the mechanisms controlling ment. Mutations in patS that affect these
expression of cyanobacterial nif genes. The nif residues result in a loss of heterocyst-inhibition
activity (105). Minigenes and heterologous transporters and periplasmic oligopeptide-

genes that encode only these residues inhibit binding proteins.
development (103). The corresponding syn- Our work with artificial minigenes indicates
thetic pentapeptide (PatS-5) inhibits heterocyst that the PatS receptor must be cytoplasmic
development at submicromolar concentrations (103). Huang et al. have now provided strong
(105). Addition of the PatS-5 peptide to the evidence that HetR is the PatS receptor (48).
growth medium of the patS null mutant They discovered that HetR homodimer is a
reduced the frequency of heterocysts but did DNA-binding protein that binds specifically to
not restore a normal pattern. However, ectopic the promoter regions of the heterocyst-specific
expression of patS from the heterocyst-specific genes hepA, patS, and hetR and that the DNA-
hepA promoter restored normal spacing in a binding activity is inhibited in vitro by the
patS null mutant. These data suggest that PatS PatS-5 pentapeptide in a dose-dependent man-
acts as a diffusible inhibitor in a cell nonau- ner (48). The heterocyst-pattern defect and
tonomous manner and that a gradient of the insensitivity to PatS produced by a hetRR223W
PatS signal, possibly a processed C-terminal allele provide genetic evidence for HetR being
peptide, originating from differentiating cells, is the PatS receptor (54). The observation that
required to establish a normal pattern (40). simultaneous overexpression of patS and hetR
RNA blot analysis and PpatS-lacZ reporter in a synthetic operon inhibits heterocyst devel-
strains show that patS is upregulated early dur- opment indicates that patS acts downstream of
ing heterocyst development (105). A PpatS-gfp hetR transcription, and is also consistent with
reporter strain showed that patS expression is PatS inhibiting HetR activity (80). These data
localized to individual cells or small groups of suggest that the PatS-to-HetR ratio is a critical
cells by 8 to 10 h after nitrogen step-down factor in developmental decisions; a high PatS-
(106). By 12 to 14 h, fluorescence is confined to-HetR ratio, enhanced by HetR autodegra-
mostly to individual cells arranged in a pattern dation, would be characteristic of vegetative
resembling that of mature heterocysts, and by cells in which differentiation is inhibited, but it
18 h the bright cells are almost exclusively pro- is less clear how HetR remains active in differ-
heterocysts (Fig. 1D).These data support a lat- entiating cells (48).
eral-inhibition model in which a PatS product An important unanswered question is: what
acts as an intercellular signal generated by dif- confers immunity to the differentiating cells
ferentiating cells to inhibit differentiation of against the inhibitory PatS signal? One explana-
neighboring cells. A normal pattern is not tion is that hetR upregulation in differentiating
restored in a patS mutant when a patS5 mini- cells precedes and controls transcription of patS,
gene is expressed in differentiating cells, sug- which results in a high HetR-to-PatS ratio in
gesting that the pentapeptide produced by the differentiating cells (48). A second possibility is
minigene is confined to the cytoplasm and can- that the full-length PatS peptide is not active
not be exported and function in cell-to-cell immediately upon translation in differentiating
signaling (103). One hypothesis is that the full- cells and that it requires cleavage or modifica-
length or processed PatS is exported from dif- tion before or after export to become active
ferentiated cells into the periplasmic space, and (108). Third, the active PatS concentration
then taken up by neighboring vegetative cells may be reduced in differentiating cells by
where it inhibits differentiation. It has been mechanisms related to export, degradation, or
proposed that the periplasmic space is continu- modification of PatS in proheterocyst. Fourth,
ous and serves as a conduit along cyanobacterial in differentiating cells, the HetR receptor may
filaments for metabolites and regulatory mole- become insensitive to the PatS signal due to a
cules (32).The uptake of the PatS signal by tar- posttranslational modification or to a hetero-
get cells presumably requires oligopeptide cyst-specific factor or protein that interacts with
permeases that consist of multisubunit ABC- HetR to relieve PatS-dependent inhibition.
84 ■ ALDEA ET AL.
A recent epistasis analysis of four genes by further differentiation of nearly all cells.
involved in pattern formation in Anabaena Thus, patS- and hetN-dependent inhibitory
Strain PCC 7120 suggests that PatA has two pathways are the major mechanisms that pre-
distinct activities, to promote differentiation vent heterocyst differentiation and influence
as well as to attenuate the negative effects of the developmental pattern (10, 80). However,
PatS and HetN on differentiation (80). PatA this conclusion does not rule out a contributing
may be involved in the immunity of differenti- role for the products of nitrogen fixation, other
ating cells to inhibitory signals by directly unknown diffusible signals, or internal regula-
interfering with the inhibitory signals them- tion of differentiation related to metabolism or
selves or by interacting with HetR to render it the cell cycle.
insensitive to inhibition. In two mutants that undergo almost com-
plete differentiation under nitrogen-limiting
The hetN Gene Is Required for conditions, hetRR223W and the patS hetN double
Maintenance of the Heterocyst Pattern mutant, the process of differentiation is asyn-
An additional inhibitory signal requires the chronous; it takes several days for nearly all cells
hetN gene, which encodes a protein similar to to differentiate (10, 54).This seems to support
ketoacyl reductases (8). HetN appears to play a previous ideas that,at any particular time,not all
role in the maintenance of the normal hetero- cells in a filament are equally competent to dif-
cyst pattern, as opposed to the role of patS in ferentiate into heterocysts. This effect may be
establishing the initial pattern (16).When hetN due to regulation based on cell lineage, specific
is overexpressed from a copper-inducible pro- signals that create a preexisting pattern,or phys-
moter, heterocyst development is completely iology such as differential stores of nitrogen
blocked. In the absence of hetN expression, fila- reserves (10, 54, 106). However, it is not known
ments develop a normal pattern in the first 24 h what factors influence the developmental com-
after nitrogen step-down, but by 48 h excessive petence of individual cells within a filament.
differentiation produces a Mch phenotype (16).
Immunoblot assays showed a low level of HetN OTHER PREDICTED
in vegetative cells under noninducing condi- CELL-TO-CELL COMMUNICATION
tions, and after nitrogen step-down, HetN lev- IN CYANOBACTERIA
els first dropped and then increased, with HetN In addition to the ability to form heterocysts,
localized to heterocysts (61). Overexpression of some cyanobacteria exhibit other types of
hetN blocks the accumulation of HetR protein development. For example, N. punctiforme vege-
(61), prevents the patterned expression of a tative cells have three possible developmental
hetR-gfp reporter, and also suppresses the alternatives: heterocysts, akinetes, and hor-
Mch phenotype caused by overexpression of mogonia (68, 70). Akinetes are spore-like cells
hetR (16). Thus, it is proposed that a putative structurally equipped to endure cold and desic-
hetN-dependent signal blocks heterocyst cation, and can remain viable for hundreds of
development upstream and downstream of years prior to germination (2).Akinetes usually
hetR transcription, possibly by blocking hetR- develop when a culture approaches stationary
positive autoregulation (80). phase, but they can be induced synchronously
Inactivation of both patS and hetN results in in a zwf mutant strain of N. punctiforme follow-
almost complete differentiation of filaments in ing dark incubation in the presence of fructose
the absence of combined nitrogen (10).A nitro- (4). Some genes that are required for heterocyst
gen source of ammonium, but not nitrate, development are also involved in akinete for-
inhibits heterocyst development in the double mation,such as hetR and hepA (60,99).Akinetes
mutant.After nitrogen step-down from ammo- form along filaments in different developmen-
nium, the double mutant strain produces a Mch tal patterns in different cyanobacterial strains
phenotype similar to a patS mutant followed (2).They can be found adjacent to heterocysts,
at intercalary positions between heterocysts, ACKNOWLEDGMENTS

and in some cases all cells differentiate into The work from this laboratory was supported by Public
akinetes. In the absence of heterocysts, the Health Service grant GM36890 from the National
Institutes of Health and Department of Energy grant
akinetes seem to form at random positions DE-FG03-ER020309.
along the filament, whereas the presence of
heterocysts influences akinete positioning, REFERENCES
implying the existence of cell-to-cell commu- 1. Adams, D. G. 1997. Cyanobacteria, p. 109–148.
nication (1).These developmental patterns may In J. A. Shapiro and M. Dworkin (ed.), Bacteria as
involve signaling mechanisms similar to those Multicellular Organisms. Oxford University Press,
found in heterocyst development, but nothing New York, N Y.
2. Adams, D. G., and N. G. Carr. 1981.The devel-
is currently known about the control of akinete opmental biology of heterocyst and akinete for-
developmental pattern formation. mation in cyanobacteria. Crit. Rev. Microbiol.
Hormogonia are short differentiated fila- 9:45–100.
ments released after fragmentation of the parent 3. Aldehni, M. F., and K. Forchhammer. 2006.
filament that often produce gas vesicles and Analysis of a non-canonical NtcA-dependent pro-
moter in Synechococcus elongatus and its regulation
possess gliding motility.These filaments consist by NtcA and PII. Arch. Microbiol. 184:378–386.
of small cells that result from rapid cell division 4. Argueta, C., K. Yuksek, R. Patel, and M. L.
in the absence of cell growth or DNA replica- Summers. 2006. Identification of Nostoc puncti-
tion, which is possible because cyanobacteria forme akinete-expressed genes using differential
have multiple genome copies in each vegetative display. Mol. Microbiol. 61:748–757.
5. Bauer, C. C., K. S. Ramaswamy, S. Endley, L.
cell (69).The role of hormogonia is to dissemi- A. Scappino, J.W. Golden, and R. Haselkorn.
nate and colonize new portions of a habitat. 1997. Suppression of heterocyst differentiation in
Eventually, hormogonial filaments return to Anabaena PCC 7120 by a cosmid carrying wild-
vegetative growth and can then produce hete- type genes encoding enzymes for fatty acid syn-
rocysts (68). Factors influencing hormogonia thesis. FEMS Microbiol. Lett. 151:23–30.
6. Black, K., W. J. Buikema, and R. Haselkorn.
development include changes in nutrient or 1995.The hglK gene is required for localization of
light conditions under free-living conditions heterocyst-specific glycolipids in the cyanobac-
and interactions with the plant partner in sym- terium Anabaena sp. strain PCC 7120. J. Bacteriol.
biotic associations (68). 177:6440–6448.
The ability to form symbiotic associations 7. Black,T.A.,Y. Cai, and C. P.Wolk. 1993.Spatial
expression and autoregulation of hetR, a gene
with plants and fungi speaks to the versatility of involved in the control of heterocyst development
heterocystous cyanobacteria. N. punctiforme can in Anabaena. Mol. Microbiol. 9:77–84.
establish symbiosis with the bryophyte Antho- 8. Black,T. A., and C. P.Wolk. 1994.Analysis of a
ceros punctatus and the angiosperm Gunnera spp. Het- mutation in Anabaena sp. strain PCC 7120
and can act as an intracellular symbiont of the implicates a secondary metabolite in the regulation
of heterocyst spacing. J. Bacteriol. 176:2282–2292.
mycorrhizal fungus Geosiphon piriforme (68). 9. Borthakur, D., M. Basche,W. J. Buikema, P. B.
The symbiotic interaction with the plant part- Borthakur, and R. Haselkorn. 1990. Expres-
ners has a profound effect on the physiology sion, nucleotide sequence and mutational analysis
and development of the cyanobacteria.Hetero- of two open reading frames in the nif gene region
cyst frequencies increase severalfold, growth of Anabaena sp. strain PCC7120. Mol. Gen. Genet.
221:227–234.
and photosynthesis are diminished, and hor- 10. Borthakur, P. B., C. C. Orozco, S. S.Young-
mogonium development is influenced by the Robbins, R. Haselkorn, and S. M. Callahan.
action of extracellular factors produced by the 2005. Inactivation of patS and hetN causes lethal
plant (68, 69). It is clear that multicellular levels of heterocyst differentiation in the filamen-
prokaryotic organisms such as heterocystous tous cyanobacterium Anabaena sp.PCC 7120.Mol.
Microbiol. 57:111–123.
cyanobacteria must be able to integrate extra- 11. Brahamsha, B., and R. Haselkorn. 1992. Iden-
cellular and internal signals to produce appro- tification of multiple RNA polymerase sigma fac-
priate responses to their environment. tor homologs in the cyanobacterium Anabaena sp.
86 ■ ALDEA ET AL.
strain PCC 7120:cloning,expression,and inactiva- differentiation in the cyanobacterium Anabaena sp.

tion of the sigB and sigC genes. J. Bacteriol. strain PCC 7120. J. Bacteriol. 188:8520–8525.
174:7273–7282. 25. Ehira, S., and M. Ohmori. 2006. NrrA, a
12. Brahamsha, B., and R. Haselkorn. 1991. Isola- nitrogen-responsive response regulator facilitates
tion and characterization of the gene encoding the heterocyst development in the cyanobacterium
principal sigma factor of the vegetative cell RNA Anabaena sp. strain PCC 7120. Mol. Microbiol.
polymerase from the cyanobacterium Anabaena sp. 59:1692–1703.
strain PCC 7120. J. Bacteriol. 173:2442–2450. 26. Ernst, A.,T. Black,Y. Cai, J. M. Panoff, D. N.
13. Buikema, W. J., and R. Haselkorn. 2001. Tiwari, and C. P.Wolk. 1992. Synthesis of nitro-
Expression of the Anabaena hetR gene from a genase in mutants of the cyanobacterium Anabaena
copper-regulated promoter leads to heterocyst sp.strain PCC 7120 affected in heterocyst develop-
differentiation under repressing conditions. Proc. ment or metabolism. J. Bacteriol. 174:6025–6032.
Natl.Acad. Sci. USA 98:2729–2734. 27. Espinosa, J., K. Forchhammer, S. Burillo, and
14. Buikema,W. J., and R. Haselkorn. 1991. Isola- A. Contreras. 2006. Interaction network in
tion and complementation of nitrogen fixation cyanobacterial nitrogen regulation:PipX,a protein
mutants of the cyanobacterium Anabaena sp. strain that interacts in a 2-oxoglutarate dependent
PCC 7120. J. Bacteriol. 173:1879–1885. manner with PII and NtcA. Mol. Microbiol.
15. Cai,Y., and C. P.Wolk. 1997. Anabaena sp. strain 61:457–469.
PCC 7120 responds to nitrogen deprivation with a 28. Fadi Aldehni, M., J. Sauer, C. Spielhaupter, R.
cascade-like sequence of transcriptional activa- Schmid, and K. Forchhammer. 2003. Signal
tions. J. Bacteriol. 179:267–271. transduction protein P(II) is required for NtcA-
16. Callahan, S. M., and W. J. Buikema. 2001.The regulated gene expression during nitrogen
role of HetN in maintenance of the heterocyst pat- deprivation in the cyanobacterium Synechococcus
tern in Anabaena sp. PCC 7120. Mol. Microbiol. elongatus strain PCC 7942. J. Bacteriol. 185:
40:941–950. 2582–2591.
17. Campbell, E. L., M. F. Cohen, and J. C. 29. Fan, Q., G. Huang, S. Lechno-Yossef, C. P.
Meeks. 1997. A polyketide-synthase-like gene is Wolk, T. Kaneko, and S. Tabata. 2005. Clus-
involved in the synthesis of heterocyst glycolipids tered genes required for synthesis and deposition
in Nostoc punctiforme strain ATCC 29133. Arch. of envelope glycolipids in Anabaena sp. strain PCC
Microbiol. 167:251–258. 7120. Mol. Microbiol. 58:227–243.
18. Carrasco, C. D., J. A. Buettner, and J. W. 30. Fan, Q., S. Lechno-Yossef, S. Ehira, T.
Golden. 1995.Programmed DNA rearrangement Kaneko, M. Ohmori, N. Sato, S. Tabata, and
of a cyanobacterial hupL gene in heterocysts. Proc. C. P. Wolk. 2006. Signal transduction genes
Natl.Acad. Sci. USA 92:791–795. required for heterocyst maturation in Anabaena sp.
19. Carrasco, C. D., S. D. Holliday, A. Hansel, P. strain PCC 7120. J. Bacteriol. 188:6688–6693.
Lindblad, and J. W. Golden. 2005. Heterocyst- 31. Fiedler, G.,A. M. Muro-Pastor, E. Flores, and
specific excision of the Anabaena sp. strain PCC I. Maldener. 2001. NtcA-dependent expression
7120 hupL element requires xisC. J. Bacteriol. of the devBCA operon, encoding a heterocyst-
187:6031–6038. specific ATP-binding cassette transporter in
20. Curatti, L., E. Flores, and G. Salerno. 2002. Anabaena spp. J. Bacteriol. 183:3795–3799.
Sucrose is involved in the diazotrophic metabolism 32. Flores, E., A. Herrero, C. P. Wolk, and I.
of the heterocyst-forming cyanobacterium Maldener. 2006. Is the periplasm continuous in
Anabaena sp. FEBS Lett. 513:175–178. filamentous multicellular cyanobacteria? Trends
21. Dixon, R., and D. Kahn. 2004. Genetic regula- Microbiol. 14:439–443.
tion of biological nitrogen fixation.Nat.Rev.Micro- 33. Forchhammer, K. 2004. Global carbon/nitro-
biol. 2:621–631. gen control by PII signal transduction in
22. Dong, Y., X. Huang, X. Y. Wu, and J. Zhao. cyanobacteria:from signals to targets.FEMS Micro-
2000. Identification of the active site of HetR pro- biol. Rev. 28:319–333.
tease and its requirement for heterocyst differenti- 34. Forchhammer, K., and N. Tandeau de
ation in the cyanobacterium Anabaena sp. strain Marsac. 1995. Phosphorylation of the PII protein
PCC 7120. J. Bacteriol. 182:1575–1579. (glnB gene product) in the cyanobacterium Syne-
23. Edwards, R., and M. Merrick. 1995.The role of chococcus sp. strain PCC 7942: analysis of in vitro
uridylyltransferase in the control of Klebsiella pneu- kinase activity. J. Bacteriol. 177:5812–5817.
moniae nif gene regulation. Mol. Gen. Genet. 35. Frias, J. E., E. Flores, and A. Herrero. 1994.
247:189–198. Requirement of the regulatory protein NtcA
24. Ehira, S., and M. Ohmori. 2006. NrrA directly for the expression of nitrogen assimilation and
regulates the expression of hetR during heterocyst heterocyst development genes in the cyanobac-
terium Anabaena sp. PCC 7120. Mol. Microbiol. 48. Huang, X.,Y. Dong, and J. Zhao. 2004. HetR
14:823–832. homodimer is a DNA-binding protein required
36. Golden, J.W., C. D. Carrasco, M. E. Mulligan, for heterocyst differentiation, and the DNA-
G. J. Schneider, and R. Haselkorn. 1988. Dele- binding activity is inhibited by PatS. Proc. Natl.
tion of a 55-kilobase-pair DNA element from the Acad. Sci. USA 101:4848–4853.
chromosome during heterocyst differentiation of 49. Ikeda,T. P.,A. E. Shauger, and S. Kustu. 1996.
Anabaena sp. strain PCC 7120. J. Bacteriol. Salmonella typhimurium apparently perceives exter-
170:5034–5041. nal nitrogen limitation as internal glutamine limi-
37. Golden, J. W., S. J. Robinson, and R. tation. J. Mol. Biol. 259:589–607.
Haselkorn. 1985. Rearrangement of nitrogen 50. Jiang, F., S.Wisen, M.Widersten, B. Bergman,
fixation genes during heterocyst differentiation in and B. Mannervik. 2000. Examination of the
the cyanobacterium Anabaena. Nature (London) transcription factor NtcA-binding motif by in vitro
314:419–423. selection of DNA sequences from a random
38. Golden, J. W., L. L. Whorff, and D. R. Wiest. library. J. Mol. Biol. 301:783–793.
1991. Independent regulation of nifHDK operon 51. Jones, K. M., W. J. Buikema, and R.
transcription and DNA rearrangement during Haselkorn. 2003. Heterocyst-specific expression
heterocyst differentiation in the cyanobacterium of patB, a gene required for nitrogen fixation in
Anabaena sp. strain PCC 7120. J. Bacteriol. Anabaena sp. strain PCC 7120. J. Bacteriol.
173:7098–7105. 185:2306–2314.
39. Golden, J.W., and H.-S.Yoon. 1998.Heterocyst 52. Kaiser, D. 2001. Building a multicellular organ-
formation in Anabaena. Curr. Opin. Microbiol. ism. Annu. Rev. Genet. 35:103–123.
1:623–629. 53. Khudyakov, I., and C. P. Wolk. 1997. hetC, a
40. Golden, J.W., and H.-S.Yoon. 2003.Heterocyst gene coding for a protein similar to bacterial ABC
development in Anabaena. Curr. Opin. Microbiol. protein exporters, is involved in early regulation of
6:557–563. heterocyst differentiation in Anabaena sp. strain
41. Gonzalez, L., V. Phalip, and C. C. Zhang. PCC 7120. J. Bacteriol. 179:6971–6978.
2001. Characterization of PknC, a Ser/Thr kinase 54. Khudyakov, I.Y., and J.W. Golden. 2004. Dif-
with broad substrate specificity from the ferent functions of HetR, a master regulator of
cyanobacterium Anabaena sp. strain PCC 7120. heterocyst differentiation in Anabaena sp. PCC
Eur. J. Biochem. 268:1869–1875. 7120, can be separated by mutation. Proc. Natl.
42. Hanson,T. E., K. Forchhammer, N.Tandeau Acad. Sci. USA 101:16040–16045.
de Marsac, and J. C. Meeks. 1998.Characteriza- 55. Khudyakov, I.Y., and J.W. Golden. 2001. Iden-
tion of the glnB gene product of Nostoc punctiforme tification and inactivation of three group 2 sigma
strain ATCC 29133:glnB or the PII protein may be factor genes in Anabaena sp. strain PCC 7120. J.
essential. Microbiology 144:1537–1547. Bacteriol. 183:6667–6675.
43. Haselkorn, R., J. W. Golden, P. J. Lammers, 56. Kill, K.,T.T. Binnewies,T. Sicheritz-Ponten,
and M. E. Mulligan. 1986. Developmental H.Willenbrock, P. F. Hallin,T. M.Wassenaar,
rearrangement of cyanobacterial nitrogen-fixation and D. W. Ussery. 2005. Genome update: sigma
genes. Trends Genet. 2:255–259. factors in 240 bacterial genomes. Microbiology
44. Hebbar, P. B., and S. E. Curtis. 2000. Charac- 151:3147–3150.
terization of devH, a gene encoding a putative 57. Laurent, S., H. Chen, S. Bedu, F. Ziarelli, L.
DNA binding protein required for heterocyst Peng, and C. C. Zhang. 2005. Nonmetaboliz-
function in Anabaena sp. strain PCC 7120. J. Bacte- able analogue of 2-oxoglutarate elicits heterocyst
riol. 182:3572–3581. differentiation under repressive conditions in
45. Herbaud, M. L., A. Guiseppi, F. Denizot, J. Anabaena sp. PCC 7120. Proc. Natl.Acad. Sci. USA
Haiech, and M. C. Kilhoffer. 1998. Calcium 102:9907–9912.
signalling in Bacillus subtilis. Biochim. Biophys. Acta 58. Laurent, S., K. Forchhammer, L. Gonzalez,T.
1448:212–226. Heulin, C. C. Zhang, and S. Bedu. 2004. Cell-
46. Herrero,A.,A. M. Muro-Pastor,A.Valladares, type specific modification of PII is involved in the
and E. Flores. 2004. Cellular differentiation and regulation of nitrogen metabolism in the
the NtcA transcription factor in filamentous cyanobacterium Anabaena PCC 7120. FEBS Lett.
cyanobacteria. FEMS Microbiol. Rev. 28:469–487. 576:261–265.
47. Huang, G., Q. Fan, S. Lechno-Yossef, E.Woj- 59. Lechno-Yossef, S., Q. Fan, S. Ehira, N. Sato,
ciuch, C. P. Wolk, T. Kaneko, and S. Tabata. and C. P. Wolk. 2006. Mutations in four regula-
2005. Clustered genes required for the synthesis of tory genes have interrelated effects on heterocyst
heterocyst envelope polysaccharide in Anabaena maturation in Anabaena sp. strain PCC 7120. J. Bac-
sp. strain PCC 7120. J. Bacteriol. 187:1114–1123. teriol. 188:7387–7395.
88 ■ ALDEA ET AL.
60. Leganes, F., F. Fernandez-Pinas, and C. P. 72. Moorhead, G. B., and C. S. Smith. 2003.
Wolk. 1994.Two mutations that block heterocyst Interpreting the plastid carbon, nitrogen, and
differentiation have different effects on akinete dif- energy status. A role for PII? Plant Physiol.
ferentiation in Nostoc ellipsosporum. Mol. Microbiol. 133:492–498.
12:679–684. 73. Mulligan, M. E., and R. Haselkorn. 1989.
61. Li, B., X. Huang, and J. Zhao. 2002. Expression Nitrogen fixation (nif) genes of the cyanobac-
of hetN during heterocyst differentiation and its terium Anabaena species strain PCC 7120. The
inhibition of hetR up-regulation in the cyanobac- nifB-fdxN-nifS-nifU operon. J. Biol. Chem.
terium Anabaena sp. PCC 7120. FEBS Lett. 264:19200–19207.
517:87–91. 74. Muro-Pastor,A. M., E. Olmedo-Verd, and E.
62. Liang, J., L. Scappino, and R. Haselkorn. Flores. 2006. All4312, an NtcA-regulated two-
1992. The patA gene product, which contains a component response regulator in Anabaena sp.
region similar to CheY of Escherichia coli, controls strain PCC 7120. FEMS Microbiol. Lett.
heterocyst pattern formation in the cyanobac- 256:171–177.
terium Anabaena 7120. Proc. Natl. Acad. Sci. USA 75. Muro-Pastor, A. M., A.Valladares, E. Flores,
89:5655–5659. and A. Herrero. 2002.Mutual dependence of the
63. Liang, J., L. Scappino, and R. Haselkorn. expression of the cell differentiation regulatory
1993.The patB gene product, required for growth protein HetR and the global nitrogen regulator
of the cyanobacterium Anabaena sp. strain PCC NtcA during heterocyst development. Mol. Micro-
7120 under nitrogen-limiting conditions, contains biol. 44:1377–1385.
ferredoxin and helix-turn-helix domains. J. Bacte- 76. Muro-Pastor, A. M., A.Valladares, E. Flores,
riol. 175:1697–1704. and A. Herrero. 1999.The hetC gene is a direct
64. Liu, D., and J.W. Golden. 2002.hetL overexpres- target of the NtcA transcriptional regulator in
sion stimulates heterocyst formation in Anabaena cyanobacterial heterocyst development. J. Bacteriol.
sp. strain PCC 7120. J. Bacteriol. 184:6873–6881. 181:6664–6669.
65. Lockau, W., B. Massalsky, and A. Dirmeier. 77. Muro-Pastor, M. I., J. C. Reyes, and F. J.
1988. Purification and partial characterization of a Florencio. 2005. Ammonium assimilation in
calcium-stimulated protease from the cyanobac- cyanobacteria. Photosynth. Res. 83:135–150.
terium, Anabaena variabilis. Eur. J. Biochem. 78. Olmedo-Verd, E., E. Flores, A. Herrero,
172:433–438. and A. M. Muro-Pastor. 2005. HetR-
66. Maheswaran, M., C. Urbanke, and K. Forch- dependent and -independent expression of
hammer. 2004. Complex formation and catalytic heterocyst-related genes in an Anabaena strain
activation by the PII signaling protein of N-acetyl- overproducing the NtcA transcription factor.
L-glutamate kinase from Synechococcus elongatus J. Bacteriol. 187:1985–1991.
strain PCC 7942. J. Biol. Chem. 279:55202–55210. 79. Olmedo-Verd, E., A. M. Muro-Pastor, E.
67. Makarova, K. S., E.V. Koonin, R. Haselkorn, Flores, and A. Herrero. 2006. Localized induc-
and M. Y. Galperin. 2006. Cyanobacterial tion of the ntcA regulatory gene in developing
response regulator PatA contains a conserved N- heterocysts of Anabaena sp. strain PCC 7120.
terminal domain (PATAN) with an alpha-helical J. Bacteriol. 188:6694–6699.
insertion. Bioinformatics 22:1297–1301. 80. Orozco, C. C., D. D. Risser, and S. M.
68. Meeks, J. C., E. L. Campbell, M. L. Summers, Callahan. 2006. Epistasis analysis of four genes
and F. C.Wong. 2002. Cellular differentiation in from Anabaena sp. strain PCC 7120 suggests a
the cyanobacterium Nostoc punctiforme.Arch. Micro- connection between PatA and PatS in heterocyst
biol. 178:395–403. pattern formation. J. Bacteriol. 188:1808–1816.
69. Meeks, J. C., and J. Elhai. 2002. Regulation of 81. Paz-Yepes, J., E. Flores, and A. Herrero. 2003.
cellular differentiation in filamentous cyanobacte- Transcriptional effects of the signal transduction
ria in free-living and plant-associated symbiotic protein P(II) (glnB gene product) on NtcA-
growth states. Microbiol. Mol. Biol. Rev. 66:94–121. dependent genes in Synechococcus sp. PCC 7942.
70. Meeks, J. C., J. Elhai, T. Thiel, M. Potts, F. FEBS Lett. 543:42–46.
Larimer, J. Lamerdin, P. Predki, and R.Atlas. 82. Ramasubramanian, T. S., T. F. Wei, A. K.
2001.An overview of the genome of Nostoc puncti- Oldham, and J.W. Golden. 1996.Transcription
forme, a multicellular, symbiotic cyanobacterium. of the Anabaena sp. strain PCC 7120 ntcA gene:
Photosynth. Res. 70:85–106. multiple transcripts and NtcA binding. J. Bacteriol.
71. Merrick, M. J., and R. A. Edwards. 1995. 178:922–926.
Nitrogen control in bacteria. Microbiol. Rev. 59: 83. Ramirez, M. E., P. B. Hebbar, R. Zhou, C. P.
604–622. Wolk, and S. E. Curtis. 2005. Anabaena sp. strain
PCC 7120 gene devH is required for synthesis of 96. Vazquez-Bermudez, M. F., A. Herrero, and
the heterocyst glycolipid layer. J. Bacteriol. E. Flores. 2002. 2-Oxoglutarate increases the
187:2326–2331. binding affinity of the NtcA (nitrogen control)
84. Risser, D. D., and S. M. Callahan. 2007. Muta- transcription factor for the Synechococcus glnA
genesis of hetR reveals amino acids necessary for promoter. FEBS Lett. 512:71–74.
HetR function in the heterocystous cyanobac- 97. Vetting, M.W., S. S. Hegde, J. E. Fajardo, A.
terium Anabaena sp. PCC 7120. J. Bacteriol. Fiser, S. L. Roderick, H. E. Takiff, and J. S.
189:2460–2467. Blanchard. 2006. Pentapeptide repeat proteins.
85. Schneider, G. J., and R. Haselkorn. 1988. Biochemistry 45:1–10.
RNA polymerase subunit homology among 98. Wei,T. F.,T. S. Ramasubramanian, and J.W.
cyanobacteria, other eubacteria and archaebacte- Golden. 1994. Anabaena sp. strain PCC 7120
ria. J. Bacteriol. 170:4136–4140. ntcA gene required for growth on nitrate and het-
86. Shi,Y.,W. Zhao,W. Zhang, Z.Ye, and J. Zhao. erocyst development.J.Bacteriol. 176:4473–4482.
2006.Regulation of intracellular free calcium con- 99. Wolk, C. P. 1996. Heterocyst formation. Annu.
centration during heterocyst differentiation by Rev. Genet. 30:59–78.
HetR and NtcA in Anabaena sp. PCC 7120. Proc. 100. Wolk, C. P.,Y. Cai, L. Cardemil, E. Flores, B.
Natl.Acad. Sci. USA 103:11334–11339. Hohn, M. Murry, G. Schmetterer, B.
87. Smith, R. J., S. Hobson, and I. R. Ellis. 1987. Schrautemeier, and R.Wilson. 1988.Isolation
Evidence for calcium-mediated regulation of het- and complementation of mutants of Anabaena sp.
erocyst frequency and nitrogenase activity in Nos- strain PCC 7120 unable to grow aerobically on
toc 6720. New Phytologist 105:531–541. dinitrogen. J. Bacteriol. 170:1239–1244.
88. Stevenson, B. S., and J. B. Waterbury. 2006. 101. Wolk, C. P., A. Ernst, and J. Elhai. 1994.
Isolation and identification of an epibiotic bac- Heterocyst metabolism and development,
terium associated with heterocystous Anabaena p. 769–823. In D. A. Bryant (ed.), The Mole-
cells. Biol. Bull. 210:73–77. cular Biology of Cyanobacteria, vol. 1. Kluwer
89. Straley, S. C., G. V. Plano, E. Skrzypek, P. L. Academic Publishers, Dordrecht, The Nether-
Haddix, and K. A. Fields. 1993. Regulation by lands.
Ca2 in the Yersinia low-Ca2 response. Mol. 102. Wong, F. C., and J. C. Meeks. 2001.The hetF
Microbiol. 8:1005–1010. gene product is essential to heterocyst differentia-
90. Su, Z., V. Olman, F. Mao, and Y. Xu. 2005. tion and affects HetR function in the cyanobac-
Comparative genomics analysis of NtcA regulons terium Nostoc punctiforme. J. Bacteriol.
in cyanobacteria: regulation of nitrogen assimila- 183:2654–2661.
tion and its coupling to photosynthesis. Nucleic 103. Wu, X., D. Liu, M. H. Lee, and J.W. Golden.
Acids Res. 33:5156–5171. 2004. patS minigenes inhibit heterocyst develop-
91. Tanigawa, R., M. Shirokane, S. Maeda Si, T. ment of Anabaena sp. strain PCC 7120. J. Bacteriol.
Omata, K. Tanaka, and H. Takahashi. 2002. 186:6422–6429.
Transcriptional activation of NtcA-dependent 104. Xu, X., and C. P.Wolk. 2001. Role for hetC in
promoters of Synechococcus sp. PCC 7942 by 2- the transition to a nondividing state during hete-
oxoglutarate in vitro. Proc. Natl. Acad. Sci. USA rocyst differentiation in Anabaena sp. J. Bacteriol.
99:4251–4255. 183:393–396.
92. Thiel, T. 1996. Isolation and characterization of 105. Yoon, H. S., and J. W. Golden. 1998. Hetero-
the VnfEN genes of the cyanobacterium Anabaena cyst pattern formation controlled by a diffusible
variabilis. J. Bacteriol. 178:4493–4499. peptide. Science 282:935–938.
93. Thiel, T., E. M. Lyons, and J. C. Erker. 1997. 106. Yoon, H. S., and J.W. Golden. 2001. PatS and
Characterization of genes for a second Mo- products of nitrogen fixation control heterocyst
dependent nitrogenase in the cyanobacterium pattern. J. Bacteriol. 183:2605–2613.
Anabaena variabilis. J. Bacteriol. 179:5222–5225. 107. Yoshimura, H., S. Okamoto,Y. Tsumuraya,
94. Tisa, L. S., B. M. Olivera, and J. Adler. 1993. and M. Ohmori. 2007. Group 3 sigma factor
Inhibition of Escherichia coli chemotaxis by omega- gene,sigJ,a key regulator of desiccation tolerance,
conotoxin, a calcium ion channel blocker. J. Bacte- regulates the synthesis of extracellular polysac-
riol. 175:1235–1238. charide in cyanobacterium Anabaena sp. strain
95. Torrecilla, I., F. Leganes, I. Bonilla, and F. PCC 7120. DNA Res. 14:13–24.
Fernandez-Pinas. 2004. A calcium signal is 108. Zhang, C. C., S. Laurent, S. Sakr, L. Peng,
involved in heterocyst differentiation in the and S. Bedu. 2006. Heterocyst differentiation
cyanobacterium Anabaena sp. PCC7120. Microbiol- and pattern formation in cyanobacteria: a chorus
ogy 150:3731–3739. of signals. Mol. Microbiol. 59:367–375.
90 ■ ALDEA ET AL.
109. Zhao,Y.,Y. Shi,W. Zhao, X. Huang, D.Wang, 111. Zhu, J., R. Kong, and C. P.Wolk. 1998.Regu-
N. Brown, J. Brand, and J. Zhao. 2005. CcbP, lation of hepA of Anabaena sp.strain PCC 7120 by
a calcium-binding protein from Anabaena sp. elements 5′ from the gene and by hepK. J. Bacte-
PCC 7120, provides evidence that calcium ions riol. 180:4233–4242.
regulate heterocyst differentiation. Proc. Natl. 112. Ziegler, K., D. P. Stephan, E. K. Pistorius, H.
Acad. Sci. USA 102:5744–5748. G. Ruppel, and W. Lockau. 2001.A mutant of
110. Zhou, R., X.Wei, N. Jiang, H. Li,Y. Dong, K. the cyanobacterium Anabaena variabilis ATCC
L. Hsi, and J. Zhao. 1998. Evidence that HetR 29413 lacking cyanophycin synthetase: growth
protein is an unusual serine-type protease. Proc. properties and ultrastructural aspects. FEMS
Natl.Acad. Sci. USA 95:4959–4963. Microbiol. Lett. 196:13–18.
DIVERSE CELL-CELL SIGNALING
MOLECULES CONTROL FORMATION
OF AERIAL HYPHAE AND SECONDARY
METABOLISM IN STREPTOMYCETES
Joanne M.Willey and Justin R. Nodwell
7
The Actinobacteria, including the genus Strepto- tion for dispersal. Spore formation involves the
myces, constitute, on average, about 13% of soil development of aerial hyphae that grow out of
bacterial communities, making them a domi- the soil and extend vertically into the air. Cell
nant form of life on Earth (31).The success of division is rare in the substrate hyphae, explain-
the streptomycetes is likely due, at least in part, ing the multigenomic nature of these cells;
to their mycelial growth habit, their capacity to however, concerted hyphal septation takes
consume otherwise refractory organic materi- place in each aerial filament, converting them
als by producing extracellular hydrolases, and into chains of unigenomic prespores. These
their ability to produce and secrete small mole- round up and acquire properties that make
cules that may modulate the growth of compet- them resistant to some environmental chal-
ing microorganisms. Indeed, the production of lenges.Their eventual release enables the initia-
so-called secondary metabolites makes these tion of the growth cycle elsewhere.
bacteria especially good sources of medicinal Our understanding of streptomycete mor-
compounds, including antibiotics, immune phological and physiological differentiation has
suppressants, chemotherapeutic drugs, and oth- been advanced by the study of several geneti-
ers (5). The streptomycete life cycle initiates cally well-defined species, in particular Strepto-
with spore germination, leading to the growth myces coelicolor and Streptomyces griseus.S.coelicolor
of vegetative, multigenomic filamentous cells has been subjected to extensive molecular
called substrate hyphae. These extend into the genetic characterization,resulting in the identi-
substratum, thereby sampling nutrients across a fication of mutants blocked in aerial hyphae
variety of microenvironments. These cells are formation (bald or bld mutants) (39), sporula-
nonmotile so if local conditions for growth tion (white or whi mutants) (3, 37, 38, 39), and
deteriorate, the organism depends on sporula- the production of secondary metabolites (e.g.,
abs, afs) (1, 21). Evidence suggests that the for-
mation of aerial hyphae and the production of
Joanne M.Willey Department of Biology, Hofstra University, secondary metabolites are linked; they occur at
Hempstead, NewYork 11549. Justin R.Nodwell Depart-
ment of Biochemistry, Health Sciences Centre, McMaster about the same time, and a number of muta-
University, Hamilton, Ontario, Canada, L8N 3Z5. tions have been identified that block both
91
92 ■ WILLEY AND NODWELL
processes at once in S. coelicolor and S. griseus discovered, it was shown that certain S. coelicolor
(14).Although it is clear that nutrient depriva- bld mutants release diffusible substances that
tion can trigger differentiation (40, 60), identi- restore the capacity of other bld mutants to
fying the specific environmental cues and their make SapB, aerial hyphae, and pigmented
corresponding signal transduction systems has antibiotics. It was inferred that the extracellu-
proven to be elusive. It is known, however, that larly complemented bld mutants were blocked
intercellular signaling plays an important role. in the ability to perceive and/or respond to the
molecules secreted by the donor bld strains (94).
STREPTOMYCES AS A MODEL FOR Investigation of this extracellular complemen-
CELL-CELL COMMUNICATION tation phenomenon demonstrated that most bld
The streptomycetes were among the first bacte- mutants could be placed into one of six discrete
ria in which intercellular communication was complementation groups, suggesting the
documented. In 1957, it was reported that the exchange of at least five signaling molecules
growth of wild-type S.griseus restored morpho- during development (59, 94). The nature of
genesis to industrial strains bearing mutations most of the signals and receptors remains
that blocked aerial hyphae formation when unknown, although the involvement of an
such strains were grown in close proximity.This oligopeptide permease encoded by the bldK
suggested the involvement of an extracellular gene cluster and an oligopeptide signal has been
signal or signals (19).Ten years later Khokhlov documented (57, 58). Finally, in recent years,
reported that the -butyrylactone, A-factor workers in Beppu’s laboratory have shown that
(2-isocapr yloyl-3R-hydroxymethyl-4- many wild isolates of Streptomyces spp. are capa-
butanolide), produced by wild-type strains of S. ble of intercellular communication (87), and in
griseus could likewise rescue the morphological one instance, the extracellular compound is a
defect in bld mutants of that species and stimu- desferrioxamine siderophore (99). This appar-
late the production of streptomycin (41). He ent diversity of intercellular signaling mecha-
thereby established the paradigm that an extra- nisms suggests that the streptomycetes rely
cellular signal could induce antibiotic produc- extensively on cell-cell communication to
tion and morphological differentiation when coordinate growth with the production of sec-
applied extracellularly to mutants defective in ondary metabolites and sporulation.
these processes (29).In subsequent work,Szabó, Because morphological and physiological
Biró, and colleagues reported the discovery of differentiation (in the case of pigmented antibi-
another molecule, called factor C, that also trig- otics) can be visually monitored, extracellular
gers differentiation in developmental mutants rescue of mutant phenotypes as described
of S.griseus (10,78).Although it has been histor- above can be quite striking (Color Plate 5).
ically reported that factor C is produced by a However,the streptomycetes present challenges
strain of S. griseus (10, 11, 78, 79), it was recently that need to be considered when investigating
established that the factor C-producing cell-cell signaling. For example, the propaga-
microbe is Streptomyces flavofungini, a member of tion of aerial hyphae is supported by pro-
the Streptomyces albidoflavus clade (7). grammed cell death of vegetative filaments (48,
About a decade after the first report of factor 49, 52), which releases cellular constituents into
C, it was found that wild-type S. coelicolor can the environment that then act as nutrients. If
restore to bld mutants the capacity to form aer- small molecules resulting from this activity are
ial hyphae (93) by secreting a diffusible peptide found to influence development, are they con-
called SapB. SapB is not a signal per se, but sidered signals, cellular by-products, or both?
rather is thought to function as a surfactant, When considering how such signals might be
lowering the surface tension at the colony sur- transduced, the investigator is confronted with
face and thereby facilitating growth of aerial fil- a genome that,for S.coelicolor, encodes 84 sensor
aments.At roughly the same time that SapB was kinases and 80 response regulators (6, 30), 44
7. DIFFERENTIATION IN THE STREPTOMYCETES ■ 93
putative Ser/Thr kinases (6), more than 120 share a conserved core structure and their
TetR-like transcription factors (70), 65 sigma receptors are highly conserved, the signaling
factors, and numerous proteins of unknown pathways in which they participate and their
function (6). Further complicating matters, at biological roles can be very different. Among
least three of the response regulators important these pathways, that controlled by A-factor in S.
for development, RamR (62), BldM (53), and griseus is the best understood (Fig.2).The capac-
WhiI (2), are “orphans” that lack apparent cog- ity of -butyrolactones to alter gene expression
nate sensor kinases. Thus, the nature of the is mediated through a receptor, ArpA in the case
signaling processes that activate these transcrip- of S. griseus (64). ArpA represses the expression
tion factors remains obscure. Given the intri- of adpA, which encodes an AraC-like transcrip-
cacy of the streptomycete life cycle, it is not tional activator (33,63).Upon binding A-factor,
surprising that an uncomplicated observation ArpA loses its capacity to bind and repress adpA
regarding the morphology of the microbe transcription. AdpA also represses its own
belies an intricate web of regulatory and expression (34), establishing a negative feedback
structural events. loop, and activates the expression of secondary
regulators that control genes and gene clusters
THE -BUTYROLACTONES involved in various aspects of S. griseus morpho-
The best-understood signaling systems exhib- logical and physiological differentiation (33).
ited by the streptomycetes are those mediated Among A-factor-responsive genes, two play
by the -butyrolactones. Khoklov’s landmark an especially important role in the S. griseus
observation regarding the capacity of A-factor developmental cycle. strR encodes a pathway-
to rescue both morphogenesis and strepto- specific activator of the streptomycin biosyn-
mycin production to S. griseus mutants (41) was thetic gene cluster (71, 92), and amfR codes for
quickly confirmed and extended by the seminal a response regulator that activates genes
work of Horinouchi and others. It was soon required for production of a SapB-like surfac-
reported that synthetic A-factor [2-(6′-methyl- tant (AmfS) involved in the formation of aerial
heptanoyl)-3R-hydroxymethyl-4-butanolide] hyphae (86, 88). In this bacterium, therefore,
is active at nanomolar concentrations, and for the linkage of secondary metabolism to mor-
this reason A-factor is considered a bacterial phological differentiation is at least partly
hormone (24). -Butyrolactones have since accounted for by the control of these transcrip-
been discovered in other streptomycetes includ- tional regulators by A-factor.
ing S. coelicolor (83), Streptomyces virginiae (42), Several other genes are activated directly by
and Streptomyces fradiae (16, 76) (Fig. 1). Like the AdpA. These include ssgA, which encodes a
more familiar acylhomoserine lactones,these - protein needed for sporulation-specific septa-
butyrolactones differ in acyl side chain length tion in aerial hyphae (101), and adsA, whose
and hydroxylation.Strikingly,while the -buty- product is an extracytoplasmic sigma factor
rolactones produced by different streptomycetes required for aerial hyphae formation, although
A. B.
FIGURE 1 Two well-characterized -butyrolactones are A-factor of S. griseus (A), which

regulates antibiotic production and sporulation, and SCB-1 of S. coelicolor (B), which regulates
antibiotic production exclusively.
FIGURE 2 Three -butyrolactone-regulated pathways.The best-characterized pathway controls antibiotic

production and sporulation in S.griseus.The A-factor receptor,ArpA,represses expression of the autoregulatory
gene adpA, which in turn controls the expression of at least six genes implicated in antibiotic production or
sporulation (33, 34, 63). Of these target genes, two, strR (71, 92) and amfR (86, 88), encode transcription factors
that activate other developmental genes, while others encode proteins directly involved in spore maturation
(ssgA) (101) or proteins involved in extracellular proteolysis (sgiA, sprT, sprU) (28, 32, 35). ArpA repression of
adpA is reversed by its interaction with A-factor. In S. fradiae, a -butyrolactone receptor called TylP represses
expression of tylS and tylQ, which control tylosin production (16, 76). In S. coelicolor, the SCB-1 receptor ScbR
directly represses kasO, an activator of the kas gene cluster, and feeds into the other antibiotic biosynthetic
pathways in a less-well-characterized manner (83,84).Repression by TylP and ScbR is relieved when their cog-
nate -butyrolactones bind (76, 83).
its precise role is unknown (100). Interestingly, govern both the generation of the spore-form-
AdpA also controls protease activity by regulating aerial hyphae and the production of one or
ing the transcription of sgiA, which encodes a more antibiotics. For instance, sporulation and
subtilisin inhibitor-like protein (28);sgmA,cod- the biosynthesis of the antibiotic tylosin by S.
ing for a zinc metallopeptidase (35);and the ser- fradiae are regulated by the direct binding of a
ine protease genes sprT and sprU (32). The -butyrolactone to its receptor, TylP, which
functions of the inhibitor and the proteases are negatively regulates expression of two genes,
not well understood at present. tylS and tylQ (Fig. 2).This pathway appears to
Although much less is known about the be more complex than that of S. griseus because
regulatory pathways controlled by the - while TylQ negatively regulates expression of
butyrolactones in other streptomycetes, many the positive regulator tylR, TylS activates it.
cases resemble S. griseus in that these molecules Ultimately both tylR and tylS serve to activate
transcription of genes involved in the produc- shown alongside that of TetR in Fig. 3. Like
tion of the antibiotic (16, 76). TetR, CprB is a homodimer, each monomer
In other species, however, including Strepto- consisting of 10 -helices separated by short
myces lavendulae (44) and S. coelicolor (25, 83), the turns or loops. The DNA-binding domain
roles of the -butyrolactones appear to be (helices 1 to 3) is linked to the ligand-binding
restricted to secondary metabolism. The best- domain (helices 5 to 10) by helix 4, which
understood example is S. coelicolor, which exhibits an elevated degree of mobility,enabling
produces at least four -butyrolacones (83). movement of the DNA-binding domain rela-
One of these molecules, SCB-1, has been tive to the ligand-binding domain (54).
shown to influence the biosynthesis of two The N-terminal helix-turn-helix DNA-
well-characterized antibiotics, actinorhodin binding motif (e.g., ArpA residues 38 to 50) is
and undecylprodigiosin, as well as a less- the most highly conserved block of sequence
well-characterized molecule produced by the among the -butyrolactone receptors. Specifi-
kas type I polyketide synthase gene cluster (83, cally, this includes helix 3 (ArpA residues 43 to
84). Despite differences in signal output, the 49) and the preceding turn (residues 39 to 42),
initial transduction pathway of -butyrolactone which by analogy with TetR is predicted to
signaling is conserved: SCB-1 binds to its insert into the major groove and recognize the
receptor, ScbR, which represses the expression operator sequence. Mutational analysis shows
of kasO. KasO, a Streptomyces antibiotic regula- that altering the valine at position 41 abolishes
tory protein, then directly activates expression the interaction of ArpA with DNA, but not its
of the cryptic kas biosynthetic gene cluster (84). capacity to bind -butyrolactone ligand, sup-
To date, it appears that all -butyrolactones porting the prediction that helix 3 contacts
bind conserved receptor proteins that act as DNA.This high degree of conservation in the
transcriptional repressors. There is significant DNA-binding helix accounts for the observa-
amino acid sequence homology among the tion that CprB can recognize the same DNA
known -butyrolactone receptors in Strepto- sequence motif as ArpA (77).
myces spp., and these receptors are related to the The CprB ligand-binding domain is made
TetR protein, which controls clinical resistance up of helices 5 to 10. Helices 4, 5, 6, 7, and 8
to the antibiotic tetracycline (27, 43). Like surround a putative ligand-binding cavity of
TetR, the -butyrolactone receptors interact ~20 by 5 Å, a size that would be a reasonable
with their target promoters by binding palin- fit for a -butyrolactone-like molecule. The
dromic sequences, and binding is relieved by inner surface of the pocket is lined mostly with
the interaction of a specific ligand (70). The hydrophobic side chains (55), though it also
greatest degree of conservation is in the N includes a number of hydrophilic residues.
termini, which include the DNA-binding Residues Q64, A94, W127, L157, G163, and
helix-turn-helix domain. This supports evi- L181 are conserved in ArpA (though L157
dence suggesting that several of these proteins and L181 correspond to I and V residues,
can bind the same DNA sequence (77). respectively). Residue W127 is universally con-
Although the C termini of the proteins are served in the ligand-binding domains of the
assembled around shared sequence motifs, they -butyrolactone receptors.A mutation altering
are not as well conserved, consistent with the this residue in ArpA blocks the relief of DNA
requirements for specific ligand interactions. binding by A factor, suggesting a possible role
The crystal structure of the CprB from S. coeli- for this residue in -butyrolactone binding
color, which is believed to be an orphan - (70). A W residue in TetR has been employed
butyrolactone receptor, has been solved (55). extensively for the development of tryptophan
CprB binds the same DNA seqence as ArpA; fluorescence assays for ligand binding (81, 82)
however, its ligand has not yet been identified and it is possible that W127 could be exploited
(77). The structure of the CprB monomer is in this way to investigate the interaction of
FIGURE 3 Structural similarities between the apo structures of (A) TetR (67) and (B) CprB (54). For clarity, fea-
tures are only identified for one monomer in each dimer;the second monomer is shown only as a C-alpha backbone
trace.An internal cavity is shown as a mesh surface representation for each protein. For TetR, this cavity shows the
tetracycline-binding region as seen in several structures of TetR bound to tetracycline or its derivatives (27, 43, 68).
The corresponding cavity in CprB also seems likely to be a ligand-binding site. Structurally analogous helices are
numbered identically in each protein; the only exception is the structurally analogous helices numbered 10 and 9 in
TetR and CprB, respectively, reflecting the numbering in the original publications.To emphasize structural similar-
ities between the two proteins, unique helices (helices 9 and 10 in CprB and TetR, respectively) are not shown.
CprB and other -butyrolactone receptors to coat the cell surface, creating a hydrophobic
with their ligands. layer overtop the cell wall.The mechanism by
which this layer drives aerial growth is not
SapB AND OTHER entirely clear, but elimination of the cohesive
HYDROPHOBIC PEPTIDES forces between water molecules covering
To date, three classes of secreted, hydrophobic filaments within a mycelium is prerequisite to
molecules have been shown to be involved in the emergence of aerial hyphae, and thus
aerial hyphae formation. In S. coelicolor, these sporulation. However, among streptomycete
include the chaplins (15, 20) and a small lan- surfactant-like molecules, only SapB (93), its S.
thionine-containing peptide, SapB (45, 85, 93), griseus orthologue,AmfS (89),and SapT (46) are
which has orthologues in S. griseus (89), Strepto- capable of restoring to bld mutants the capacity
myces avermitilis, and Streptomyces scabies (95), as to erect aerial filaments; the chaplins lack this
well as a functional homologue,SapT,produced ability (15, 20). Additionally, SapB and SapT
by Streptomyces tendae (46). In addition, a have been shown to accelerate sporulation in
peptide called goadsporin, produced by Strepto- wild type (J. M. Willey, unpublished data); the
myces sp. TP-A0584, stimulates antibiotic SapB orthologues from S. avermitilis and S.
production and sporulation in a number of scabies have not yet been purified.
other Streptomyces spp. (66). Both the chaplins, SapB, AmfS, and SapT do not appear to
which include a family of eight secreted function as signals as there is no evidence that
hydrophobic proteins (15, 20), and SapB (85) they mediate cellular behavior by binding to a
are highly surface active, and their extracellular specific receptor. Instead, they function in a
accumulation decreases the surface tension at manner analogous to the fungal hydrophobins,
the colony-air interface (13). SapB is believed a conserved group of proteins produced by the
basidiomycetes. All are highly surface active, to a variety of new stimuli that trigger the
amphilphilic, and capable of self-assembly (85, expression of sporulation genes (46, 95).
98). Structurally, however, SapB (45), AmfS The identification of the signal(s) and path-
(89), and SapT (46) bear no resemblance to the way that lead to SapB production and the for-
hydrophobins. SapB is the product of the ramS mation of an aerial mycelium is an important
gene, which encodes a prepeptide that is goal. In S. coelicolor, ramS is the second gene in
posttranslationally modified by the formation the ramCSAB operon.This operon is positively
of two lanthionine bridges and cleavage of an controlled by the response regulator RamR,
N-terminal leader sequence (45). The fungal which lacks a cognate signal kinase (36, 56, 61,
hydrophobins, on the other hand, are much 62). In S. griseus, the orthologous amfTSAB
larger proteins that lack posttranslational operon (88) is similarly governed by the RamR
modification (96). Nonetheless, extracellular orthologue, AmfR, production of which is
complementation tests using the fungal under the control of AdpA,ArpA, and A-factor.
hydrophobin SC3, SapT (which contains three An additional layer of complexity in S. griseus
methyllanthionine residues and a single lan- is that the amfTSAB operon is negatively
thionine [46]), and SapB have been key to our regulated by the DNA-binding protein BldD
understanding of how these morphogenetic (90).This appears not to be the case in S. coeli-
peptides function in the streptomycetes. Devel- color, where ramCSAB expression appears to be
opmental mutants of the fungus Schizophyllum BldD independent.
commune regain the capacity to form aerial In 2001, Onaka and colleagues reported the
structures when streptomycete peptide is added discovery of a 19-amino-acid peptide they
exogenously (97), while SC3 extracellularly named goadsporin, discovered while screening
complements S. coelicolor bld mutants (85).This for extracellular factors that could stimulate
phenomenon is not nonspecific; a variety actinorhodin production by Streptomyces livi-
of other bacterial surfactants have no effect dans (66). Goadsporin, like SapB and SapT, is
(72) and the Bacillus subtilis peptide surfactin a hydrophobic oligopeptide that is posttransla-
inhibits morphogenesis, raising questions tionally modified. However, unlike SapB and
about the interaction of these two soil microbes SapT, the pregoadsporin peptide is modified to
in nature (75). include four oxazole and two thiazole rings and
The role of SapB and SapT as surfactants is two dehydroserine residues. This molecule is
demonstrated by the ultrastructure of aerial therefore more closely related to the Escherichia
structures produced by peptide- and hydro- coli bacteriocin microcin; indeed, at high
phobin-treated bld mutants.These filaments are concentrations (60 M), goadsporin inhibits
undifferentiated vegetative hyphae that grow the growth of other streptomycetes but has
into the air, rather than the coiled aerial hyphae no effect on a variety of other bacteria. By
and spore chains characteristic of wild-type contrast, when 42 streptomycete strains were
colonies undergoing sporulation (85). On the exposed to the peptide at lower concentrations,
contrary, when these molecules are applied to a 20 responded by producing pigment and 32
ramS null mutant, which is blocked late in the initiated sporulation. However, the function of
developmental program, sporulation results goadsporin in the producing organism remains
(46). Taken together, these results suggest that unclear,as it does not appear to have a morpho-
the role of the surfactants is to release nascent genetic function, nor does it induce the pro-
aerial hyphae from the aqueous confines of the duction of other secondary metabolites (66).
substrate mycelium. It is hypothesized that by Interestingly, the goadsporin biosynthetic clus-
changing the growth environment of hyphae ter includes a gene encoding an immunity pro-
from the colony surface to the aerial mycelium, tein, suggesting that the peptide is produced in
the diffusion of SapB (and by extension, other inhibitory levels (65). The genetics and struc-
such streptomycete peptides) exposes hyphae ture of goadsporin thus seem to incorporate
features of both microcin and the lantibiotics: that factor C expression in an S.griseus A-factor-
modification enzymes encoded by genes deficient mutant restored the production of
within a biosynthetic cluster catalyze the for- several secreted proteins belonging to the A-
mation of microcin-like (oxazole and thiozole) factor regulon (7). Thus, it appears that there
and lantibiotic-like (dehydroalanine) residues. may be a physiological connection between
Its ABC transporter also serves to cleave the these two very different signaling molecules.
N-terminal leader (65, 73).
OTHER EXAMPLES OF CELL-CELL
FACTOR C COMMUNICATION IN THE
Factor C was initially isolated from the spent STREPTOMYCETES
medium of S. griseus 45H, now known to be S. It is striking that when streptomycetes are
flavofungini (7). Unlike S. coelicolor, both of grown on solid medium, colonies at high den-
these species readily sporulate in liquid (78). sity routinely display a more rapid rate of mor-
Addition of factor C in concentrations as low phogenesis than those at lower density. Recent
as 0.5 ng/ml of culture fluid is sufficient to cytological examination of surface grown Strep-
induce sporulation in susceptible, sporulation- tomyces spp. demonstrates that the pattern (not
deficient strains such as S. griseus 52-1 (10). In just the rate) of differentiation is density
addition, extracellular application of factor C dependent and that differentiation is associated
also stimulates sporulation by S. griseus bld with two rounds of programmed cell death
mutants grown on solid medium (11). Surpris- (50). An early death round occurs in substrate
ingly, factor C is a relatively large, basic protein hyphae shortly after spore germination and is
of 31,038 Da after cleavage of a 38-amino-acid characterized by alternating live and dead
leader (8, 9) that features a TAT secretion signal regions along compartmentalized filaments.
sequence (7). Its production is developmentally This occurs during the first 8 h of high-density
regulated, so that it is produced following the (106) spore germination. By contrast, at low
nutrient downshift required to stimulate sporu- density (6 103 spores), cell death does not
lation (11). Factor C appears to be present in occur until newly germinated hyphae begin to
certain Streptomyces spp. and some other bacter- touch one another. In both cases, live regions of
ial species, as shown by monoclonal antibody filaments give rise to successive waves of viable
raised against factor C (80). hyphae. At high density, these hyphae form
Factor C is unusual for a bacterial signaling macroscopic patches, whereas sparsely plated
molecule as it is a relatively large protein, spores produce viable hyphae only in micro-
unlike the more familiar oligopeptide and - scopic “islands.” The second death phase occurs
butyrolactone signals. Furthermore, amino acid at the periphery of these viable patches or
sequence analysis suggests that the N-terminal islands and is concurrent with the continued
residues 69–90 may span the membrane such growth of the viable, central regions that ulti-
that an N-terminal domain is intracellular and mately overspread the mycelium and undergo
the C terminus is extracellular (11). Interest- sporulation (47, 49). The heterogeneous, yet
ingly, exogenously added factor C disappears orderly, pattern of cell-density-dependent
rapidly from cultures to which it is added.There growth and death suggests that diffusible signals
are several possible explanations for this.It could may be involved. Here, two classes of signals
be taken up actively by cells as suggested by might be envisioned: signal(s) analogous to the
Szeszák et al. (80), spontaneously inserted into killing factor produced by B. subtilis cells com-
the cell membrane, or cleaved into smaller pep- mitted to sporulation that trigger cell death in
tides.The capacity of factor C to trigger sporu- sister cells (23), and molecules that stimulate
lation in a variety of S. griseus strains suggests cellular growth and/or differentiation.
that it may function via a conserved regulatory Although most studies of streptomycete
pathway (9, 10, 11), and it was recently reported intercellular signaling have focused on a single
species, it is clear that, in nature, interspecies the capacity of subinhibitory antibiotic concen-
communication is almost certainly occurring. trations to modulate the activity of nearly 5% of
This was shown by Ueda et al. (87) who per- bacterial promoters,including some that govern
formed extensive cross-feeding experiments quorum sensing (22). Furthermore, the discov-
with laboratory strains and natural isolates. ery that over 450 wild soil-dwelling bacterial
They found that 32 out of 33 freshly isolated isolates carry the resistance genes for at least
wild streptomycetes stimulate antibiotic pro- seven antibiotics (18) seems to contradict the
duction in 11 such strains and morphogenesis notion that antibiotics are produced only to
in 19 strains. Significantly, in no case could inhibit other, competing microbes. As species-
complementation be induced by the applica- specific (and sometimes strain-specific) extra-
tion of A-factor or the -butyrylactones of cellular chemicals, antibiotics may have evolved
other Streptomyces spp.Thus, it would seem that to enable communication between bacterial
interspecies communication among these populations.
microbes is common and may be mediated by
novel molecules and mechanisms. IMPORTANT FUTURE DIRECTIONS
Intercellular communication among wild The morphological and physiological com-
streptomycetes has been explored in more plexity of the streptomycetes is reflected in the
depth in S. griseus and Streptomyces tanashiensis. diversity of extracellular signaling molecules
Desferrioxamine E was identified as the com- that have been identified to date. A key ques-
pound produced by S. griseus that accelerates tion that needs to be addressed is the identity of
morphological differentiation in S. tanashiensis the predicted sporulation signals in S. coelicolor
(99). This molecule belongs to a class of (39, 58, 59, 94). For example, previous efforts
siderophores that is widely produced by a vari- suggested that an oligopeptide, imported by the
ety of Streptomyces spp. and other bacteria. BldK oligopeptide permease, serves to activate
Because the exchange of siderophores is known the formation of the aerial mycelium. At the
to be interspecific (74), these molecules may be time, however, it was not possible to purify
well suited for communication across species. enough of this material to determine its iden-
In a more general sense, this finding serves to tity (57).With the recent improvements in mass
remind us that in the economy of natural sys- spectrometry, the identification of this mole-
tems, a single molecule may serve more than cule should now be possible. In addition, iden-
one distinct function (17, 102). tifying the genes that transduce the various
The pamamycins, a group of macrolide signals is a high priority.There are no obvious
antibiotics produced by Streptomyces alboniger, orthologues of the rap genes, which serve to
demonstrate this multifunctionality. At subin- integrate peptide signaling into a phosphorelay
hibitory concentrations, they stimulate the in B. subtilis (4, 12, 91), so this work is likely to
formation of aerial hyphae, while at higher yield important new mechanistic insights about
concentrations they inhibit the growth of chemical communication and signal transduc-
nonproducing streptomycetes and other gram- tion in bacteria.
positive bacteria (51, 69). Like siderophore-
induced development, pamamycin signaling is REFERENCES
interspecific, rescuing or stimulating aerial 1. Aceti, D. J., and W. C. Champness. 1998.
hyphae formation in 20 different Streptomyces Transcriptional regulation of Streptomyces coelicolor
spp. (26). Indeed, it has been suggested that at pathway-specific antibiotic regulators by the absA
low concentrations (i.e.,those likely to be found and absB loci. J. Bacteriol. 180:3100–3106.
in nature), antibiotics are more likely to mediate 2. Ainsa, J. A., H. D. Parry, and K. F. Chater.
1999. A response regulator-like protein that
microbial metabolic activities and cellular functions at an intermediate stage of sporulation
behavior than efficiently kill neighboring cells in Streptomyces coelicolor A3(2). Mol. Microbiol.
(17, 101). This notion is further supported by 34:607–619.
3. Ainsa, J. A., N. J. Ryding, N. Hartley, K. C. 12. Bongiorni, C., S. Ishikawa, S. Stephenson, N.

Findlay, C. J. Bruton, and K. F. Chater. 2000. Ogasawara, and M. Perego. 2005. Synergistic
WhiA, a protein of unknown function conserved regulation of competence development in Bacillus
among gram-positive bacteria, is essential for subtilis by two Rap-Phr systems. J. Bacteriol.
sporulation in Streptomyces coelicolor A3(2). J. Bacte- 187:4353–4361.
riol. 182:5470–5478. 13. Capstick, D. S., J. M.Willey, M. J. Buttner, and
4. Auchtung, J. M., C. A. Lee, and A. D. M. A. Elliot. 2007. SapB and the chaplins: con-
Grossman. 2006. Modulation of the ComA- nections between morphogenetic proteins in
dependent quorum response in Bacillus subtilis by Streptomyces coelicolor. Mol. Microbiol. 64:602–611.
multiple Rap proteins and Phr peptides. J. Bacteriol. 14. Chater, K. F., and S. Horinouchi. 2003. Sig-
188:5273–5285. nalling early developmental events in two highly
5. Baltz, R. H. 1998. Genetic manipulation of diverged Streptomyces species. Mol. Microbiol.
antibiotic-producing Streptomyces.Trends Microbiol. 48:9–15.
6:76–83. 15. Claessen, D., R. Rink,W. de Jong, J. Siebring,
6. Bentley, S. D., K. F. Chater, A. M. Cerdeno- P. de Vreugd, F. G. Boersma, L. Dijkhuizen,
Tarraga, G. L. Challis, N. R.Thomson, K. D. and H.A.Wösten. 2003.A novel class of secreted
James, D. E. Harris, M.A. Quail, H. Kieser, D. hydrophobic proteins is involved in aerial hyphae
Harper, A. Bateman, S. Brown, G. Chandra, formation in Streptomyces coelicolor by forming
C. W. Chen, M. Collins, A. Cronin, A. Fraser, amyloid-like fibrils. Genes Dev. 17:1714–1726.
A. Goble, J. Hidalgo,T. Hornsby, S. Howarth, 16. Cundliffe, E., N. Bate, A. Butler, S. Fish, A.
C. H. Huang, T. Kieser, L. Larke, L. Murphy, Gandecha, and L. Merson-Davies. 2001.The
K. Oliver, S. O’Neil, E. Rabbinowitsch, M.A. tylosin-biosynthetic genes of Streptomyces fradiae.
Rajandream, K. Rutherford, S. Rutter, K. Antonie van Leeuwenhoek 79:229–234.
Seeger, D. Saunders, S. Sharp, R. Squares, S. 17. Davies, J., G. B. Spiegelman, and G. Yim.
Squares, K. Taylor, T. Warren, A. Wietzorrek, 2006.The world of subinhibitory antibiotic con-
J.Woodward, B. G. Barrell, J. Parkhill, and D. centrations. Curr. Opin. Microbiol. 9:445–453.
A. Hopwood. 2002.Complete genome sequence 18. D’Costa, V. M., K. M. McGrann, D. W.
of the model actinomycete Streptomyces coelicolor Hughes, and G. D.Wright. 2006. Sampling the
A3(2). Nature 417:141–147. antibiotic resistome. Science 311:374–377.
7. Birkó, Z., S. Bialek, K. Buzás, E. Szájli, B. A. 19. Dondero, N. C., and T. Scotti. 1957. Excretion
Traag, K. F. Medzihradszky, S. Rigali, E. by streptomycetes of factors causing formation of
Vijgenboom, A. Penyige, Z. Kele, G. P. van aerial hyphae by old cultures. J. Bacteriol. 73:584–
Wezel, and S. Biró. 2007. Functional mimicry: 585.
the secreted signaling protein factor C triggers the 20. Elliot, M. A., N. Karoonuthaisiri, J. Huang,
A-factor response regulon in Streptomyces griseus. M. J. Bibb, S. N. Cohen, C. M. Kao, and M. J.
Mol. Cell. Proteom. 6: 1248–1256. Buttner. 2003. The chaplins: a family of
8. Birkó, Z., F. Schauwecker, F. Pfennig, F. hydrophobic cell-surface proteins involved in aer-
Szeszák, S.Vitális, U. Keller, and S. Biró. 2001. ial mycelium formation in Streptomyces coelicolor.
Expression and rapid one-step purification of bio- Genes Dev. 17:1727–1740.
logically active His-tagged factor C by Ni(2) affin- 21. Fernandez-Moreno, M. A., A. J. Martin-
ity column chromatography. FEMS Microbiol. Lett. Triana, E. Martinez, J. Niemi, H. M. Kieser,
196:223–227. D. A. Hopwood, and F. Malpartida. 1992.
9. Birkó, Z.,A. Sumegi,A.Vinnai, G. van Wezel, abaA, a new pleiotropic regulatory locus for
F. Szeszák, S. Vitális, P. T. Szabó, Z. Kele, T. antibiotic production in Streptomyces coelicolor.
Janáky, and S. Biró. 1999. Characterization of J. Bacteriol. 174:2958–2967.
the gene for factor C,an extracellular signal protein 22. Goh, E. B., G.Yim,W.Tsui, J. McClure, M. G.
involved in morphological differentiation of Strep- Surette, and J. Davies. 2002. Transcriptional
tomyces griseus. Microbiology 145:2245–2453. modulation of bacterial gene expression by subin-
10. Biró, S., I. Békési, S. Vitális, and G. Szabó. hibitory concentrations of antibiotics. Proc. Natl.
1980. A substance effecting differentiation in Acad. Sci. USA 99:17025–17030.
Streptomyces griseus. Purification and properties. 23. Gonzalez-Pastor, J. E., E. C. Hobbs, and R.
Eur. J. Biochem. 103:359–363. Losick. 2003. Cannibalism by sporulating bacte-
11. Biró, S., Z. Birkó, and G. P. van Wezel. 2000. ria. Science 301:510–523.
Transcriptional and functional analysis of the gene 24. Hara, O., and T. Beppu. 1982. Mutants
for factor C, an extracellular signal protein blocked in streptomycin production in Strepto-
involved in cytodifferentiation of Streptomyces myces griseus—the role of A-factor. J. Antibiot.
griseus.Antonie van Leeuwenhoek 78:277–285. 35:349–358.
25. Hara, O., S. Horinouchi, T. Uozumi, and T. 37. Kelemen, G. H., P. Brian, K. Flardh, L.
Beppu. 1983. Genetic analysis of A-factor synthe- Chamberlin, K. F. Chater, and M. J. Buttner.
sis in Streptomyces coelicolor A3(2) and Streptomyces 1998. Developmental regulation of transcription
griseus. J. Gen. Microbiol. 129:2939–2944. of whiE, a locus specifying the polyketide spore
26. Hashimoto, M., T. Kondo, I. Kozone, H. pigment in Streptomyces coelicolor A3 (2). J. Bacteriol.
Kawaide, H. Abe, and M. Natsume. 2003. 180:2515–2521.
Relationship between response to and production 38. Kelemen, G. H., G. L. Brown, J. Kormanec, L.
of the aerial mycelium-inducing substances Potuckova, K. F. Chater, and M. J. Buttner.
pamamycin-607 and A-factor. Biosci. Biotechnol. 1996.The positions of the sigma-factor genes,whiG
Biochem. 67:803–808. and sigF, in the hierarchy controlling the develop-
27. Hinrichs, W., C. Kisker, M. Duvel, A. Muller, ment of spore chains in the aerial hyphae of Strepto-
K. Tovar, W. Hillen, and W. Saenger. 1994. myces coelicolor A3(2). Mol. Microbiol. 21:593–603.
Structure of the Tet repressor-tetracycline complex 39. Kelemen, G. H., and M. J. Buttner. 1998. Initi-
and regulation of antibiotic resistance. Science ation of aerial mycelium formation in Streptomyces.
264:418–420. Curr. Opin. Microbiol. 1:656–662.
28. Hirano, S., J. Y. Kato, Y. Ohnishi, and S. 40. Kendrick, K. E., and J. C. Ensign. 1983.Sporu-
Horinouchi. 2006. Control of the Streptomyces lation of Streptomyces griseus in submerged culture.
subtilisin inhibitor gene by AdpA in the A-factor J. Bacteriol. 155:357–366.
regulatory cascade in Streptomyces griseus.J.Bacteriol. 41. Khokhlov, A. S.,Tovarova, II, L. N. Borisova,
188:6207–6216. S.A. Pliner, L. N. Shevchenko, E. Kornitskaia,
29. Horinouchi, S. 2002. A microbial hormone, A- N. S. Ivkina, and I.A. Rapoport. 1967. [The A-
factor, as a master switch for morphological differ- factor,responsible for streptomycin biosynthesis by
entiation and secondary metabolism in Streptomyces mutant strains of Actinomyces streptomycini]. Dokl.
griseus. Front. Biosci. 7:2045–2057. Akad. Nauk. SSSR 177:232–235.
30. Hutchings, M. I., P.A. Hoskisson, G. Chandra, 42. Kinoshita, H., H. Ipposhi, S. Okamoto, H.
and M. J. Buttner. 2004. Sensing and responding Nakano,T. Nihira, and Y.Yamada. 1997. Buty-
to diverse extracellular signals? Analysis of the sen- rolactone autoregulator receptor protein (BarA) as
sor kinases and response regulators of Streptomyces a transcriptional regulator in Streptomyces virginiae.
coelicolor A3(2). Microbiology 150:2795–2806. J. Bacteriol. 179:6986–6993.
31. Janssen, P. H. 2006. Identifying the dominant 43. Kisker, C., W. Hinrichs, K. Tovar, W. Hillen,
soil bacterial taxa in libraries of 16S rRNA and W. Saenger. 1995. The complex formed
and 16S rRNA genes. Appl. Environ. Microbiol. between Tet repressor and tetracycline-Mg2
72:1719–1728. reveals mechanism of antibiotic resistance. J. Mol.
32. Kato, J.Y., W. J. Chi,Y. Ohnishi, S. K. Hong, Biol. 247:260–280.
and S. Horinouchi. 2005. Transcriptional con- 44. Kitani, S., H. Kinoshita, T. Nihira, and Y.
trol by A-factor of two trypsin genes in Streptomyces Yamada. 1999. In vitro analysis of the butyrolac-
griseus. J. Bacteriol. 187:286–295. tone autoregulator receptor protein (FarA) of
33. Kato, J. Y., I. Miyahisa, M. Mashiko, Y. Streptomyces lavendulae FRI-5 reveals that FarA
Ohnishi, and S. Horinouchi. 2004.A single tar- acts as a DNA-binding transcriptional regu-
get is sufficient to account for the biological effects lator that controls its own synthesis. J. Bacteriol.
of the A-factor receptor protein of Streptomyces 181:5081–5084.
griseus. J. Bacteriol. 186:2206–2211. 45. Kodani, S., M. E. Hudson, M. C. Durrant, M.
34. Kato, J. Y., Y. Ohnishi, and S. Horinouchi. J. Buttner, J. R. Nodwell, and J. M. Willey.
2005. Autorepression of AdpA of the AraC/XylS 2004. The SapB morphogen is a lantibiotic-like
family, a key transcriptional activator in the A- peptide derived from the product of the develop-
factor regulatory cascade in Streptomyces griseus. J. mental gene ramS in Streptomyces coelicolor. Proc.
Mol. Biol. 350:12–26. Natl.Acad. Sci. USA 101:11448–11453.
35. Kato, J. Y., A. Suzuki, H. Yamazaki, Y. 46. Kodani, S., M. A. Lodato, M. C. Durrant,
Ohnishi, and S. Horinouchi. 2002. Control by F. Picart, and J. M. Willey. 2005. SapT, a
A-factor of a metalloendopeptidase gene involved lanthionine-containing peptide involved in aerial
in aerial mycelium formation in Streptomyces hyphae formation in the streptomycetes. Mol.
griseus. J. Bacteriol. 184:6016–6025. Microbiol. 58:1368–1380.
36. Keijser, B. J., G. P. van Wezel, G.W. Canters,T. 47. Manteca, A., M. Fernandez, and J. Sanchez.
Kieser, and E. Vijgenboom. 2000. The ram- 2005. A death round affecting a young compart-
dependence of Streptomyces lividans differentiation mentalized mycelium precedes aerial mycelium
is bypassed by copper. J. Mol. Microbiol. Biotechnol. dismantling in confluent surface cultures of Strep-
2:565–574. tomyces antibioticus. Microbiology 151:3689–3697.
48. Manteca, A., M. Fernandez, and J. Sanchez. 60. Ochi, K. 1987.Metabolic initiation of differentia-
2006. Cytological and biochemical evidence tion and secondary metabolism by Streptomyces
for an early cell dismantling event in surface griseus: significance of the stringent response
cultures of Streptomyces antibioticus. Res. Microbiol. (ppGpp) and GTP content in relation to A factor.J.
157:143–152. Bacteriol. 169:3608–3616.
49. Manteca, A., M. Fernandez, and J. Sanchez. 61. O’Connor, T. J., P. Kanellis, and J. R.
2005. Mycelium development in Streptomyces Nodwell. 2002. The ramC gene is required for
antibioticus ATCC11891 occurs in an orderly pat- morphogenesis in Streptomyces coelicolor and
tern which determines multiphase growth curves. expressed in a cell type-specific manner under the
BMC Microbiol. 5:51. direct control of RamR. Mol. Microbiol. 45:45–57.
50. Manteca, A., U. Mader, B. A. Connolly, and J. 62. O’Connor,T. J., and J. R. Nodwell. 2005. Piv-
Sanchez. 2006. A proteomic analysis of Strepto- otal roles for the receiver domain in the mecha-
myces coelicolor programmed cell death. Proteomics nism of action of the response regulator RamR of
22:6008–6022. Streptomyces coelicolor. J. Mol. Biol. 351:1030–1047.
51. McCann, P. A., and B. M. Pogell. 1979. 63. Ohnishi, Y., H. Yamazaki, J. Y. Kato, A.
Pamamycin:a new antibiotic and stimulator of aer- Tomono, and S. Horinouchi. 2005. AdpA,
ial mycelia formation. J.Antibiot. 32:673–678. a central transcriptional regulator in the A-
52. Miguelez, E. M., C. Hardisson, and M. B. factor regulatory cascade that leads to morpholog-
Manzanal. 1999. Hyphal death during colony ical development and secondary metabolism in
development in Streptomyces antibioticus: morpho- Streptomyces griseus. Biosci. Biotechnol. Biochem.
logical evidence for the existence of a process of 69:431–439.
cell deletion in a multicellular prokaryote. J. Cell 64. Onaka, H., N. Ando, T. Nihira, Y. Yamada,
Biol. 145:515–525. T. Beppu, and S. Horinouchi. 1995. Cloning
53. Molle, V., and M. J. Buttner. 2000. Different and characterization of the A-factor recep-
alleles of the response regulator gene bldM arrest tor gene from Streptomyces griseus. J. Bacteriol.
Streptomyces coelicolor development at distinct 177:6083–6092.
stages. Mol. Microbiol. 36:1265–1278. 65. Onaka, H., M. Nakaho, K. Hayashi, Y.
54. Natsume, R., Y. Ohnishi, T. Senda, and S. Igarashi, and T. Furumai. 2005. Cloning and
Horinouchi. 2004. Crystal structure of a characterization of the goadsporin biosynthetic
gamma-butyrolactone autoregulator receptor gene cluster from Streptomyces sp.TPA0584. Micro-
protein in Streptomyces coelicolor A3(2). J. Mol. Biol. biology 151:3923–3933.
336:409–419. 66. Onaka, H., H.Tabata,Y. Igarashi,Y. Sato, and
55. Natsume, R., R. Takeshita, M. Sugiyama, Y. T. Furumai. 2001. Goadsporin, a chemical sub-
Ohnishi, T. Senda, and S. Horinouchi. 2003. stance which promotes secondary metabolism and
Crystallization of CprB, an autoregulator-receptor morphogenesis in streptomycetes. I. Purification
protein from Streptomyces coelicolor A3(2).Acta Crys- and characterization. J.Antibiot. 54:1036–1044.
tallogr. D Biol. Crystallogr. 59:2313–2315. 67. Orth, P., F. Cordes, D. Schnappinger, W.
56. Nguyen, K. T., J. M. Willey, L. D. Nguyen, Hillen, W. Saenger, and W. Hinrichs. 1998.
L. T. Nguyen, P. H. Viollier, and C. J. Conformational changes of the Tet repressor
Thompson. 2002. A central regulator of mor- induced by tetracycline trapping. J. Mol. Biol.
phological differentiation in the multicellular 279:439–447.
bacterium Streptomyces coelicolor. Mol. Microbiol. 68. Orth, P., D. Schnappinger, W. Hillen, W.
46:1223–1238. Saenger, and W. Hinrichs. 2000. Structural basis
57. Nodwell, J. R., and R. Losick. 1998. Purifica- of gene regulation by the tetracycline inducible Tet
tion of an extracellular signaling molecule involved repressor-operator system. Nat. Struct. Biol.
in production of aerial mycelium by Streptomyces 7:215–219.
coelicolor. J. Bacteriol. 180:1334–1337. 69. Pogell, B. M. 1998. The pamamycins: develop-
58. Nodwell, J. R., K. McGovern, and R. Losick. mental autoregulators and antibiotics from Strepto-
1996. An oligopeptide permease responsible for myces alboniger. A review and update. Cell. Mol. Biol.
the import of an extracellular signal governing aer- 44:461–463.
ial mycelium formation in Streptomyces coelicolor. 70. Ramos, J. L., M. Martinez-Bueno, A. J.
Mol. Microbiol. 22:881–893. Molina-Henares, W. Teran, K. Watanabe, X.
59. Nodwell, J. R., M. Yang, D. Kuo, and R. Zhang, M. T. Gallegos, R. Brennan, and R.
Losick. 1999. Extracellular complementation and Tobes. 2005. The TetR family of transcriptional
the identification of additional genes involved in repressors. Microbiol. Mol. Biol. Rev. 69:326–356.
aerial mycelium formation in Streptomyces coelicolor. 71. Retzlaff, L., and J. Distler. 1995.The regulator
Genetics 151:569–584. of streptomycin gene expression, StrR, of Strepto-
myces griseus is a DNA binding activator protein 83. Takano, E., R. Chakraburtty, T. Nihira, Y.
with multiple recognition sites. Mol. Microbiol. Yamada, and M. J. Bibb. 2001. A complex role
18:151–162. for the gamma-butyrolactone SCB1 in regulating
72. Richter, M., J. M. Willey, R. Sussmuth, G. antibiotic production in Streptomyces coelicolor
Jung, and H. P. Fiedler. 1998. Streptofactin, a A3(2). Mol. Microbiol. 41:1015–1028.
novel biosurfactatn with aerial mycelium inducing 84. Takano, E., H. Kinoshita, V. Mersinias, G.
activity from Streptomyces tendae Tu 901/8c. FEMS Bucca, G. Hotchkiss, T. Nihira, C. P. Smith,
Microbiol. Lett. 163:165–172. M. Bibb,W.Wohlleben, and K. Chater. 2005.A
73. Sahl, H. G., and G. Bierbaum. 1998. Lantibi- bacterial hormone SCB1 directly controls the
otics: biosynthesis and biological activities of expression of a pathway-specific regulatory gene
uniquely modified peptides from gram-positive in the cryptic type I polyketide biosynthetic gene
bacteria. Annu. Rev. Microbiol. 52:41–79. cluster of Streptomyces coelicolor. Mol. Microbiol.
74. Sobis-Glinkowska, M., P. Lisiecki, and 56:465–479.
J. Mikucki. 1995. The interchangeability of 85. Tillotson, R. D., H. A. Wösten, M. Richter,
siderophores in staphylococci. Acta Microbiol. Pol. and J. M. Willey. 1998. A surface active protein
44:127–134. involved in aerial hyphae formation in the fila-
75. Straight, P. D., J. M. Willey, and R. Kolter. mentous fungus Schizophillum commune restores the
2006. Interactions between Streptomyces coelicolor capacity of a bald mutant of the filamentous bac-
and Bacillus subtilis: role of surfactants in raising terium Streptomyces coelicolor to erect aerial struc-
aerial structures. J. Bacteriol. 188:4918–4925. tures. Mol. Microbiol. 30:595–602.
76. Stratigopoulos, G., A. R. Gandecha, and E. 86. Ueda, K., C. W. Hsheh, T. Tosaki, H.
Cundliffe. 2002.Regulation of tylosin production Shinkawa, T. Beppu, and S. Horinouchi.
and morphological differentiation in Streptomyces 1998. Characterization of an A-factor-responsive
fradiae by TylP, a deduced gamma-butyrolactone repressor for amfR essential for onset of aerial
receptor. Mol. Microbiol. 45:735–744. mycelium formation in Streptomyces griseus. J. Bacte-
77. Sugiyama, M., H. Onaka, T. Nakagawa, and riol. 180:5085–5093.
S. Horinouchi. 1998. Site-directed mutagenesis 87. Ueda, K., S. Kawai, H. Ogawa, A. Kiyama,T.
of the A-factor receptor protein:Val-41 important Kubota, H. Kawanobe, and T. Beppu. 2000.
for DNA-binding and Trp-119 important for lig- Wide distribution of interspecific stimula-
and-binding. Gene 222:133–144. tory events on antibiotic production and sporula-
78. Szabó, G.,T.Vályi-Nagy, and S.Vitális. 1962. tion among Streptomyces species. J. Antibiot. 53:
A new factor regulating life cycle of Streptomyces 979–982.
griseus. State Publishing House of Medical Litera- 88. Ueda, K., K. Miyake, S. Horinouchi, and T.
ture, Moscow, Russia. Beppu. 1993. A gene cluster involved in aerial
79. Szabó, P. T., Z. Kele, Z. Birkó, F. Szeszák, S. mycelium formation in Streptomyces griseus encodes
Biró, and T. Janáky. 1999.Identification of factor proteins similar to the response regulators of two-
C protein from Streptomyces griseus by microelec- component regulatory systems and membrane
trospray mass spectrometry. J. Mass Spectrom. translocators. J. Bacteriol. 175:2006–2016.
34:1312–1316. 89. Ueda, K., K. Oinuma, G. Ikeda, K. Hosono,
80. Szeszák, F., S. Vitális, and G. Szabó. 1991. Y. Ohnishi, S. Horinouchi, and T. Beppu.
Presence of factor C in streptomyces and other 2002. AmfS, an extracellular peptidic mor-
bacteria, p. 11–18. In S. Baumberg, H. Krügel, and phogen in Streptomyces griseus. J. Bacteriol. 184:
D. Noack (ed.), Genetics and Product Formation in 1488–1492.
Streptomyces. Plenum, New York, N Y. 90. Ueda, K., H.Takano, M. Nishimoto, H. Inaba,
81. Takahashi, M., L. Altschmied, and W. Hillen. and T. Beppu. 2005. Dual transcriptional control
1986. Kinetic and equilibrium characterization of of amfTSBA, which regulates the onset of cellular
the Tet repressor-tetracycline complex by fluores- differentiation in Streptomyces griseus. J. Bacteriol.
cence measurements. Evidence for divalent metal 187:135–142.
ion requirement and energy transfer. J. Mol. Biol. 91. Veening, J. W., L. W. Hamoen, and O. P.
187:341–348. Kuipers. 2005. Phosphatases modulate the
82. Takahashi, M., J. Degenkolb, and W. bistable sporulation gene expression pattern in
Hillen. 1991. Determination of the equilibrium Bacillus subtilis. Mol. Microbiol. 56:1481–1494.
association constant between Tet repressor and 92. Vujaklija, D., K. Ueda, S. K. Hong,T. Beppu,
tetracycline at limiting Mg2 concentrations: and S. Horinouchi. 1991. Identification of an A-
a generally applicable method for effector- factor-dependent promoter in the streptomycin
dependent high-affinity complexes. Anal. Biochem. biosynthetic gene cluster of Streptomyces griseus.
199:197–202. Mol. Gen. Genet. 229:119–128.
93. Willey, J., R. Santamaria, J. Guijarro, M. 98. Wösten, H.A., and J. M.Willey. 2000.Surface-
Geistlich, and R. Losick. 1991. Extracellular active proteins enable microbial aerial hyphae to
complementation of a developmental mutation grow into the air. Microbiology 146:767–773.
implicates a small sporulation protein in aerial 99. Yamanaka, K., H. Oikawa, H. O. Ogawa, K.
mycelium formation by S. coelicolor. Cell 65: Hosono, F. Shinmachi, H. Takano, S.
641–650. Sakuda,T. Beppu, and K. Ueda. 2005.Desfer-
94. Willey, J., J. Schwedock, and R. Losick. 1993. rioxamine E produced by Streptomyces griseus
Multiple extracellular signals govern the produc- stimulates growth and development of Strepto-
tion of a morphogenetic protein involved in aerial myces tanashiensis. Microbiology 151:2899–2905.
mycelium formation by Streptomyces coelicolor. 100. Yamazaki, H., Y. Ohnishi, and S. Hori-
Genes Dev. 7:895–903. nouchi. 2000. An A-factor-dependent extracy-
95. Willey, J. M., A.Willems, S. Kodani, and J. R. toplasmic function sigma factor ((AdsA)) that is
Nodwell. 2006. Morphogenetic surfactants and essential for morphological development in
their role in the formation of aerial hyphae in Streptomyces griseus. J. Bacteriol. 182:4596–4605.
Streptomyces coelicolor. Mol. Microbiol. 59:731–742. 101. Yamazaki, H., Y. Ohnishi, and S. Hori-
96. Wösten, H. A., and M. L. de Vocht. 2000. nouchi. 2003.Transcriptional switch on of ssgA
Hydrophobins,the fungal coat unravelled.Biochim. by A-factor, which is essential for spore septum
Biophys.Acta 1469:79–86. formation in Streptomyces griseus. J. Bacteriol. 185:
97. Wösten, H. A., M. A. van Wetter, L. G. 1273–1283.
Lugones, H. C. van der Mei, H. J. Busscher, 102. Yim, G., H. H.Wang, and J. Davies. 2006.The
and J. G.Wessels. 1999.How a fungus escapes the truth about antibiotics. Int. J. Med. Microbiol.
water to grow into the air. Curr. Biol. 9:85–88. 296:163–170.
METABOLITES AS INTERCELLULAR
SIGNALS FOR REGULATION OF
COMMUNITY-LEVEL TRAITS
Russell D. Monds and George A. O’Toole
8
The concept of bacteria as solitary nomadic description of the response and the nature of
individuals that act without recourse to the the signaling will be instructive. In the simplest
activities of their neighbors has been overturned model, quorum sensing refers to the process by
in recent years by research demonstrating that which cells sense overall population size and
bacteria can exist in organized groups that regulate activities based on achieving a thresh-
exhibit community-level traits and that bacteria old density of cells in their environment.Mech-
are also capable of cell-to-cell communication anistically,this is achieved by each cell excreting
in the form of diffusible extracellular signals, a basal level of signaling molecule, often a
allowing coordination of activities through homoserine lactone, into a diffusion-limited
space and time (32). In this chapter we review environment. In this way, the concentration of
evidence for the role of chemical-dependent signal in the extracellular space is roughly pro-
signaling in the regulation of community-level portional to the number of bacteria in that
traits such as microbial biofilm formation. space producing that signal. Cells are also able
However, instead of more canonical signaling to specifically detect the levels of this molecule,
systems, we describe work on new and emerg- such that when a certain threshold concentra-
ing systems that describe roles for excreted cel- tion is reached, indicating a certain population
lular metabolites as intercellular signals. size, specific regulatory pathways are activated.
The current paradigm for cell-to-cell com- The biological characteristics of homoserine
munication in bacteria is largely based on the lactones that characterize them as an intercellu-
signaling molecules homoserine lactones and lar signaling molecule are of consequence
their role in the phenomenon referred to as because they undoubtedly reflect the bounds
quorum sensing (94).The specifics of this sig- for criteria used to identify and evaluate new
naling system are well reviewed elsewhere in putative signaling molecules and systems. For
this book; however, for our purposes a brief instance, homoserine lactones are not thought
to play roles in central processes of cell metabo-
lism, such that they are viewed as being pro-
Russell D. Monds and George A. O’Toole Department of
Microbiology and Immunology, Dartmouth Medical School, duced and excreted for the sole purpose of
Hanover, New Hampshire 03755. providing the chemical means for intercellular
105
106 ■ MONDS AND O’TOOLE
signaling. Historically, signaling pathways are with perturbations to core metabolic networks.
seen as overarching networks that control As a consequence, it is much harder to show
metabolic networks but are not part of them. causal relationships between genes and pheno-
We would argue that this constitutes the types.Why is it that mutants in metabolic path-
canonical view of a bona fide signaling system, ways are often reported in genetic screens but
which as a consequence has limited the scope rarely pursued further? We would argue it is
and breadth of biological processes and mole- because such mutants are less tractable, rather
cules that have been investigated under the than less interesting.The works we discuss here
framework of intercellular signaling. were chosen because they provide different
Metabolic processes are indeed the heart perspectives on metabolites and their biologi-
of any microbe’s existence.They are responsible cal roles, as well as extend the scope with
for the generation of energy and carbon which we think about communication among
equivalents required to drive cellular biosyn- microbes.The validity of these claims will, in all
thetic processes.As a result, metabolic processes cases, need further rigorous experimentation.
and the products of these processes literally Throughout, we have made an effort to discuss
reflect the energy and physiological status of the critical experiments that are pertinent to par-
cell. The extension from this logic is that ticular cases, and at the end discuss ecological,
metabolites also represent good candidates for evolutionary, and empirical criteria for the
directly relaying information about a particular characterization of intercellular signals,address-
cell’s microniche to other cells in adjacent but ing specific criticisms voiced over the role of
potentially distinct microniches. The use of metabolites as intercellular signals.
molecules as chemical signals that are directly
connected with the relevant internal physiology ANTIBIOTICS
could, in many ways, be more energy and time Most microbial small molecules have been stud-
efficient. Rather than the levels of a metabolite ied because of their ability to inhibit the growth
having to activate synthesis and export of a dis- of another microorganism. Molecules that pos-
tinct and dedicated signaling molecule, the sess these activities are commonly referred to as
metabolite itself (or a derivative of that metabo- antibiotics.The therapeutic value of antibiotics
lite) plays dual roles, thereby eliminating the is unquestionable; however, understanding the
need for intermediate processes. biological role of these compounds in natural
The goal of this chapter is to reconceptualize settings may not be as intuitive as the definition
the current view of metabolism and intercellu- of antibiotics would suggest (99). Classically, we
lar signaling as mechanistically independent have inferred that the properties of antibiotics
processes, rather than provide a comprehensive seen as useful for humans are also the properties
overview of all signaling pathways.To this end, selected for through microbial evolution.
we discuss work from numerous laboratories Indeed, the ability to produce compounds that
that present varying degrees of evidence sup- inhibit the growth of one’s competitors seems
porting a role for excreted metabolites in regu- of obvious benefit within niches with limited
lating the group behavior of microbes, such as resources. However, in recent years experimen-
biofilm formation, swarming, and filamentous tal evidence has been published indicating new
growth. It is important to note that our goal is roles for antibiotics as signaling molecules that
not to convince you that all of the processes we can modify gene expression when present at
describe are bona fide examples of metabolic subinhibitory concentrations (22, 33). These
intercellular sensing. Ultimately, separating and results suggest that, depending on the concen-
characterizing two distinct biological roles for a tration, antibiotics have different biological
metabolic process are experimentally difficult. consequences. In this section we discuss recent
This is largely because there is a good chance work indicating roles for antibiotics as global
that secondary indirect effects will be associated regulators of biofilm formation.
8. METABOLITE SIGNALS FOR COMMUNITY TRAIT REGULATION ■ 107
Rachid and colleagues reported that subin- tions to a response regulator (referred to as arr)
hibitory concentrations of tetracycline and that inhibited the ability of tobramycin to
the streptogramin antibiotic quinupristin- enhance biofilm formation.
dalfopristin promote in vitro biofilm formation Arr is predicted to be an inner membrane
of Staphylococcus epidermidis 220 (76). Increases protein that encodes a cytoplasmic EAL
in biofilm formation were correlated with a domain. These domains have been implicated
10-fold increase in expression of the ica operon in specifically degrading the intracellular sec-
upon treatment with antibiotic. The ica locus ond messenger c-di-GMP, and the levels of this
encodes enzymes that are responsible for syn- molecule are often associated with modulation
thesis of a polysaccharide intercellular adhesin of surface-associated properties (15, 80). In this
(PIA), and previous studies have shown an way,Arr could form the regulatory connection
important role for PIA in promoting multilay- between extracellular concentrations of amino-
ered biofilm formation (34, 40). In this case it is glycosides and modulation of biofilm forma-
not known whether treatment with tetracy- tion via changes to the levels of intracellular
cline or quinupristin-dalfopristin is eliciting c-di-GMP. One point of note is that mutations
specific or global changes in transcription; to arr were also seen to increase susceptibility of
however, some level of specificity of interaction P. aeruginosa to tobramycin treatment. This is
was shown by the demonstration that other interesting as it further supports the idea that
antibiotics such as oxacillin, chloramphenicol, antibiotic resistance within biofilms has a
and vancomycin had no effect on transcription genetic component, but it does complicate the
of the ica operon (76). The potential for these authors’ analysis since it is unclear to what
mechanisms to be clinically relevant was degree the increased susceptibility of the arr
demonstrated by the ability of quinupristin- mutant to tobramycin indirectly leads to reduc-
dalfopristin to substantially enhance biofilm tions in biofilm formation.
formation of two clinical isolates in a dose- In a subsequent study, Linares et al. looked at
dependent fashion. the changes in gene expression of planktonic
A role for antibiotics as specific signals regu- cultures of P. aeruginosa exposed to subin-
lating surface commitment was further elabo- hibitory concentrations of the antibiotics
rated in work presented by Hoffman et al. tobramycin, tetracycline, and norfloxacin (56).
The authors demonstrated that subinhibitory All three antibiotics conferred numerous yet
concentrations of aminoglycoside antibiotics distinct alterations to gene expression relative to
induced biofilm formation in both Escherichia an untreated control.These changes in expres-
coli and Pseudomonas aeruginosa. In particular, sion profiles were correlated with phenotypic
growth in the presence of tobramycin resulted differences. Tobramycin as well as tetracycline
in 3.4 times the amount of surface-attached and norfloxacin were demonstrated to induce
biomass for P. aeruginosa biofilms (41). Impor- biofilm formation by P. aeruginosa. In addition,
tantly, the concentration of tobramycin was tobramycin was seen to increase cell motility,
only a third of the minimum inhibitory con- whereas tetracycline was observed to activate
centration (MIC) and did not result in any inhi- expression of type III secretion. All antibiotics
bition of growth or global protein synthesis. were used at concentrations below their MIC.
Similar trends were seen with other aminogly- Phenazines constitute a broad range of hete-
cosides; however, antibiotics such as chloram- rocyclic compounds that have been largely
phenicol and carbenicillum had no observed studied on the basis of their antibiotic effects
effect, arguing that a general and nonspecific (62, 73). Pyocyanin is one class of phenazine
stress response was not the cause of the pheno- that is known to be secreted by P. aeruginosa and
type. The existence of specific pathways is a known virulence factor (54). Recent work
required for sensing and responding to amino- in the Newman laboratory has shown that
gylcosides was supported by isolation of muta- pyocyanin can alter the transcription of a
specific subset of P. aeruginosa genes, suggesting understand the role of antibiotics in natural
an alternative role for pyocyanin as an intercel- environments. The ability of a small molecule
lular signaling molecule (24). Intriguingly, sev- to inhibit the growth of other microorganisms
eral of the genes regulated by pyocyanin were was easily translated into an obvious biological
previously shown to be under quorum-sensing benefit that could explain their adaptive benefit
control. Combined with the fact that and basis of selection through evolution.
phenazine synthesis itself is controlled by ter- Inhibiting the growth of competitors could be
minal effectors of the quorum-sensing net- of significant advantage in many niches with
work, these data suggest that pyocyanin limited resources. However, questions have
represents a new terminal physiological signal been raised as to whether this is the correct
in the quorum-sensing network of P. aeruginosa. interpretation. For instance, the concentration
It was also demonstrated that a known tran- of antibiotics in soil is often very low, and the
scriptional regulator, soxR, was required for fact that many antibiotics are produced pre-
transcriptional modulation of mexGHI-opmD dominantly in stationary phase seems at odds
and PA2274 by pyocyanin. This finding sup- with a role in excluding competitors from
ports the idea that transcriptional modulation niche-specific resources (99). The possibility
by pyocyanin is not a consequence of nonspe- that antibiotics can act as intercellular signaling
cific mechanisms; however, it is still not known molecules has catalyzed new ways of thinking
how pyocyanin levels alter soxR activity. about their roles in natural communities as well
Together these studies offer a different per- as the selective forces behind their evolution.
spective on the biological roles of antibiotics in Investigations of antibiotics as intercellular
natural settings. What we are seeing is that signals are at their early stages. Many questions
antibiotics can have different biological effects need to be addressed to solidify this new bio-
at different concentrations.Whereas high con- logical description. For instance, it will be
centrations of antibiotics can inhibit cell important to investigate whether mechanisms
growth, lower concentrations can act as signals exist to connect the specific sensing of antibi-
leading to specific changes to gene expression. otics in the environment to regulation of spe-
This biphasic response has been suggested as cific transcriptional outputs. In this regard,
appropriately considered under the toxicologi- Bader et al. demonstrated that the PhoQ/PhoS
cal framework of hormesis (99).We also know two-component system of Salmonella enterica
from the studies above that activation of general serovar Typhimurium is directly activated by
stress pathways is insufficient to explain the spe- the binding of antimicrobial peptides produced
cific impact on transcription mediated by by the innate immune system (4).
treatment with subinhibitory concentrations In the above case, the antimicrobial agent is
of antibiotics. This observation is important, detected and the signal transduced via a path-
because it reinforces the idea that antibiotics are way that is distinct from that serving as the tar-
likely interacting with cellular targets distinct get for the antimicrobial’s toxicity. In other
from those associated with the mechanisms by cases it may be that the response to subin-
which they are known to inhibit cell growth. hibitory levels of antibiotics is elicited via
Antibiotics represent a small subset of the mechanisms similar to those required for inhi-
vast array of small molecules produced and bition of growth at higher concentrations. For
excreted by microorganisms. Their identifica- example, low concentrations of antibiotics may
tion and categorization were based on their simply alter the kinetics of the interaction,lead-
ability to inhibit the growth of other microor- ing to effects subtle enough to not inhibit
ganisms (73). In this way the term antibiotic is growth but still able to alter gene transcription
useful as it relays information about a mole- rates. Probing the molecular basis of physiolog-
cule’s properties, but it may also have been a ical responses to subinhibitory levels of antibi-
hindrance by dominating how we have tried to otics will be important for developing and
evaluating a new biological perspective of mutations to enzymes required for synthesis of

antibiotics and their evolution. putrescine from arginine resulted in a substantial
delay in swarmer cell differentiation, as well as
POLYAMINES an overall decrease in the speed of translocation
Polyamines are a group of polycationic mole- across the surface (85).The defects in swarming
cules that are known to play important and could be cross-complemented by a wild-type
diverse roles in the cellular biology of both strain, suggesting putrescine could be acting as
eukaryotes and prokaryotes (88). The most an extracellular signal.In support of this conclu-
common polyamines in microbial systems are sion, addition of exogenous putrescine at physi-
putrescine and spermidine, but others such as ologically relevant concentrations was able to
cadeverine and norspermidine are also known restore normal differentiation as well as normal
to play roles. Polyamines have been demon- translocation velocities (85).Furthermore,addi-
strated to have effects on both transcription and tion of exogenous putriscene was seen to pro-
translation, which is consistent with the fact mote swarmer cell differentiation in the wild
that in vivo most polyamines have been shown type. Together, these results suggest a role for
to exist predominantly as complexes with putrescine as an extracellular signal that can
RNA.The levels of polyamines in the cell are affect swarming motility in P. mirabilis.
generally in the millimolar range but are known At present the mechanism by which
to vary markedly in response to different putrescine modulates swarming is not known.
growth conditions (88, 101). Based on polyamine signaling in other systems,
The internal concentration of polyamines is we may anticipate that putrescine directly
tightly regulated through a combination of modulates transcription or translation of factors
biosynthesis, transport, and excretion. In E. coli, important for cellular differentiation. Ulti-
putrescine is synthesized either by decarboxyla- mately, the identification and characterization
tion of L-ornithine or decarboxylation of L- of mutants recalcitrant to stimulation by
arginine followed by removal of a urea putrescine will be required to gain insight into
molecule (88).As well as dedicated enzymes for whether an active cellular mechanism exists for
synthesis, E. coli is also known to have special- control of swarming by polyamines.
ized transport systems for polyamine uptake
and excretion (42,46,84).Current research also Biofilm Formation by Yersinia pestis
suggests that conserved mechanisms exist for Levels of the polyamine putrescine have also
biosynthesis and transport of polyamines in been demonstrated as important for biofilm
other bacterial systems (43, 58). formation by Yersinia pestis (70). Single muta-
tions to arginine decarboxylase (speA) and
Swarming in Proteus mirabilis ornithine decarboxylase (speC) led to decreases
One of the first reports linking polyamines to in the concentration of intracellular putrescine,
control of community-level traits came from whereas the double mutant did not produce
research on swarming by Proteus mirabilis. detectable amounts of either putrescine or sper-
Swarming is a form of motility used by a range midine. After confirming the biosynthetic role
of bacteria to translocate across surfaces and is of speA and speC, the authors went on to show
typified by a high level of multicellular coordi- that reduction in biofilm formation,as analyzed
nation (79). Swarming in P. mirabilis is known to by CSLM and crystal violet staining, correlated
require differentiation between two distinct cell with the intracellular levels of putrescine. For
types referred to as vegetative and swarmer. example, the speC, speA, and speAC mutants
Morphologically, swarmer cells can be 10 to 40 showed increasing reductions in biofilm forma-
times longer than vegetative cells and show a tion that correlated with the relative levels of
large increase in the number of flagella. In 2004 putrescine. Growth of speA and speC mutants
Sturgill and Rather presented data showing that was comparable to wild type, whereas the
double mutant grew at 65% of the exponential tional experiments went on to show that the
rate of the wild type. Interestingly, addition of gene mbaA was also required for stimulation of
exogenous putrescine was able to restore biofilm formation by norsperimdine.The mbaA
biofilm formation to the speC speA double gene encodes a predicted inner membrane pro-
mutant; however, no restoration was seen with tein that was previously implicated as a negative
addition of either spermidine or agmatine. So regulator of biofilm formation in V. cholerae (7).
the requirement for putrescine for competent The nspS mbaA double mutant was shown to
biofilm formation seems reasonably specific. have the same enhanced biofilm phenotype as
Patel et al. consider their results in the context the mbaA single mutant, indicating that nspS
of putrescine as an intracellular signal; however, acts upstream of mbaA in a genetic pathway.
we would suggest that a role for putrescine as an Mba is a putative EAL/GGDEF protein, which
extracellular signal in Y. pestis is also possible. It implicates c-di-GMP signaling networks in
is likely that intracellular levels of putrescine are norspermidine-dependent biofilm regulation.
determined by the fine balance among synthe- Norspermidine is produced by V. cholerae as
sis, transport, and excretion. In this way, the lev- well as a range of other bacteria, archaea (37),
els of extracellular putrescine affect intracellular and eukarya (36), suggesting that the potential
levels, and in turn the levels of extracellular for interdomain signaling exists. Karatan et al.
putrescine are controlled by the balance of syn- note that the concentration of norspermidine
thesis, transport, and excretion exhibited by in seawater is at levels unlikely to activate nspS-
organisms in that immediate environment. dependent biofilm pathways;however,norsper-
midine signaling may promote association of V.
Biofilm Formation by Vibrio cholerae cholerae with other organisms that maintain a
Norspermidine differs from spermidine by one higher concentration gradient of norspermi-
carbon residue, and in Vibrio cholerae norsper- dine in their immediate environment. One
midine constitutes the major polyamine (97). important question is whether intracellular
Studies by Karatan et al. demonstrated a role for pools of norspermidine are important for con-
norspermidine as an intercellular signal trol of biofilm formation in V. cholerae. For
involved in the stimulation of biofilm forma- instance, does the balance of transport and
tion of V. cholerae. Mutants in a gene referred to secretion play an important role in modulating
in the study as nspS (norspermidine sensor) norspermidine intercellular signaling,as may be
resulted in defects in biofilm formation relative the case for putrescine signaling in P.mirabilis? It
to the wild type (45). Specifically, the nspS would be interesting to look at the phenotypes
mutant, although forming a monolayer in a of norspermidine synthetic and transport/
fashion similar to the wild type, did not accu- efflux mutants in regard to biofilm formation
mulate biomass on the surface to the same by V. cholerae.
degree. Sequence similarity of nspS to potD, the
periplasmic binding component of the E. coli Summary
spermidine transport system, suggested a role Polyamines have been long recognized as
for polyamines in regulating V. cholerae biofilm important cellular components; however,
development. Concordant with this notion, knowledge of their specific role in the cell has
addition of norspermidine was shown to remained elusive.Evidence that polyamines can
enhance wild-type biofilm formation in a effect transcription and translation through
dose-dependent fashion. Maximal biofilm for- interactions with RNA has favored a view of
mation (threefold increase) was seen with a polyamines as intracellular regulators. In this
norspermidine concentration of 100 M.The chapter we have discussed evidence showing
effect of norspermidine was dependent on the potential for polyamines to also play roles as
nspS, which supports the idea that norspermi- extracellular signals.The existence of synthetic
dine may be the signal sensed by NspS. Addi- pathways, as well as active mechanisms for
import and export of polyamines, indicates the L-rhamnose. P. aeruginosa is known to make and
potential for fine regulation of the polyamine excrete both monorhamnolipids (mono-
levels, both inside and outside the cell. Studies RHLs) and dirhamnolipids (di-RHLs), which
described here have begun to link the regula- have either one or two attached rhamnose
tion of multicellular phenomena to levels of groups, respectively (52, 60). Interestingly, the
polyamines; however, several areas of investiga- RHL precursor HAA is also secreted into the
tion will be important for solidifying extracellular environment (23). RHLs are
polyamines as extracellular signals that modu- amphipathic in nature, having a hydrophobic
late group behavior versus intracellular signals lipid and a hydrophillic sugar residue.As a con-
that modulate the interaction of an individual sequence, RHLs are described as having ten-
with a group. sioactive properties, which facilitate their
For instance, determining whether it is the action as surface-wetting agents and emulsi-
intracellular pool or extracellular pool of fiers. RHLs have been reported to have
polyamines that are sensed and responded to numerous biological activities, and recently this
will be informative. In the case of V. cholerae, the has been extended to the modulation of
requirement for NspS suggests that sensing of community-level traits.RHLs have been shown
extracellular levels is required; however, in the to be involved in maintenance of biofilm archi-
case of P. mirabilis and Y. pestis, it is not clear tecture (21), biofilm detachment (6), and
whether the export of polyamines serves only swarming motility (50). In these instances the
as a mechanism for an individual to regulate biological activity of RHLs has generally been
intracellular levels or as a means to signal to explained by their physical interaction with
other bacteria by controlling the extracellular environments (i.e., direct perturbation of cell-
concentration of polyamines. In this regard, it cell contacts); however, recent work from our
will be important to understand the relation- laboratory has raised the distinct possibility that
ship between extracellular concentrations and rhamnolipids can also act as extracellular signals
de novo synthesis with regard to regulation of that modulate group behavior, namely swarm-
intracellular concentrations of polyamines. A ing motility in P. aeruginosa (14).
specific example would be to ask whether P. aeruginosa swarming motility is typified by
mutation of the putrescine import system in P. the coordinated movement of bacteria across a
mirabilis disrupts the putrescine-signaling path- semisolid medium. Previous work has shown
way. Intercellular signaling would be supported that rhamnolipid biosynthesis is required for
by a role for importation of putrescine, whereas swarming,presumably by facilitating the reduc-
nonrequirement of importation to promote tion of surface tension,thereby allowing flagella
efficient swarming may suggest that a model of to propel bacteria across the semisolid surface
intracellular signaling is more appropriate. Fur- (50). One of the more pronounced features of
thermore, if an active mechanism is required to swarming by P. aeruginosa PA14 is the produc-
sense extracellular levels of putrescine, then a tion of tendrils that emanate from the origin in
mutant defective for both synthesis and import a pattern of radial spokes (90). Tendrils, once
of putrescine should not be responsive to initiated, do not cross into each other.This pol-
chemical complementation with exogenous icy of avoidance also applies for swarms origi-
putrescine. nating from separate colonies (13).
From these observations we hypothesized
RHAMNOLIPIDS that an extracellular factor produced by P. aerug-
Rhamnolipids (RHLs) are a group of biosur- inosa was modulating the direction of swarming.
factants first identified in cultures of P. A mutagenesis screen was carried out that iso-
aeruginosa (39, 44). Structurally they consist of lated several mutants that were unable to inhibit
a glycolipid core of hydroxylalkanoic acid the swarming of the wild-type strain (14).The
(HAA) with subsequent additions of the sugar majority of these mutants mapped to genes
involved in RHL synthesis or known regulators If mono-RHLs act as signals, then genetic
of rhamnolipid synthesis. The involvement of networks should exist that sense and respond to
rhamnolipids was verified in several ways. First, the presence of mono-RHLs. In support of this
filtered supernatant from an RHL biosynthetic idea, Caiazza et al. reported that sadB mutants
mutant could not suppress swarming of the Wt, are insensitive to RHL-dependent inhibition of
and second, addition of purified RHL could swarming (14). SadB was previously character-
recapitulate swarming retardation when spotted ized as a cytoplasmic protein important in the
in the vicinity of a swarm.These data indicate a early stages of biofilm formation by P. aeruginosa
dual role for RHLs in swarming,both as positive (13). Combined with its role in regulating
factors promoting movement via its surface- swarming, the sadB gene product may coordi-
wetting properties and as negative factors mod- nate and regulate multiple community-level
ulating the form of the swarming community. activities.
One of the intriguing characteristics Currently it is not known how rhamnolipid
of rhamnolipid biosynthesis is that the inter- concentration is linked to SadB function.
mediates in this pathway, namely HAA and Potentially, a membrane protein binds rham-
monorhamnolipid, are excreted in addition to nolipids and transduces this signal to SadB,
the end product, dirhamnolipid (23). Caiazza or alternatively, rhamnolipids act more gener-
et al. investigated the possibility that intermedi- ally to alter membrane composition or physical
ates in rhamnolipid biosynthesis had different properties, which subsequently leads to acti-
roles in swarming: surface wetting to physically vation of SadB. An analogous example might
promote swarming and as a signal that is sensed be the cpx system in E. coli. Cpx is a two-
to modulate direction of the swarm (14). component system that is known to be acti-
The biosynthesis of di-RHLs requires the vated by perturbations to the outer membrane
sequential action of RhlA, RhlB, and RhlC. and is required for efficient biofilm formation
RhlA is required for formation of HAA from a (29,81).Interestingly,Cpx was also shown to be
cellular lipid precursor pool, whereas RhlB and activated in response to interaction with sur-
RhlC are required for the subsequent addition faces, raising the possibility that perturbations
of rhamnose moieties to form mono- and di- in membranes due to surface collisions were
RHLs, respectively (67, 68, 77).Therefore, indi- sensed by Cpx as signals for regulating the tran-
vidual mutations to rhlA, rhlB, or rhlC result in sition to a surface-committed lifestyle (69).
the accumulation of different intermediates of In summary, the biosurfactant properties of
RHL biosynthesis. An rhlA mutant does not rhamnolipids have dominated how we think
excrete any RHLs or intermediates, whereas an about their biological roles, wherein we have
rhlB mutant still excretes HAAs and an rhlC tended to consider them largely in the realms of
mutant excretes both HAAs and mono-RHLs, their propensity to alter the physical interac-
but not di-RHLs. Caiazza et al., confirming tions between biological interfaces. Recent
work by Dezeil, demonstrated that HAAs are work described here points to the possibility
the minimal requirement to facilitate surface that RHLs can play roles as diffusible signals
wetting and swarming due to the fact that an regulating group activities, once again suggest-
rhlB mutant could still swarm, whereas an rhlA ing that a biomolecule may perform separate
mutant was not able to swarm (14, 23). A role and discrete biological functions.
for mono-RHLs as signaling molecules was
indicated by the fact that filtered supernatants INDOLE
from the rhlC mutant that contain mono- Indole is an aromatic secondary metabolite pro-
RHLs were still capable of repressing wild-type duced through hydrolysis of tryptophan by the
swarming, whereas rhlA and rhlB mutant superenzyme tryptophanase (tnaA). In E. coli, indole
natants, lacking mono-RHLs, had no impact is imported from the extracellular environment
on swarming motility. predominantly by the Mtr permease (98),
whereas efflux of indole out of the cell is per- duced by E.coli.The fusions then served as valu-
formed by the AcrEF pump (47). Initially the able biosensors to pursue the identity of the
action of tryptophanase was seen as a way for activating signals presumably excreted by E.coli.
the cell to synthesize tryptophan from environ- In a followup to this screen,Wang et al. report
mental sources of indole; however, the equilib- the identification and validation of indole
rium of this reaction favors the breakdown of as an extracellular signal required for activation
tryptophan to form indole. The intracellular of three of the four biosensor fusions. Specifi-
role of indole has remained unclear, especially cally, a fusion to gabT, encoding a glutamate:
considering that (i) high concentrations of succinate semialdehyde dehydrogenase, was
indole (5 mM) in the environment are toxic and used as the biosensor to follow the activating
(ii) high intracellular concentrations of indole signal through successive fractions of E. coli
can inhibit cell division (16, 93). For some, no stationary-phase supernatant. From this proce-
explanation of cellular significance would seem dure the activating signal was purified to
necessary; that is, indole is a toxic by-product of homogeneity and identified by mass spectrom-
a metabolic reaction to generate pyruvate and etry as indole (93).This result was further cor-
ammonia, so excretion of indole serves to roborated by demonstration that synthetic
detoxify the cell and facilitate generation of indole by itself was able to elicit premature acti-
metabolizable sources of carbon and nitrogen. vation of gabT, astD, and tnaB expression in a
The presence of mechanisms for indole import, dose-dependent manner. In contrast, indole did
however, would seem contrary to this explana- not activate expression of the cysK fusion, sug-
tion. In the next paragraph we outline research gesting that other signals were present in E. coli
indicating a novel role for indole as an extracel- stationary-phase supernatants.
lular signaling molecule in E. coli that has been What is the biological response of E. coli to
linked to the regulation of biofilm formation. indole-dependent signaling? Wang et al. note
The identity of extracellular signaling mole- that gabT, astD, and tnaA are involved in path-
cules in E. coli has been pursued with some ways that degrade amino acids to produce
rigor by Phillip Rather and his group. They pyruvate (93). One role for indole signaling
implemented an elegant genetic screen to iden- may be to prepare the cell for nutrient-limiting
tify genes that were transcriptionally activated conditions, when catabolism of amino acids
by the accumulation of an extracellular signal in will provide important resources; however, this
the growth medium (3). Briefly, the transposon has yet to be demonstrated. To clearly define
Tn5 carrying a promoterless lacZ and pro- indole as a relevant extracellular signal, it is
moterless tetracycline resistance gene was used important to demonstrate that physiologically
to generate random transcriptional fusions in relevant flux in the levels of indole elicits a spe-
E. coli. Those cells that were blue on X-Gal cific biological response that is distinct from
indicated transcription of the lacZ gene at any the metabolism of the signal. In this respect,
time during the growth of the colony;however, research from two groups has linked indole sig-
those cells in which transcription was activated naling to regulation of biofilm formation.
later in the growth phase would be sensitive to Di Martino and colleagues identified that
tetracycline when restreaked. Cells that dis- mutation of the E. coli S17 tryptophanase
played these two phenotypes were then (tnaA) resulted in reduced biofilm formation
screened for premature activation by the addi- on polystyrene as well as reduced attachment
tion of preconditioned supernatant prepared to human pneumocyte cells (26). Subsequent
from stationary-phase cells. Cells with fusions studies clearly correlated deficiencies in indole
that satisfied the last criterion led to the identi- production with defective biofilm formation
fication of four genes (gabT, astD, tnaB, and by the tryptophanase mutant. Moreover,
cysK) whose activation was likely dependent on oxindolyl-L-alanine, a potent competitive
the concentration of extracellular signals pro- inhibitor of tryptophanase, also repressed
biofilm formation and indole production (25). shown to have significantly lower levels of
Consistent with description as an extracellular intracellular and extracellular indole. Further-
signal, supplementation of media with indole more, addition of exogenous indole (250 M)
(625 M) resulted in restoration of tnaA decreased biofilm formation of both mutants
biofilm formation back to 84% that of the wild to levels approximately the same as that of the
type. The authors suggest that the levels of wild type. Importantly, this concentration of
indole used for chemical complementation indole was reported as not inhibiting growth
were physiologically relevant. Based on pub- of the E. coli and also falls within the range
lished values of indole concentration in E. coli of indole concentrations observed in E. coli
supernatants (150 to 350 M),this claim would stationary-phase cultures (93). The case for
seem somewhat inappropriate (93). However, it indole serving as a repressor of E. coli biofilm
should be noted that the concentration of formation was strengthened by subsequent
indole realized within a biofilm is not well work in Wood’s group showing that addition of
understood, making it difficult to interpret the exogenous indole decreases biofilm formation
authors’ claim. The concept of physiological by the wild type and that mutants in the indole
relevance is important to our characterization synthetic pathway show enhanced biofilm
of putative novel intercellular signals, and we formation (55).
will return to it later in the chapter. Currently it is unresolved as to why studies
Di Martino et al. also extended their studies from two different groups have reported oppo-
to several other indole-producing bacteria (25). site effects of indole on E.coli biofilms;however,
Interestingly,they demonstrated that oxindolyl- it could be that indole is an important biological
L-alanine could negatively affect the biofilm cue that can be integrated into regulation of
formation of Klebsiella oxytoca, Citrobacter koseri, biofilm pathways in different ways, depending
Providencia stuartii, and Morganella morganii but on the niche-specific selection experienced by
did not affect biofilms by a non-indole-produc- a given species or strain of bacteria over evolu-
ing strain of Klebsiella pneumoniae.These results tionary time.Consistent with this,the two stud-
suggest that the indole-dependent modulation ies discussed used different strains of E. coli with
of biofilm formation may be broadly applicable different evolutionary histories. Furthermore,
to indole-producing organisms. Wood’s group recently showed that addition of
Research from Thomas Wood’s group has indole could enhance the biofilm formation of
also reported a role for indole in regulating two pseudomonads (55), an opposite pheno-
biofilm formation by E. coli (28, 55). However, type to that they have shown for E. coli. The
in contrast to the results of Di Martino et al., potential for an environmentally relevant mole-
indole was demonstrated to be a repressor of cule to act as a signal for both inhibition and
biofilm formation by E. coli. Initially, Domka et enhancement of biofilm formation is not with-
al. were interested in understanding the molec- out precedent. Our work and that of Clay
ular basis of enhanced biofilm formation con- Fuqua’s group have shown that low levels of
ferred by mutations to yliH and yceP, two inorganic phosphate inhibit the biofilms of
hypothetical proteins of unknown function Pseudomonas fluorescens (63) and Pseudomonas
(28). Microarray analysis indicated broad aureofaciens (65) but promote biofilm formation
changes in gene expression profiles between by the Agrobacterium tumefaciens (19).
each mutant relative to the wild type. Among The molecular mechanisms underpinning
these, genes for indole export were upregulated regulation by indole are just beginning to
and genes for indole synthesis and import were be understood (27, 55, 104). Recently, muta-
downregulated, suggesting that yliH and yceP tions to the luxR homolog sdiA were shown to
may regulate the concentration of intracellular suppress the loss of biofilm formation by E.
or extracellular indole. Concordant with this coli in response to exogenous indole. Impor-
hypothesis, both yliH and yceP mutants were tantly, sdiA expression and other known sdi-
dependent functions were shown to be modu- extracellular signal in Providencia stuartii and E.
lated by indole concentrations, thereby making coli that regulates biofilm formation (86). Previ-
a good case for SdiA as a member of the indole- ous work by Phillip Rather’s group had identi-
biofilm signaling pathway (55). Further work fied fusions in P. stuartii that were activated by
needs to be done to fully ascertain how SdiA conditioned supernatant (3). One of these
activity affects biofilm formation, but regula- fusions (cma37::lacZ) was then used to perform
tion of motility seems to be at least part of the a genetic screen for suppressor mutations that
answer.It is of note that E.coli does not produce reduced activation of the fusion, with the hope
homoserine lactones but can recognize and of identifying the nature of the activating signal.
respond to foreign homoserine lactones, likely A mutation was recovered in the gene cysE that
through binding to SdiA. It will be interesting reduced activation of cma37::lacZ by over 50%
to see if SdiA can bind indole,and if so,whether (86). CysE catalyzes the conversion of L-serine
signal-independent (sidA) mutants can be to O-acetylserine (OAS), which is an interme-
made. Such experiments will establish SdiA as diate required for production of cysteine.
an important sensor and regulator of the indole- Sturgill et al. went on to demonstrate that OAS
dependent regulation of biofilm formation. was sufficient to activate cma37::lacZ, thereby
indicating that the enzymatic function of CysE
AMINO ACID METABOLISM is required for the production of an extracellular
Amino acids are literally building blocks of life. signal.The authors note that OAS is not likely
Their central role in core biosynthetic processes to be the exact signal as higher concentrations
has forged the scientific context with which of OAS were required to activate cma37::lacZ
they have largely been studied.The majority of than are found in conditioned media. Also, the
bacteria synthesize most amino acids de novo signal in conditioned media was stable at a
but are also known to possess transport systems higher pH than that of OAS.The true nature of
to import and export amino acids, as well as the signal is not known; however, it could be a
various amino acid derivatives (12). Importa- derivative of OAS or a peptide containing OAS.
tion of environmental amino acids makes sense It may be possible to distinguish between these
in that it would reduce the need for de novo options by assessing whether transporters of
synthesis and conserve biosynthetic energy. OAS or derivatives are required for cysE-
The biological role of efflux pumps in 9-9 dependent signaling. Several genes have been
metabolism has not been well studied, and ini- identified that actively export or augment
tially it may seem counterproductive to excrete export of OAS from the cell that would provide
valuable cellular resources. Most explanations good candidates for this analysis (31). If such
of this phenomenon center on metabolic ratio- transporters were required, it would provide
nales such as toxicity associated with intracellu- evidence against the involvement of a peptide
lar accumulation of amino acids; however, it is signal,which likely would not be transported by
unclear what, if any, role the phenomenon of these systems. In any respect, it is clear that cysE
amino acid toxicity plays in bacterial physiol- is an important determinant in producing an
ogy (12). In this section we consider evidence extracellular signal.
for another role of amino acids as extracellular Evidence supporting a broader role for cysE-
signals allowing communication between cells. dependent signaling networks was provided by
In some ways this should not be an entirely for- demonstrating that a subset of E. coli fusions
eign concept.Acetylated derivatives of methio- known to be activated by extracellular signals
nine are widely accepted as intercellular were also activated by OAS and that super-
signaling molecules, called homoserine lac- natants from cysE mutants were unable to acti-
tones (32, 38). vate these same fusions.A physiological role for
Cysteine biosynthetic pathways have re- cysE-dependent signaling was investigated by
cently been implicated in the production of an testing the biofilm formation of E. coli cysE
mutants. Interestingly, cysE strains showed Another explanation we believe worth

enhanced biofilm formation relative to the wild investigating is that a proportion of these efflux
type, with the mutant covering more surface systems excrete amino acid derivatives that
area and having greater biomass. Importantly, function in intercellular signaling networks
addition of OAS could reduce biofilms to levels (103). Levels of redundancy for particular sub-
similar to that of the wild type, consistent with strates may suggest differential functions. For
the idea that cysE mutants could not produce an example, in E. coli both YdeD and YfiK are
extracellular signal that negatively regulates thought to represent alternative efflux pumps
biofilm formation. It will be important to iden- for cysteine and/or OAS (31). Similarly, RhtA
tify the true nature of the activating signal to was recently shown to facilitate homoserine
allow more robust analysis of its physiological and threonine export,processes also carried out
role; however, these studies provide evidence by the RhtB and RhtC transporters (57).
that the excretion of amino acids and/or deriv- Indeed, it should be realized that the natural
atives can play roles in regulating community- substrates of RhtB family exporters are not
level traits such as biofilm formation. known.The finding that YfiK can export OAS,
Is there a broader role for amino acids as in addition to cysteine, points to the possibility
intercellular signals? Several new families of that native substrates may actually be amino
inner membrane proteins have been identified acid derivatives or intermediates.This possibil-
that function as amino acid efflux pumps (2, 30, ity is intriguing because research on homoser-
102). One of these is the RhtB family, a group ine lactones provides good precedent for the
of inner membrane proteins thought to facili- possibility that amino acid derivatives can func-
tate active export of small charged molecules, tion as signaling molecules in bacteria.
such as amino acids. RhtB was initially studied Ultimately, significantly more work is
in E. coli, where it was found to be important in needed to lend empirical weight to the notion
the export of both homoserine and homoser- that a family of novel efflux pumps play roles
ine lactone (102).E.coli actually contains at least in intercellular signaling. The correlation of
five paralogs of rhtB (2), four of which have community-level traits to the activity of spe-
been experimentally shown to export other cific exporters will be crucial,as will identifying
amino acids and derivatives. RhtC was shown the native substrates and the cellular pathways
to export threonine (102),YfiK can export cys- with which these substrates interact.
teine or OAS (31),YeaS can export leucine (51),
and LysE is known to excrete lysine (30, 91). MICROBE-HOST SIGNALING
Why is it that E. coli has a range of export WITH METABOLITES
machinery that can discriminate between spe- So far in this chapter we have discussed exam-
cific groups of amino acids? One argument is ples within the conceptual framework of meta-
that these exporters function in homeostatic bolic signaling between bacteria, either with
roles, ensuring that amino acids do not reach different individuals of a clonal population or
toxic levels in the cell; however, several lines of with other members of a more diverse commu-
reasoning challenge the sufficiency of this nity. In this section we extend our discussion to
explanation.First,although high concentrations instances of metabolic signaling between
of exogenous amino acids have been correlated microbes and eukaryotic hosts. Below we
with growth inhibition, the concentrations briefly discuss three examples that offer support
required are in the millimolar range (30), which for such signaling.
sheds some doubt on the physiological rele-
vance of this phenomenon. Second, the speci- Nitric Oxide Signaling in P. aeruginosa
ficity and number of these transporters seem at Nitric oxide (NO) is an important biomolecule
odds with a role in a general stress response to that is known to act as a signal controlling
amino acid levels. an array of cellular processes in many eukary-
otes and multicellular organisms. In humans, between NO concentrations and regulation of

NO is produced by cells throughout the body cellular processes.
and is known to be important for processes Another interesting aspect to the work pre-
such as vasoconstriction, neural communica- sented by Barraud et al.is the role of P.aeruginosa
tion, and the innate immune response. NO is NO anaerobic metabolism in biofilm forma-
also a reactive oxygen species (ROS) such that tion. P. aeruginosa is capable of utilizing oxidized
imbalances in NO metabolism and NO level nitrogen species, such as nitrate (NO3), nitrite
are implicated as causal factors in various (NO2) or nitrous oxide (N2O), as alternative
human diseases (1, 11). terminal electron acceptors (105).Barraud et al.
Work by Barraud et al. has recently provided demonstrated that mutation of the nitrite
evidence that exogenous NO acts as a signal reductase gene nirS led to delays in dispersal of a
to regulate P. aeruginosa biofilm formation. mature biofilm relative to that of the wild type,
Specifically, they demonstrated that at low whereas mutations in norCB,the NO reductase,
subinhibitory concentrations of extracellular resulted in increased detachment from the
NO, P. aeruginosa biofilm formation was inhib- biofilm.These results are consistent with higher
ited. In fact, NO was even capable of inducing levels of intracellular NO promoting detach-
dispersion of a preformed biofilm when used at ment from the biofilm, since nirS mutants can-
these same low concentrations (5). Interest- not convert NO2 to NO, whereas norCB
ingly,high levels of NO,in the millimolar range, mutants are unable to further process NO to
promoted biofilm formation with up to four- N2O. Even in aerobic culturing conditions,
fold more biomass than untreated biofilms. anaerobic zones are known to exist at the cen-
Currently nothing is understood about the ter of mature biofilms, which should support
mechanisms responsible for either of these nitrate reduction and the generation of nitric
NO-dependent responses, but it is tempting to oxide (95,100).Nitric oxide is a freely diffusible
speculate that they could represent differential gas, such that individual cells should contribute
responses to host-dependent variations in the to an extracellular pool of NO as a conse-
level of NO. P. aeruginosa is an opportunistic quence of the level of their own anaerobic
pathogen that is a major disease factor for metabolism.This scenario once again provides
patients with cystic fibrosis or immunocom- the opportunity for intercellular signaling
promised patients (10, 61). The ability to mediated by a metabolic intermediate. Incre-
respond to NO levels could represent an adap- mental accumulation of NO among many cells
tation to utilizing a human metabolic signal as leads to increases in the local concentration of
an environmental sensor aiding the regulation NO to levels that activate pathways for biofilm
of its own life strategies: whether to form or dispersal among the group.
maintain life as a biofilm.Intriguingly,members In summary, NO can promote dispersal in P.
of the H-NOX (heme-nitric oxide- and/or aeruginosa biofilms but the source of NO could
oxygen-binding domain) protein family have arise via distinct mechanisms: (i) NO produced
been identified in several prokaryotes (9).This is by other eukaryotes such as humans or (ii)
of interest because eukaryotic NO receptors are anaerobic metabolism of P. aeruginosa itself. It is
also H-NOX family proteins, suggesting the difficult to know which of these scenarios has
possibility that bacteria have the mechanisms driven the evolution of NO-dependent biofilm
for sensing NO concentrations.This idea is sup- regulation.That is to say, was it interactions with
ported by work from Boon et al. demonstrating other eukaryotes that provided the selective
that three H-NOX proteins, two from environment for adaptation,or was it more gen-
Legionella pneumophila and one from Nostoc eral environmental settings where internal pro-
punctiforme, could selectively bind NO while duction of NO from anaerobic metabolism
discriminating against O2 (8).H-NOX proteins provided the raw material for adaptation? Given
are widespread and may provide the link that high concentrations of NO elicited
increases in biofilm formation, it may be possi- trations of IAA than the wild type.These results
ble that both environments have contributed to suggest that Yap1 is a key component for medi-
the evolution of NO-dependent signaling. ating the response to IAA. Interestingly, IAA
High NO concentrations, like those associated was highly specific, with other related com-
with host-mediated events such as macrophage- pounds such as indole having no effect on for-
mediated killing, could have led to the mation of pseudohyphae or invasive growth.
evolution of two related yet distinct adaptive This suggested that mechanisms existed to
responses. Clearly, more work is required to specifically transport IAA into the cell. On the
address these hypotheses. basis of this hypothesis, Prusty et al. identified a
group of seven genes (referred to as AVT1-7)
Yeast-Plant Signaling with Auxin based on the similarity of their predicted gene
Dimorphic transitions are an integrative part of products to transporters implicated in auxin
many fungal lifestyles and are often associated transport in plants. Consistent with their role in
with changes in virulence of pathogens (82). IAA transport, mutation of any avt gene
The yeast Saccharomyces cerevisiae undergoes reduced IAA uptake and abolished the ability of
transition to a filamentous form termed IAA to induce morphological transitions.
pseudohyphal growth and has served as a good In summary, this work shows that a plant
model to study the signals and molecular hormone, the metabolite IAA, is specifically
mechanisms involved in this transition (59). used as a signal to induce coordinated imple-
Pseudohyphal growth occurs in response to mentation of a differential developmental
specific environmental cues such as starvation program at a community level. The presence
for nitrogen or carbon.After sensing the appro- of AVT-like genes in most fungi sequenced to
priate signals, diploid cells form multicellular date suggests that the use of a plant hormone
aggregates that form invasive filaments capable to regulate cellular phenomena such as dimor-
of burrowing into the substratum. This so- phic transitions could be a broad property of
called invasive growth is thought of as a means plant-associated fungi. Moreover, IAA-based
for sessile yeast cells to forage for more nutri- signaling may also extend to the realm of
ents, but it also has much relevance to the rhizosphere-colonizing bacteria, many of
mechanism by which pathogenic fungi com- which are known to be able to synthesize and
promise host cell integrity (59). excrete IAA (18, 71). This potentially opens
Recent work from the laboratory of Gerry up a myriad of interactions based on IAA
Fink has provided evidence that the plant hor- signaling pathways across multiple domains of
mone indole-3-acetic acid (IAA), commonly life, involving both symbiotic (72, 89) and
known as auxin, acts as a signal to stimulate pathogenic (66) relationships.
pseudohyphal growth by S. cerevisiae (75).
Specifically, addition of IAA at concentrations Plant-Derived Salicylic Acid
of 50 M resulted in diploid filamentation and Attenuates P. aeruginosaVirulence
haploid invasive growth on agar plates. The Salicylic acid (SA) is a phenolic metabolite pro-
molecular basis of this response was investigated duced by plants and an important phytohor-
and shown to require Flo11, a cell-surface pro- mone that plays a key role in the induction of
tein required for nutrition-based induction of plant defenses (35, 49, 83). Most studies have
both pseudohyphal and invasive growth.Tran- focused on the effects of SA on plant-associated
scriptional profiling of cells treated with IAA processes of resistance induction and have not
identified a group of genes with consensus considered the possibility that SA can interact
binding sites to Yap1, which is a known tran- directly with the pathogen. In an interesting
scription factor.Consistent with a role forYap1, article, Prithiviraj et al. report that SA can also
a yap1 deletion mutant was hypersensitive to protect plants from infection by modulating the
IAA, forming filaments at much lower concen- virulence properties of the pathogen that is
commencing its attack (74). Initially, Prithiviraj specific cell processes have provided the cur-
et al. noted that mutants of Arabidopsis that pro- rency for adaptation of individuals and their
duce more SA show less colonization by P. interactions. With this in mind, Keller and
aeruginosa and a lower rate of mortality postin- Surette have presented a poignant and relevant
fection. The authors then examined the effect evolution-driven perspective on the constraints
of SA on a range of known virulence factors, implicit with characterization of a cell-cell sig-
such as biofilm formation, exoenzyme produc- naling event as communication (48). For Keller
tion, and pyocyanin production. Interestingly, and Surette, the distinction between various
in vitro biofilm formation was reduced by SA types of intercellular signaling rests largely on
in a dose-dependent manner when a range of differences in the benefits and costs realized
physiologically relevant concentrations (0.1 by the “transmitter” of the signal versus the
to 1 mM) were used. Concomitant with inhi- “receiver” of the signal. Communication is a
bition of biofilm formation, SA treatment two-way interaction that requires an individual
resulted in reductions in pyocyanin production to generate a signal,as well as another individual
as well as decreased production of extracellular to receive and act on that signal. Furthermore,
elastases and proteases, all of which have been maintenance of communicative relationships
reported as important virulence factors in the over evolution requires that both transmitter
Arabidopsis disease model (78, 92). and receiver benefit from the interaction.This
In an attempt to gain insight into the cellular logic extends from the foundations of natural
targets with which SA is interacting, transcrip- selection.That is to say, signal generation by the
tional profiles of Wt and Wt treated with SA transmitter will be selected against unless the
were compared. A range of genes were found cost of production is offset by the benefits of
to be both activated and repressed in response the signal being sensed and acted on by the pop-
to SA treatment. Some of these genes suggested ulation as a whole.For the receiver,mechanisms
direct links to the observed phenotypes, such as to specifically sense and act on the signal will be
downregulation of quorum-sensing genes and selected against if no direct benefit is associated
various genes in protein export machinery like with the elicited response. On the basis of this
exoT and exsC.Therefore,transcriptional analy- criterion, we suggest that communication is a
sis provided some interesting candidates that specific cell-cell interaction that falls under the
will require further functional analysis to assess larger umbrella of intercellular signaling.
their role in SA attenuation of virulence. Other types of intercellular signaling can also
Although the mechanistic basis of the effects of be defined in evolutionary terms and largely
SA are not understood, this study provides a differ in the selected pressures experienced by
good platform to pursue the possibility that SA the transmitter and receiver. For instance, the
functions as a metabolic intercellular signaling fact that an organism produces and excretes a
molecule that can interfere with P. aeruginosa molecule that can elicit an adaptive response in
regulatory pathways controlling expression of a receiver does not necessitate that this is the
virulence determinants and multicellular traits, function of the molecule that was selected for in
such as biofilm formation. the transmitter. Alternatively, the molecule
could simply be a metabolic end product that
EVOLUTION OF SIGNALING has been usurped by the receiver as a descriptor
RELATIONSHIPS: COMMUNICATION of its environment and an appropriate cue for
OR ENVIRONMENTAL BIOTIC CUE expression of various traits. In this scenario the
The incorporation of evolutionary criteria and requirement that signal production and receipt
perspectives into characterization of signaling of the signal be of benefit to both organisms is
systems is important because it speaks to the relaxed. In fact, there is likely no benefit to the
biological foundations of the underlying inter- transmitter in this scenario. Conceptually, this
action. The benefits and costs associated with type of interaction is similar to many well-
known bacterial systems that have evolved to because auxin production in plants seems likely
sense environmental cues.Take, for instance, the to have evolved primarily for plant-specific pur-
Pho regulon, which is a group of genes that are poses, where sensing of auxin has been selected
expressed in response to concentrations of inor- for in microbes due to an advantage of correlat-
ganic phosphate and has also been shown to ing expression of a trait with the presence of the
impact biofilm formation (63–65). The main plant-derived metabolite. Furthermore, the
difference is, of course, that the phosphate is an benefits to the plant for actively participating in
abiotic signal, whereas the systems we refer to this signaling interaction are less clear, especially
under the umbrella description of intercellular in the case of host-pathogen interactions.
signaling use biotic signals produced as a conse- SA signaling between P. aeruginosa and plants
quence of an organism’s metabolism. could be a good candidate for description as
Another type of signaling relationship dis- chemical manipulation. In this case, SA inter-
cussed in depth by Keller and Surette is one of feres with regulatory networks of P. aeruginosa
chemical manipulation. Similar to sensing of that are important for infection and pathogene-
biotic cues, a process of chemical manipulation sis in planta.This leads to likely benefits for the
does not benefit both members of the interac- transmitter but a cost to the receiver.It is impor-
tion; however, in the case of chemical manipu- tant to note that not all host-microbe interac-
lation, benefit is ascribed to the transmitter, not tions are necessarily prohibited from being
the receiver. This is due to the ability of the categorized as communication. For instance,
transmitter to produce a signal that elicits an in the case of mutualistic symbiosis, signals
unintended response by the receiver. This exchanged between host and microbe may well
response in turn is beneficial to the transmitter, be required for appropriate coordination of cel-
but neutral, or more likely of cost, to the lular events and maintenance of a productive
receiver. For a more thorough discussion of the relationship.
evolution of communication versus other In summary, intercellular signaling is a broad
intercellular signaling systems, we direct the description of interactions between organisms
reader to the thoughtful review by Keller and mediated by chemical signals. On the basis of
Surette (48). evolutionary and ecological considerations,
In framing the scope and context of this intercellular signaling can be described as either
chapter, we have purposely avoided general use communication (benefits to both transmitter
of the term communication, largely due to the and receiver), response to a biotic cue (benefit
evolutionary perspectives discussed above. to receiver), and chemical manipulation (bene-
Instead, we have generally described putative fit to transmitter).
novel systems of intercellular signaling in
microbes, of which some are good candidates CRITERIA AND LOGICAL LIMITS FOR
for description as communication systems, THINKING ABOUT INTERCELLULAR
whereas others may be better described as sys- SIGNALING MOLECULES
tems for responding to biotic cues. From the Defining intercellular signaling broadly as inter-
cases we have presented, systems such as amino actions between organisms mediated by chemi-
acid-, rhamnolipid-, and indole-dependent sig- cal signals is good in that it accurately describes
naling are good candidates for description as the biological scope of this class of phenome-
communication. This is because signaling is non. But at the experimental level, this descrip-
likely to occur between groups of bacteria that tion is less useful for determining the standard of
regulate similar responses through production empirical evidence required as proof that an
and sensing of the same signal. interaction is accurately characterized as inter-
Other systems, such as auxin signaling in cellular signaling. Without a priori defined
yeast, are likely better candidates for description empirical standards for biological proof, the
as signaling responses to biotic cues. This is question inappropriately becomes “what isn’t
intercellular signaling?” To this end, a recent they require that the signal is produced during
commentary by Winzer et al. raises concerns as “specific stages of growth,under certain physio-
to whether descriptions of metabolite-based logical conditions, or in response to changes in
intercellular signaling systems have been held the environment.”To us, these conditions seem
up to appropriate experimental rigor (96).The more like potential properties of a given inter-
main issue presented by Winzer et al. pertains to cellular signaling system than useful criteria for
the perceived failure of experimenters to defining required characteristics of an intercel-
demonstrate that the putative signaling mole- lular signal. Whether the signal is produced
cule generates a cellular response that is separate throughout the growth phase of an organism or
and distinct from processes related to the metab- only in stationary phase is reflective of the biol-
olism of the putative signal, or toxicity associ- ogy that links environment to signal production
ated with high concentrations of the signal. For and system response, not the validity of its
instance, changes in gene expression that corre- description as an intercellular signal.
late with the accumulation of a toxic product The identification of the signal is important
may be the indirect consequence of a general because it provides the necessary tools to
stress response rather than a specific response to demonstrate specificity of the signaling path-
the metabolite per se.Similarly,if the response of way and investigate causal mechanism.We agree
a cell to the presence of a metabolite is to induce with Winzer et al. that correlation of a physio-
expression of import machinery and degrada- logical response with the addition of spent
tive enzymes targeted at the metabolite, is there supernatant to a fresh culture is not sufficient
the need to explain this phenomenon in terms evidence for the existence of a specific cell-to-
other than the adaptive response of the individ- cell signaling system.We would point out that
ual to the presence of a metabolizable substrate? for most of the work discussed in this chapter,
We agree with Winzer et al. that these are the signal or a close derivative of the signal has
valid concerns and that high standards must be been identified.
set for what constitutes proof of metabolite-
mediated intercellular signaling. Mechanisms Exist To Sense and
Below we present six properties that we sug- Respond Specifically to the Signal
gest are complicit with the description of a mol- The requirement for understanding mecha-
ecule as an intercellular signal.These criteria are nism is critical to establishing causality.Without
meant to be generally applicable and of practical clarification of mechanism, there is no way to
value.Therefore, rather than setting forth a defi- rule out that the response is due to indirect
nition that fits the answer “homoserine lac- effects and is not a specific biological response
tone,” we suggest these criteria as a standard of to that signal.Winzer et al. further specify that
proof, or experimental benchmark, that should the signal should be recognized by a specific
be met before a metabolite, or any molecule, is receptor that is presumably located on the cell
classified as an intercellular signal.In the follow- surface. Once again, we believe this condition
ing section we discuss each of these criteria with to be unnecessarily restrictive. In many of the
respect to the criticisms of Winzer et al.and the examples we have presented, the signal may
research discussed throughout this chapter. well be imported into the cell, such that it is the
intracellular concentration that is sensed.
The Signal Is Secreted Mechanistic studies will be required to validate
and Has Been Identified these claims; however, in principle, it is entirely
This is the most important yet basic require- feasible that cells could monitor the intracellu-
ment for any intercellular signaling.The signal lar levels of a metabolite as an indirect means of
must be accessible to other cells. Winzer et al. assessing its extracellular concentration. In fact,
have argued for additional constraints in regard there is at least one line of logic to suggest
to the nature of signal production. Specifically, that this may be the more likely scenario for
metabolic signals. Metabolic enzymes are often NO. At low concentrations its ability to diffuse
subject to allosteric control from end products across cell membranes makes it an ideal signal-
of the overall pathway.As a result of this, mech- ing molecule for mediating rapid responses.
anisms to recognize the concentration of a However,at high concentration these properties
metabolic signal may already be established are offset by the inherent toxicity of NO (1,11).
within the very architecture of the metabolic The evolution of mechanisms to tightly control
pathways themselves. Gene duplication and the cellular concentrations of NO has allowed
subsequent divergent evolution could have widespread use of NO as a valuable signaling
resulted in conversion of an allosterically regu- molecule, despite its highly toxic nature.
lated metabolic enzyme into an intracellular
sensor for a metabolic signal. Alternatively, the The Response Evoked
metabolic enzyme could have evolved dual Is Separable from the Primary
functions, carrying out catalysis and regulation Metabolism of the Signal
of unlinked cellular processes. The presence of This criterion is important for distinguishing
dedicated import systems for several of the puta- between the responses of individuals to
tive intercellular signaling metabolites we have molecules for their metabolic potential versus
discussed is consistent with these possibilities. using metabolites as community-generated
physiological signals to regulate processes dis-
The Concentration of Signal tinct from the metabolic pathway that created
Required To Elicit the Response them.In this chapter we have discussed putative
Is Not Toxic to the Cell intercellular signaling systems that regulate
Many biological molecules are known to inhibit community-level traits such as biofilm forma-
growth of microorganisms when present at high tion and swarming. In this way, all of the exam-
concentrations in the environment. Such is the ples discussed link metabolite concentrations
appropriateness of the old adage “dose makes with the regulation of a phenotype distinct
the poison,” often attributed to the “father of from the metabolism of the signal.
toxicology,”Paracelsus (53).If the concentration
of the signal required to elicit the response Purified Signal Can Recapitulate the
begins to inhibit the growth of the cell, then it Biological Response at Physiologically
becomes much more likely that the response is a Relevant Concentrations
nonspecific consequence of the stress imposed The ability to demonstrate both necessity and
by the molecule.That said,it is still formally pos- sufficiency is a powerful component of the
sible that a specific signaling event is occurring proof of explanation for any biological phe-
that is distinct yet concomitant with the initia- nomenon. Identification of the signal is ulti-
tion of a general stress response. However, in mately confirmed by its ability to recapitulate
these instances, a detailed understanding of the biological response without the addition of
mechanism would be required to make a strong supplementary factors. Also of importance is
case for a specific signaling response.In regard to that the concentrations of the signal required to
the work discussed in this chapter, it is impor- recapitulate the biological response correlate
tant to note that although many of the putative well with the concentration of that metabolite
metabolic signals we have discussed are toxic to in the original test conditions.The usefulness of
the cell at higher concentrations, in all cases the this criterion was demonstrated by studies on
biological response was elicited at concentra- the basis of cysE-dependent signaling. In this
tions of the signal that did not inhibit growth. case, OAS was thought to be the signal, but
Managing the fine relationship between the higher concentrations were required to elicit
utility and toxicity of a given signal may be a the response than are found in conditioned
common scenario in biological systems. Take, media,suggesting that the real signal was likely a
for instance, the bona fide intercellular signal derivative of OAS, but not OAS itself.
The Signal Network Is Adaptive limited in persuasive power due to the difficul-
at the Level of the Community ties associated with determining what con-
Biologists often ascribe a particular function to stitutes a physiologically and evolutionarily
a cellular process. The measured properties of relevant experimental environment.
the physical process are used to infer the true In summary, although experimentally chal-
physiological role of that process in a biologi- lenging, we suggest that it is important for
cally relevant context.What are also implicit in researchers to consider the evolutionary con-
these statements are inferences about the evolu- straints implicit with ascribing biological func-
tion of the trait.By ascribing a biological role to tions to molecules or systems. As we discussed
a process, one is also indirectly making state- in the previous section, it is the evolutionary
ments about the biological basis for the traits and ecological relationships underpinning a
selection through evolution. At some level this signaling phenomenon that dictate which
is unavoidable; however, it is important that this individuals or communities benefit from the
relationship be acknowledged. signaling interaction. These considerations
For instance, in the cases we have been dis- will determine whether it is important to test if
cussing here, stating that a metabolite is an a given trait is adaptive for transmitter, receiver,
intercellular signal that regulates biofilm forma- or both.
tion also implies regulation of biofilm forma-
tion is a specific function of the metabolite that
THE WORLD IS NOT A QUORUM:
was selected through evolution. The term INTEGRATING NEW CONCEPTS
“selected” in turn implies that the trait was OF INTERCELLULAR SIGNALING
adaptive within the context of the bacterium’s WITH THE CURRENT PARADIGM
niche. In an ideal situation we would be able to The concept of quorum sensing as outlined in
test this prediction empirically. That is, we the overview is elegant and captivating in its
would be able to test whether regulation of simplicity.The discovery and subsequent fervor
biofilm formation by a particular metabolite of investigation that surrounded homoserine
was an adaptive trait at the level of the commu- lactones and their role in density-dependent
nity. In reality, this is experimentally very regulation provided the tangible basis for a
difficult, and certainly is the most rigorous belief that microbes really could communicate
of the criteria we propose. One experimental with each other. Since these initial findings, the
approach might be to assess whether a func- term quorum sensing has become synonymous
tional signaling network confers a fitness with intercellular signaling in microbes. In fact,
advantage during competition studies within so intimately associated have these terms
the context of a biofilm. Mutations in the sens- become that the latter is at risk of being
ing machinery would be most appropriate for mistaken as the definition of the former. That
inhibiting the signaling network, as this would is to say, intercellular signaling is at risk of
be most likely to avoid pleiotrophic effects asso- being conceptually confined to description of
ciated with general disruption of core meta- quorum-sensing-related phenomena.In a simi-
bolic processes (i.e., mutation of the enzymes lar fashion, the risk also exists that homoserine
responsible for synthesis of the metabolite). lactones will become the surrogate definition
Coinoculation of wild-type and “sensing” of a quorum-sensing molecule rather than sim-
mutant in a biofilm model system, followed by ply an example of a specific type of quorum-
analysis of population dynamics over time, sensing molecule.
could facilitate identification of an experimen- Much of this chapter has been directed at
tal environment that confers a biological value introducing emerging models of intercellular
to maintenance of a functional signaling net- signaling that can potentially utilize an array of
work, thereby providing support for the claims microbial metabolites to transmit information
of system function.Admittedly, this approach is between cells. These systems also reflect the
opportunities that exist to break new intellec- from description as quorum-sensing systems
tual ground by studying phenomena within a simply because they do not involve homoserine
process-oriented and functionally defined lactones. By utilizing functional criteria as our
experimental framework. yardstick for defining density-dependent regu-
In the previous section we outlined a set of lation, the opportunity arises to discover new
generally applicable criteria that functionally and distinct biological processes,even new roles
define the characteristics inherent in intercel- for apparently “well-defined” cellular processes
lular signaling systems.We would like to rein- such as those of metabolism.The validity of this
force that quorum sensing is but one category perspective is embodied by the recent discovery
of a larger set of phenomena that would fit into that phenazines constitute the new terminal
our functional criteria for intercellular signal- effector of the quorum-sensing network in P.
ing systems. Biological roles other than moni- aeruginosa (24), as well as the discovery of
toring population density are likely dependent autoinducer-2 (17, 87). Also of relevance is a
on the characteristics of the signaling network. report suggesting that homoserine lactones can
For instance, quorum sensing in theory relies function as biosurfactants that are important in
on the low basal yet constant production of the swarming of Rhizobium etli (20). Ironically,
an excreted signal, such that the total extracel- we may have come full circle, realizing that the
lular concentration is approximately propor- archetypal quorum-sensing molecule may itself
tional to the density of the population. But have multiple biological roles.
what about a scenario where production of the
signal is regulated by the environment or the CONCLUSION
physiological status of the organism? In this sit- Historically, we have defaulted to metabolic
uation, accumulation of the signal is uncoupled explanations to explain phenomena involving
from a relationship with cell number, but molecules with well-known roles in metabolic
instead reflects either the immediate environ- pathways. In this chapter we have pursued an
ment of that organism or its physiological state. investigation of the propensity of metabolites to
An organism in a juxtaposed but different display multiple biological functions, to act not
microniche may be granted access to this infor- just within the confines of energy and biosyn-
mation as a consequence of sensing and thetic pathways, but also as intercellular signals
responding to the accumulation of the excreted communicating information about their physi-
signal.This signal could be a metabolite, similar ological status to members of their community
to situations we have discussed here, or a more who are adapted to receive and process that
traditional signaling molecule. The key differ- information. In the course of this chapter, we
ence of this model lies in the coupling of the have outlined evidence that supports the
physiological state of the organism to produc- role of various metabolites in the regulation of
tion of the signal, not in the difference in type community-level traits, such as biofilm forma-
or nature of the signal per se. Furthermore, tion, swarming, and filamentation. Much work
physiologically or environmentally regulated is needed to rigorously validate and explore
signal production raises the possibility that these claims, but we at least hope to have con-
fewer cells are necessary to generate a threshold vinced the reader that there is much fertile
level of signal. In this way, coordination among ground to explore and that metabolites are
communities of microbes could be achieved likely to play many distinct roles in the biology
through authoritative action of the few, rather of an organism that are separate from their roles
than by committee. in primary metabolism.
Functional definitions are also important
when it comes to thinking about what consti- ACKNOWLEDGMENTS
tutes a quorum-sensing system. Metabolically We thank Judith Merritt and Pete Newell for critical
based signaling systems should not be excluded and thoughtful discussion.
This work is funded by the National Science Foun- by Pseudomonas aeruginosa PA14. J. Bacteriol.
dation (CAREER9984521) and the National Institutes 186:4476–4485.
of Health (AI051360). 14. Caiazza, N. C., R. M. Shanks, and G. A.
O’Toole. 2005. Rhamnolipids modulate swarm-
ing motility patterns of Pseudomonas aeruginosa. J.
REFERENCES Bacteriol. 187:7351–7361.
1. Achike, F. I., and C. Y. Kwan. 2003. Nitric 15. Camilli, A., and B. L. Bassler. 2006. Bac-
oxide, human diseases and the herbal products that terial small-molecule signaling pathways. Science
affect the nitric oxide signaling pathway. Clin. Exp. 311:1113–1116.
Pharmacol. Physiol. 30:605–615. 16. Chant, E. L., and D. K. Summers. 2007. Indole
2. Aleshin, V. V., N. P. Zakataeva, and V. A. signalling contributes to the stable maintenance of
Livshits. 1999.A new family of amino acid-efflux Escherichia coli multicopy plasmids. Mol. Microbiol.
proteins. Trends Biochem. Sci. 24:133–135. 63:35–43.
3. Baca-DeLancey, R. R., M. M. South, X. 17. Chen, X., S. Schauder, N. Potier, A. Van
Ding, and P. N. Rather. 1999. Escherichia coli Dorsselaer, I. Pelczer, B. L. Bassler, and F. M.
genes regulated by cell-to-cell signaling. Proc. Natl. Hughson. 2002. Structural identification of a
Acad. Sci. USA 96:4610–4614. bacterial quorum-sensing signal containing boron.
4. Bader, M. W., S. Sanowar, M. E. Daley, A. R. Nature 415:545–549.
Schneider, U. Cho,W. Xu, R. E. Klevit, H. Le 18. Costacurta, A., and J. Vanderleyden. 1995.
Moual, and S. I. Miller. 2005. Recognition of Synthesis of phytohormones by plant-associated
antimicrobial peptides by a bacterial sensor kinase. bacteria. Crit. Rev. Microbiol. 21:1–18.
Cell 122:461–472. 19. Danhorn, T., M. Hentzer, M. Givskov, M. R.
5. Barraud, N., D. J. Hassett, S. H. Hwang, Parsek, and C. Fuqua. 2004. Phosphorus limita-
S. A. Rice, S. Kjelleberg, and J. S. Webb. tion enhances biofilm formation of the plant
2006. Involvement of nitric oxide in biofilm pathogen Agrobacterium tumefaciens through the
dispersal of Pseudomonas aeruginosa. J. Bacteriol. PhoR-PhoB regulatory system. J. Bacteriol.
188:7344–7353. 186:4492–4501.
6. Boles, B. R., M. Thoendel, and P. K. Singh. 20. Daniels, R., S. Reynaert, H. Hoekstra, C.Ver-
2005. Rhamnolipids mediate detachment of reth, J. Janssens, K. Braeken, M. Fauvart, S.
Pseudomonas aeruginosa from biofilms. Mol. Micro- Beullens, C. Heusdens, I. Lambrichts, D. E.
biol. 57:1210–1223. De Vos, J. Vanderleyden, J. Vermant, and J.
7. Bomchil, N., P.Watnick, and R. Kolter. 2003. Michiels. 2006. Quorum signal molecules as bio-
Identification and characterization of a Vibrio surfactants affecting swarming in Rhizobium etli.
cholerae gene, mbaA, involved in maintenance of Proc. Natl.Acad. Sci. USA 103:14965–14970.
biofilm architecture. J. Bacteriol. 185:1384–1390. 21. Davey, M. E., N. C. Caiazza, and G. A.
8. Boon, E. M., J. H. Davis, R.Tran, D. S. Karow, O’Toole. 2003. Rhamnolipid surfactant produc-
S. H. Huang, D. Pan, M. M. Miazgowicz, R. tion affects biofilm architecture in Pseudomonas
A. Mathies, and M. A. Marletta. 2006. Nitric aeruginosa PAO1. J. Bacteriol. 185:1027–1036.
oxide binding to prokaryotic homologs of the sol- 22. Davies, J., G. B. Spiegelman, and G. Yim.
uble guanylate cyclase beta1 H-NOX domain. J. 2006.The world of subinhibitory antibiotic con-
Biol. Chem. 281:21892–21902. centrations. Curr. Opin. Microbiol. 9:445–453.
9. Boon, E. M., and M.A. Marletta. 2005. Ligand 23. Deziel, E., F. Lepine, S. Milot, and R.
specificity of H-NOX domains: from sGC to bac- Villemur. 2005. rhlA is required for the produc-
terial NO sensors. J. Inorg. Biochem. 99:892–902. tion of a novel biosurfactant promoting swarming
10. Boucher, R. C. 2004. New concepts of the motility in Pseudomonas aeruginosa: 3-(3-hydrox-
pathogenesis of cystic fibrosis lung disease. Eur. yalkanoyloxy)alkanoic acids (HAAs), the precur-
Respir. J. 23:146–158. sors of rhamnolipids. Microbiology 149:2005–2013.
11. Bredt, D. S. 1999. Endogenous nitric oxide syn- 24. Dietrich, L. E.,A. Price-Whelan,A. Petersen,
thesis: biological functions and pathophysiology. M. Whiteley, and D. K. Newman. 2006. The
Free Radic. Res. 31:577–596. phenazine pyocyanin is a terminal signalling factor
12. Burkovski, A., and R. Kramer. 2002. Bacterial in the quorum sensing network of Pseudomonas
amino acid transport proteins: occurrence, func- aeruginosa. Mol. Microbiol. 61:1308–1321.
tions, and significance for biotechnological appli- 25. Di Martino, P., R. Fursy, L. Bret, B. Sun-
cations. Appl. Microbiol. Biotechnol. 58:265–274. dararaju, and R. S. Phillips. 2003. Indole can
13. Caiazza, N. C., and G.A. O’Toole. 2004. SadB act as an extracellular signal to regulate biofilm
is required for the transition from reversible to formation of Escherichia coli and other indole-
irreversible attachment during biofilm formation producing bacteria. Can. J. Microbiol. 49:443–449.
26. Di Martino, P., A. Merieau, R. Phillips, N. 40. Heilmann, C., O. Schweitzer, C. Gerke, N.
Orange, and C. Hulen. 2002. Isolation of an Vanittanakom, D. Mack, and F. Gotz. 1996.
Escherichia coli strain mutant unable to form biofilm Molecular basis of intercellular adhesion in the
on polystyrene and to adhere to human pneumo- biofilm-forming Staphylococcus epidermidis. Mol.
cyte cells: involvement of tryptophanase. Can. J. Microbiol. 20:1083–1091.
Microbiol. 48:132–137. 41. Hoffman, L. R., D. A. D’Argenio, M. J. Mac-
27. Domka, J., J. Lee, T. Bansal, and T. Wood. Coss, Z. Zhang, R. A. Jones, and S. I. Miller.
2007.Temporal gene-expression in Escherichia coli 2005.Aminoglycoside antibiotics induce bacterial
K-12 biofilms. Environ. Microbiol. 9:332–346. biofilm formation. Nature 436:1171–1175.
28. Domka, J., J. Lee, and T. K. Wood. 2006.YliH 42. Igarashi, K., K. Ito, and K. Kashiwagi. 2001.
(BssR) and YceP (BssS) regulate Escherichia coli K- Polyamine uptake systems in Escherichia coli. Res.
12 biofilm formation by influencing cell signaling. Microbiol. 152:271–278.
Appl. Environ. Microbiol. 72:2449–2459. 43. Igarashi, K., and K. Kashiwagi. 1999.
29. Dorel, C., P. Lejeune, and A. Rodrigue. 2006. Polyamine transport in bacteria and yeast. Biochem.
The Cpx system of Escherichia coli, a strategic sig- J. 3:633–642.
naling pathway for confronting adverse conditions 44. Jarvis, F. G., and M. J. Johnson. 1949.A glycol-
and for settling biofilm communities? Res. Micro- ipid produced by Pseudomonas aeruginosa. J. Am.
biol. 157:306–314. Chem. Soc. 71:4124–4126.
30. Eggeling, L., and H. Sahm. 2003. New ubiqui- 45. Karatan, E.,T. R. Duncan, and P. I.Watnick.
tous translocators:amino acid export by Corynebac- 2005. NspS, a predicted polyamine sensor, medi-
terium glutamicum and Escherichia coli. Arch. ates activation of Vibrio cholerae biofilm formation
Microbiol. 180:155–160. by norspermidine. J. Bacteriol. 187:7434–7443.
31. Franke, I., A. Resch,T. Dassler,T. Maier, and 46. Kashiwagi, K., S. Shibuya, H. Tomitori, A.
A. Bock. 2003.YfiK from Escherichia coli promotes Kuraishi, and K. Igarashi. 1997. Excretion and
export of O-acetylserine and cysteine. J. Bacteriol. uptake of putrescine by the PotE protein in
185:1161–1166. Escherichia coli. J. Biol. Chem. 272:6318–6323.
32. Fuqua, C., M. R. Parsek, and E. P. Greenberg. 47. Kawamura-Sato, K., K. Shibayama,T. Horii,
2001. Regulation of gene expression by cell-to- Y. Iimuma, Y. Arakawa, and M. Ohta. 1999.
cell communication: acyl-homoserine lactone Role of multiple efflux pumps in Escherichia coli
quorum sensing. Annu. Rev. Genet. 35:439–468. in indole expulsion. FEMS Microbiol. Lett. 179:
33. Goh, E. B., G.Yim,W.Tsui, J. McClure, M. G. 345–352.
Surette, and J. Davies. 2002. Transcriptional 48. Keller, L., and M. G. Surette. 2006. Communi-
modulation of bacterial gene expression by subin- cation in bacteria: an ecological and evolutionary
hibitory concentrations of antibiotics. Proc. Natl. perspective. Nat. Rev. Microbiol. 4:249–258.
Acad. Sci. USA 99:17025–17030. 49. Klessig, D. F., J. Durner, R. Noad, D. A.
34. Gotz, F. 2002. Staphylococcus and biofilms. Mol. Navarre, D. Wendehenne, D. Kumar, J. M.
Microbiol. 43:1367–1378. Zhou, J. Shah, S. Zhang, P. Kachroo,Y.Trifa,
35. Halim,V.A.,A.Vess, D. Scheel, and S. Rosahl. D. Pontier, E. Lam, and H. Silva. 2000. Nitric
2006.The role of salicylic acid and jasmonic acid in oxide and salicylic acid signaling in plant defense.
pathogen defence. Plant Biol. 8:307–313. Proc. Natl.Acad. Sci. USA 97:8849–8855.
36. Hamana, K., and S. Matsuzaki. 1985. Further 50. Kohler, T., L. K. Curty, F. Barja, C. van
study on polyamines in primitive unicellular Delden, and J. C. Pechere. 2000. Swarming of
eukaryotic algae. J. Biochem. 97:1311–1315. Pseudomonas aeruginosa is dependent on cell-to-cell
37. Hamana, K.,T.Tanaka, R. Hosoya, M. Niitsu, signaling and requires flagella and pili. J. Bacteriol.
and T. Itoh. 2003. Cellular polyamines of the aci- 182:5990–5996.
dophilic, thermophilic and thermoacidophilic 51. Kutukova, E. A., V. A. Livshits, I. P. Altman,
archaebacteria,Acidilobus,Ferroplasma,Pyrobaculum, L. R. Ptitsyn, M. H. Zyiatdinov, I. L.
Pyrococcus, Staphylothermus,Thermococcus,Thermodis- Tokmakova, and N. P. Zakataeva. 2005. The
cus and Vulcanisaeta. J. Gen. Appl. Microbiol. yeaS (leuE) gene of Escherichia coli encodes an
49:287–293. exporter of leucine, and the Lrp protein regulates
38. Hanzelka, B. L., and E. P. Greenberg. 1996. its expression. FEBS Lett. 579:4629–4634.
Quorum sensing in Vibrio fischeri: evidence that S- 52. Lang, S., and D. Wullbrandt. 1999. Rhamnose
adenosylmethionine is the amino acid substrate for lipids—biosynthesis, microbial production and
autoinducer synthesis. J. Bacteriol. 178:5291–5294. application potential. Appl. Microbiol. Biotechnol.
39. Hauser, G., and M. L. Karnovsky. 1957. 51:22–32.
Rhamnose and rhamnolipide biosynthesis by 53. Langman, L. J., and B. M. Kapur. 2006.Toxi-
Pseudomonas aeruginosa. J. Biol. Chem. 224:91–105. cology: then and now. Clin. Biochem. 39:498–510.
54. Lau, G., D. Hassett, H. Ran, and F. Kong. 67. Ochsner, U. A., A. Fiechter, and J. Reiser.
2004.The role of pyocyanin in Pseudomonas aerugi- 1994. Isolation, characterization, and expression in
nosa infection. Trends Mol. Med. 10:599–606. Escherichia coli of the Pseudomonas aeruginosa rhlAB
55. Lee, J., A. Jayaraman, and T. K. Wood. 2007. genes encoding a rhamnosyltransferase involved in
Indole is an inter-species biofilm signal mediated rhamnolipid biosurfactant synthesis. J. Biol. Chem.
by SdiA. BMC Microbiol. 7:42. 269:19787–19795.
56. Linares, J., I. Gustafsson, F. Baquero, and J. 68. Ochsner, U.A.,T. Hembach, and A. Fiechter.
Martinez. 2006.Antibiotics as intermicrobial sig- 1996. Production of rhamnolipid biosurfactants.
naling agents instead of weapons. Proc. Natl. Acad. Adv. Biochem. Eng. Biotechnol. 53:89–118.
Sci. USA 103:19484–19489. 69. Otto, K., and T. J. Silhavy. 2002. Surface sensing
57. Livshits, V. A., N. P. Zakataeva, V. V. Aleshin, and adhesion of Escherichia coli controlled by the
and M. V. Vitushkina. 2003. Identification and Cpx-signaling pathway. Proc. Natl. Acad. Sci. USA
characterization of the new gene rhtA involved in 99:2287–2292.
threonine and homoserine efflux in Escherichia coli. 70. Patel, C. N., B. W. Wortham, J. L. Lines, J. D.
Res. Microbiol. 154:123–135. Fetherston, R. D. Perry, and M. A. Oliveira.
58. Lu, C. D., Y. Itoh, Y. Nakada, and Y. Jiang. 2006. Polyamines are essential for the formation of
2002. Functional analysis and regulation of the plague biofilm. J. Bacteriol. 188:2355–2363.
divergent spuABCDEFGH-spuI operons for 71. Patten, C. L., and B. R. Glick. 1996. Bacterial
polyamine uptake and utilization in Pseudomonas biosynthesis of indole-3-acetic acid. Can. J. Micro-
aeruginosa PAO1. J. Bacteriol. 184:3765–3773. biol. 42:207–220.
59. Madhani, H. D., and G. R. Fink. 1998. The 72. Patten, C. L., and B. R. Glick. 2002. Role of
control of filamentous differentiation and viru- Pseudomonas putida indoleacetic acid in develop-
lence in fungi. Trends Cell Biol. 8:348–353. ment of the host plant root system. Appl. Environ.
60. Maier, R. M., and G. Soberon-Chavez. 2000. Microbiol. 68:3795–3801.
Pseudomonas aeruginosa rhamnolipids: biosynthesis 73. Price-Whelan, A., L. E. Dietrich, and D. K.
and potential applications. Appl. Microbiol. Biotech- Newman. 2006. Rethinking ‘secondary’ metabo-
nol. 54:625–633. lism: physiological roles for phenazine antibiotics.
61. Matsui, H., V. E. Wagner, D. B. Hill, U. E. Nat. Chem. Biol. 2:71–78.
Schwab,T. D. Rogers, B. Button, R. M.Taylor 74. Prithiviraj, B., H. P. Bais, T. Weir, B. Suresh,
II, R. Superfine, M. Rubinstein, B. H. E. H. Najarro, B.V. Dayakar, H. P. Schweizer,
Iglewski, and R. C. Boucher. 2006. A physical and J. M. Vivanco. 2005. Down regulation of
linkage between cystic fibrosis airway surface virulence factors of Pseudomonas aeruginosa by sali-
dehydration and Pseudomonas aeruginosa biofilms. cylic acid attenuates its virulence on Arabidopsis
Proc. Natl.Acad. Sci. USA 103:18131–18136. thaliana and Caenorhabditis elegans. Infect. Immun.
62. Mavrodi, D., W. Blankenfeldt, and L. 73:5319–5328.
Thomashow. 2006. Phenazine compounds in 75. Prusty, R., P. Grisafi, and G. R. Fink. 2004.The
fluorescent Pseudomonas spp. biosynthesis and reg- plant hormone indoleacetic acid induces invasive
ulation. Annu. Rev. Phytopathol. 44:417–445. growth in Saccharomyces cerevisiae. Proc. Natl. Acad.
63. Monds, R. D., P. D. Newell, R. H. Gross, and Sci. USA 101:4153–4157.
G.A. O’Toole. 2007.Phosphate-dependent mod- 76. Rachid, S., K. Ohlsen,W.Witte, J. Hacker, and
ulation of c-di-GMP levels regulates Pseudomonas W. Ziebuhr. 2000. Effect of subinhibitory antibi-
fluorescens Pf0-1 biofilm formation by controlling otic concentrations on polysaccharide intercellular
secretion of the adhesin LapA. Mol. Micro- adhesin expression in biofilm-forming Staphylococ-
biol.63:656–679. cus epidermidis. Antimicrob. Agents Chemother.
64. Monds, R. D., P. D. Newell, J. A. Schwartz- 44:3357–3363.
man, and A. O’Toole G. 2006. Conservation 77. Rahim, R., U. A. Ochsner, C. Olvera, M.
of the Pho regulon in Pseudomonas fluorescens Graninger, P. Messner, J. S. Lam, and G.
Pf0-1. Appl. Environ. Microbiol. 72:1910–1924. Soberon-Chavez. 2001. Cloning and functional
65. Monds, R. D., M.W. Silby, and H. K. Mahanty. characterization of the Pseudomonas aeruginosa rhlC
2001. Expression of the Pho regulon negatively gene that encodes rhamnosyltransferase 2, an
regulates biofilm formation by Pseudomonas aureo- enzyme responsible for di-rhamnolipid biosynthe-
faciens PA147–2. Mol. Microbiol. 42:415–426. sis. Mol. Microbiol. 40:708–718.
66. Navarro, L., P. Dunoyer, F. Jay, B. Arnold, N. 78. Rahme, L. G., F. M. Ausubel, H. Cao, E.
Dharmasiri, M. Estelle, O.Voinnet, and J. D. Drenkard, B. C. Goumnerov, G. W. Lau, S.
Jones. 2006.A plant miRNA contributes to anti- Mahajan-Miklos, J. Plotnikova, M. W. Tan,
bacterial resistance by repressing auxin signaling. J. Tsongalis, C. L. Walendziewicz, and R.
Science 312:436–439. G. Tompkins. 2000. Plants and animals share
functionally common bacterial virulence factors. Vivanco. 2004.Pseudomonas aeruginosa-plant root

Proc. Natl. Acad. Sci. USA 97:8815–8821. interactions. Pathogenicity, biofilm formation,
79. Rather, P. N. 2005. Swarmer cell differentiation and root exudation. Plant Physiol. 134:320–331.
in Proteus mirabilis. Environ. Microbiol. 7:1065– 93. Wang, D., X. Ding, and P. N. Rather. 2001.
1073. Indole can act as an extracellular signal in
80. Romling, U., M. Gomelsky, and M. Y. Escherichia coli. J. Bacteriol. 183:4210–4216.
Galperin. 2005. C-di-GMP: the dawning of a 94. Waters, C. M., and B. L. Bassler. 2005. Quo-
novel bacterial signalling system. Mol. Microbiol. rum sensing: cell-to-cell communication in bac-
57:629–639. teria. Annu. Rev. Cell Dev. Biol. 21:319–346.
81. Ruiz, N., and T. J. Silhavy. 2005. Sensing exter- 95. Werner, E., F. Roe, A. Bugnicourt, M. J.
nal stress:watchdogs of the Escherichia coli cell enve- Franklin,A. Heydorn, S. Molin, B. Pitts, and
lope. Curr. Opin. Microbiol. 8:122–126. P. S. Stewart. 2004. Stratified growth in
82. Sanchez-Martinez, C., and J. Perez-Martin. Pseudomonas aeruginosa biofilms. Appl. Environ.
2001. Dimorphism in fungal pathogens: Candida Microbiol. 70:6188–6196.
albicans and Ustilago maydis—similar inputs, differ- 96. Winzer, K., K. R. Hardie, and P. Williams.
ent outputs. Curr. Opin. Microbiol. 4:214–221. 2002. Bacterial cell-to-cell communication:
83. Shah, J. 2003. The salicylic acid loop in plant sorry, can’t talk now—gone to lunch! Curr. Opin.
defense. Curr. Opin. Plant Biol. 6:365–371. Microbiol. 5:216–222.
84. Soksawatmaekhin,W.,A. Kuraishi, K. Sakata, 97. Yamamoto, S., M. A. Chowdhury, M.
K. Kashiwagi, and K. Igarashi. 2004. Excretion Kuroda, T. Nakano, Y. Koumoto, and S.
and uptake of cadaverine by CadB and its physio- Shinoda. 1991. Further study on polyamine
logical functions in Escherichia coli. Mol. Microbiol. compositions in Vibrionaceae. Can. J. Microbiol.
51:1401–1412. 37:148–153.
85. Sturgill, G., and P. N. Rather. 2004. Evidence 98. Yanofsky, C.,V. Horn, and P. Gollnick. 1991.
that putrescine acts as an extracellular signal Physiological studies of tryptophan transport and
required for swarming in Proteus mirabilis. Mol. tryptophanase operon induction in Escherichia
Microbiol. 51:437–446. coli. J. Bacteriol. 173:6009–6017.
86. Sturgill, G., C. M.Toutain, J. Komperda, G.A. 99. Yim, G., H. H.Wang, and J. Davies. 2006.The
O’Toole, and P. N. Rather. 2004. Role of CysE truth about antibiotics. Int. J. Med. Microbiol.
in production of an extracellular signaling mole- 296:163–170.
cule in Providencia stuartii and Escherichia coli: loss of 100. Yoon, S. S., R. F. Hennigan, G. M. Hilliard,
CysE enhances biofilm formation in Escherichia U.A. Ochsner, K. Parvatiyar, M. C. Kamani,
coli. J. Bacteriol. 186:7610–7617. H. L.Allen,T. R. DeKievit, P. R. Gardner, U.
87. Surette, M. G., M. B. Miller, and B. L. Bassler. Schwab, J. J. Rowe, B. H. Iglewski, T. R.
1999.Quorum sensing in Escherichia coli,Salmonella McDermott, R. P. Mason, D. J.Wozniak, R.
typhimurium, and Vibrio harveyi: a new family of E. Hancock, M. R. Parsek,T. L. Noah, R. C.
genes responsible for autoinducer production. Boucher, and D. J. Hassett. 2002. Pseudomonas
Proc. Natl.Acad. Sci. USA 96:1639–1644. aeruginosa anaerobic respiration in biofilms: rela-
88. Tabor, C.W., and H.Tabor. 1985. Polyamines in tionships to cystic fibrosis pathogenesis. Dev. Cell
microorganisms. Microbiol. Rev. 49:81–99. 3:593–603.
89. Theunis, M., H. Kobayashi,W. J. Broughton, 101. Yoshida, M., K. Kashiwagi,A. Shigemasa, S.
and E. Prinsen. 2004. Flavonoids, NodD1, Taniguchi, K.Yamamoto, H. Makinoshima,
NodD2, and nod-box NB15 modulate expres- A. Ishihama, and K. Igarashi. 2004.A unifying
sion of the y4wEFG locus that is required for model for the role of polyamines in bacterial cell
indole-3-acetic acid synthesis in Rhizobium growth, the polyamine modulon. J. Biol. Chem.
sp. strain NGR234. Mol. Plant Microbe Interact. 279:46008–46013.
17:1153–1161. 102. Zakataeva, N. P., V. V. Aleshin, I. L. Tok-
90. Toutain, C. M., M. E. Zegans, and G. A. makova, P. V. Troshin, and V. A. Livshits.
O’Toole. 2005. Evidence for two flagellar stators 1999. The novel transmembrane Escherichia coli
and their role in the motility of Pseudomonas aerug- proteins involved in the amino acid efflux. FEBS
inosa. J. Bacteriol. 187:771–777. Lett. 452:228–232.
91. Vrljic, M., H. Sahm, and L. Eggeling. 1996.A 103. Zakataeva, N. P., E.A. Kutukova, S.V. Gron-
new type of transporter with a new type of cellular ski, P. V. Troshin, V. A. Livshits, and V. V.
function: L-lysine export from Corynebacterium glu- Aleshin. 2006.Export of metabolites by the pro-
tamicum. Mol. Microbiol. 22:815–826. teins of the DMT and RhtB families and its pos-
92. Walker, T. S., H. P. Bais, E. Deziel, H. P. sible role in intercellular communication.
Schweizer, L. G. Rahme, R. Fall, and J. M. Mikrobiologiia 75:509–520.
104. Zhang, X. S., R. Garcia-Contreras, and 105. Zumft, W. G. 2004. Denitrification by

T. K. Wood. 2007. YcfR (BhsA) influences pseudomonads:control and assemply processes,p.
Escherichia coli biofilm formation through stress 193–227. In J. L. Ramos (ed.), Pseudomonas, vol. 3.
response and surface hydrophobicity. J. Bacteriol. Kluwer Academic/Plenum Publishers, New
189:3051–3062. York, NY.
CELL-CELL SIGNALING
IN MUTUALISTIC AND
PATHOGENIC
ASSOCIATIONS WITH
HUMANS,ANIMALS,
AND PLANTS
II
LuxR-TYPE PROTEINS IN PSEUDOMONAS
AERUGINOSA QUORUM SENSING: DISTINCT
MECHANISMS WITH GLOBAL
IMPLICATIONS
Martin Schuster and E. P. Greenberg
9
PSEUDOMONAS AERUGINOSA: A in association with a broad range of eukaryotic
VERSATILE,TALKATIVE PATHOGEN hosts. As a human pathogen, P. aeruginosa is
Intercellular communication by the exchange known for the ability to cause a variety of acute
of chemical signals, termed quorum sensing, and chronic persistent infections in immuno-
allows a bacterial population to coordinate compromised individuals. Notably, it chroni-
gene expression in response to cell density (3, cally colonizes the lungs of patients with the
75).This process is often important in the colo- genetic disorder cystic fibrosis.
nization of animal and plant hosts by symbiotic Quorum-sensing-regulated expression of an
or pathogenic bacteria, and it can influence the array of virulence factors such as extracellular
development and persistence of bacterial enzymes (LasB elastase, LasA protease, alkaline
biofilms. Although different bacterial groups protease), secondary metabolites (pyocyanin,
have different mechanisms for monitoring their hydrogen cyanide, pyoverdin), and toxins (exo-
own abundance in a local environment, one toxin A) plays an important role in the infection
mechanism that has emerged as common in process (64, 71, 87). These quorum-sensing-
Proteobacteria is that governed by diffusible acyl- regulated factors are involved in tissue damage
homoserine lactone (acyl-HSL) signals. and systemic spread.In mouse and rat models of
The opportunistic pathogen Pseudomonas acute and chronic infection, and in several
aeruginosa provides one of the most intensely invertebrates, strains deficient in quorum sens-
studied examples of acyl-HSL-controlled gene ing show decreased colonization and reduced
expression. This metabolically versatile organ- or no pathology compared to the respective
ism thrives in diverse habitats, including fresh- parent strains. Quorum sensing is also involved
water, soil, and nosocomial environments, and in biofilm development (13). Biofilm bacteria
are up to a thousand times more resistant to
E. P. Greenberg Department of Microbiology, University of antibiotics than their planktonic counterparts,
Washington, Seattle, Washington 98195. Martin Schuster
Department of Microbiology, Oregon State University, Cor- and they are thought to play a significant role in
vallis, Oregon 97331. many persistent infections (12).There is micro-
133
134 ■ SCHUSTER AND GREENBERG
scopic and physiological evidence that P. aerugi- genes under RhlRI control is predicted to
nosa cells exist as biofilms in the lungs of cystic occur subsequent to the activation of genes
fibrosis patients (11). At least under some in under LasRI control.Such a temporally ordered
vitro conditions P. aeruginosa biofilms form sequence of gene expression may be essential for
structured groups with stalked mushroom- the coordination of early and late events in a
shaped aggregates more than 100 m in thick- successful infection. Target promoters respond
ness. Quorum-sensing-deficient mutants of to each system with varying degrees of speci-
P. aeruginosa in contrast form thin and unstruc- ficity (68, 88), but little is known about the
tured biofilms. Biofilms treated with a com- sequence determinants for this specificity.
pound that specifically inhibits quorum sensing Genetic evidence suggests that LasR and RhlR
are similarly flat and unstructured and are more bind to conserved palindromic sequences of
susceptible to antibiotic treatment than fully some quorum-controlled promoters, and more
differentiated biofilms (32). such sites have been located upstream of other
quorum-controlled genes (2, 58, 65, 88).These
THE P. AERUGINOSA so-called las-rhl boxlike sequences show similar-
QUORUM-SENSING CIRCUITRY ity to the lux box promoter element required for
Acyl-HSL signaling was initially discovered and quorum control of the V. fischeri luminescence
characterized in the luminescent marine bac- genes (14). Our recent in vitro characterization
terium Vibrio fischeri (24, 25, 87).The diffusible of the promoter-binding sequences of purified
signal, also called autoinducer, was identified LasR provided further insights. Unlike genetic
as N-3-(oxo-hexanoyl)-HSL (3-O-C6-HSL) studies initially suggested, LasR-binding sites
(18). It is enzymatically synthesized by the showed little overall sequence conservation and
product of the luxI gene (21, 35). The signal did not require dyad symmetry (69). LasR
receptor encoded by the luxR gene is a cyto- bound to some promoters cooperatively,
plasmic acyl-HSL-responsive transcriptional and to others noncooperatively, and although
activator (21, 22). LuxR-3-O-C6-HSL binds sequence similarity was low overall, it was com-
to a 20-bp palindromic promoter element paratively high within each binding class. How-
(termed the lux-box) to activate transcription ever, the physiological significance of these
of the lux operon containing luxI and the genes distinct modes of binding and the exact deter-
required for light production (14, 20). minants of LasR promoter specificity remain to
In P.aeruginosa,quorum-sensing gene regula- be determined.
tion is accomplished by two complete acyl-HSL Another component of the P. aeruginosa
systems, LasR-LasI and RhlR-RhlI, and by an quorum-sensing network is QscR, a third
orphan receptor,QscR (Color Plate 6).LasI and LuxR homolog (Color Plate 6). Initially, QscR
RhlI are the LuxI-type synthases that produce was found to repress the transcription of several
the signal molecules N-3-(oxododecanoyl)- quorum-controlled genes (9).This process has
HSL (3-O-C12-HSL) (54) and N-butyryl-HSL been suggested to involve heterodimer forma-
(C4-HSL) (55, 90), respectively. LasR and RhlR tion between QscR and the two other regula-
are the LuxR-type regulators that respond to tors (41), but recent work shows that QscR can
their cognate signals and activate transcription also act as a transcriptional activator in the pres-
(28, 60). LasR and RhlR also induce the tran- ence of the LasI-generated signal, 3-O-C12-
scription of their cognate synthase genes, creat- HSL (42). QscR regulates its own set of target
ing a positive feedback loop that allows a rapid genes besides controlling the expression of sev-
increase in signal production and dissemination eral QS genes (43). Purified QscR has been
(40,70).Both systems are connected in a hierar- shown to bind to two individual promoters that
chical fashion, as LasR, at least under some have elements similar in sequence to those
conditions, controls the expression of rhlI and found in LasR- or RhlR-dependent promot-
rhlR (40, 47, 57).Therefore, the activation of the ers, but QscR does not bind to LasR- and
9. LuxR-TYPE PROTEINS IN P. AERUGINOSA QUORUM SENSING ■ 135
RhlR-specific promoters that have been tested to all three studies, and most of the genes that
(42). A LasR-specific promoter, however, can had been identified by transposon analysis
be converted to a QscR-specific promoter were confirmed. QS-box elements were found
when two bases in the LasR-binding site are in up to 25% of all predicted quorum-
changed to bases that conform with the QscR- controlled promoters, and it was concluded that
binding sequence, suggesting that these posi- most quorum-controlled genes are regulated
tions are important specificity determinants indirectly by quorum sensing (68, 83). How-
(42).To acknowledge the fact that LasR, RhlR, ever, these searches did not include sequence
and QscR all bind to sequences of similar com- information from studies with purified LasR,
position, we propose the term “QS-box” and further biochemical analyses of RhlR and
instead of “las-rhl box” to describe these LuxR- QscR promoter binding may reveal additional
type protein binding sites in P. aeruginosa. insights. It is therefore possible that a refined
search would reveal many more candidate pro-
GLOBAL GENE REGULATION moters under direct quorum-sensing control.
AND REGULATORY NETWORKS The most overrepresented functional class
Since the discovery of cell-cell signaling in P. among the quorum-sensing genes identified by
aeruginosa in the early 1990s (53), the list of either study was, as expected, secreted factors
genes reported to be controlled by quorum (including toxins and extracellular enzymes),
sensing has increased steadily. Whiteley et al. reinforcing the important role of quorum-
took the first global approach to identify a larger sensing gene expression in virulence. However,
set of quorum-controlled genes (89).They gen- many genes involved in general metabolic
erated random lacZ transcriptional fusions in functions such as central intermediary metabo-
the chromosome of a P. aeruginosa lasI rhlI signal lism, biosynthesis of cofactors, and fatty acid
generation mutant. By screening the resulting metabolism were also affected above average.
library for acyl-HSL-dependent induction of This suggested that acyl-HSL signaling triggers
-galactosidase,35 genes were identified.How- major physiological changes in the cell that
ever, on the basis of the number of mutants reach far beyond virulence functions and that
screened, it was estimated that there were over may contribute to adaptation to a high cell
200 additional quorum-controlled genes. density environment. Many of the quorum-
The availability of high-density DNA controlled target genes, however, were identi-
microarrays made a more comprehensive iden- fied by only one or two of the three groups.
tification of quorum-controlled genes feasible. Apart from differences in data analysis, this dis-
Three groups independently utilized this tech- parity is likely due to differences in experimen-
nology to characterize acyl-HSL-dependent tal conditions (79). Consistent with this notion,
gene expression in a P. aeruginosa lasI rhlI mutant Wagner et al. found that the transcription of
(32, 68, 83). One group also compared tran- many quorum-controlled genes varied with
scription of a lasR rhlR signal receptor mutant respect to growth medium and oxygen avail-
to that of the parent strain,and target genes were ability (83).The three transcriptome studies also
considered to be quorum controlled if they identified several genes that were repressed by
showed differential expression with both strain quorum sensing.There was little overlap among
combinations (68). Schuster et al. (68) and Wag- the repressed genes found in the three inde-
ner et al. (83) each identified more than 300 pendent studies.
quorum-induced genes, which represent more One microarray study provided additional
than 6% of the P. aeruginosa genome. Hentzer insights into the signal requirements for the
et al. (32) reported a considerably smaller set of activation of individual quorum-controlled
genes as they used more stringent criteria for genes (68). Comparison of the transcript
differential expression.About 20% of the identi- profiles of a signal generation mutant in the
fied quorum-controlled genes were common presence of 3-O-C12-HSL alone and in the
presence of both signals, 3-O-C12-HSL and activate expression of the quorum-controlled

C4-HSL, revealed that signal specificities are on hydrogen cyanide biosynthetic operon hcnABC
a continuum;i.e.,some genes respond no better (58). Activation occurs together with LasR
to both signals than to 3-O-C12-HSL alone and RhlR and requires a conserved sequence
(“las-specific” genes), most genes show pro- element, termed the FNR/ANR box, in
gressively greater responses to both signals than the upstream regulatory region hcnABC. A
to 3OC12-HSL alone, and some genes respond search for sequences with similarity to the
well to both signals but not at all to 3-O-C12- FNR/ANR consensus sequence obtained
HSL alone (“rhl-specific” genes).This elaborate from the PRODORIC database (52) identified
expression pattern appears to be the result of such sites in about 25% of all predicted
the differential responsiveness, in essence bind- quorum-controlled promoters (66). Thus,
ing strength,of individual QS-box sequences to ANR may be an important component in the
active LasR and/or RhlR (69), although much coregulation of quorum-controlled genes
remains to be learned about the molecular under anaerobic conditions.
details of these interactions. The quinolone signaling pathway is another
Microarray studies also recently identified a example of a regulatory circuitry that affects the
regulon for QscR (43).Transcript profiles of a P. expression of quorum-sensing target genes,
aeruginosa qscR mutant and the parental strain although in this case, quinolone signaling itself
were compared throughout growth in batch is also regulated by acyl-HSL signaling. The
culture. QscR affected the expression of more Pseudomonas quinolone signal (PQS), 2-heptyl-
than 400 genes; about 20% were induced, and 3-hydroxy-4-quinolone, was originally identi-
80% were repressed.Of these genes,150 are also fied as a third signal besides 3-O-C12-HSL and
controlled by the LasR-LasI and RhlR-RhlI C4-HSL that controls quorum-sensing gene
systems. Complementary transcriptome analy- expression (56).The genes required for the syn-
sis of a P. aeruginosa mutant expressing a qscR thesis of a direct precursor of PQS (pqsABCD
allele without the predicted DNA-binding and phnAB) are activated by the transcriptional
domain identified a subset of at least 40 of regulator MvfR (also called PqsR) (16, 27).
the genes that are candidates for direct activa- MvfR itself is under the control of LasR-3-O-
tion by QscR. C12-HSL (32, 68), and accordingly, there is a
It has become evident that P. aeruginosa quo- large overlap between the MvfR and quorum-
rum-sensing gene expression is also significantly sensing regulons (15). Under certain culture
influenced by other regulatory circuits, which conditions, however, PQS can also be produced
explains its dependence on environmental con- in the absence of LasR (17). MvfR mostly reg-
ditions.There appear to be two major levels of ulates rhl-dependent genes without affecting
signal integration within the quorum-sensing the production of acyl-HSL signals or the
network, the regulator LasR and individual expression of lasR or rhlR.These findings sug-
quorum-controlled genes (66). LasR represents gest that MvfR/PQS and rhl-dependent quo-
a central checkpoint as it is highly intercon- rum sensing are parallel pathways that converge
nected with other regulatory pathways (82). at the promoters of their target genes. Interest-
The catabolite repressor homolog Vfr is one ingly, PQS has been shown to be packaged into
example of a global regulator in P. aeruginosa membrane vesicles for extracellular transport
that affects the expression of LasR.Vfr is acti- between bacterial cells (46).
vated by binding the alarmone cyclic AMP Several other pathways affect quorum-
(91).Vfr directly induces transcription of lasR at sensing gene expression at both levels of signal
the transition from the logarithmic to the sta- integration. The stationary-phase sigma factor
tionary phase of growth (1). RpoS, for example, is a global regulator that
The anaerobic regulator ANR seems to modulates the expression of numerous quo-
affect quorum-sensing gene expression at the rum-controlled genes at the onset of stationary
target gene level. ANR has been shown to phase (67). Some genes appear to be regulated
directly as they possess a putative RpoS- secretion functions and exopolysaccharide

binding site that is similar to the Escherichia coli genes involved in biofilm formation (29, 80).As
RpoS consensus sequence.Others genes appear such, it has been proposed to be a major switch
to be regulated indirectly because RpoS also in the control of acute versus chronic infection.
affects LasR and RhlR expression.
Another example is the GacA/GacS two- LuxR-TYPE PROTEINS
component regulatory system. It modulates LuxR-type polypeptides can be subdivided
expression of LasR and several quorum- into two functional domains (Fig. 1) based on
controlled target genes posttranscriptionally sequence conservation, genetic and biochemi-
through the small regulatory RNA RsmZ and cal analysis of representative members, and the
the RNA-binding protein RsmA (33, 59, 63). crystal structure of TraR from Agrobacterium
It functions as a pleiotropic regulatory system tumefaciens (78, 92). The acyl-HSL-binding
of P. aeruginosa virulence. Together with two region comprises a conserved cluster of residues
other histidine kinases, RetS and LadS, the in the amino-terminal domain of LuxR homo-
GacAS/RsmAZ pathway is involved in con- logues, and mutations in this region abolish the
trolling the reciprocal expression of type III binding of 3-O-C6-HSL to LuxR (31, 73).
FIGURE 1 Structure and function of LuxR-type transcription factors. (A) Key regions
based on sequence conservation, biochemical analysis of LuxR, TraR and other family
members, and the recent TraR crystal structure.The indicated multimerization region has
been genetically defined as being required for LuxR multimerization. Several details men-
tioned in the text are not included in this cartoon. (B) A 3-oxo-C8-HSL molecule showing
the specific TraR residues that are thought to coordinate each position of this acyl-HSL
(Tyr61 interacts along the acyl chain). (C) The palindromic tra box (the DNA-binding site
for TraR) is shown with the two residues (Arg206 and Arg210) that make specific base con-
tacts with this site.Symmetrical contacts are made with the DNA sequence in each half-site.
LuxR-type proteins also contain a helix-turn- RNAP in the interaction with LuxR and TraR.
helix motif (HTH) in their carboxy-terminal Site-directed mutagenesis identified amino
domain that is required for DNA binding (73). acid residues within the N-terminal and C-
Specific residues in the HTH and flanking terminal domains of LuxR and TraR that are
sequences are conserved in the LuxR family. impaired in transcription activation but not
Precise removal of the N-terminal domains of DNA binding. These residues form surface
LuxR and LasR results in proteins that activate patches that are important for positive control
their target genes constitutively (2, 7).This sug- of transcription,most likely by directly contact-
gests that acyl-HSL binding at the amino termi- ing the a-subunit C-terminal domain of
nus relieves inhibition of the DNA-binding RNAP (19, 61, 86). LuxR and TraR are suffi-
domain. Several amino acid substitutions cient to activate transcription of their target
within the C-terminal domain, including the promoters in vitro through purified RNAP,
HTH, have been shown to abolish DNA bind- suggesting that no other factor is necessary for
ing (19, 44, 76, 86). transcript activation (74, 77, 93).
Another important consequence of acyl- Several LuxR homologues from Erwinia and
HSL binding is multimerization, as has been Pantoea function as repressors. ExpR from
shown in vivo with LuxR from V. fischeri and Erwinia chrysanthemi and EsaR from Pantoea
LasR from P. aeruginosa, and in vitro with TraR stewartii have been characterized in consider-
from A. tumefaciens and CarR from the plant able detail (Table 1).Accumulation of the acyl-
pathogen Erwinia carotovora (8, 37, 62, 85, 94). HSL results in target gene expression, but it is
Results with RhlR are more controversial.One accomplished by a different mechanism com-
study suggests that RhlR dimerizes in vitro pared with LuxR-type activators. The tran-
in the absence of its cognate signal, C4-HSL scription factor binds to its DNA sequence in
(81), but another shows that C4-HSL induces the absence of acyl-HSL, and interaction with
multimerization in vivo (39). For LuxR,TraR, the signal causes dissociation from the DNA,
LasR, and RhlR, acyl-HSL is required for thereby derepressing target genes (5, 49, 50).
DNA binding and activation of transcription, Studies with EsaR have also shown that it exists
whereas CarR can bind DNA as a stable dimer as a dimer in the absence of acyl-HSL (50).
in the absence of acyl-HSL.Acyl-HSL binding The TraR crystal structure in complex with
results in the formation of higher-order its cognate acyl-HSL (3-O-C8-HSL) and
oligomers of CarR that activate target gene canonical DNA-binding sequence provides
expression (85). detailed insights into DNA and acyl-HSL bind-
The multimeric LuxR homologues recog- ing (78, 92) (Fig. 1). The carboxy-terminal
nize conserved palindromic sequence elements HTH of TraR is positioned within the major
upstream of target genes. These sequences are groove of the DNA-binding site,allowing base-
located proximal to the regulated promoter,and specific contacts. Strikingly, the acyl-HSL is
they share considerable similarity with the completely buried within the protein, which is
LuxR target sequence upstream of the V. fischeri consistent with biochemical studies indicating
lux operon (14, 20, 26, 65).The dyad symmetry that the acyl-HSL ligand associates tightly with
of most lux-box-type sequences likely reflects TraR during protein synthesis but cannot bind
the symmetrical binding domain of its corre- to prefolded apo-TraR (93, 94). In addition to
sponding regulator,in analogy to TraR.Binding several hydrophobic interactions,TraR binds 3-
of the multimeric LuxR-type protein to its O-C8-HSL through four hydrogen bonds (Fig.
DNA sequence directly upstream of the 35 2). The corresponding amino acid residues are
region allows interaction with RNA poly- highly conserved among LuxR family mem-
merase. Mutational analyses identified roles for bers and are required for TraR function (6).
the -subunit carboxy-terminal domain (23, Molecular modeling predicts that 3-O-C12-
61, 72) as well as for the 70-subunit (34) of HSL binding to LasR is similar to 3-O-C8-HSL
binding to TraR (51), but that 3-O-C6-HSL HSL are bound reversibly and are not required
binding to LuxR is different (38). Characteriza- for folding. QscR is the only exception. The
tion of site-directed mutants supports the observation that QscR binds 3-O-C12-HSL
notion that the acyl-HSL ligand is flipped in reversibly is based on purified His-tagged
LuxR compared with TraR (38). QscR.Although it cannot be excluded that the
Based on the recent biochemical characteri- acyl-HSL-binding properties differ in the
zation of other LuxR-type proteins, general native protein, additional in vivo studies show
patterns of acyl-HSL/receptor interaction that the native protein exhibits relaxed acyl-
emerge that allow us to distinguish three sepa- HSL binding specificity (42). QscR binds to
rate classes (Table 1 and Color Plate 7). The and can be activated by several different acyl-
three P. aeruginosa receptors LasR, RhlR, and HSL, consistent with a rather loose protein-
QscR each represent one such class: LasR is a acyl-HSL interaction.
class 1 receptor. Purified LasR, like TraR, binds What might be the consequences of these
its ligand very tightly and requires 3-O-C12- different interactions between the various
HSL for the expression of fully soluble, active LuxR-type proteins and their cognate acyl-
protein (69). QscR is a class 2 receptor. It also HSL signals? If LuxR-type proteins such as
requires its ligand for the expression of soluble, RhlR do not require ligand for folding and
active protein, but once folded, the ligand binds bind ligand as mature apo-proteins (Color Plate
reversibly (42). RhlR is a class 3 receptor. It 7), they should induce target genes relatively
does not require C4-HSL for the expression of quickly because protein activation through lig-
stable, soluble apo-protein (48), suggesting that and binding would occur instantly. In contrast,
this ligand is not an integral part of RhlR struc- if LuxR-type proteins such as LasR and QscR
ture and is bound reversibly. However, C4-HSL require their ligand for proper folding, they
binding has yet to be assessed with purified should induce transcription of target genes rel-
RhlR in vitro. atively slowly because protein activation
The data in Table 1 suggest that there is a through translation would be slow. This is
correlation between the length of the acyl-side because the process of translation itself takes
chain and the type of acyl-HSL/receptor inter- time (e.g.,about 15 amino acids per s in E.coli at
action: Long-chain acyl-HSL are bound tightly 37C) and occurs largely sequentially due to
and are required for the proper folding of their the generally high protein-to-mRNA ratio (on
cognate receptors, whereas short-chain acyl- average,100 to 1,000 times more protein copies
TABLE 1 Characteristics of acyl-HSL/LuxR-type receptor interaction
LuxR-type Acyl-HSL required Binding

receptor a Organism Acyl-HSL for folding strength Classb Reference(s)
LasR P. aeruginosa 3-O-C12 Yes Very tight 1 69
His-QscR P. aeruginosa 3-O-C12 Yes Reversible 2 42
TraR A. tumefaciens 3-O-C8 Yes Tight 1 93, 94
CepR B. cenocepacia C8 Yes Tight 1 84
LuxR V. fischeri 3-O-C6 Yes Reversible 2 77
His-CarRc E. carotovora 3-O-C6 ? Reversible ? 85
EsaR (repressor) P. stewartii 3-O-C6 No Reversible 3 49
ExpR (repressor) E. chrysanthemi 3-O-C6 No Reversible 3 5
RhlR P. aeruginosa C4 No Reversible 3 48
a
P. aeruginosa receptors are in boldface.
b
We define individual classes as follows.Class 1 proteins require acyl-HSL for folding (protein synthesis) and,once folded,bind acyl-HSL
tightly. Class 2 proteins require acyl-HSL for folding, but the mature protein binds acyl-HSL reversibly. Class 3 proteins do not require acyl-
HSL for folding, and the mature protein binds acyl-HSL reversibly.
c
His-CarR was purified by refolding from inclusion bodies, because the protein was insoluble with and without added 3-O-C6-HSL.
Thus, acyl-HSL requirement for folding cannot be addressed.
than mRNA copies per cell) (4, 30). Conse- iniscent of a developmental program in more
quently, for class 1 and class 2 proteins, a brief complex systems. Consistent with this idea, P.
rise (“spike”) in signal concentration would not aeruginosa lasI but not rhlI mutants form flat
lead to a significant increase in active protein biofilms that do not differentiate into complex,
and therefore would not induce target genes. mushroom- and pillar-shaped structures (13).
This would synchronize and coordinate the The loose signal binding of QscR combined
commitment to entering a multicellular life- with the finding that it can be activated by sig-
style. In network engineering terms, such a nals other than 3-O-C12-HSL (C12-HSL, C10-
mechanism would constitute a low-pass filter.It HSL, and 3-O-C10-HSL) has also led to the
would reduce the sensitivity of the system by speculation that this LuxR protein might par-
filtering fast acyl-HSL signal increases to avoid ticipate in interspecies signaling (42).This is an
untimely upregulation of target genes. intriguing possibility, although at this point it is
A second, related property of LuxR-type not clear whether P. aeruginosa commonly
proteins is their distinct ligand-binding. LasR encounters or lives in association with bacterial
and TraR, for example, appear to bind their species that produce such acyl-HSL signals.
ligands irreversibly (69, 93, 94), whereas QscR, Certain Burkholderia cepacia complex isolates
LuxR, and RhlR bind their ligands reversibly produce C10-HSL (10), and members of the B.
(42, 77) (Table 1). Based on these ligand- cepacia complex and P. aeruginosa sometimes
binding properties, another prediction would coexist in the lungs of cystic fibrosis patients.
be that ligand-bound class 1 LuxR-type pro-
teins are much more resistant to sudden drops in CONCLUSIONS
external signal concentration than are ligand- Acyl-HSL quorum sensing by P. aeruginosa rep-
bound class 2 or class 3 proteins.In other words, resents one of the best-understood cell-cell
class 1 proteins might be inherently more robust communication systems in bacteria.The system
to “noise.” Indeed, TraR, a class 1 protein, has is highly intertwined with other cellular path-
been shown to activate a TraR-dependent ways, rendering it responsive to a multitude of
reporter fusion in E. coli several hours after environmental signals. Despite this complexity,
removal of the cognate signal 3OC8-HSL from many of its features can be explained solely by
the culture medium (45).With LuxR, a class 2 the compact circuitry of two synthases and
protein, lux gene transcription drops within three receptors. For example, the broad spec-
minutes of 3-O-C8-HSL removal (36). trum of signal specificities of quorum-
Both properties—protein activation during controlled genes may be caused solely by the
translation and irreversible ligand binding— binding strength of quorum-sensing receptors
have the potential to improve the robustness of to promoters.The distinct interactions between
the quorum-sensing system by avoiding noise- acyl-HSL ligands and individual receptors have
triggered up- or downregulation of target the potential for specific signal reception and
genes. It makes sense that LasR would exhibit response patterns. Such simplicity within com-
both of these properties because it is the master plexity, the large number of comparatively
regulator in the P. aeruginosa quorum-sensing well-characterized quorum-regulated genes,
circuitry, controlling the induction of a large and the presence of distinctly different LuxR
regulon comprising hundreds of target genes. proteins in one organism all make P. aeruginosa
With respect to bacterial social behavior, an ideal model system for further studies.
that is,the formation of biofilms,the irreversible Although much remains to be learned, detailed
binding of 3-O-C12-HSL to LasR could impart insights into the molecular mechanisms of
a long-term commitment to a community quorum-sensing gene regulation may provide
lifestyle. Once a quorum is reached, a change in an answer to a larger question: Why does
the gene expression profile is triggered that one cell or population of cells need so many
leads to the formation of a mature biofilm,rem- quorum-sensing systems and not just one?
REFERENCES 14. Devine, J. H., G. S. Shadel, and T. O. Baldwin.

1. Albus, A. M., E. C. Pesci, L. J. 1989. Identification of the operator of the lux reg-
Runyen-Janecky, S. E. West, and B. H. ulon from the Vibrio fischeri strain ATCC7744.
Iglewski. 1997. Vfr controls quorum sensing in Proc. Natl.Acad. Sci. USA 86:5688–5692.
Pseudomonas aeruginosa.J.Bacteriol. 179:3928–3935. 15. Deziel, E., S. Gopalan, A. P. Tampakaki, F.
2. Anderson, R. M., C. A. Zimprich, and L. Lepine, K. E. Padfield, M. Saucier, G. Xiao,
Rust. 1999. A second operator is involved in and L. G. Rahme. 2005. The contribution of
Pseudomonas aeruginosa elastase (lasB) activation. J. MvfR to Pseudomonas aeruginosa pathogenesis and
Bacteriol. 181:6264–6270. quorum sensing circuitry regulation: multiple
3. Bassler, B. L. 2002. Small talk. Cell-to-cell com- quorum sensing-regulated genes are modulated
munication in bacteria. Cell 109:421–424. without affecting lasRI, rhlRI or the production of
4. Bremer, H., and P. P. Dennis. 1996. Modula- N-acyl-L-homoserine lactones. Mol. Microbiol.
tion of chemical composition and other parame- 55:998–1014.
ters of the cell by growth rate, p. 1553–1569. In F. 16. Deziel, E., F. Lepine, S. Milot, J. He, M. N.
C. Neidhardt, R. Curtiss, J. L. Ingraham, E. E. C. Mindrinos, R. G. Tompkins, and L. G.
Lin, K. B. Low, B. Magasanik,W. S. Reznikoff, M. Rahme. 2004. Analysis of Pseudomonas aeruginosa
Riley, M. Schaechter, and H. E. Umbarger (ed.), 4-hydroxy-2-alkylquinolines (HAQs) reveals a
Escherichia coli and Salmonella: Cellular and Mole- role for 4-hydroxy-2-heptylquinoline in cell-to-
cular Biology, 2nd ed.ASM Press,Washington, DC. cell communication. Proc. Natl. Acad. Sci. USA
5. Castang, S., S. Reverchon, P. Gouet, and W. 101:1339–1344.
Nasser. 2006. Direct evidence for the modulation 17. Diggle, S. P., K.Winzer, S. R. Chhabra, K. E.
of the activity of the Erwinia chrysanthemi quorum- Worrall, M. Camara, and P. Williams. 2003.
sensing regulator ExpR by acylhomoserine lac- The Pseudomonas aeruginosa quinolone signal mol-
tone pheromone. J. Biol. Chem. 281:29972–29987. ecule overcomes the cell density-dependency of
6. Chai,Y., and S. C. Winans. 2004. Site-directed the quorum sensing hierarchy, regulates rhl-
mutagenesis of a LuxR-type quorum-sensing dependent genes at the onset of stationary phase
transcription factor: alteration of autoinducer and can be produced in the absence of LasR. Mol.
specificity. Mol. Microbiol. 51:765–776. Microbiol. 50:29–43.
7. Choi, S. H., and E. P. Greenberg. 1991.The C- 18. Eberhard,A.,A. L. Burlingame, C. Eberhard,
terminal region of the Vibrio fischeri LuxR protein G. L. Kenyon, K. H. Nealson, and N. J.
contains an inducer-independent lux gene Oppenheimer. 1981. Structural identification of
activating domain. Proc. Natl. Acad. Sci. USA autoinducer of Photobacterium fischeri luciferase.
88:11115–11119. Biochemistry 20:2444–2449.
8. Choi, S. H., and E. P. Greenberg. 1992.Genetic 19. Egland, K. A., and E. P. Greenberg. 2001.
evidence for multimerization of LasR, the tran- Quorum sensing in Vibrio fischeri: analysis of the
scriptional activator of Vibrio fischeri luminescence. LuxR DNA binding region by alanine-scanning
Mol. Mar. Biol. Biotechnol. 1:408–413. mutagenesis. J. Bacteriol. 183:382–386.
9. Chugani, S. A., M.Whiteley, K. M. Lee, D. D. 20. Egland, K. A., and E. P. Greenberg. 1999.
Argenio, C. Manoil, and E. P. Greenberg. Quorum sensing in Vibrio fischeri: elements of the
2001. QscR, a modulator of quorum-sensing sig- luxl promoter. Mol. Microbiol. 31:1197–1204.
nal synthesis and virulence in Pseudomonas aerugi- 21. Engebrecht, J., K. Nealson, and M.
nosa. Proc. Natl.Acad. Sci. USA 98:2752–2757. Silverman. 1983. Bacterial bioluminescence:
10. Conway, B.A.,V.Venu, and D. P. Speert. 2002. isolation and genetic analysis of functions from
Biofilm formation and acyl homoserine lactone Vibrio fischeri. Cell 32:773–781.
production in the Burkholderia cepacia complex. J. 22. Engebrecht, J., and M. Silverman. 1984. Iden-
Bacteriol. 184:5678–5685. tification of genes and gene products necessary for
11. Costerton, J.W. 2001. Cystic fibrosis pathogene- bacterial bioluminescence. Proc. Natl. Acad. Sci.
sis and the role of biofilms in persistent infection. USA 81:4154–4158.
Trends Microbiol. 9:50–52. 23. Finney, A. H., R. J. Blick, K. Murakami, A.
12. Costerton, J.W., P. S. Stewart, and E. P Green- Ishihama, and A. M. Stevens. 2002. Role of the
berg. 1999. Bacterial biofilms: a common cause of C-terminal domain of the alpha subunit of RNA
persistent infection. Science 284:1318–1322. polymerase in LuxR-dependent transcriptional
13. Davies, D. G., M. R. Parsek, J. P. Pearson, activation of the lux operon during quorum sens-
B. H. Iglewski, J. W. Costerton, and E. P. ing. J. Bacteriol. 184:4520–4528.
Greenberg. 1998.The involvement of cell-to-cell 24. Fuqua, C., and E. P. Greenberg. 2002. Listen-
signals in the development of a bacterial biofilm. ing in on bacteria: acyl-homoserine lactone sig-
Science 280:295–298. nalling. Nat. Rev. Mol. Cell. Biol. 3:685–695.
25. Fuqua, C., M. R. Parsek, and E. P. Greenberg. the Vibrio fischeri luminescence system. J. Bacteriol.
2001. Regulation of gene expression by cell-to- 163:1210–1214.
cell communication: acyl-homoserine lactone 37. Kiratisin, P., K. D. Tucker, and L. Passador.
quorum sensing. Annu. Rev. Genet. 35:439–468. 2002. LasR, a transcriptional activator of
26. Fuqua, W. C., and S. C. Winans. 1994. A Pseudomonas aeruginosa virulence genes,functions as
LuxR-LuxI type regulatory system activates a multimer. J. Bacteriol. 184:4912–4919.
Agrobacterium Ti plasmid conjugal transfer in the 38. Koch, B.,T. Liljefors,T. Persson, J. Nielsen, S.
presence of a plant tumor metabolite. J. Bacteriol. Kjelleberg, and M. Givskov. 2005. The LuxR
176:2796–2806. receptor: the sites of interaction with quorum-
27. Gallagher, L. A., S. L. McKnight, M. S. sensing signals and inhibitors. Microbiology
Kuznetsova, E. C. Pesci, and C. Manoil. 2002. 151:3589–3602.
Functions required for extracellular quinolone sig- 39. Lamb, J. R., H. Patel, T. Montminy, V. E.
naling by Pseudomonas aeruginosa. J. Bacteriol. Wagner, and B. H. Iglewski. 2003. Functional
184:6472–6480. domains of the RhlR transcriptional regulator of
28. Gambello, M. J., and B. H. Iglewski. 1991. Pseudomonas aeruginosa.J.Bacteriol. 185:7129–7139.
Cloning and characterization of the Pseudomonas 40. Latifi, A., M. Foglino, K.Tanaka, P.Williams,
aeruginosa lasR gene, a transcriptional activator of and A. Lazdunski. 1996.A hierarchical quorum-
elastase expression. J. Bacteriol. 173:3000–3009. sensing cascade in Pseudomonas aeruginosa links the
29. Goodman,A. L., B. Kulasekara,A. Rietsch, D. transcriptional activators LasR and RhIR (VsmR)
Boyd, R. S. Smith, and S. Lory. 2004.A signal- to expression of the stationary-phase sigma factor
ing network reciprocally regulates genes associated RpoS. Mol. Microbiol. 21:1137–1146.
with acute infection and chronic persistence in 41. Ledgham, F., I.Ventre, C. Soscia, M. Foglino,
Pseudomonas aeruginosa. Dev. Cell 7:745–754. J. N. Sturgis, and A. Lazdunski. 2003. Interac-
30. Gygi, S. P.,Y. Rochon, B. R. Franza, and R. tions of the quorum sensing regulator QscR:inter-
Aebersold. 1999. Correlation between protein action with itself and the other regulators of
and mRNA abundance in yeast. Mol. Cell. Biol. Pseudomonas aeruginosa LasR and RhlR. Mol.
19:1720–1730. Microbiol. 48:199–210.
31. Hanzelka, B. L., and E. P. Greenberg. 1995. 42. Lee, J. H., Y. Lequette, and E. P. Greenberg.
Evidence that the N-terminal region of the Vibrio 2006. Activity of purified QscR, a Pseudomonas
fischeri LuxR protein constitutes an autoinducer- aeruginosa orphan quorum-sensing transcription
binding domain. J. Bacteriol. 177:815–817. factor. Mol. Microbiol. 59:602–609.
32. Hentzer, M., H.Wu, J. B.Andersen, K. Riedel, 43. Lequette, Y., J. H. Lee, F. Ledgham, A. Laz-
T. B. Rasmussen, N. Bagge, N. Kumar, M. A. dunski, and E. P. Greenberg. 2006. A distinct
Schembri, Z. Song, P. Kristoffersen, M. QscR regulon in the Pseudomonas aeruginosa quo-
Manefield, J.W. Costerton, S. Molin, L. Eberl, rum-sensing circuit. J. Bacteriol. 188:3365–3370.
P. Steinberg, S. Kjelleberg, N. Hoiby, and M. 44. Luo, Z. Q., and S. K. Farrand. 1999. Signal-
Givskov. 2003.Attenuation of Pseudomonas aerugi- dependent DNA binding and functional domains
nosa virulence by quorum sensing inhibitors. of the quorum-sensing activator TraR as identified
EMBO J. 22:3803–3815. by repressor activity. Proc. Natl. Acad. Sci. USA
33. Heurlier, K., F. Williams, S. Heeb, C. 96:9009–9014.
Dormond, G. Pessi, D. Singer, M. Camara, 45. Luo, Z. Q., S. Su, and S. K. Farrand. 2003. In
P.Williams, and D. Haas. 2004. Positive control situ activation of the quorum-sensing transcrip-
of swarming, rhamnolipid synthesis, and lipase tion factor TraR by cognate and noncognate acyl-
production by the posttranscriptional RsmA/ homoserine lactone ligands: kinetics and
RsmZ system in Pseudomonas aeruginosa PAO1. consequences. J. Bacteriol. 185:5665–5672.
J. Bacteriol. 186:2936–2945. 46. Mashburn, L. M., and M. Whiteley. 2005.
34. Johnson, D. C., A. Ishihama, and A. M. Membrane vesicles traffic signals and facili-
Stevens. 2003. Involvement of region 4 of the tate group activities in a prokaryote. Nature
sigma70 subunit of RNA polymerase in transcrip- 437:422–425.
tional activation of the lux operon during quorum 47. Medina, G., K. Juarez, R. Diaz, and G.
sensing. FEMS Microbiol. Lett. 228:193–201. Soberon-Chavez. 2003. Transcriptional regula-
35. Kaplan, H. B., and E. P. Greenberg. 1985. Dif- tion of Pseudomonas aeruginosa rhlR, encoding a
fusion of autoinducer is involved in regulation of quorum-sensing regulatory protein. Microbiology
the Vibrio fischeri luminescence system. J. Bacteriol. 149:3073–3081.
163:1210–1214. 48. Medina, G., K. Juarez, B. Valderrama, and
36. Kaplan, H. B., and E. P. Greenberg. 1985. Dif- G. Soberon-Chavez. 2003. Mechanism of
fusion of autoinducer is involved in regulation of Pseudomonas aeruginosa RhlR transcriptional
regulation of the rhlAB promoter. J. Bacteriol. regulator RsmA modulates production of viru-
185:5976–5983. lence determinants and N-acylhomoserine lac-
49. Minogue, T. D., A. L. Carlier, M. D. tones in Pseudomonas aeruginosa. J. Bacteriol.
Koutsoudis, and S. B. von Bodman. 2005.The 183:6676–6683.
cell density-dependent expression of stewartan 60. Piper, K. R., S. Beck von Bodman, and S. K.
exopolysaccharide in Pantoea stewartii ssp.stewartii is Farrand. 1993. Conjugation factor of Agrobac-
a function of EsaR-mediated repression of the rcsA terium tumefaciens regulates Ti plasmid transfer by
gene. Mol. Microbiol. 56:189–203. autoinduction. Nature 362:448–450.
50. Minogue, T. D., M. Wehland-von Trebra, F. 61. Qin, Y., Z. Q. Luo, and S. K. Farrand. 2004.
Bernhard, and S. B. von Bodman. 2002. The Domains formed within the N-terminal region of
autoregulatory role of EsaR,a quorum-sensing reg- the quorum-sensing activator TraR are required
ulator in Pantoea stewartii ssp. stewartii: evidence for a for transcriptional activation and direct interaction
repressor function. Mol. Microbiol. 44:1625–1635. with RpoA from Agrobacterium. J. Biol. Chem.
51. Muh, U., B. J. Hare, B. A. Duerkop, M. 279:40844–40851.
Schuster, B. L. Hanzelka, R. Heim, E. R. 62. Qin, Y., Z. Q. Luo, A. J. Smyth, P. Gao, S.
Olson, and E. P. Greenberg. 2006. A struc- Beck von Bodman, and S. K. Farrand. 2000.
turally unrelated mimic of a Pseudomonas aeruginosa Quorum-sensing signal binding results in dimeri-
acyl-homoserine lactone quorum-sensing signal. zation of TraR and its release from membranes into
Proc. Natl.Acad. Sci. USA 103:16948–16952. the cytoplasm. EMBO J. 19:5212–5221.
52. Munch, R., K. Hiller, H. Barg, D. Heldt, S. 63. Reimmann, C., M. Beyeler, A. Latifi, H.
Linz, E. Wingender, and D. Jahn. 2003. Winteler, M. Foglino, A. Lazdunski, and D.
PRODORIC: prokaryotic database of gene regu- Haas. 1997. The global activator GacA
lation. Nucleic Acids Res. 31:266–269. of Pseudomonas aeruginosa PAO positively
53. Passador, L., J. M. Cook, M. J. Gambello, controls the production of the autoinducer N-
L. Rust, and B. H. Iglewski. 1993. Expres- butyryl-homoserine lactone and the formation of
sion of Pseudomonas aeruginosa virulence genes the virulence factors pyocyanin, cyanide, and
requires cell-to-cell communication. Science lipase. Mol. Microbiol. 24:309–319.
260:1127–1130. 64. Rumbaugh, K. P., J. A. Griswold, and A. N.
54. Pearson, J. P., K. M. Gray, L. Passador, K. D. Hamood. 2000. The role of quorum sensing in
Tucker,A. Eberhard, B. H. Iglewski, and E. P. the in vivo virulence of Pseudomonas aeruginosa.
Greenberg. 1994. Structure of the autoinducer Microbes Infect. 2:1721–1731.
required for expression of Pseudomonas aerugi- 65. Rust, L., E. C. Pesci, and B. H. Iglewski. 1996.
nosa virulence genes. Proc. Natl. Acad. Sci. USA Analysis of the Pseudomonas aeruginosa elastase (lasB)
91:197–201. regulatory region. J. Bacteriol. 178:1134–1140.
55. Pearson, J. P., L. Passador, B. H. Iglewski, 66. Schuster, M., and E. P. Greenberg. 2006.
and E. P. Greenberg. 1995.A second N-acylho- A network of networks: quorum-sensing gene
moserine lactone signal produced by Pseudomonas regulation in Pseudomonas aeruginosa. Int. J. Med.
aeruginosa. Proc. Natl.Acad. Sci. USA 92:1490–1494. Microbiol. 296:73–81.
56. Pesci, E. C., J. B. Milbank, J. P. Pearson, S. 67. Schuster, M., A. C. Hawkins, C. S. Harwood,
McKnight, A. S. Kende, E. P. Greenberg, and E. P. Greenberg. 2004. The Pseudomonas
and B. H. Iglewski. 1999. Quinolone signaling aeruginosa RpoS regulon and its relationship to
in the cell-to-cell communication system of quorum sensing. Mol. Microbiol. 51:973–985.
Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA 68. Schuster, M., C. P. Lohstroh,T. Ogi, and E. P.
96:11229–11234. Greenberg. 2003. Identification, timing and sig-
57. Pesci, E. C., J. P. Pearson, P. C. Seed, and B. H. nal specificity of Pseudomonas aeruginosa quorum-
Iglewski. 1997. Regulation of las and rhl quorum controlled genes: a transcriptome analysis. J.
sensing in Pseudomonas aeruginosa. J. Bacteriol. Bacteriol. 185:2066–2079.
179:3127–3132. 69. Schuster, M., M. L. Urbanowski, and E. P.
58. Pessi, G., and D. Haas. 2000. Transcriptional Greenberg. 2004. Promoter specificity in
control of the hydrogen cyanide biosynthetic Pseudomonas aeruginosa quorum sensing revealed by
genes hcnABC by the anaerobic regulator ANR DNA binding of purified LasR. Proc. Natl. Acad.
and the quorum-sensing regulators LasR and Sci. USA 101:15833–15839.
RhlR in Pseudomonas aeruginosa. J. Bacteriol. 70. Seed, P. C., L. Passador, and B. H. Iglewski.
182:6940–6949. 1995.Activation of the Pseudomonas aeruginosa lasI
59. Pessi, G., F. Williams, Z. Hindle, K. Heurlier, gene by LasR and the Pseudomonas autoinducer
M. T. Holden, M. Camara, D. Haas, and P. PAI: an autoinduction regulatory hierarchy. J. Bac-
Williams. 2001. The global posttranscriptional teriol. 177:654–659.
71. Smith, R. S., and B. H. Iglewski. 2003. P. aerug- 84. Weingart, C. L., C. E.White, S. Liu,Y. Chai, H.
inosa quorum-sensing systems and virulence. Curr. Cho, C. S. Tsai, Y. Wei, N. R. Delay, M. R.
Opin. Microbiol. 6:56–60. Gronquist, A. Eberhard, and S. C. Winans.
72. Stevens, A. M., N. Fujita, A. Ishihama, and E. 2005. Direct binding of the quorum sensing regu-
P. Greenberg. 1999. Involvement of the RNA lator CepR of Burkholderia cenocepacia to two target
polymerase alpha-subunit C-terminal domain in promoters in vitro. Mol. Microbiol. 57:452–467.
LuxR-dependent activation of the Vibrio fischeri 85. Welch, M., D. E.Todd, N. A.Whitehead, S. J.
luminescence genes. J. Bacteriol. 181:4704–4707. McGowan, B.W. Bycroft, and G. P. Salmond.
73. Stevens, A. M., and E. P. Greenberg. 1999. 2000. N-acyl homoserine lactone binding to the
p. 231–242. In G. M. Dunny and S. C.Winans (ed.), CarR receptor determines quorum-sensing speci-
Cell-Cell Signaling in Bacteria. ASM Press,Washing- ficity in Erwinia. EMBO J. 19:631–641.
ton, DC. 86. White, C. E., and S. C.Winans. 2005. Identifi-
74. Stevens, A. M., and E. P. Greenberg. 1997. cation of amino acid residues of the Agrobacterium
Quorum sensing in Vibrio fischeri: essential ele- tumefaciens quorum-sensing regulator TraR that are
ments for activation of the luminescence genes. J. critical for positive control of transcription. Mol.
Bacteriol. 179:557–562. Microbiol. 55:1473–1486.
75. Taga, M. E., and B. L. Bassler. 2003. Chemical 87. Whitehead, N. A., A. M. Barnard, H. Slater,
communication among bacteria. Proc. Natl. Acad. N. J. Simpson, and G. P. Salmond. 2001.
Sci. USA 100 (Suppl 2):14549–14554. Quorum-sensing in gram-negative bacteria.
76. Trott, A. E., and A. M. Stevens. 2001. Amino FEMS Microbiol. Rev. 25:365–404.
acid residues in LuxR critical for its mechanism of 88. Whiteley, M., and E. P. Greenberg. 2001. Pro-
transcriptional activation during quorum sensing moter specificity elements in Pseudomonas aerugi-
in Vibrio fischeri. J. Bacteriol. 183:387–392. nosa quorum-sensing-controlled genes. J. Bacteriol.
77. Urbanowski, M. L., C. P. Lostroh, and E. P. 183:5529–5234.
Greenberg. 2004. Reversible acyl-homoserine 89. Whiteley, M., K. M. Lee, and E. P. Greenberg.
lactone binding to purified Vibrio fischeri LuxR 1999. Identification of genes controlled by quo-
protein. J. Bacteriol. 186:631–637. rum sensing in Pseudomonas aeruginosa. Proc. Natl.
78. Vannini, A., C. Volpari, C. Gargioli, E. Acad. Sci. USA 96:13904–13909.
Muraglia, R. Cortese, R. De Francesco, P. 90. Winson, M. K., M. Camara, A. Latifi, M.
Neddermann, and S. D. Marco. 2002.The crys- Foglino, S. R. Chhabra, M. Daykin, M. Bally,
tal structure of the quorum sensing protein TraR V. Chapon, G. P. Salmond, B.W. Bycroft, et al.
bound to its autoinducer and target DNA. EMBO 1995. Multiple N-acyl-L-homoserine lactone
J. 21:4393–4401. signal molecules regulate production of virulence
79. Vasil, M. L. 2003. DNA microarrays in analysis of determinants and secondary metabolites in
quorum sensing: strengths and limitations. J. Bacte- Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA
riol. 185:2061–2065. 92:9427–9431.
80. Ventre, I., A. L. Goodman, I. Vallet-Gely, P. 91. Wolfgang, M. C., V. T. Lee, M. E. Gilmore,
Vasseur, C. Soscia, S. Molin, S. Bleves, A. and S. Lory. 2003. Coordinate regulation of
Lazdunski, S. Lory, and A. Filloux. 2006. bacterial virulence genes by a novel adenylate
Multiple sensors control reciprocal expression cyclase-dependent signaling pathway. Dev. Cell
of Pseudomonas aeruginosa regulatory RNA and 4:253–263.
virulence genes. Proc. Natl. Acad. Sci. USA 92. Zhang, R. G., T. Pappas, J. L. Brace, P. C.
103:171–176. Miller, T. Oulmassov, J. M. Molyneaux, J. C.
81. Ventre, I., F. Ledgham,V. Prima, A. Lazdun- Anderson, J. K. Bashkin, S. C.Winans, and A.
ski, M. Foglino, and J. N. Sturgis. 2003.Dimer- Joachimiak. 2002. Structure of a bacterial
ization of the quorum sensing regulator RhlR: quorum-sensing transcription factor complexed
development of a method using EGFP fluores- with pheromone and DNA. Nature 417:971–974.
cence anisotropy. Mol. Microbiol. 48:187–198. 93. Zhu, J., and S. C. Winans. 1999. Autoinducer
82. Viretta, A. U., and M. Fussenegger. 2004. binding by the quorum-sensing regulator TraR
Modeling the quorum sensing regulatory network increases affinity for target promoters in vitro and
of human-pathogenic Pseudomonas aeruginosa. decreases TraR turnover rates in whole cells. Proc.
Biotechnol. Prog. 20:670–678. Natl.Acad. Sci. USA 96:4832–4837.
83. Wagner,V. E., D. Bushnell, L. Passador, A. I. 94. Zhu, J., and S. C. Winans. 2001.The quorum-
Brooks, and B. H. Iglewski. 2003. Microarray sensing transcriptional regulator TraR requires its
analysis of Pseudomonas aeruginosa quorum-sensing cognate signaling ligand for protein folding, pro-
regulons:effects of growth phase and environment. tease resistance, and dimerization. Proc. Natl.Acad.
J. Bacteriol. 185:2080–2095. Sci. USA 98:1507–1512.
QUORUM SENSING IN VIBRIO
CHOLERAE PATHOGENESIS
Fiona R. Stirling, Zhi Liu, and Jun Zhu
10
Vibrio cholerae is a free-living aquatic bacterium two of the major quorum-sensing controlled
that is also able to colonize the small intestine of phenotypes in V. cholerae, virulence gene
humans, where it can cause cholera. For V. expression and biofilm formation, are repressed
cholerae to exist in these two diverse habitats it at high cell density.To understand why quorum
must be able to sense its environment and sensing negatively regulates virulence factor
respond accordingly to express the required expression and biofilm formation at high cell
survival factors in an appropriate temporal and density, it is important to consider the require-
spatial pattern. One of the environmental sig- ments for these phenotypes at different stages in
nals measured by V. cholerae is its own cell den- the life cycle of V. cholerae.
sity, which it achieves by a quorum-sensing
mechanism. V. cholerae produces at least two dif- THE TWO LIFESTYLES OF THE
ferent autoinducer molecules whose concen- HUMAN PATHOGEN V. CHOLERAE
tration provides a measure of bacterial cell V. cholerae is a gram-negative curved rod
density. Once the concentration of autoinduc- bacterium that is motile by means of a single
ers has reached a critical threshold, a phospho- polar flagellum. This bacterium normally
rylation cascade is initiated in V. cholerae that resides in the aquatic environment, but certain
results in the coordinated expression and serogroups are also able to survive in the human
repression of an array of quorum-sensing con- small intestine, where they can cause the acute
trolled genes. Phenotypes controlled by the dehydrating diarrheal disease cholera. Of the
quorum-sensing system of V. cholerae include over 200 known V. cholerae serogroups, only O1
virulence gene expression, biofilm formation, and O139 have been associated with epidemic
protease production, the manufacture of an and pandemic outbreaks (48).These toxigenic
antiprotozoal factor, competence, the stress V. cholerae serogroups are capable of causing
response, chemotaxis, and motility. However, in massive fluid loss that can be fatal in up to 50%
contrast to many other bacteria studied to date, of severe cholera cases if left untreated. Fortu-
nately a simple treatment involving the rapid
replacement of lost fluids and ions can decrease
Fiona R. Stirling, Zhi Liu, and Jun Zhu Department of Micro-
biology, University of Pennsylvania School of Medicine, mortality to less than 1%. However, cholera
Philadelphia, Pennsylvania 19104-6076. still remains a global killer that results in an
145
146 ■ STIRLING ET AL.
estimated 120,000 deaths per year, mainly in the host as a biofilm may also facilitate the sur-
the developing countries of South Asia,Africa, vival of the bacteria as they travel through the
and Latin America. In addition to causing acidic environment of the stomach as V. cholerae
death,V.cholerae also causes high morbidity rates biofilms are more than 1,000-fold more resist-
and has a severe social and economic impact on ant to acid shock than planktonic cells that are
the already-poor communities that it affects. highly sensitive to low pH (62).
The transition of V. cholerae from its aquatic Once V. cholerae has entered the host and tra-
reservoir to the human host is initiated by the versed the hostile stomach environment,it must
ingestion of contaminated food or water. For penetrate the mucous layer and adhere to and
disease to occur, the ingested bacteria must first colonize the epithelial cells of the small intes-
survive the acidic environment of the stomach tine.To achieve this, V. cholerae produces a num-
and then penetrate the mucosal layer and ber of virulence factors, including the cholerae
adhere to and colonize the epithelium of the toxin (CT) and the toxin coregulated pilus
small intestine. At this site V. cholerae is able to (TCP).TCP is a type IV pilus encoded by the
multiply extensively and stimulate the intestinal Vibrio pathogenicity island (VPI) whose proba-
epithelial cells to secrete large amounts of water ble function is to mediate adherence to the
and solutes into the lumen of the small intes- intestinal mucosal cells (22, 23, 50). It has been
tine.The resulting diarrhea then facilitates the demonstrated that this structure is absolutely
exit of the bacterium from its host back into the required for colonization as its mutation abol-
environment where it is able to infect subse- ishes the ability of V. cholerae to colonize the
quent hosts.This cycle between two habitats as small intestine of both mice and humans (1,16).
diverse as the aquatic environment and the Once V. cholerae has adhered to the intestinal
human intestine necessitates that V. cholerae epithelial cells, it is able to multiply extensively
expresses a broad range of different factors in a while producing the potent CT that causes the
carefully controlled fashion. severe diarrhea characteristic of cholera (21).
During inhabitation of aquatic environ- The genes required for the production of this
ments,V.cholerae lives in association with various toxin, ctxA and ctxB, are not integral compo-
species of phytoplankton and zooplankton, nents of the V. cholerae genome but are located
often in the form of biofilms. Biofilms are in the genome of a lysogenic bacteriophage
surface-associated structures that contain bacte- (CTX) (55). CT acts by ADP-ribosylating
ria embedded in a self-produced, extracellular and constitutively activating the adenylate
polymeric matrix,and it has been demonstrated cyclase enzyme in intestinal cells that leads to a
that the ability of V.cholerae to form these organ- rise in intracellular cyclic AMP levels. The
ized structures on biotic and abiotic surfaces abnormally high cyclic AMP levels stimulate
enhances their survival in the environment (54, the epithelial cells to pump chloride and other
61). In addition to increasing bacterial fitness in ions into the lumen of the intestine, leading to
the natural environment,V.cholerae biofilms may an expulsion of water to maintain the osmotic
also be important in initiating human infection. equilibrium. Water and ions lost from the
It has been suggested that V. cholerae enters the epithelial cells into the intestinal lumen are
host within biofilms as these structures enable replaced from the blood and tissues, and thus
the bacteria to reach a high enough density, massive fluid and electrolyte loss from the host
approximately 104 to 106 total cells, to initiate occurs. In addition to the well-characterized
symptomatic cholera (7). In addition, it was virulence factors CT and TCP, V. cholerae pro-
observed that filtration of drinking water duces a number of additional proteins that may
through sari cloth led to a dramatic decline in be involved in infection. For example, nonfim-
cholera cases in a Bangladeshi village, and this brial adhesins such as hemagglutinin and a
may be a consequence of the removal of V. group of surface proteins encoded by the acf
cholerae biofilms from the water (7). Entering (accessory colonization factor) genes may
10. QUORUM SENSING IN VIBRIO CHOLERAE PATHOGENESIS ■ 147
mediate tighter binding to the epithelial cells sensing, which is a phenomenon exhibited by
than is possible by TCP alone (47). many gram-positive and gram-negative bacteria
The human intestine provides an ideal envi- whereby they monitor their population density
ronment for V. cholerae to increase its biomass, by producing and responding to increasing con-
and from this site the bacteria can disseminate centrations of signaling molecules called
back into the environment and subsequently autoinducers (56) (see chapters 2 and 20).
infect other hosts. It has been proposed that the There is also a third type of quorum-sensing
hemagglutin protease (HapA), although not regulation,exemplified by Vibrio harveyi, which
required for colonization, may contribute to combines components of both the gram-
bacterial detachment from epithelial cells and negative and gram-positive systems. Like the
assist in the exit of V. cholerae from the host (10). gram-negative bacteria, the autoinducer signals
It has also been recently observed that late in produced by V. harveyi are small chemicals, not
infection chemotaxis and motility genes are peptides, which are able to diffuse out of the
coordinately upregulated to facilitate detach- cell. However, the signal detection and relay
ment from the epithelial surface and migration apparatus consists of two-component phos-
into the luminal fluid (44). Bacteria in the photransferase proteins similar to those found
lumen of the intestine are then able to exit the in the quorum-sensing systems of gram-
host in the vast quantities of stools passed by positive bacteria. V. harveyi produces at least
infected individuals. In addition to providing a three different autoinducer molecules that are
habitat for replication, the human host also detected by separate membrane-bound recep-
appears to prime V. cholerae so that it is more tors (15). The cell density information from
infectious than V. cholerae from stationary-phase these three receptors is then channeled into a
cultures grown in vitro (41). V. cholerae cells may single phosphorylation cascade pathway that
be shed as biofilms in human stools, therefore controls the luciferase operon and thus light
accounting for this enhanced infectivity (9). production in V.harveyi. Details are described in
For V. cholerae to produce the necessary fac- chapter 20.
tors for survival in either the aquatic habitat or Although pathogenic V. cholerae does not
the human intestine, it must be able to sense contain a luciferase operon, it is able to produce
these different environments and translate this light in a density-dependent fashion when
information into appropriate gene expression transformed with the V. harveyi lux operon (2,
patterns. Although little is known about the 43).These data suggest that not only is V.cholerae
environmental signals sensed by V. cholerae in its able to participate in quorum sensing but also
natural aquatic environment, in vitro studies that it shares many components of the quorum-
have identified a number of environmental sig- sensing system of the closely related V. harveyi.
nals that affect virulence gene expression.These Indeed, it had been shown that V. cholerae pro-
environmental cues include pH, osmolarity, duces two of the same autoinducer molecules as
temperature, bile salts, and iron availability (8, V. harveyi and the proteins involved in the
17, 28, 46, 49, 52). detection and relaying of these chemicals are
conserved in both species. Initially, the work on
THE MECHANISMS OF QUORUM V. harveyi helped elucidate the details of the
SENSING IN V. CHOLERAE quorum-sensing system in V. cholerae (Fig. 1),
In addition to the multiple environmental sig- and more recently, information gained from
nals known to regulate virulence gene expres- experiments on V.cholerae has provided a deeper
sion, V. cholerae is also able to sense its own insight into the mechanism of V. harveyi quo-
population density and respond accordingly to rum sensing.
control a number of phenotypes including viru- V. cholerae produces two known autoinduc-
lence gene expression and biofilm formation ers, CAI-1 (cholerae autoinducer 1) and AI-2
(12, 62, 63). V. cholerae achieves this by quorum (autoinducer 2) (43). AI-2 is a furanosyl borate
FIGURE 1 Current model for quorum sensing in V. cholerae.At low cell density, LuxQ, CqsA, and LuxU act as
autophosphorylating kinases that cause LuxO phosphorylation. Phosphorylated LuxO, in conjunction with 54
and Fis,induces the synthesis of the Qrr1–4 sRNAs that act with Hfq to repress HapR production.CsrA also func-
tions via an unknown component (X) to activate LuxO. At high cell density, the autoinducers AI-2 and CAI-1
(produced by LuxS and CqsA,respectively) accumulate and bind to their cognate receptors,LuxP and CqsS.LuxQ,
CqsS,and LuxU function as phosphatases,and LuxO is dephosphorylated.Dephosphorylated LuxO is inactive and
cannot repress HapR; thus, HapR is produced. CsrA is also repressed by the VarS/VarA/CsrB, C, and D sRNA
pathway and thus cannot activate LuxO.VqmA further activates HapR, and HapR functions as an autorepressor.
OM, outer membrane; IM, inner membrane; P, phosphate group; gray arrows, direction of phosphate flow; dashed
arrows, hypothetical interaction.
diester whose synthesis is dependent on the homoserine lactone synthesized by LuxLM (3).
LuxS protein.AI-2 is also produced by V. harveyi V. cholerae, however, does not contain the genes
and many other gram-negative and gram- required for synthesis or detection of AI-1,
positive bacteria and is postulated to function in although it may also produce a third unrelated
interspecies communication (42). In contrast, autoinducer molecule, as discussed later.
CAI-1 is only produced by V. cholerae and V. har- Once produced by their respective syn-
veyi and is thus believed to be responsible for thases, CAI-1 and AI-2 diffuse across the bacte-
providing species-specific cell density informa- rial cell membrane of V. cholerae and interact
tion. The structure of CAI-1 remains to be with their cognate sensors. CAI-1 binds to
determined,but its synthesis is known to require CqsS, and AI-2 binds to the periplasmic LuxP
CqsA, a protein predicted to possess amino- protein of the LuxPQ membrane complex.
transferase activity (15).V.harveyi also produces a CqsA and LuxQ are hybrid sensor/kinases that
third autoinducer, AI-1, which is an acyl- are able to switch from functioning as kinases to
phosphatases depending on their autoinducer- Phosphorylated LuxO is a transcriptional

bound state.At low cell density,when CqsA and activator and as such cannot function to directly
LuxPQ have no autoinducer bound, they act as repress hapR expression but must instead
autophosphorylating kinases that phosphory- induce the expression of a repressor of HapR.It
late the cytoplasmic phosphotransferase protein was recently discovered that this repressor in V.
LuxU, which in turn phosphorylates LuxO. cholerae comprises four small RNAs (sRNAs)
However, at high cell density when CqsS is that act with the sRNA-binding protein Hfq to
bound by CAI-1 and LuxPQ is bound by AI-2, destabilize hapR mRNA (34).The four sRNAs,
these proteins are converted into phosphatases named Qrr1–4 (quorum regulatory RNA
that cause the loss of phosphate from LuxU and 1–4), are transcribed from LuxO-dependent,
ultimately LuxO. The parallel CqsSA and 54-dependent promoters at low cell density
LuxSPQ systems therefore integrate through and in concert with Hfq bind to hapR
LuxU to regulate the levels of phosphorylated mRNA and cause the degradation of both
LuxO in response to cell density (43). themselves and the hapR mRNA. sRNAs can
LuxO is a 54-dependent transcriptional be made quickly, and their ability to stimulate
activator whose activity is dependent on its the codegradation of both themselves and their
phosphorylation status. Only phosphorylated targets provides a rapid and ultrasensitive switch
LuxO is active; therefore, at low cell density for the all-or-nothing responses required in
when CqsS and LuxQ are acting as kinases, quorum sensing. In V. cholerae it was observed
LuxO is active,and at high cell density when the that to eliminate the degradation of hapR
autoinducer-bound CqsS and LuxPQ are act- mRNA by the Hfr-Qrr repressor, complexes
ing as phosphatases,LuxO is inactive.LuxO was required the simultaneous deletion of all four
demonstrated to be a pivotal component of the Qrrs, indicating that these sRNAs act in a
quorum-sensing system of both V. cholerae and redundant manner with the expression of only
V. harveyi; its deletion completely abolishes one being sufficient for HapR repression. It has
density-dependent gene expression.The obser- been suggested that this use of multiple sRNAs
vations that deletion of LuxO results in consti- may enable additional regulatory inputs, such as
tutive light production and that the expression metabolic status, to influence the transition
of a constitutively active variant of LuxO results between the low cell density and high cell
in a dark phenotype when the V. harveyi lux density states in V. cholerae.
operon is used as a reporter suggest that phos- In addition to the two parallel CAI-1-CqsS
phorylated LuxO negatively regulates quorum- and AI-2-LuxPQ quorum-sensing systems that
sensing-activated phenotypes at low cell density act through the common LuxU-LuxO-
and its inactivation at high cell density allows Qrr/Hfq-HapR pathway,there are a number of
expression of these phenotypes (43).LuxO does additional components known to regulate the
not act directly to repress lux expression in V. level of HapR and therefore the quorum-
cholerae at low cell density but mediates its effect sensing-controlled phenotypes of V. cholerae.
through repression of the transcriptional regula- These include HapR itself, the transcriptional
tor HapR (homologous to LuxR in V. harveyi). regulator VqmA, a third phosphorelay system
HapR belongs to the TetR family of transcrip- that acts in parallel to CAI-1-CqsS and AI-2-
tional regulators and is able to function as both LuxPQ, and the small nucleoid protein Fis. At
an activator, as is the case with lux expression, high cell density when the concentration of
and a repressor. Therefore, at low cell density, HapR is elevated, HapR is able to bind to a sin-
phosphorylated LuxO is active and represses gle site in its promoter region and repress its
HapR, making it unable to induce lux expres- own expression (36). This autorepression may
sion,but at high cell density LuxO is inactivated function to prevent the runaway expression of
by dephosphorylation and thus HapR is pro- hapR at high cell density and thus enable V.
duced and stimulates light production. cholerae to respond rapidly to a drop in cell
density.VqmA is also able to bind directly to the an extracellular autoinducer or whether the
hapR promoter,but in contrast to HapR,VqmA trigger is an endogenously produced intracellu-
activates hapR transcription (37). If VqmA is lar signal.
artificially expressed at low cell density, it is able Recently the small nucleoid protein Fis was
to partially overcome the inhibition of hapR by identified as an additional component of the
LuxO and the Qrrs, probably by increasing the quorum-sensing system of V. cholerae (32). Fis is
concentration of hapR mRNA. However, able to regulate expression of a number of genes
vqmA expression is autoregulated in a cell- by binding to their promoters and either elicit-
density-dependent manner, and therefore it ing changes in DNA topology or contacting
would only be present and able to activate hapR the -subunit of RNA-polymerase. In the V.
expression at high cell density. cholerae quorum-sensing system, Fis is proposed
When the quorum-sensing pathways of V. to act by binding to the qrr promoters and caus-
cholerae were being dissected at the molecular ing DNA bending that enhances the contact
level, it was noted that the simultaneous muta- between 54 and phosphorylated LuxO. As fis
tion of both the CAI-1 and AI-2 systems did expression is only induced during early expo-
not abolish density-dependent light induction nential phase in V. cholerae, it is believed to act in
from the lux operon (43). Furthermore, concert with phosphorylated LuxO and 54 at
although mutation of LuxO eliminated all low cell density to activate production of Qrr
density-dependent regulation, strains with 1–4 and therefore contribute to the inhibition
LuxU deleted still maintained the ability to reg- of HapR production.
ulate luciferase production in a cell-density-
dependent manner.These data suggested that V. QUORUM-SENSING-CONTROLLED
cholerae contains at least one additional signal PHENOTYPES IN V. CHOLERAE
transduction pathway that acts through LuxO As previously discussed,the V.harveyi lux operon
but independently of LuxU. Recently the was instrumental in dissecting the mechanisms
VarS/VarA-CsrA/BCD system was identified, of quorum sensing in V. cholerae because it pro-
and it is thought that this pathway may form vided an easily assayed phenotype when trans-
part of the predicted third quorum-sensing- formed into this species. However, although
dependent system (33).VarS is a two-compo- some nonpathogenic V. cholerae strains do con-
nent sensor kinase that transfers phosphate to tain lux genes or the vestiges of lux genes (45), it
the VarA response regulator.VarA activates the is evident that this is not the primary quorum-
expression of three sRNAs—CsrB, CrsC, and sensing target, and a number of studies have
CrsD—which bind to and inhibit the activity therefore been carried out to elucidate what
of the sRNA-binding protein CsrA. Under genes are regulated by the quorum-sensing sys-
conditions where csrBCD expression is not tem of V. cholerae.These studies have identified a
induced by VarS/VarA, CsrA is produced and number of quorum-sensing-controlled targets,
activates LuxO in a LuxU-independent man- most notably the genes encoding the virulence
ner. It is unclear how CsrA activates LuxO as factors CT and TCP and the genes required for
it does not appear to induce luxO expression; biofilm formation (12, 62, 63). In addition, a
it has therefore been postulated that an as yet growing number of other phenotypes such as
unknown component couples signaling from protease production, the manufacture of an
VarS/VarA-CsrA/BCD to LuxO.Although the antiprotozoal factor, competence, the stress
ability of the VarS/VarA-CsrA/BCD system response, chemotaxis, and motility have also
to operate in parallel to the CAI-1-CqsS and been shown to be under V. cholerae quorum-
AI-2-LuxPQ quorum-sensing systems of V. sensing regulation (39, 40, 53, 63).
cholerae has been firmly established, it remains For a number of pathogenic bacteria, for
unclear whether this third pathway represents a example, Pseudomonas aeruginosa and Staphly-
true quorum-sensing system that responds to ococcus aureus, quorum sensing activates viru-
lence factor expression at high cell density and mutant.This surprising result suggested that vir-
as a consequence quorum-sensing-deficient ulence gene expression was in fact repressed at
mutants exhibit reduced virulence in mam- high cell density by the quorum-sensing system
malian infection models.To assess whether this of V. cholerae, and thus the luxO mutant, which
was also the case for V. cholerae, Zhu and col- constitutively expresses hapR and is therefore
leagues investigated what effect mutating hapR locked in a high cell density state, is unable to
had on the virulence of V. cholerae in the infant produce virulence factors and colonize mice.
mouse model of infection (63). As explained HapR is capable of repressing virulence
above, hapR expression is repressed at low cell gene expression at high cell density through the
density and activated at high cell density; thus, a indirect repression of the AraC-like transcrip-
hapR mutant is locked in a low cell density state. tion activator,ToxT (Fig. 2).ToxT is the central
If HapR was required to activate virulence fac- regulator of virulence gene expression in V.
tor expression at high cell density, it follows that cholerae that coordinately activates the tran-
the hapR mutant would be attenuated for viru- scription of a number of genes, including those
lence. However, it was shown that the hapR required for CT and TCP biosynthesis and the
mutant was able to colonize infant mice to the acf genes (5, 47).The expression of toxT is regu-
same extent as wild-type V. cholerae. In contrast, lated in part by two membrane complexes,
the luxO mutant of V. cholerae was profoundly ToxRS and TcpPH, which are believed to sense
defective in its ability to colonize the small environmental cues and modulate toxT expres-
intestine of infant mice. Furthermore, microar- sion via their transcriptional activator compo-
ray experiments revealed that a number of nents, ToxR and TcpP, respectively (13). In
known virulence genes, including those addition to activating the genes required for
encoded by the ctx, tcp, and acf operons, were CT and TCP synthesis, ToxT is also able to
repressed between 2.7-and 45-fold in the luxO upregulate its own expression and that of the
FIGURE 2 Repression of virulence factors by HapR. Under conditions that are conducive for virulence factor
expression,TcpPH and ToxRS activate the expression of toxT, which in turn leads to expression of genes required
for the synthesis of cholera toxin and toxin coregulated pili.When HapR is produced at high cell density,it represses
the transcription of aphA.The repression of AphA leads to the downregulation of TcpPH,ToxT, and subsequently
virulence factor expression.TCP, toxin coregulated pili; CT, cholera toxin.
TcpPH membrane complex, resulting in a pos- regulates quorum-sensing-controlled pheno-

itive feedback mechanism. Many further layers types through LuxO but independently of
of complexity impinge on these basic signal LuxU. Cross-feeding experiments revealed that
transduction cascades to influence virulence virulence factor repression by HapR is indeed a
gene expression; for example, tcpPH expression quorum-sensing-controlled mechanism requir-
is also induced by the transcriptional activators ing the presence of extracellular autoinducers.
AphA and AphB (27). It is this regulatory path- When the cell-free culture fluid from wild-type
way, in which AphA activates tcpPH expression V. cholerae was added to the cqsAluxS double
and TcpPH induces expression of toxT, that is mutant that is unable to synthesize either CAI-
negatively regulated by quorum sensing in V. 1 or AI-2, tcpP expression was significantly
cholerae (26).At high cell density, unphosphory- reduced and no TCP or CT expression was evi-
lated LuxO is inactive and therefore cannot dent. This effect was abolished when the cell-
repress hapR expression. HapR binds to the free culture supernatant was supplied by a
promoter of the transcriptional activator AphA cqsAluxS double mutant or the recipient strain
and represses its transcription. As AphA is contained a hapR mutation. It was also shown
required for tcpPH expression and TcpPH acti- that although either autoinducer is sufficient to
vates the global virulence gene regulator ToxT, partially repress tcpP expression, a more than
repressing AphA results in the repression of vir- additive repressive effect occurs when both
ulence gene expression. At low cell density, CAI-1 and AI-2 are present, indicating that
phosphorylated LuxO downregulates HapR these molecules function synergistically.
through the action of Hfq and the sRNAs, In addition to repressing virulence gene
Qrr1–4. The repression of aphA by HapR is expression, the quorum-sensing system of V.
therefore lifted and AphA can bind to the pro- cholerae also inhibits biofilm formation at high
moter of tcpPH, recruit AphB, and activate cell density (12, 63), a phenomenon that again
tcpPH expression.TcpPH, presumably in com- contrasts with other bacteria that use quorum
bination with the appropriate environmental sensing to activate biofilm formation at high
signals, then induces toxT expression, which in cell density. The ability of V. cholerae to form
turn activates transcription of the genes biofilms is achieved by a multistep develop-
required for CT and TCP production. mental process that is controlled by a number of
To investigate the requirement of compo- interacting regulatory pathways (57). The for-
nents upstream of LuxO in the quorum- mation of these surface-attached microbial
sensing-mediated virulence repression in V. communities is initiated by a bacterium
cholerae, infant mouse colonization assays and approaching a surface, reducing its motility, and
TCP and CT production assays were carried forming a transient association with the surface.
out with a number of different quorum-sensing The bacterium then multiplies and produces
mutants (43). These experiments revealed that exopolysaccharide to form a microcolony.
although mutation of luxO abolished TCP and Mature biofilm structures, characterized by pil-
CT production and reduced the colonization lars and channels, are subsequently generated
efficiency of V. cholerae, when components of after additional bacterial growth and continued
the CAI-1-CqsS or AI-2-LuxPQ quorum- exopolysaccharide production.Thus, central to
sensing systems were mutated, either separately the development of a biofilm structure is the
or together, the resulting mutants were still production of the extracellular polymeric
wild type for virulence and capable of synthe- matrix, and in V. cholerae, this substance is called
sizing TCP and CT. Furthermore, mutation of Vibrio polysaccharide (VPS). VPS synthesis
luxU did not significantly affect intestinal colo- requires the concerted action of 17 proteins
nization or production of TCP or CT. These that are encoded by two unlinked vps operons,
results provide further evidence for the third vpsA-K and vpsL-Q, located on the large chro-
sensory system, VarS/VarA-CsrA/BCD, that mosome of V. cholerae (59, 61).
It is important that vps gene expression is concentrations because CytR, which regulates
tightly regulated in V. cholerae to ensure biofilm nucleoside catabolism genes in response to cyti-
formation only occurs when conditions are dine concentrations in Escherichia coli, represses
favorable for this mode of living.The regulatory vps gene expression in V. cholerae (14).
networks that achieve this control are starting to The first indication that quorum sensing was
be unraveled, and it is evident that a number of also involved in the regulation of biofilm for-
different extracellular and intracellular cues are mation in V. cholerae arose from the observation
integrated by complex interacting pathways to that a hapR mutant was enhanced in its ability
control VPS production. Two positive regula- to form a biofilm, whereas a luxO mutant was
tors,VpsR and VpsT, have been identified that deficient in its ability to form a biofilm (12,63).
are required for vps gene expression (6, 59). In When HapR is not expressed, either in the
addition to activating vps gene expression,VpsR hapR mutant or in the constitutively active
and VpsT also positively regulate their own luxO mutant, biofilms are generated that are
expression and that of each other,creating com- thicker and denser and contain more extracel-
plex feedback loops (6). VpsR and VpsT are lular material (12, 62). In contrast, when HapR
homologous to two-component response reg- is overexpressed by the luxO mutant, significant
ulators and are therefore thought to sense and biofilm structures fail to form. Microarray
respond to environmental stimuli.Although no analysis demonstrated that the enhanced
external signal has been identified for VpsT, it biofilm formation by the hapR mutant resulted
has recently been shown that bile acids stimu- from a 2.3- to 8.1-fold overexpression of the
late vps production and biofilm formation via vps genes.Also, oligo-based S1 nuclease protec-
activation ofVpsR (18).Another environmental tion assays showed that while vps is unde-
signal that may regulate V. cholerae biofilm for- tectable in wild-type planktonic cells and
mation through an as yet unknown pathway is produced by wild-type biofilms primarily at
the presence of a surface. It appears that V. earlier time points, vps is strongly expressed in
cholerae can monitor flagellar torque to sense both hapR planktonic cells and hapR biofilms
when a surface is encountered as mutations in even at late time points. Taken together, these
genes required for the synthesis of flagellar or a results illustrate that HapR acts to repress
sodium-driven motor affect both vps expression biofilm formation at high cell density by
and biofilm formation (30, 58). In addition to repressing vps gene expression. To determine
responding to external cues, V. cholerae also the contribution of CAI-1 and AI-2 to the reg-
senses internal conditions such as the concen- ulation of biofilm formation in V. cholerae, vps
tration of cyclic diguanylate (c-diGMP). c- expression and biofilm formation by the cqsA
diGMP is an intracellular signaling molecule mutant and the luxS mutant were examined.
that can modulate the cell surface properties of The luxS mutant produced normal biofilms,
several bacterial species, including VPS produc- whereas the cqsA mutant produced thicker
tion and biofilm formation in V. cholerae (51). biofilms and higher levels of vps expression,
The concentration of c-diGMP in a cell is suggesting that while AI-2 is largely dispensa-
oppositely regulated by proteins that contain a ble, the CAI-1 autoinducer is important for
GGDEF domain and function as diguanylate regulating biofilms. Furthermore, when the
cyclases and proteins that contain an EAL cqsA mutant was grown in the presence of CAI-
domain and function as phosphodiesterases. V. 1, the resulting biofilms were wild type in
cholerae is predicted to contain 53 genes that appearance, confirming that CAI-1 controls
encode proteins with either a GGDEF or EAL biofilm formation in V. cholerae.
domain or both,and a number of these enzymes The negative regulation of vps expression
have been shown to either activate or repress and subsequent biofilm formation by HapR
biofilm formation (11, 35). Biofilm formation may either result from direct repression of the
may also be regulated by intracellular cytidine two vps operons or indirect repression through
one of the regulators of vps expression, such as by modulating the concentration of c-diGMP
VpsR,VpsT, or the GGDEF and EAL domain in the cell. HapR represses the expression of
proteins (Fig. 3). Although a number of studies cdgA, which is predicted to encode a GGDEF
have suggested that vpsR expression is not con- domain protein that may act as a diguanylate
trolled by HapR (12, 62), a microarray experi- cyclase (4). Repression of a diguanylate cyclase
ment utilizing a different V. cholerae strain by HapR would result in a decrease in c-
demonstrates that, at least in some strains, either diGMP concentration and subsequent repres-
the transcription or message abundance of vpsR sion of vps gene expression at high cell density.It
and vpsT is negatively regulated by HapR (60). has also been suggested that AphA, the protein
Furthermore, hapR itself appears to be nega- that links HapR with virulence gene regulation
tively regulated by both VpsT and VpsR in this in V. cholerae, is involved in the regulation of c-
strain, creating a complex feedback mechanism diGMP concentrations by HapR (24). AphA
(4). In addition to repressing VpsT and VpsR, regulates the expression of acgA and acgB, which
HapR may also regulate the abundance of encode proteins that contain an EAL domain
diguanylate cyclases and phosphodiesterases and and a GGDEF domain, respectively. Overex-
therefore indirectly control vps gene expression pression of the predicted phosphodiesterase,
FIGURE 3 Repression of biofilms by HapR.At high cell density, HapR represses Vibrio poly-
saccharide expression and therefore biofilm formation.This repression may be via direct repres-
sion of the vps genes, by repression of the positive regulators VpsR or VpsT, or by modulating the
level of c-diGMP in the cell. High levels of c-diGMP activate Vibrio polysaccharide expression,
and HapR may function to reduce the c-diGMP concentration in the cell by activating EAL
domain containing phosphodiesterases, for example, AcgA, or by repressing GGDEF domain
containing diguanylate cyclases, for example, CdgA. The interactions between HapR and the
other proteins are not necessarily direct.VPS, Vibrio polysaccharide.
AcgA, reduces biofilm formation and overex- directly upregulate transcription of the global
pression of AcgB,which is predicted to function stress response sigma regulatory factor, RpoS
as a diguanylate cyclase, and enhances biofilm (unpublished results). The involvement of
formation. Therefore, as HapR represses aphA quorum sensing in the chemotaxis and motility
expression, it may indirectly lead to a decrease of V. cholerae was indicated by microarray data
in c-diGMP levels through the regulation of showing that the expression levels of several
AcgA and AcgB by AphA. genes involved in these processes are altered
The mechanisms of quorum-sensing con- in the luxO mutant compared to the wild-type
trol of biofilm formation in V. cholerae is further strain (63). Further experiments revealed that
complicated by a recent finding that the con- the luxO mutant was indeed less motile than
centration of the autoinducer CAI-1 is higher either the hapR mutant or wild-type V. cholerae.
in biofilms than in planktonic cultures (38). The ability of HapR to repress aphA expression
Although expression of cqsA, the CAI-1 syn- and subsequently tcpPH expression indicates
thase gene, is equivalent in bacteria found in that any genes regulated by either of these two
either biofilms or planktonic cultures,it is likely factors may also be quorum sensing regulated.
that the biofilm matrix restricts diffusion TcpPH is known to control genes involved
of the autoinducer and therefore increases its in metabolism and nutrient uptake, and
local concentration. The higher concentration AphA has been shown to repress both the
of CAI-1 results in an earlier induction of penicillin amidase gene and genes required for
quorum-sensing in biofilm-associated cells,and acetoin biosynthesis; therefore the spectrum
this timing is essential for modulating biofilm of quorum-sensing-controlled genes in V.
thickness and associated phenotypes.Therefore, cholerae may be much broader than currently
V. cholerae appears capable of regulating its appreciated (24, 25).
biofilm architecture by temporal induction of
the quorum-sensing system. CONTRIBUTION OF QUORUM
In addition to modulating the expression of SENSING TO THE LIFE CYCLE
virulence factors and biofilm formation, the OF V. CHOLERAE
quorum-sensing system of V. cholerae also con- The generally accepted paradigm is that the
trols a number of other phenotypes, including phenotypes controlled by quorum sensing are
protease production, the manufacture of an those that are only beneficial when produced
antiprotozoal factor, competence, the stress coordinately by a population of bacteria rather
response, chemotaxis, and motility. HapR was than by an individual bacterium. For example,
originally identified as a regulator of the HA quorum sensing has been shown to control bio-
protease encoded by hapA, thus explaining the luminescence,virulence,biofilm formation,and
origins of its name (19).HapR directly activates natural competence in a variety of species, and
hapA expression, enabling quorum sensing to these processes are normally only useful when
positively regulate protease production at high expressed by bacteria at high cell density. In the
cell density (63). Furthermore, additional case of V. cholerae, however, it has been demon-
secreted proteases may be negatively regulated strated that quorum sensing acts in the opposite
by LuxO as the protease levels in cell-free cul- manner to repress virulence factors and biofilm
ture fluids from the luxOhapR double mutant formation at high cell density. To explain this
are higher than those from the hapR single apparent paradox, it is important to refer to the
mutant. HapR also positively regulates the pro- life cycle of this pathogen and examine the con-
duction and secretion of an antiprotozoal factor tribution quorum sensing makes to its survival
in biofilms at high cell density and is required in the different habitats it encounters.
for chitin-induced natural competence in V. In its natural aquatic environment, the abil-
cholerae (39, 40). Recent experiments from our ity of V.cholerae to exist in biofilms may enhance
laboratory also show that HapR is able to its fitness by providing a protective barrier
against both physical-chemical and biological tive regulation of biofilm formation by quorum
insults. In a number of cases the presence of a sensing in V. cholerae may be important in pre-
functional quorum-sensing system has also venting the formation of biofilms that are overly
been implicated in the increased environmental thick and thus inhibit bacterial detachment in
fitness of V. cholerae biofilms. For example, bac- the host. This hypothesis is supported by the
teria in wild-type V. cholerae biofilms survive observation that while the hapR mutant exhibits
significantly better than bacteria in hapR a similar colonization efficiency compared to
mutant biofilms when exposed to seawater wild-type V. cholerae, when planktonic cells are
(20). In addition, the ability of HapR to induce used as the inoculum in the infant mouse model
expression and secretion of an antiprotozoal of infection, hapR mutant biofilms show a 10-
factor in biofilms at high cell density enables V. fold colonization defect compared to wild-
cholerae to survive protozoan grazing in its natu- type biofilms (62). This result suggests that
ral habitat (39).Quorum sensing could also play quorum sensing may affect intestinal coloniza-
a role in environmental survival through its tion by a mechanism that involves a HapR-
control of chitin-induced natural competence, dependent phenotype expressed in biofilms,
as the induction of horizontal gene transfer may such as detachment.
enable the acquisition of genes that increase the Detachment of V.cholerae from biofilm struc-
survival potential of V. cholerae (40). tures may result in a decrease in cell density that
As previously discussed, V. cholerae biofilms removes the quorum-sensing repression of vir-
may represent the appropriate bacterial inocu- ulence genes. The subsequent production of
lum required to initiate human infection.These TCP would then enable colonization of the
densely packed, acid-resistant structures proba- small intestine, and CT production would
bly enable a high enough dose of bacteria to induce diarrhea. Although initial colonization
enter the host,traverse the hostile stomach envi- of the intestinal epithelial surface requires
ronment,and reach the upper intestine for colo- expression of these virulence factors, the subse-
nization to occur (62). However, if V. cholerae quent bacterial growth at this site is likely to
does enter the host in a high cell-density result in an increase in autoinducer concentra-
biofilm, the virulence genes required for colo- tion and quorum-sensing-mediated inhibition
nization will be repressed by HapR.Therefore, of virulence factor production. Indeed, tran-
once in the small intestine it is important that scriptome analysis of V. cholerae isolated from
individual cells disperse from the biofilm struc- patient stool samples failed to detect ctx and tcp
ture in order to relieve the quorum-sensing- transcripts, indicating that these genes are not
mediated repression of TCP, CT, and other expressed during the later stages of infection
virulence factors required for colonization and (29). However, although HapR represses viru-
disease. It is unclear what causes the bacteria to lence gene expression, it induces the expression
disperse from biofilms in the small intestine— of the hemagglutinin protease, HapA, at high
whether it is an unknown host signal or the cell density. As HapA is proposed to act
consequence of the natural dynamic equilib- as a detachase during V. cholerae infection, its
rium of bacterial cells entering and exiting quorum-sensing-regulated induction at later
biofilm structures, or a combination of these stages of infection may permit individual cells
two mechanisms. However, it is likely that the to detach from the epithelium and either estab-
ability of quorum sensing to negatively regulate lish new infection foci in the intestine or exit
biofilm formation in V. cholerae contributes to the host. Although the above model explains
this detachment. In addition to producing a how quorum sensing may act at a number of
thicker biofilm, in vitro experiments show that stages to enhance V. cholerae survival in both the
hapR quorum-sensing-deficient mutants detach aquatic environment and human host and to
from these biofilms to a much lower extent than facilitate the transition between these two habi-
wild-type V. cholerae (62). Therefore, the nega- tats, it is interesting to note that not all virulent
V. cholerae strains have a functional quorum- of virulence factors and biofilms is influenced
sensing system.Indeed,analysis of 16 geograph- to a greater extent by the concentration of
ically diverse V. cholerae strains revealed that 50% CAI-1 than AI-2 (62), it is of critical impor-
contained a mutated hapR sequence (20). For tance to characterize the chemical properties of
example, the V. cholerae sequence strain N16961 this autoinducer and determine its exact role in
has a naturally occurring frameshift mutation in regulating virulence.Furthermore,since CAI-1
hapR that eliminates the quorum-sensing is able to negatively regulate virulence genes in
input. N16961 is, however, fully virulent and V. cholerae, this autoinducer may represent a
biofilm proficient, indicating that quorum potential therapeutic molecule, and to fulfill
sensing is not an absolute requirement for V. this role, its structure must be known.
cholerae survival and infection. The retention From previous work it is evident that quo-
of a functional quorum-sensing system by some rum sensing in V. cholerae is very complex,
V. cholerae strains and its mutation in others sug- with a number of different pathways converg-
gest that there are certain niches where quorum ing to control the expression of HapR, which
sensing is advantageous and others where it is in turn regulates a number of different pheno-
detrimental to survival. For example, although types. Within this network there are some
quorum sensing conferred a survival advantage components that remain to be identified, for
on V. cholerae in biofilms when exposed to example, the hypothetical third autoinducer
seawater, planktonic cells appeared to survive that may act through the VarS/VarA-
better when hapR was mutated. Therefore, CsrA/BCD pathway to regulate LuxO activity
the apparently high occurrence of quorum- and the molecule that links CsrA and LuxO
sensing-deficient V. cholerae strains may result (33). Also, although there have been a number
from the ability of these bacteria to survive bet- of suggestions as to how HapR negatively reg-
ter as planktonic cells in the environment. ulates vps expression and thus biofilm forma-
tion (4, 24, 60), the exact mechanism of this
PERSPECTIVES AND repression remains to be elucidated. Recent
FUTURE STUDIES work in our laboratory has shown that quorum
In the last five years, a great deal has been dis- sensing is further controlled in V. cholerae by the
covered about quorum sensing in the human presence of a secretion-competent flagellum
pathogen V. cholerae. Many of the details of the (unpublished results), adding yet a further layer
quorum-sensing pathway have been elucidated of complexity to this highly regulated system.
at the molecular level, and it has been demon- In addition to these mechanistic questions,
strated that V. cholerae utilizes this method of much still remains to be discovered about the
gene regulation to control the expression of a relevance of quorum sensing to V. cholerae in its
number of different phenotypes. In contrast to natural environments.As previously mentioned,
many other bacteria studied to date, V. cholerae several toxigenic V. cholerae strains lack a func-
uses quorum sensing to repress virulence factor tional quorum-sensing system, which indicates
expression and biofilm formation at high cell that in certain environments quorum-sensing
density, and it has been speculated that this reg- mutants have a survival advantage (20). How-
ulation contributes to the life cycle of this ever, the retention of this complex regulatory
pathogen. network in other strains of V. cholerae suggests
Despite these recent advances in the under- that in particular situations quorum sensing is
standing of quorum sensing in V. cholerae, an advantageous phenotype.To assess the signif-
numerous unanswered questions remain. For icance of quorum sensing, it is important to
example, although the structure of the autoin- carry out experiments under conditions that
ducer AI-2 is known,the chemical nature of the mimic as closely as possible the natural habitat
other autoinducer CAI-1 still remains to be of V. cholerae. For example, most of the work on
determined.As the quorum-sensing regulation quorum sensing in V. cholerae has been carried
out in vitro,and it remains unclear how quorum 4. Beyhan, S., K. Bilecen, S. R. Salama, C.
signals and other environmental cues are Casper-Lindley, and F. H.Yildiz. 2007. Regu-
lation of rugosity and biofilm formation in Vibrio
integrated to regulate phenotypes such as viru- cholerae: comparison of VpsT and VpsR regulons
lence and biofilm formation in vivo. Recombi- and epistasis analysis of vpsT, vpsR, and hapR. J.
nation-based in vivo expression technology has Bacteriol. 189:388–402.
demonstrated that the temporal expression pat- 5. Bina, J., J. Zhu, M. Dziejman, S. Faruque, S.
terns of critical V. cholerae virulence genes in Calderwood, and J. Mekalanos. 2003. ToxR
regulon of Vibrio cholerae and its expression in vib-
vivo are substantially different from those rios shed by cholera patients. Proc. Natl. Acad. Sci.
exhibited in vitro (31), and therefore, it is essen- USA 100:2801–2806.
tial to examine the temporal profiles of quorum 6. Casper-Lindley, C., and F. H. Yildiz. 2004.
sensing and virulence gene regulation in the VpsT is a transcriptional regulator required for
context of a true infection. Furthermore, the expression of vps biosynthesis genes and the devel-
opment of rugose colonial morphology in Vibrio
contribution that quorum sensing makes to cholerae O1 El Tor. J. Bacteriol. 186:1574–1578.
the survival of V. cholerae in its marine environ- 7. Colwell, R. R., A. Huq, M. S. Islam, K. M.
ment must be further studied to appreciate Aziz, M.Yunus, N. H. Khan, A. Mahmud, R.
the advantages that quorum sensing confers B. Sack, G. B. Nair, J. Chakraborty, D.A. Sack,
on a bacterium that usually lives outside its and E. Russek-Cohen. 2003. Reduction of
cholera in Bangladeshi villages by simple filtration.
human host. In an attempt to address this Proc. Natl.Acad. Sci. USA 100:1051–1055.
question, our laboratory has recently found 8. DiRita,V. J. 1992. Co-ordinate expression of vir-
that HapR upregulates RpoS, which enhances ulence genes by ToxR in Vibrio cholerae. Mol. Micro-
V. cholerae viability under oxidative and nutri- biol. 6:451–458.
tional stress conditions and thus may facilitate 9. Faruque, S. M., K. Biswas, S. M. Udden, Q. S.
Ahmad, D. A. Sack, G. B. Nair, and J. J.
survival in the aquatic environment (unpub- Mekalanos. 2006. Transmissibility of cholera: in
lished results). vivo-formed biofilms and their relationship to
In conclusion, additional study of quorum infectivity and persistence in the environment.
sensing in V. cholerae is required to facilitate a Proc. Natl.Acad. Sci. USA 103:6350–6355.
deeper understanding of the natural role that 10. Finkelstein, R. A., M. Boesman-Finkelstein,
Y. Chang, and C. C. Hase. 1992. Vibrio cholerae
quorum sensing plays during the various phases hemagglutinin/protease, colonial variation, viru-
of the life cycle of this pathogen. It is our hope lence, and detachment. Infect. Immun. 60:472–478.
that further insight into the interplay that exists 11. Galperin, M. Y., A. N. Nikolskaya, and E. V.
between quorum sensing, biofilm formation, Koonin. 2001. Novel domains of the prokaryotic
infectivity, and pathogenesis will contribute to two-component signal transduction systems.
FEMS Microbiol. Lett. 203:11–21.
the discovery of novel preventions and treat- 12. Hammer, B. K., and B. L. Bassler. 2003. Quo-
ments for cholera. rum sensing controls biofilm formation in Vibrio
cholerae. Mol. Microbiol. 50:101–104.
REFERENCES 13. Hase, C. C., and J. J. Mekalanos. 1998. TcpP
1. Attridge, S. R., E. Voss, and P. A. Manning. protein is a positive regulator of virulence gene
1993. The role of toxin-coregulated pili in the expression in Vibrio cholerae. Proc. Natl. Acad. Sci.
pathogenesis of Vibrio cholerae O1 El Tor. Microb. USA 95:730–734.
Pathog. 15:421–431. 14. Haugo, A. J., and P. I. Watnick. 2002. Vibrio
2. Bassler, B. L., E. P. Greenberg, and A. M. cholerae CytR is a repressor of biofilm develop-
Stevens. 1997.Cross-species induction of lumines- ment. Mol. Microbiol. 45:471–483.
cence in the quorum-sensing bacterium Vibrio 15. Henke, J. M., and B. L. Bassler. 2004. Three
harveyi. J. Bacteriol. 179:4043–4045. parallel quorum-sensing systems regulate gene
3. Bassler, B. L., M.Wright, R. E. Showalter, and expression in Vibrio harveyi. J. Bacteriol. 186:6902–
M. R. Silverman. 1993. Intercellular signalling 6914.
in Vibrio harveyi: sequence and function of genes 16. Herrington, D.A., R. H. Hall, G. Losonsky, J.
regulating expression of luminescence. Mol. J. Mekalanos, R. K.Taylor, and M. M. Levine.
Microbiol. 9:773–786. 1988.Toxin, toxin-coregulated pili, and the toxR
regulon are essential for Vibrio cholerae pathogenesis environmental signals in Vibrio cholerae. Curr. Opin.
in humans. J. Exp. Med. 168:1487–1492. Microbiol. 6:186–190.
17. Hung, D. T., and J. J. Mekalanos. 2005. Bile 29. Larocque, R. C., J. B. Harris, M. Dziejman,
acids induce cholera toxin expression in Vibrio X. Li, A. I. Khan, A. S. Faruque, S. M.
cholerae in a ToxT-independent manner. Proc. Natl. Faruque, G. B. Nair, E.T. Ryan, F. Qadri, J. J.
Acad. Sci. USA 102:3028–3033. Mekalanos, and S. B. Calderwood. 2005.Tran-
18. Hung, D. T., J. Zhu, D. Sturtevant, and J. J. scriptional profiling of Vibrio cholerae recovered
Mekalanos. 2006. Bile acids stimulate biofilm directly from patient specimens during early and
formation in Vibrio cholerae. Mol. Microbiol. late stages of human infection. Infect. Immun.
59:193–201. 73:4488–4493.
19. Jobling, M. G., and R. K. Holmes. 1997. Char- 30. Lauriano, C. M., C. Ghosh, N. E. Correa, and
acterization of hapR, a positive regulator of the K. E. Klose. 2004. The sodium-driven flagellar
Vibrio cholerae HA/protease gene hap, and its motor controls exopolysaccharide expression in
identification as a functional homologue of the Vibrio cholerae. J. Bacteriol. 186:4864–4874.
Vibrio harveyi luxR gene. Mol. Microbiol. 26: 31. Lee, S. H., D. L. Hava, M. K. Waldor, and A.
1023–1034. Camilli. 1999. Regulation and temporal expres-
20. Joelsson, A., Z. Liu, and J. Zhu. 2006. Genetic sion patterns of Vibrio cholerae virulence genes dur-
and phenotypic diversity of quorum-sensing sys- ing infection. Cell 99:625–634.
tems in clinical and environmental isolates of Vibrio 32. Lenz, D. H., and B. L. Bassler. 2007.The small
cholerae. Infect. Immun. 74:1141–1147. nucleoid protein Fis is involved in Vibrio cholerae
21. Kaper, J. B., J. G. Morris, Jr., and M. M. quorum sensing. Mol. Microbiol. 63:859–871.
Levine. 1995. Cholera. Clin. Microbiol. Rev. 33. Lenz, D. H., M. B. Miller, J. Zhu, R.V. Kulka-
8:48–86. rni, and B. L. Bassler. 2005. CsrA and three
22. Karaolis, D. K., J. A. Johnson, C. C. Bailey, E. redundant small RNAs regulate quorum sensing
C. Boedeker, J. B. Kaper, and P. R. Reeves. in Vibrio cholerae. Mol. Microbiol. 58:1186–1202.
1998.A Vibrio cholerae pathogenicity island associ- 34. Lenz, D. H., K. C. Mok, B. N. Lilley, R. V.
ated with epidemic and pandemic strains. Proc. Kulkarni, N. S. Wingreen, and B. L. Bassler.
Natl.Acad. Sci. USA 95:3134–3139. 2004.The small RNA chaperone Hfq and multiple
23. Karaolis, D. K., S. Somara, D. R. Maneval, Jr., small RNAs control quorum sensing in Vibrio har-
J.A. Johnson, and J. B. Kaper. 1999.A bacterio- veyi and Vibrio cholerae. Cell 118:69–82.
phage encoding a pathogenicity island, a type-IV 35. Lim, B., S. Beyhan, J. Meir, and F. H.Yildiz.
pilus and a phage receptor in cholera bacteria. 2006. Cyclic-diGMP signal transduction systems
Nature 399:375–379. in Vibrio cholerae: modulation of rugosity and
24. Kovacikova, G., W. Lin, and K. Skorupski. biofilm formation. Mol. Microbiol. 60:331–348.
2005. Dual regulation of genes involved in acetoin 36. Lin, W., G. Kovacikova, and K. Skorupski.
biosynthesis and motility/biofilm formation by 2005. Requirements for Vibrio cholerae HapR
the virulence activator AphA and the acetate- binding and transcriptional repression at the hapR
responsive LysR-type regulator AlsR in Vibrio promoter are distinct from those at the aphA pro-
cholerae. Mol. Microbiol. 57:420–433. moter. J. Bacteriol. 187:3013–3019.
25. Kovacikova, G., W. Lin, and K. Skorupski. 37. Liu, Z., A. Hsiao, A. Joelsson, and J. Zhu.
2003.The virulence activator AphA links quorum 2006. The transcriptional regulator VqmA
sensing to pathogenesis and physiology in Vibrio increases expression of the quorum-sensing
cholerae by repressing the expression of a penicillin activator HapR in Vibrio cholerae. J. Bacteriol. 188:
amidase gene on the small chromosome.J.Bacteriol. 2446–2453.
185:4825–4836. 38. Liu, Z., F. R. Stirling, and J. Zhu. 2007.Tempo-
26. Kovacikova, G., and K. Skorupski. 2002. Reg- ral quorum-sensing induction regulates Vibrio
ulation of virulence gene expression in Vibrio cholerae biofilm architecture. Infect. Immun.
cholerae by quorum sensing: HapR functions at the 75:122–126.
aphA promoter. Mol. Microbiol. 46:1135–1147. 39. Matz, C., D. McDougald, A. M. Moreno, P.Y.
27. Kovacikova, G., and K. Skorupski. 1999.A Vib- Yung, F. H. Yildiz, and S. Kjelleberg. 2005.
rio cholerae LysR homolog, AphB, cooperates with Biofilm formation and phenotypic variation
AphA at the tcpPH promoter to activate expres- enhance predation-driven persistence of Vibrio
sion of the ToxR virulence cascade. J. Bacteriol. cholerae. Proc. Natl. Acad. Sci. USA 102:16819–
181:4250–4256. 16824.
28. Krukonis, E. S., and V. J. DiRita. 2003. From 40. Meibom, K. L., M. Blokesch, N.A. Dolganov,
motility to virulence: sensing and responding to C. Y. Wu, and G. K. Schoolnik. 2005. Chitin
induces natural competence in Vibrio cholerae. Sci- 53. Vance, R. E., J. Zhu, and J. J. Mekalanos. 2003.
ence 310:1824–1827. A constitutively active variant of the quorum-
41. Merrell, D. S., S. M. Butler, F. Qadri, N. A. sensing regulator LuxO affects protease produc-
Dolganov, A. Alam, M. B. Cohen, S. B. tion and biofilm formation in Vibrio cholerae. Infect.
Calderwood, G. K. Schoolnik, and A. Immun. 71:2571–2576.
Camilli. 2002. Host-induced epidemic spread of 54. Wai, S. N.,Y. Mizunoe, A.Takade, S. I. Kawa-
the cholera bacterium. Nature 417:642–645. bata, and S. I.Yoshida. 1998. Vibrio cholerae O1
42. Miller, M. B., and B. L. Bassler. 2001. Quorum strain TSI-4 produces the exopolysaccharide
sensing in bacteria. Annu. Rev. Microbiol. 55:165– materials that determine colony morphology,stress
199. resistance, and biofilm formation. Appl. Environ.
43. Miller, M. B., K. Skorupski, D. H. Lenz, R. K. Microbiol. 64:3648–3655.
Taylor, and B. L. Bassler. 2002. Parallel quorum 55. Waldor, M. K., and J. J. Mekalanos. 1996.Lyso-
sensing systems converge to regulate virulence in genic conversion by a filamentous phage encoding
Vibrio cholerae. Cell 110:303–314. cholera toxin. Science 272:1910–1914.
44. Nielsen,A.T., N.A. Dolganov, G. Otto, M. C. 56. Waters, C. M., and B. L. Bassler. 2005. Quo-
Miller, C. Y. Wu, and G. K. Schoolnik. 2006. rum sensing: cell-to-cell communication in bacte-
RpoS controls the Vibrio cholerae mucosal escape ria. Annu. Rev. Cell. Dev. Biol. 21:319–346.
response. PLoS Pathog 2:e109. 57. Watnick, P. I., and R. Kolter. 1999. Steps in the
45. Palmer, L. M., and R. R. Colwell. 1991.Detec- development of a Vibrio cholerae El Tor biofilm.
tion of luciferase gene sequence in nonlumines- Mol. Microbiol. 34:586–595.
cent Vibrio cholerae by colony hybridization and 58. Watnick, P. I., C. M. Lauriano, K. E. Klose, L.
polymerase chain reaction.Appl.Environ.Microbiol. Croal, and R. Kolter. 2001.The absence of a fla-
57:1286–1293. gellum leads to altered colony morphology,
46. Parsot, C., and J. J. Mekalanos. 1990. Expres- biofilm development and virulence in Vibrio
sion of ToxR, the transcriptional activator of the cholerae O139. Mol. Microbiol. 39:223–235.
virulence factors in Vibrio cholerae, is modulated by 59. Yildiz, F. H., N. A. Dolganov, and G. K.
the heat shock response. Proc. Natl.Acad. Sci. USA Schoolnik. 2001.VpsR,a member of the response
87:9898–9902. regulators of the two-component regulatory
47. Peterson, K. M., and J. J. Mekalanos. 1988. systems, is required for expression of vps biosyn-
Characterization of the Vibrio cholerae ToxR thesis genes and EPS(ETr)-associated phenotypes
regulon: identification of novel genes involved in in Vibrio cholerae O1 El Tor. J. Bacteriol. 183:
intestinal colonization. Infect. Immun. 56:2822– 1716–1726.
2829. 60. Yildiz, F. H., X. S. Liu,A. Heydorn, and G. K.
48. Sack, D. A., R. B. Sack, G. B. Nair, and A. K. Schoolnik. 2004. Molecular analysis of rugosity
Siddique. 2004. Cholera. Lancet 363:223–233. in a Vibrio cholerae O1 El Tor phase variant. Mol.
49. Skorupski, K., and R. K.Taylor. 1997. Control Microbiol. 53:497–515.
of the ToxR virulence regulon in Vibrio cholerae by 61. Yildiz, F. H., and G. K. Schoolnik. 1999. Vibrio
environmental stimuli. Mol. Microbiol. 25:1003– cholerae O1 El Tor: identification of a gene cluster
1009. required for the rugose colony type, exopolysac-
50. Taylor, R. K.,V. L. Miller, D. B. Furlong, and J. charide production, chlorine resistance, and
J. Mekalanos. 1987. Use of phoA gene fusions to biofilm formation. Proc. Natl. Acad. Sci. USA
identify a pilus colonization factor coordinately 96:4028–4033.
regulated with cholera toxin. Proc. Natl. Acad. Sci. 62. Zhu, J., and J. J. Mekalanos. 2003. Quorum
USA 84:2833–2837. sensing-dependent biofilms enhance colonization
51. Tischler, A. D., and A. Camilli. 2004. Cyclic in Vibrio cholerae. Dev. Cell 5:647–656.
diguanylate (c-di-GMP) regulates Vibrio cholerae 63. Zhu, J., M. B. Miller, R. E.Vance, M. Dziej-
biofilm formation. Mol. Microbiol. 53:857–869. man, B. L. Bassler, and J. J. Mekalanos. 2002.
52. Tischler, A. D., and A. Camilli. 2005. Cyclic Quorum-sensing regulators control virulence
diguanylate regulates Vibrio cholerae virulence gene gene expression in Vibrio cholerae. Proc. Natl. Acad.
expression. Infect. Immun. 73:5873–5882. Sci. USA 99:3129–3134.
SIGNAL INTEGRATION AND
VIRULENCE GENE REGULATION
IN STAPHYLOCOCCUS AUREUS
Edward Geisinger and Richard P. Novick
11
The etiologic agent of a spectrum of disease exoprotein genes (93, 109, 118). In Table 1 are
states ranging from the superficial to the sys- listed the most important of these genes and
temic and invasive, Staphylococcus aureus is a their products. The demonstration that agr
pathogen of extraordinary versatility. Its adapt- mutants are attenuated for virulence in several
ability and flexibility allow it to respond to animal models (1, 33, 42) has established agr as a
numerous environmental obstacles and thus global regulator of staphylococcal virulence.
thrive in a variety of inhospitable host niches. Further support for the concept of global regu-
This versatility depends on a tremendous range lation of virulence was obtained by the subse-
of interacting accessory gene systems, many of quent isolation of other pleiotropic mutants,
which participate in pathogenesis. The great including sarA (24), sae (45), sarS (also known as
majority of accessory genes involved in patho- sarH1 [136]), and rot (90). In Table 2 is a list of
genesis, referred to collectively as the virulon, the most important regulatory and transcrip-
encode proteins that are either displayed on the tion factors identified thus far.
bacterial surface or released into the surround- These regulatory systems sense and integrate
ings. These enable the organism to evade host various extracellular and intracellular inputs—
defenses, to adhere to cells and the tissue cell density, energy availability, environmental
matrix, to spread within the host, and to signals, and superantigens (SAgs)—to control
degrade cells and tissues, for both nutrition and the production of exoproteins when they are
protection. required. In vitro, exoprotein production fol-
Evidence for global regulation of the staphy- lows a specific temporal program in which
lococcal virulon materialized with the identifi- adhesins are made before hemolysins or pro-
cation of the agr locus, a quorum-sensing teases and other degradative enzymes,and this is
system that controls the expression of most likely also the case during infection ( J.S.Wright
and R. P. Novick, unpublished data).Transcript
Edward Geisinger and Richard P. Novick The Helen L. and profiling has shown that genes encoding surface
Martin S. Kimmel Center for Biology and Medicine, Skirball proteins are downregulated early in growth,
Institute for Biomolecular Medicine, Molecular Pathogenesis
Program, New York University School of Medicine, 540 First whereas those encoding secreted proteins are
Avenue, New York, New York. 10016. upregulated postexponentially (31;Y. Fang and
161
TABLE 1 Staphylococcal extracellular accessory proteinsa
162 ■ GEISINGER AND NOVICK

Action of regulatory genes
Gene Location Product Activity/ Timing Reference(s)
function agr saeRS rot sarA sarS sarT tst
Superantigens
sea Phage Enterotoxin A Food xp 0 0 140, 144
poisoning,TSS
seb SaPI3 Enterotoxin B Food pxp C 38, 141, 144
poisoning,TSS
sec SaPI4 Enterotoxin C Food pxp 120
poisoning,TSS
sed Plasmid Enterotoxin D Food pxp 159
poisoning,TSS
eta ETA phage Exfoliatin A Scalded pxp 130
skin syndrome
etb Plasmid Exfoliatin B Scalded pxp
skin syndrome
tst SaPI1,2,bov1 Toxic shock Toxic shock pxp C 119, 144
toxin-1 syndrome
Cytotoxins
hla Chrom -hemolysin Hemolysin, pxp C 46, 70, 119, 124,
cytotoxin 127, 136, 144
hlb Chrom -hemolysin Hemolysin, pxp C 46, 70, 119, 124
cytotoxin
hld Chrom -hemolysin Hemolysin, xp 0 0 46, 119, 127, 136
cytotoxin
hlg Chrom -hemolysin Hemolysin, pxp C 17, 70, 122, 124
cytotoxin
lukS/F PVL phage P-V leukocidin Leukolysin pxp 124, 144
Enzymes
splA-F Chrom Serine Putative 124
protease-like protease
sspA Chrom V8 protease Spreading pxp 3, 20, 105, 122, 144
factor
aur Chrom Metalloprotease Processing pxp 3, 20, 105, 122
(aureolysin) enzyme?
sspB Chrom Cysteine Processing 0 3, 124
protease enzyme?
scp Chrom Staphopain Spreading, pxp 3
(protease II) nutrition
geh Chrom Glycerol Spreading, pxp 0 C 124, 132
ester hydrolase nutrition
lip Chrom Lipase Spreading, pxp 0 C 19
(butyryl esterase) nutrition
fme Chrom FAME Fatty acid pxp C 19
esterification
11. SIGNAL INTEGRATION AND VIRULENCE GENE REGULATION IN S.AUREUS ■ 163

plc Chrom PI-phospholipase C pxp 3
nuc Chrom Nuclease Nutrition pxp 122
hal Chrom Hyaluronic acid lyase Spreading factor pxp C
coa Chrom Coagulase Clotting, exp 46, 68, 122,
clot digestion 124, 149
sak Phage Staphylokinase Plasminogen pxp 0 119
activator
Surface proteins
spa Chrom Protein A Anti-immune, exp C 28, 46,
anti-PMN 119, 124
cna PT islet Collagen BP Collagen pxp 0 13
binding
fnbA Chrom Fibronectin Fibronectin exp 122, 126
BPA binding
fnbB Chrom Fibronectin Fibronectin exp 122, 126
BPB binding
clfA Chrom Clumping Fibrinogen exp 0 149
factor A binding
clfB Chrom Clumping Fibrinogen exp 0 0 87, 124
factor B binding
SA1000 Chrom Hypothetical Fibrinogen exp 15, 70
fibrinogen binding
binding protein
Capsular polysaccharides
cap5 Chrom Polysaccharide Antiphagocytosis? pxp 110
capsule type 5
cap8 Chrom Polysaccharide Antiphagocytosis? pxp 110
capsule type 8
a
Chrom, chromosomal; xp, throughout exponential phase; exp, early exponential phase only; pxp, postexponential phase; 0, no effect of gene on expression; , upregulated; , downregulated; C,
controversial; P-V, Panton-Valentine; FAME, fatty acid methyl ester.
TABLE 2 Known accessory gene regulation and transcription units in S. aureus

Regulatory unit Description Role Reference(s)
agrACDB/rna III TCS, autoinduced by peptide Regulates many extracellular and cytoplasmic 103
protein accessory genes in response to cell
density
saePQRS TCS, autoinduced Regulates many extracellular protein genes in 47
response to environmental stimuli
arlRS TCS Regulates autolysis, capsule production, and 36, 74
certain accessory genes
srrAB TCS Regulates certain accessory genes in response 155
to PO2
B Rpo alternative sigma factor Active in late exponential phase; regulates 66
many accessory genes
rot Transcription factor Pleiotropic regulator; affects accessory gene 90, 124
transcription in general antagonism to agr
sarA Transcription factor Pleiotropic repressor; assists in agr 27
autoinduction under certain conditions
sarS(sarH1) Transcription factor Activates transcription of spa and possibly 136
other surface protein genes
tcaR Transcription factor Activates transcription of spa through sarS 88
sarT Transcription factor Represses transcription of hla and possibly 127
other exoprotein genes
sarR Transcription factor Transcription factor for sarA and possibly sarS 79, 82
sarU Transcription factor Activates transcription of agr RNAIII 81
mgrA (NorR, Rat) Transcription factor Affects expression of large number of genes 73, 75
involved in virulence and autolysis;
regulates exoproteins in direction
analogous to agr
sarV Transcription factor Affects expression of many genes involved in 83
virulence and autolysis
sarX Transcription factor Represses exoprotein synthesis, likely via agr 80
repression
R. P. Novick, unpublished data). This shift in function (21, 40) have a general inhibitory
expression pattern is correlated with the agr sys- effect on the synthesis of the exoproteins listed
tem, which is the single known quorum sensor in Table 1. In other cases, the response may be
of the staphylococci. Agr is activated in mid- related to specific environmental exigencies;
exponential phase and is known to upregulate a thus, acid pH (≈ 5.8) downregulates tst (see
number of secreted genes and to downregulate Table 1 for gene abbreviations) and hla but
several surface genes (31, 103). However, when upregulates ssp and spa (147).Added to this elab-
energy or biosynthetic metabolites are insuffi- orate set of regulatory patterns is the genetic
cient, as with respiratory chain (91) or citric variability of SAg genes; as the SAgs in and of
acid cycle mutants (134),many of these proteins themselves cause toxinoses, a strain expressing
are not synthesized, enabling the organism to one has no need of most other exoproteins.
maintain its housekeeping functions. Several Because these only complicate matters by evok-
environmental signals affect the production of ing host defenses, the toxin shuts down their
extracellular proteins, in some cases perhaps in genes (144). In this chapter, we outline what is
relation to the need to conserve resources under known about the regulatory strategies underly-
detrimental conditions.Thus, 1 M NaCl, subin- ing this complex set of behaviors and begin to
hibitory ethanol, and subinhibitory concentra- model the organization of the overall regulatory
tions of antibiotics (50) that inhibit ribosome network.
TWO-COMPONENT SYSTEMS (Fig. 1).These variations have resulted in at least

Information on the external environment is four agr specificity groups in S. aureus and prob-
read by the cell via signal receptors,which for S. ably one or more in each of 15 other staphylo-
aureus appear to be the primary regulatory coccal species examined (30, 55, 57, 106).The
modality for expression of the virulon. In addi- groups are defined by the mutual inhibition by
tion to the quorum-sensing two-component their peptides of the agr response in heterolo-
system (TCS) agr, three other distinct TCSs are gous pairings, resulting in a novel type of bacte-
presently known to be involved: sae (44, 47), srr rial interference in which the agr regulon,rather
(155) (independently analyzed by Throup et al. than growth, is blocked (57).The ability of an
[138] and referred to as srh), and arl (34) (Fig. AIP to activate its cognate receptor is highly
1B–D).These four represent one-quarter of the sequence specific; a single amino acid substitu-
putative TCSs identified by examination of the tion can change group specificity (AIPs I and
S. aureus genome (18, 85; D. McDevitt, personal IV in Color Plate 8).The N-terminal one-third
communication), and there is every reason to of agrB and the C-terminal (cytoplasmic) histi-
believe that some of the others are also involved. dine protein kinase domain of AgrC are highly
conserved, whereas the intervening sequences
The agr System are highly divergent, constituting the hyper-
We begin with the well-studied agr locus, ≈3 kb variable region indicated in Fig. 1. The diver-
in length and consisting of divergent transcrip- gent regions determine group specificity and
tion units, driven by promoters P2 and P3 (Fig. must therefore have evolved in concert.
1A).The P2 operon encodes a two-component Functional variants within the agr locus are
system and its autoinducing ligand (102). The presumably designed for cross-group and cross-
primary function of the operon is to activate the species interference, so that they serve to isolate
two agr promoters.The agr-activating ligand is a populations and may represent a fundamental
posttranslationally modified, autoinducing pep- determinant of strain divergence and specia-
tide (AIP), seven to nine amino acids in length, tion. Indeed, agr groupings broadly correlate
which is processed from a propeptide encoded with strain genotypes (56), and analysis of sev-
by agrD (58).The AIP binds to the N-terminal eral S. aureus strain sets revealed that genotypic
transmembrane domain of the agr signal recep- class, with rare exceptions, associated with a sin-
tor, agrC (58, 72, 76), activating the agr TCS, of gle agr group, pointing to agr group differentia-
which AgrA is the response regulator.Activated tion as a primary evolutionary event that
AgrA then upregulates promoters P2 and P3. preceded genotypic divergence (153). In keep-
The P3 transcript, RNAIII, rather than AgrA, is ing with this idea,agr groups,at least in S.aureus,
the intracellular effector of target gene regula- are predicted to correlate with specific biotypes,
tion (54, 103).As agr is autoinduced by an extra- and evidence supporting this prediction has
cellular ligand that is encoded within the emerged with respect to clinical features.Thus,
operon, the density-sensing circuit is doubly most menstrual toxic shock syndrome (TSS)
autocatalytic, resulting in a very rapid burst of strains belong to agr group III (57), as do all 16
activity once the autoinduction threshold has strains found to cause leukocidin-induced
been reached. The expression of this system necrotizing pneumonia (43); most of the
entails a tremendous metabolic burden,resulting recently encountered VISA strains belong to agr
in frequent spontaneous agr mutants in the labo- group II (125), and most exfoliatin-producing
ratory (12,133),especially in strains that lack B, strains belong to agr group IV (55, 89).A similar
which modulates the agr regulon (see below). species-specific divergence has been described
for the identically organized comAP locus,
agr SPECIFICITY GROUPS responsible for transformation competence in
agr is conserved throughout the staphylococci bacilli (139).Intergroup interference in S.aureus
with interesting variations in the B-D-C region has enabled a test of the effects of blocking agr
on an experimental staphylococcal infection, that targets the peptide to the membrane (158),
the skin abscess model of Barg et al. (4), in where its C-terminal domain may interact
which coadministration of the synthetic group specifically with AgrB (157).The putative cat-
II AIP along with or immediately after the bac- alytic core of AgrB,identified by mutagenesis to
teria sharply attenuated an infection caused by a contain conserved histidine and cysteine
group I strain (86, 150).The growing realization residues, mediates the C-terminal processing of
that agr groups are biologically and clinically AgrD, whose conserved cysteine residue could
significant has prompted the development of agr subsequently catalyze formation of the thiolac-
typing methods. Several of these use the PCR tone ring. For at least group I, and likely other
to generate agr group-specific products that are groups and species, the N-terminal processing
identified by restriction site polymorphisms of AgrD is carried out by the signal peptidase,
(108) or sequencing (30).In the author’s labora- SpsB (64).The temporal order of these events is
tory, activation or inhibition of an agrP3- unclear, and how AgrD traverses the membrane
luciferase reporter during adjacent coculture on is yet to be discovered.
an agar surface is used (151). The group specificity of AgrB is less strin-
gent than that of the AIP-receptor interaction.
AIP SYNTHESIS, STRUCTURE, Thus, AgrB-I and AgrB-III will each process
AND ACTIVITY AgrD-I and AgrD-III with equal efficiency, but
The agrD-encoded propeptide is both N- and neither will process AgrD-II or AgrD-sl and
C-terminally processed to form a unique thio- vice versa (57). The agrD sequence has been
lactone ring between the conserved central determined for nearly 30 different strains,
cysteine and the peptide’s C-terminal carboxyl including representatives of some 15 different
(57) (Color Plate 8). This cyclic thiolactone staphylococcal species (30, 55, 57, 59, 106, 143)
is essential for AIP activity and is the hallmark (Fig. 2). The AIPs from strains of all four S.
of these peptides.The only exception to this is a aureus agr specificity groups and from represen-
serine in Staphylococcus intermedius (30). S. inter- tative strains of Staphylococcus lugdunensis,
medius strains produce an AIP, shown by mass Staphylococcus warneri, Staphylococus epidermidis,
spectrometry to be a nonapeptide containing a and S. intermedius have been sequenced and/or
cyclic lactone (62), with both autoinducing synthesized in vitro (55, 59, 78). AIPs I and IV
and cross-inhibiting activities (59). A detailed are octapeptides,AIPs II and Si-I are nonapep-
picture of AIP biosynthesis has emerged, in tides (57, 62), and AIPs III and Sl-I and -II are
which AgrB, a polytopic transmembrane pro- heptapeptides, suggesting that most staphylo-
tein with demonstrated endopeptidase activity, coccal AIPs are seven to nine amino acids
plays a central role (57, 115, 123, 156). AgrD in length. The AIPs form a coherent group
possesses an N-terminal, amphipathic leader with generally conserved structural features,
FIGURE 1 The two-component systems known to affect the virulon. (A) The agr quorum-sensing system.The
pro-AIP peptide is processed and secreted by AgrB, binds to an extracellular loop in the receptor-HPK,AgrC, acti-
vating autophosphorylation, followed by phosphorylation of the response regulator, AgrA, which, in conjunction
with SarA, activates the two agr promoters, P2 and P3, leading to the production of RNAIII, which controls tran-
scription of the target genes via intracellular regulatory mediators, including Rot and a second two-component
module, saeRS, and possibly others. (B). sae.The sae locus, about 3.5 kb, contains four open reading frames, P, Q, R,
and S. R and S form a classical two-component signaling module.The functions of P and Q are unknown.sae is tran-
scribed from two or three promoters, one of which is active in an agr-null strain and the other(s) is activated by
RNAIII.All three major transcripts,A, B, and C, end at ter. D may be independently transcribed or derived from C
by processing. PCR probes used to map the transcripts are shown. (C). arlRS (adapted from Fournier et al. [36]).The
arlRS locus encodes a receptor-HPK (arlS) and a response regulator (arlR), driven by a single promoter and followed
by a terminator stem-loop. (D). srrAB (adapted from Yarwood et al. [155]).The srrAB locus encodes a receptor-HPK
(srrB) and a response regulator (srrA), driven by a single promoter that generates two transcripts whose relative sig-
nificance is unknown.
FIGURE 2 Comparison of agrD sequences. Sequences were aligned visually. Predicted AIPs are in bold and are
set between spaces.Those for which sequence has been confirmed by in vitro synthesis or by mass spectroscopy are
indicated with footnote. Saur, S. aureus; Sarc, S. auricularis; Sarl, S. arletta; Scap, S. capitis; Scapr, S. capri; Scarn, S.
carnosus; Sconc, S. cohneii cohneii; Sconu, S. cohneii urealyticum; Sepi, S. epidermidis; Sgal, S. gallinarum; Sint, S. inter-
medius; Slug, S. lugdunensis; Ssim, S. simulans; Swar, S. warneri; Sxyl, S. xylosus.
including a strong gradient of increasing and inhibiting peptides at the receptor is strictly
hydrophobicity from N to C termini, culmi- competitive (77).
nating in two bulky hydrophobic residues, lim- Cyclic peptide autoinducers have begun to
ited to FLVY, plus an occasional M. be identified in other genera; for example, a
Detailed structure-function analyses have tailless thiolactone ring autoinducer has been
been performed for AIPs II and I. The results characterized and demonstrated to regulate
from these studies as well as the general struc- adherence in Lactococcus plantarum (135), and in
ture of the AIPs are summarized in Color Plate Enterococcus faecalis, an 11-residue peptide with
8. The cyclical structure is generally required an 8-member lactone ring activates the fsr
for both agr activation or inhibition (76,86,89). quorum-sensing system, a regulator of gelati-
Replacement of the thiolactone by a lactone or nase synthesis (98, 114). In this case, the native
lactam bond virtually eliminates the activation AIP was more stable to protease degradation in
but not the cross-group inhibition function of culture supernatants than a linear peptide of the
the peptide, although these variant peptides are same primary sequence (J. Nakayama, personal
not self-inhibitors. For AIP II, the tail region is communication).Thus,the cyclic structure may
critical for activation, as its removal converted it provide stability and may also be well suited to
into a universal inhibitor of S. aureus agr func- the dual activation-inhibition role required for
tion (76). The interaction between activating these peptides. Nevertheless, a variety of linear
peptides serve as signal receptor ligands in many interactions. In particular, the chimeras whose
gram-positive bacteria, including Streptococcus promoter-distal subdomain was of group III
pneumoniae, Bacillus subtilis, and lactobacilli. identity could not be inhibited but rather were
activated by a variety of AIPs, including some
AgrC,THE agr SIGNAL RECEPTOR that are strong inhibitors of other AgrCs.
AgrC has a C-terminal domain containing the Revealingly, these promiscuous mutant recep-
conserved features of a histidine protein kinase tors (152), as well as wild-type receptors (86),
and a polytopic N-terminal sensor domain could not be activated (or inhibited) by AIPs
shown by phoA fusions to consist of five or six whose C-terminal hydrophobic residues were
transmembrane helices (72). Pull-down studies replaced with alanine, implicating the role of
identifying AgrC as the only cellular protein these ring residues in generalized binding to the
capable of binding the AIP confirmed the pro- receptor.
tein’s role as the agr signal receptor (58). Muta- Together, these results have given rise to a
tional analysis in the authors’ laboratory has model in which interaction with the receptor
suggested that agr activation follows the classical involves two distinct events. First, the peptide
TCS paradigm, that is, AIP-dependent trans- interacts with a hydrophobic pocket of the
autophosphorylation of AgrC dimers followed receptor in a non-sequence-specific manner,
by forward flow of phosphate to AgrA (E. mediated by the two bulky C-terminal residues
Geisinger, E. A. George, T. W. Muir, and R. P. of the AIP.Second,it makes one or more group-
Novick, unpublished data). specific, polarity-dependent contacts with spe-
Predictably, group specificity resides in the cific sites in the receptor, leading to activation;
N-terminal transmembrane domain of AgrC, as the absence of such a contact would result in
demonstrated by switching the two domains inhibition.The broadening of specificity in the
between AgrCs of different groups (78). Addi- I-III and IV-III chimeras could thus represent a
tionally,switching the proximal and distal halves situation in which the receptor is misfolded in
of the N-terminal sensor domain of the AgrCs such a way as to be poised for activation by the
localized the specific recognition of AIPs I and binding of a hydrophobic AIP without the need
IV to the distal subdomain (152), suggesting for any specific activating contacts.
that the single amino acid that differs between
these two AIPs (aspartate versus tyrosine at posi- AgrA
tion 5) makes a specific contact in this region of AgrA has the sequence features of a response
the receptor. Sequence comparison of AgrCs I regulator (100) and is required for activation of
and IV in this region highlighted a small set of the two agr promoters, completing the autoin-
hydrophilic (for group I) or hydrophobic (for duction circuit (102, 103). AgrA was demon-
group IV) residues that matched the polarity of strated to bind to the agr promoters with high
the unique residue in the AIP cognate. Indeed, affinity (65), and the interaction was localized
exchange of as few as two of these residues in via DNase I protection assays to paired direct
AgrC-IV for those of group I resulted in a repeats in the P2 and P3 promoter regions that
nearly full switch in receptor specificity (E. fit or closely resembled the consensus binding
Geisinger and R. P. Novick, unpublished data), sequence of the LytTR response regulator fam-
more precisely localizing the specificity region ily, of which AgrA is a member.Addition of the
of AgrC-I and -IV to the second extracellular small phosphate donor acetyl phosphate
loop region and suggesting that polarity is a key enhanced this binding, and at least for its inter-
feature of the AIP-receptor interaction. action with agrP2, appeared to influence the
Proximal and distal chimeras involving other oligomeric state of AgrA, as inferred from
groups,however,have given results that are very electrophoretic mobility shift assays. Interest-
difficult to explain on the basis of the classical ingly, AgrA bound the P2 promoter more
“lock-and-key” model for intermolecular strongly than P3, leading to the speculation that
autoinduction precedes P3 induction. It is pleiotropic transcription factor Rot (39) that

worth noting that AgrA homologs, PlnC, PlnD, has broad regulatory effects on agr-regulated
and SppR in Lactobacillus plantarum and Lacto- genes and generally acts counter to agr, down-
bacillus sake (29, 121), respond to autoinducing regulating genes encoding secreted proteins
peptides and bind to heptanucleotide repeats and upregulating surface protein genes (124)
that are similar to those in the agr intergenic (Table 1). RNAIII makes two loop-loop con-
region,also with differential affinity and timing. tacts with the rot mRNA between its C-rich
AgrA may interact with or displace the regula- loops and two loops of the rot transcript, one of
tor SarA, also known to bind agrP2 and activate which contains its SD sequence, preventing
expression of both promoters under certain ribosome binding, as revealed by enzymatic
conditions (23, 95, 117). probing and toeprinting experiments (15).
RNAIII thus indirectly affects the expression of
agr-RNAIII:THE agr EFFECTOR a large set of virulence genes. It is possible that
Agr autoinduction leads to production of the RNAIII acts similarly on other regulatory gene
P3 transcript, RNAIII, which is the intracellu- transcripts.
lar effector of the agr regulon (103).RNAIII is a RNAIII also acts directly at the level of
highly abundant and stable regulatory RNA, translation for at least three virulence gene
with a half-life likely greater than 45 min (52;E. products, -hemolysin (94, 103), protein A
Geisinger, unpublished data). RNAIII contains (52), and a novel fibrinogen-binding protein
an open reading frame encoding the virulence SA1000 (15), but almost certainly, other viru-
factor, -hemolysin, but this 26-amino-acid lence factors will be shown to be similarly
peptide does not appear to have any regulatory regulated.The 5′ region of RNAIII is comple-
function (54). RNAIII has a complex second- mentary to the hla leader (Color Plate 9),which
ary structure (7)(Color Plate 9) that is well con- folds into an untranslatable configuration unless
served (although the sequence is not) among prevented from doing so by RNAIII; it is likely
several staphylococcal species (7, 137), resulting that translation of the hld reading frame is
in interspecific cross-reactivity of the molecule. required for this interaction. The 3′ end of
RNAIII acts reciprocally, upregulating tran- RNAIII is complementary to the translation
scription of most of the extracellular protein initiation site of the spa and SA1000 mRNAs
genes and downregulating that of many surface and blocks their translation (Color Plate 9). In
protein genes (103, 126). Structure-function all of the above examples,target gene regulation
analyses revealed that (i) the 3 end of RNAIII by RNAIII involves multiple loop-loop con-
is necessary and sufficient for repression of spa tacts or the formation of an extended duplex.
transcription (7) and (ii) nonoverlapping 5 and The translational repression events involve an
3 subregions of RNAIII were independently additional regulatory component, cleavage by
active in stimulating hla transcription (103), the endoribonuclease RNaseIII (15, 52), which
implying functional redundancy with regard to renders the repression of these target transcripts
its ability to activate target gene transcription. irreversible. Interestingly, the chaperone pro-
This most probably involves nearly identical, tein, Hfq, which facilitates the action of sRNAs
C-rich sequences in the unpaired regions of in Escherichia coli and other species (69), does
stem-loops 7, 13, and 14 (Color Plate 9), which not appear to be involved in RNAIII-depend-
are complementary to the canonical Shine- ent gene regulation in S.aureus (14,15,39),per-
Dalgarno (SD) sequence. RNAIII was hypoth- haps due to the abundance and multifunctional
esized to function by interfering with the nature of RNAIII.
translation of other regulatory proteins by
virtue of these anti-SD sequences.This hypoth- agr IN VIVO
esis was recently confirmed when RNAIII The importance of agr in pathogenesis has been
was shown to repress the translation of the demonstrated by examining agr mutants in a
variety of animal models, including skin defect in exoprotein production; the mutation
abscesses (86, 150), endocarditis (25), septic had no effect on the production of RNAIII and
arthritis (1), and osteomyelitis (42). Interest- so was either downstream or epistatic to agr.
ingly, agr is not expressed in the peritoneal sep- Transcriptomic and proteomic approaches have
sis model (6; R. Jin and R. P. Novick, identified the generally upregulatory effect of
unpublished data) and does not appear to affect the TCS on a relatively limited subset of viru-
virulence (J.Wei and R. P. Novick, unpublished lence genes that partially overlaps with the agr
data), probably because the agr autoinducer set, including fibronectin- and fibrinogen-
does not accumulate locally. In a recent study binding proteins and secreted exoproteins (70,
employing the mouse subcutaneous abscess 122). sae has a complex transcriptional pattern
model, agr activity, tracked over time using an that is profoundly influenced by agr and envi-
agrP3-lux fusion, was shown to peak very early ronmental stimuli (101). It is initially tran-
(3 h) before a neutrophil-induced eclipse (150). scribed as a 2-kb mRNA that disappears
Combined with the observation that a sterile postexponentially. Immediately after the onset
abscess could be formed by injection of filtered of RNAIII synthesis in mid-exponential phase,
culture supernatants from an agr, but not agr, two larger transcripts appear that are not seen in
strain, these results suggested that robust release the sae mutant and are greatly reduced in an agr-
of exotoxins is crucial very early during infec- null strain and in a sarA mutant.The longest of
tion. The temporal program of agr and other these transcripts includes two additional open
accessory gene systems are further considered reading frames 5′ to saeR that are likely to be
below. important for sae function and perhaps for
The functionality of agr in staphylococcal transducing environmental signals.
biofilm development and growth has been Several environmental stimuli, such as high
studied. Although inactive during biofilm for- salt, low pH, glucose, and subinhibitory antibi-
mation,possibly related to the increased expres- otics, affect the sae transcription pattern (101).
sion of adhesins (145), agr is active in the For example, in the presence of 1 M NaCl or a
detachment process, in which it could facilitate subinhibitory concentration of clindamycin,
dissemination to other sites (154). In addition, hla and spa transcripts as well as the larger sae
intracellular growth has been implicated as a transcripts are greatly decreased, whereas with
significant component of S. aureus pathogenesis subinhibitory concentrations of -lactam
(2, 5, 61), and as AIP would accumulate to acti- antibiotics, all four are upregulated.These envi-
vating concentrations in intracellular compart- ronmental cues act independently of agr but
ments, the role of agr has been examined in this may act through SarA or one of its homologs.
context. agr was shown to be active after cell Thus, sae appears to lie at a convergence of cell
internalization and to be required for endo- density and environmental signals. Its impor-
some escape (113, 131) as well as for induction tance in S. aureus infection has been demon-
of apoptosis (49,148) in a variety of mammalian strated in several animal models (8, 48, 70, 116).
epithelial and endothelial cells. While -
hemolysin was suggested to mediate apoptosis arlRS
induction (49), the role of other agr-regulated arlRS (Fig. 1C) is a third TCS involved in regu-
exoproteins in these processes remains to be lation of the staphylococcal virulon, identified
determined. on the basis of its control of autolysis and of the
norA polyvalent export pump (34). agr mutants
The sae TCS are reported to be defective in arlRS expression,
saeRS (45, 47) (Fig. 1B) represents the second whereas an arlS mutant apparently overex-
major TCS involved in global regulation of the presses agr, especially the agr P2 transcript, con-
staphylococcal virulon, identified as a transpo- sistent with independent regulation of P2 and
son insertion mutation with a pleiotropic P3 (36).agr and arlRS thus formally represent an
autorepression circuit such that arlRS counters resent a central regulatory logic, but one whose
agr autoinduction. Consistent with this is the biology remains to be determined.
reported downregulation by arlRS of overall
exoprotein synthesis, presumably consequent ALTERNATIVE SIGMA FACTORS
to downregulation of agr (36) and/or upregula- A second major mechanism of response to envi-
tion of rot (71). ArlRS has also been demon- ronmental stimuli is via alternative sigma fac-
strated to be involved in regulating capsule tors, which are generally activated directly
synthesis (74) and responding to agents that within the cell rather than through signal trans-
modulate DNA topology (35). duction. S. aureus possesses two of these,
homologs of B. subtilis B and H (96), the latter
srrAB of which does not seem to be involved in viru-
The fourth TCS involved in expression of the lence. B is required for the expression of genes
staphylococcal virulon is srrAB (155) (Fig. 1D), involved in pigment synthesis, defense against
also known as srhSR (138), a homolog of the oxidative stress, and other functions, as well as
O2-responsive resDE system of B. subtilis (9). expression of one of the B transcripts. It is acti-
The srrAB mutants are profoundly growth vated by environmental stress and energy deple-
defective in the absence of oxygen, although tion (reduced ATP/ADP ratio), as well as by
they grow normally under aerobic conditions. environmental stimuli such as ethanol (21) and
The role of srrAB was recently analyzed using salicylic acid (107), and its activity is regulated
an antisense RNA approach to repress its ex- by a complex posttranslational pathway consist-
pression conditionally, with which it was ing of rsbU, V, and W (129). B is usually bound
demonstrated that this TCS differentially regu- by RsbW, an antisigma factor, that phosphory-
lates virulence genes such as tst and spa in aero- lates RsbV, an antiantisigma factor. Under con-
bic and anaerobic conditions (111). srrAB also ditions of environmental stress, RsbV ~P is
appears to inhibit agr activation (155), possibly dephosphorylated by either of two phos-
through direct binding to the agr promoters phatases, RsbU or RsbP, and then binds RsbW,
(112), and is itself downregulated by agr releasing and activating B. B recognizes a
(unpublished data; P. Schlievert, personal com- unique promoter [GTTT(N14-17)GGGTAT],
munication). agr and srrAB thus represent a which has been identified for 23 different S.
mutual cross-inhibition circuit that would nec- aureus genes (41), including one of the three
essarily have to respond to extrinsic regulatory sarA and one of the three sarS promoters, plus
inputs. srrAB also regulates many genes genes encoding transport functions and others
involved in energy metabolism and evidently involved in generating NADH2. B is also
regulates energy transduction under anaerobic required for certain genes that lack a B pro-
conditions (138). It may be activated by moter; these are presumably regulated by B-
menaquinone or a derivative, one of the interdependent transcription factors.
mediates in the oxidative respiratory pathway. Microarray analysis has identified the B reg-
Thus, this TCS may connect the agr signaling ulon (10), which includes genes putatively
pathway with the overall energy metabolism of involved in cell envelope biosynthesis and
the cell. It appears that each of these three TCSs turnover, intermediary metabolism, signaling
exerts its effects on the virulon and other acces- pathways, and virulence. Many adhesin genes,
sory genes largely, although not entirely, such as coa and fnbB, of which the former has a
through its interaction with agr. Thus, saeRS B-dependent promoter (99), are upregulated
and agr appear to be mutually upregulatory, while exoprotein and toxin genes are repressed.
srrAB and agr are reported to be mutually B thus appears to be antagonistic to agr. While
downregulatory,and arlRS and agr seem to con- some studies suggest that B is important in
stitute an autorepression circuit.These different pathogenesis (60),especially in cell adhesion and
interactions, probably indirect, presumably rep- uptake (32, 97), others suggest that it is not (22,
51,99).Furthermore,a small fraction of S.aureus homologs appear to belong to the group of reg-
clinical isolates are nonpigmented and overpro- ulatory proteins that bind DNA with limited
duce various exoproteins,some owing to a defi- sequence specificity (e.g., H-NS and HU).
ciency in B(63), supporting the idea that B Represented in Fig. 3 is an example of the
may not be required for pathogenesis. Strains of type of interweaving regulatory circuitry in
the 8325 lineage, which are B deficient owing which SarA and its homologs are involved.
to an 11-base deletion in rsbU, show important Intermediary transcription factors have been
differences in overall biology from their rsbU- shown to affect the transcription of accessory
repaired derivatives (11), including a reduction genes, other SarA homologs, and RNAIII (3,
in the lag phase of growth, an increase in overall 81, 128), generating complex activation cas-
growth yield and in starvation survival (51), as cades and feedback loops and defining subsets of
well as in the expression of both regulatory and accessory genes. RNAIII interacts directly or
exoprotein genes (see below).The exact contri- indirectly with several factors, by regulating
bution of B to pathogenesis remains to be elu- either their synthesis or their action. For exam-
cidated, as does its precise role in the overall ple,Rot,whose pleiotropic regulatory effects on
regulatory network. the agr regulon are in direct opposition to those
of agr, is posttranscriptionally downregulated by
TRANSCRIPTION FACTORS agr-RNAIII (39). In the case of the spa activator
In general, transmission of environmental sig- SarS (37) (Fig. 3), it is known that RNAIII
nals recognized by transmembrane and intracel- blocks its expression (136), likely through Rot,
lular receptors to effector (or target) genes thereby effecting repression of spa transcription.
involves pleiotropic intracellular transcription While SarA has been demonstrated to be an agr
factors, including the Sar (staphylococcal acces- activator, other homologs such as SarX have
sory regulation) family of homologous winged recently been suggested to repress agr transcrip-
helix-turn-helix DNA-binding proteins, tion directly (53). It has also been observed that
namely SarA, R, S, T, U,V, X (80); Rot; TcaR SarT downregulates agr (128), apparently acting
(88); and MgrA (75).These regulatory factors, via SarU (81), an agr upregulator.The downreg-
listed in Table 2 with a summary of their general ulation of sarT by agr thus generates a negative
properties (reviewed by Cheung and Zhang feedback loop as shown in Fig. 3. Additional
[26] and Bronner et al. [16]), affect a wide vari- data such as those obtained by transcript profil-
ety of genes,most of which encode virulence or ing (31, 124) will fill out this circuitry, perhaps
other accessory functions. They interact with defining regulatory subsets of target genes that
one another,with the TCSs,and with B,as well could have biological or clinical relevance.
as with the target genes themselves, generating
an extraordinarily complex regulatory net- Superantigens
work. The DNA-binding segments of these Remarkably, at least two of the major staphylo-
proteins are well conserved, many containing coccal SAg toxins, toxic shock syndrome toxin
the motif KXRXXXDER, whereas other parts 1 (TSST-1) and staphylococcal enterotoxin B
of the proteins are less well conserved.The pro- (SEB), are themselves transcription factors, act-
totype, SarA, is a 14.7-kDa DNA-binding pro- ing as global repressors of most exoprotein
tein, distantly related to VirF of Shigella fiexneri. genes at the level of transcription (see Table 1),
SarA binds as a dimer, whereas at least three of and are also autorepressors (144). These two
its homologs, SarS, SarU, and SarY, appear to be proteins, as well as the other staphylococcal and
the result of duplications and therefore intrinsi- streptococcal SAgs, are structurally very closely
cally dimeric.Given the degree of structural and related, as shown by X-ray crystallography
sequence similarity among the members of this (104). However, there is no striking sequence
family (26), the possibility of heterodimeric similarity corresponding to the inhibitory
combinations has been suggested (136).The Sar regions of the two proteins. The data establish
FIGURE 3 Regulatory interactions involving SarA and its homologs.Arrows represent upregulation, bars rep-
resent downregulation. Gray lines represent translation; black lines represent interactions that are probably, but not
always certainly, transcriptional.The interactions illustrated are based on reviews by Arvidson and Tegmark (3) and
Cheung and Zhang (26) and on papers by Manna and Cheung (80–82,84),Ingavale et al.(53),and Said-Salim et al.
(124).Although the abbreviations are mostly in italics, on the assumption that the interactions are likely to be at the
transcriptional level, there is actually very little evidence to indicate whether they are direct or indirect or at what
level they occur. Question marks represent the most speculative. B is shown entering the system via sarS and sarA,
which have B-dependent promoters and are likely to represent important intermediates in the pathways by which
environmental signals are handled.
clearly that the protein itself, rather than the manifest as early as one can detect the toxins. It
mRNA or the DNA, is the inhibitor, and that remains unclear precisely how this regulatory
a region ending about the middle of the C- paradigm fits into the overall regulatory net-
terminal half of the protein is necessary (144).As work; however, it is evidently of major clinical
the purified toxin has no effect when added to a importance.In postsurgical TSS,resulting from a
culture,and a deletion derivative lacking the sig- contaminated wound,the infection is often very
nal peptide retains its inhibitory activity (N. difficult to detect as, unlike the typical staphylo-
Vojtov, H. Ross and R. P. Novick, unpublished coccal lesion, the wound is neither purulent
data), it is clear that an intracellular form, most nor inf lamed. It is well known that TSST-1
likely the precursor, is the effector. As neither has major effects on the production of cytokines
protein appears to bind DNA directly, they pre- and probably inf luences the inf lamma-
sumably act through an intermediate transcrip- tory response by this means. It is additionally
tion factor, one that must be present early hypothesized that one or more of the exopro-
enough to account for the inhibitory effects that teins, possibly lipase, the synthesis of which is
inhibited by the SAg,is responsible for attracting thesis of some surface proteins, coagulase, and
polymorphonuclear leukocytes and stimulating possibly other accessory proteins.The synthesis
the inf lammatory response. We note that the of these is probably initiated during the transi-
pore-forming Panton-Valentine leukocidin tion from stationary phase to exponential phase
toxin, associated with staphylococcal necrotiz- and may come under the general metabolic
ing pneumonia, has recently also been demon- program governing this transition. The nature
strated to downregulate the expression of several of the signals acting at this stage represents a key
exoproteins at the transcriptional level (67), and area for study. Other surface protein genes are
it is predicted that yet other variable genes switched on shortly after the onset of exponen-
encoding toxins that cause toxinoses will be tial growth and, as typified by spa, are switched
shown to act in the manner described above. off shortly thereafter, concomitantly with the
appearance of agr-RNAIII (142).This clear rec-
iprocity, however, is not seen with all strains and
REGULATORY ORGANIZATION
under all conditions (136), and may be related
Temporal Program to B activity (S. Herbert and R. P. Novick,
Much of the facultative gene expression system, unpublished data) or to media or other growth
especially including the virulon, is temporally variables. The agr AIP reaches its threshold
organized so that the component genes must around mid-exponential phase, activating agr
contain regulatory sequences that are activated expression. In 8325 derivatives, however, cer-
combinatorially in a time-dependent manner tain exoprotein genes, such as coa, are sharply
by different incoming signals acting through downregulated well before the appearance of
intracellular response elements.As suggested by RNAIII, suggesting that some other inhibitory
the model in Color Plate 10, the entire acces- signal is responsible.
sory gene regulatory network must also be cou- The second transition, between the expo-
pled to the overall energy metabolism of the nential and postexponential phases (possibly a
cell,and it has been suggested that there must be consequence of decreasing availability of oxy-
a key coupling parameter, such as the levels of gen owing to increasing population density), is,
nucleotide polyphosphates or of other energy- in most strains, accompanied by upregulation of
transducing cofactors such as NADH2 (R.Proc- the genes encoding secreted proteins. agr,
tor, personal communication). Key enzymes of which, in 8325 derivatives, is activated 2 or
intermediary metabolism, such as aconitase more hours earlier,sets the level of expression of
(134), could be involved in this coupling, possi- most of these proteins,but not the timing (142);
bly acting through one or more TCSs, such as in fact,upregulation of these genes occurs at the
srr. On the basis of results obtained with in vitro onset of the postexponential phase, regardless of
cultures, it appears that the surface proteins are when,or even whether,RNAIII transcription is
probably required earlier in the course of an activated (unpublished data). This is consistent
infection than the secreted enzymes, immuno- with the results of temporal activation studies,
toxins, cytotoxins, and the above-mentioned in which activation of hla transcription may
intracellular metabolic enzymes.This sequential occur as much as 6 h after RNAIII (cloned to
activation seems to be, at least in part, a function the -lactamase promoter and induced) reaches
of population density. Starting with stationary its maximum level (142). In a sarA mutant,
phase,there would appear to be three key transi- RNAIII transcription is delayed by an hour and
tion points in the in vitro growth cycle, possibly is closely coordinated with the onset of hla tran-
occurring in response to intracellular signals, scription. This effect can be attributed to SarS
such as GTP levels. because, in the double SarA/SarS mutant,
First is the transition to exponential phase, RNAIII is not delayed and there is again a tim-
which involves not only the revival of biosyn- ing differential (136). In some strains, such as
thetic and other metabolic pathways required those of agr group IV, hla and other exoprotein
for growth and cell division, but also the syn- genes are upregulated earlier, concomitantly
with RNAIII synthesis (55), suggesting that the The overall metabolic machinery of the
exponential to postexponential phase transition staphylococcal cell is affected by energy
may not be a critical regulatory point for hla and resources dependent on oxygenation and
other exoprotein genes in these strains. A fur- nutrition. An unknown energy signal, possibly
ther complication is the apparent postexponen- nucleotide polyphosphate (NPP) or NADH2
tial upregulation of DNA gyrase by agr (≈ level, is transmitted to the agr locus through
sixfold) (31),raising the possibility that agr regu- a black box (BB-1), which up- or downregu-
lation could involve changes in superhelix den- lates agr according to energy resources, whereas
sity that are well known to occur during housekeeping functions are fueled preferen-
postexponential growth and to affect a variety tially by metabolic energy and nutritional
of promoters (although there are very few data resources. agr autoactivates by means of the
on this in staphylococci). AIP and interacts with at least three other
The third transition, from postexponential TCSs, sae, arl and srr, which may up- or down-
to stationary phase, is accompanied by a major regulate it, establishing one or more feedback
metabolic rearrangement that prepares the cell loops. These interactions may or may not
for long-term survival by shutting down most occur at the level of RNAIII. RNAIII, in turn,
housekeeping and facultative genes and by acti- feeds into BB-2, generating signals that regu-
vating genes required for long-term survival late transcription factors, especially the SarA
(146), by mechanisms that are not well under- homologs; highlighted in Color Plate 10 is
stood in gram-positive bacteria. its possibly distinct action on Rot, the only
The environmental factors that affect transcription factor whose expression is
expression of various components of the viru- known to be directly influenced by RNAIII.
lon would exert their effects whenever they are The other known TCSs also transmit signals
encountered. Certain of these (pH, O2 tension, through BB-2, either to the transcription fac-
CO2 concentration) would typically vary dur- tors or directly to the target genes. Environ-
ing growth in laboratory cultures and would be mental inputs signal either through BB-5,
expected to have increased importance late in which activates (or deactivates) B, again by an
growth. Others would be encountered only unknown mechanism. B, in turn, acts on those
under special circumstances and are not transcription factors and other genes that
regarded as elements of the temporal program. have B promoters. Other environmental
inputs signal through BB-3 to the transcription
Overall Regulatory Strategy— factors; they may or may not involve B. The
a Black-Box Model transcription factors, viewed as a pool, receive
Finally, we have attempted to conceptualize inputs from various sources, determining how
in Color Plate 10 the overall strategy used for they will interact with the target genes (repre-
the regulation of accessory genes as a grand sented here by the virulon and the responses to
metabolic scheme, viewed as a condensation of stresses such as heat, cold, etc.). Transcription
the temporal program just described. In this factors also feed back to agr, establishing
scheme, there are major unknown pathways, additional feedback loops, and probably inter-
indicated by black boxes (BB) and black arrows act similarly with the other TCSs. Finally, at
in Color Plate 10, and much of the information least two of the SAgs, signaling through BB-4,
implied by the colored arrows is also rather transmit information through the transcription
sketchy at best. The largest segment of the factors for downregulation of the various exo-
regulatory system is the transcription factors, protein genes.
including the SarA homologs.They are viewed
collectively here; Fig. 3. represents a prelimi- ADDENDUM IN PROOF
nary attempt to detail some of their individual Reports that two upstream genes, svrA (Garvis,
activities. S., J. M. Mei, J. Ruiz-Albert, and D.W. Holden,
Microbiology 148:3235–3243, 2002) and traP agr regulatory RNA, and identification of the RNA
(Gov,Y.,I.Borovok,M.Korem,V.K.Singh,R.K. domain involved in repression of protein A expres-
sion. RNA 6:668–679.
Jayaswal,B. J.Wilkinson,S.M.Rich,and N.Bal- 8. Benton, B. M., J. P. Zhang, S. Bond, C. Pope,
aban,J.Biol.Chem.279:14665–14672,2004),are T. Christian, L. Lee, K. M.Winterberg, M. B.
independently required for agr activation have Schmid, and J. M. Buysse. 2004. Large-
recently been found to be erroneous, in that scale identification of genes required for full
adventitious agr mutations were responsible for virulence of Staphylococcus aureus. J. Bacteriol.
186:8478–8489.
the reported phenotypes in both cases (Chen,J., 9. Birkey, S. M.,W. Liu, X. Zhang, M. F. Duggan,
and R. P. Novick, Microbiology 153:1604–1608, and F. M. Hulett. 1998. Pho signal transduction
2007; Shaw, L. N., I. M. Jonsson,V. K. Singh, A. network reveals direct transcriptional regulation of
Tarkowski, and G. C. Stewart, Infect. Immun. one two-component system by another two-com-
75:4519–4527, 2007; Tsang, L. H., S.T. Daily, E. ponent regulator:Bacillus subtilis PhoP directly reg-
ulates production of ResD. Mol. Microbiol.
C. Weiss, and M. S. Smeltzer, Infect. Immun. 30:943–953.
75:4528–4533, 2007; Adhikari, R. P., S. Arvid- 10. Bischoff, M., P. Dunman, J. Kormanec, D.
son, and R. P. Novick, Infect. Immun. 75:4534- Macapagal, E. Murphy,W. Mounts, B. Berger-
4540, 2007). Consequently, there is no evidence Bachi, and S. Projan. 2004. Microarray-based
for “SQS 1,” a proposed second signaling path- analysis of the Staphylococcus aureus sigmaB regulon.
J. Bacteriol. 186:4085–4099.
way for agr activation (Gov,Y.,A. Bitler, G. Dell’ 11. Bischoff, M., J. M. Entenza, and P. Giachino.
Acqua, J. V. Torres, and N. Balaban, Peptides 2001. Influence of a functional sigB operon on the
22:1609–1620, 2001). global regulators sar and agr in Staphylococcus aureus.
J. Bacteriol. 183:5171–5179.
REFERENCES 12. Bjorklind, A., and S. Arvidson. 1980. Mutants
1. Abdelinour, A., S. Arvidson, T. Bremell, C. of Staphylococcus aureus affected in the regulation of
Ryden, and A. Tarkowski. 1993. The accessory exoprotein synthesis. FEMS Microbiol. Lett.
gene regulator (agr) controls Staphylococcus aureus 7:203–206.
virulence in a murine arthritis model. Infect. Immun. 13. Blevins, J. S., A. F. Gillaspy,T. M. Rechtin, B.
61:3879–3885. K. Hurlburt, and M. S. Smeltzer. 1999. The
2. Almeida, R. A., K. R. Matthews, E. Cifrian, A. staphylococcal accessory regulator (sar) represses
J. Guidry, and S. P. Oliver. 1996. Staphylococcus transcription of the Staphylococcus aureus collagen
aureus invasion of bovine mammary epithelial cells. adhesin gene (cna) in an agr-independent manner.
J. Dairy Sci. 79:1021–1026. Mol. Microbiol. 33:317–326.
3. Arvidson, S., and K.Tegmark. 2001. Regulation 14. Bohn, C., C. Rigoulay, and P. Bouloc. 2007.
of virulence determinants in Staphylococcus aureus. No detectable effect of RNA-binding protein Hfq
Int. J. Med. Microbiol. 291:159–170. absence in Staphylococcus aureus. BMC Microbiol.
4. Barg, N., C. Bunce, L. Wheeler, G. Reed, and 7:10.
J. Musser. 1992. Murine model of cutaneous 15. Boisset, S., T. Geissmann, E. Huntzinger, P.
infection with gram-positive cocci. Infect. Immun. Fechter, N. Bendridi, M. Possedko, C. Cheva-
60:2636–2640. lier, A. C. Helfer, Y. Benito, A. Jacquier, C.
5. Bayles, K. W., C. A. Wesson, L. E. Liou, L. K. Gaspin, F. Vandenesch, and P. Romby. 2007.
Fox, G. A. Bohach, and W. R. Trumble. 1998. Staphylococcus aureus RNAIII coordinately
Intracellular Staphylococcus aureus escapes the endo- represses the synthesis of virulence factors and the
some and induces apoptosis in epithelial cells. Infect. transcription regulator Rot by an antisense mech-
Immun. 66:336–342. anism. Genes Dev. 21:1353–1366.
6. Bellinger-Kawahara, C., K. Francis, P.Winter- 16. Bronner, S., H. Monteil, and G. Prevost. 2004.
berg, and J.Yu. 2002. Disruption of Staphylococcus Regulation of virulence determinants in Staphylo-
aureus RNAIII agr locus does not cause virulence coccus aureus: complexity and applications. FEMS
attenuation in a mouse sepsis model.Abstracts of the Microbiol. Rev. 28:183–200.
102nd General Meeting of the American Society for 17. Bronner, S., P. Stoessel, A. Gravet, H. Mon-
Microbiology. American Society for Microbiology, teil, and G. Prevost. 2000.Variable expressions
Washington, DC. of Staphylococcus aureus bicomponent leuco-
7. Benito, Y., F. A. Kolb, P. Romby, G. Lina, J. toxins semiquantified by competitive reverse
Etienne, and F. Vandenesch. 2000. Probing transcription-PCR. Appl. Environ. Microbiol. 66:
the structure of RNAIII, the Staphylococcus aureus 3931–3938.
18. Brunskill, E., and K. Bayles. 1996. Identifica- agr locus in Staphylococcus species. J. Bacteriol.
tion of LytSR-regulated genes from Staphylococcus 184:1180–1186.
aureus. J. Bacteriol. 178:5810–5816. 31. Dunman, P. M., E. Murphy, S. Haney, D. Pala-
19. Chamberlain, N. R., and B. Imanoel. 1996. cios, G. Tucker-Kellogg, S. Wu, E. L. Brown,
Genetic regulation of fatty acid modifying enzyme R. J. Zagursky, D. Shlaes, and S. J. Projan.
from Staphylococcus aureus. J. Med. Microbiol. 44: 2001. Transcription profiling-based identification
125–129. of Staphylococcus aureus genes regulated by the agr
20. Chan, P. F., and S. J. Foster. 1998. Role of SarA and/or sarA loci. J. Bacteriol. 183:7341–7353.
in virulence determinant production and environ- 32. Entenza, J. M., P. Moreillon, M. M. Senn, J.
mental signal transduction in Staphylococcus aureus J. Kormanec, P. M. Dunman, B. Berger-Bachi,
Bacteriol. 180:6232–6241. S. Projan, and M. Bischoff. 2005. Role of
21. Chan, P. F., and S. J. Foster. 1998.The role of sigmaB in the expression of Staphylococcus aureus
environmental factors in the regulation of viru- cell wall adhesins ClfA and FnbA and contribution
lence- determinant expression in Staphylococcus to infectivity in a rat model of experimental endo-
aureus 8325-4. Microbiology 144:2469–2479. carditis. Infect. Immun. 73:990–998.
22. Chan, P. F., S. J. Foster, E. Ingham, and M. O. 33. Foster, T. J., M. O’Reilly, P. Phonimdaeng, J.
Clements. 1998.The Staphylococcus aureus alterna- Cooney, A. H. Patel, and A. J. Bramley. 1990.
tive sigma factor sigmaB controls the environmen- Genetic studies of virulence factors of Staphylococ-
tal stress response but not starvation survival or cus aureus. Properties of coagulase and gamma-
pathogenicity in a mouse abscess model.J.Bacteriol. toxin, alpha-toxin, beta-toxin and protein A in the
180:6082–6089. pathogenesis of S. aureus infections, p. 403–420. In
23. Cheung,A., M. Bayer, and J. Heinrichs. 1997. R. P. Novick (ed.), Molecular Biology of the Staphylo-
sar genetic determinants necessary for transcrip- cocci.VCH Publishers, New York, NY.
tion of RNAII and RNAIII in the agr locus of 34. Fournier, B., and D. C. Hooper. 2000. A
Staphylococcus aureus. J. Bacteriol. 179:3963–3673. new two-component regulatory system involved
24. Cheung, A. L., J. M. Coomey, C. A. Butler, S. in adhesion, autolysis, and extracellular prote-
J. Projan, and V. A. Fischetti. 1992. Regulation olytic activity of Staphylococcus aureus. J. Bacteriol.
of exoprotein expression in Staphylococcus aureus by 182:3955–3964.
a locus (sar) distinct from agr. Proc. Natl. Acad. Sci. 35. Fournier, B., and A. Klier. 2004. Protein A gene
USA 89:6462–6466. expression is regulated by DNA supercoiling
25. Cheung, A. L., K. J. Eberhardt, E. Chung, M. which is modified by the ArlS-ArlR two-compo-
R.Yeaman, P. M. Sullam, M. Ramos, and A. S. nent system of Staphylococcus aureus. Microbiology
Bayer. 1994. Diminished virulence of a sar-/agr- 150:3807–3819.
mutant of Staphylococcus aureus in the rabbit model 36. Fournier, B.,A. Klier, and G. Rapoport. 2001.
of endocarditis. J. Clin. Invest. 94:1815–1822. The two-component system ArlS-ArlR is a regu-
26. Cheung,A. L., and G. Zhang. 2002.Global reg- lator of virulence gene expression in Staphylococcus
ulation of virulence determinants in Staphylococcus aureus. Mol. Microbiol. 41:247–261.
aureus by the SarA protein family. Front. Biosci. 37. Gao, J., and G. C. Stewart. 2004. Regulatory
1:D1825–D1842. elements of the Staphylococcus aureus protein A
27. Cheung, A. L., J. H. Heinrichs, and M. G. (Spa) promoter. J. Bacteriol. 186:3738–3748.
Bayer. 1996. Characterization of the sar locus and 38. Gaskill, M. E., and S. A. Khan. 1988. Regula-
its interaction with agr in Staphylococcus aureus. J. tion of the enterotoxin B gene in Staphylococcus
Bacteriol. 178:418–423. aureus. J. Biol. Chem. 263:6276–6280.
28. Chien, Y. 1999. SarA, a global regulator of viru- 39. Geisinger, E., R. P. Adhikari, R. Jin, H. F.
lence determinants in Staphylococcus aureus,binds to Ross, and R. P. Novick. 2006. Inhibition of rot
a conserved motif essential for sar-dependent gene translation by RNAIII, a key feature of agr func-
regulation. J. Biol. Chem. 274:37169–37176. tion. Mol. Microbiol. 61:1038–1048.
29. Diep, D. B., R. Myhre, O. Johnsborg, A. 40. Gemmel, C. G., and A. M. A. Shibl. 1976.The
Aakra, and I. F. Nes. 2003. Inducible bacteriocin control of toxin and enzyme biosynthesis in
production in Lactobacillus is regulated by differen- staphylococci by antibiotics, p. 657–664. In J.
tial expression of the pln operons and by two anto- Jeljaszewicz (ed.), Staphylococci and Staphylococal
gonizing response regulators, the activity of which Diseases. Gustav FischerVerlag,Stuttgart,Germany.
is enhanced upon phosphorylation. Mol. Microbiol. 41. Gertz, S., S. Engelmann, R. Schmid, A.-K.
47:483–494. Ziebandt, K. Tischer, C. Scharf, J. Hacker,
30. Dufour, P., S. Jarraud, F. Vandenesch, T. and M. Hecker. 2000. Characterization of the
Greenland, R. P. Novick, M. Bes, J. Etienne, sigma B regulon in Staphylococcus aureus. J. Bacteriol.
and G. Lina. 2002. High genetic variability of the 182:6983–6991.
42. Gillaspy, A. F., S. G. Hickmon, R. A. Skinner, Staphylococcus aureus RNAIII and the endoribonu-
J. R. Thomas, C. L. Nelson, and M. S. clease III coordinately regulate spa gene expres-
Smeltzer. 1995. Role of the accessory gene regu- sion. EMBO J. 24:824–835.
lator (agr) in pathogenesis of staphylococcal 53. Ingavale, S., W. van Wamel, T. T. Luong, C.Y.
osteomyelitis. Infect. Immun. 63:3373–3380. Lee, and A. L. Cheung. 2005.Rat/MgrA,a regu-
43. Gillet, Y., B. Issartel, P. Vanhems, J. C. lator of autolysis,is a regulator of virulence genes in
Fournet, G. Lina, M. Bes, F. Vandenesch, Y. Staphylococcus aureus. Infect. Immun. 73:1423–1431.
Piemont, N. Brousse, D. Floret, and J. Eti- 54. Janzon, L., and S. Arvidson. 1990.The role of
enne. 2002. Association between Staphylococcus the delta-lysin gene (hld) in the regulation of viru-
aureus strains carrying gene for Panton-Valentine lence genes by the accessory gene regulator (agr) in
leukocidin and highly lethal necrotising pneumo- Staphylococcus aureus . EMBO J. 9:1391–1399.
nia in young immunocompetent patients. Lancet 55. Jarraud, S., G. J. Lyon, A. M. S. Figueiredo, L.
359:753–759. Gerard, F.Vandenesch, J. Etienne,T.W. Muir,
44. Giraudo, A., G. Martinez, A. Calzolari, and and R. P. Novick. 2000. Exfoliatin-producing
R. Nagel. 1994. Characterization of a transposi- strains define a fourth agr specificity group in
tional mutant of Staphylococcus aureus underpro- Staphylococcus aureus. J. Bacteriol. 182:6517–6522.
ducing exoproteins. Rev. Latinoam Microbiol. 56. Jarraud, S., C. Mougel, J.Thioulouse, G. Lina,
36:171–176. H. Meugnier, F. Forey, X. Nesme, J. Etienne,
45. Giraudo,A., C. Raspanti,A. Calzolari, and R. and F.Vandenesch. 2002. Relationships between
Nagel. 1994. Characterization of a Tn551-mutant Staphylococcus aureus genetic background,virulence
of Staphlococcus aureus defective in the production of factors, agr groups (alleles), and human disease.
several exoproteins. Can. J. Microbiol. 40:677–681. Infect. Immun. 70:631–641.
46. Giraudo, A.T. 1997.The sae locus of Staphylococ- 57. Ji, G., R. Beavis, and R. P. Novick. 1997. Bac-
cus aureus controls exoprotein synthesis at the tran- terial interference caused by autoinducing peptide
scriptional level. Arch. Microbiol. 168:53–58. variants. Science 276:2027–2030.
47. Giraudo, A.T., A. Calzolari, A. A. Cataldi, C. 58. Ji, G., R. C. Beavis, and R. P. Novick. 1995.
Bogni, and R. Nagel. 1999. The sae locus of Cell density control of staphylococcal virulence
Staphylococcus aureus encodes a two-component mediated by an octapeptide pheromone.Proc.Natl.
regulatory system. FEMS Microbiol. Lett. Acad. Sci. USA 92:12055–12059.
177:15–22. 59. Ji, G., W. Pei, L. Zhang, R. Qiu, J. Lin, Y.
48. Goerke, C., U. Fluckiger, A. Steinhuber, V. Benito, G. Lina, and R. P. Novick. 2005.
Bisanzio, M. Ulrich, M. Bischoff, J. M. Patti, taphylococcus intermedius produces a functional agr
and C. Wolz. 2005. Role of Staphylococcus aureus autoinducing peptide containing a cyclic lactone.
global regulators sae and sigmaB in virulence gene J. Bacteriol. 187:3139–3150.
expression during device-related infection. Infect. 60. Jonsson, I. M., S. Arvidson, S. Foster, and
Immun. 73:3415–3421. A. Tarkowski. 2004. Sigma factor B and RsbU
49. Haslinger-Loffler, B., B. C. Kahl, M. are required for virulence in Staphylococcus
Grundmeier, K. Strangfeld, B.Wagner, U. Fis- aureus-induced arthritis and sepsis. Infect. Immun.
cher, A. L. Cheung, G. Peters, K. Schulze- 72:6106–6111.
Osthoff, and B. Sinha. 2005. Multiple virulence 61. Kahl, B. C., M. Goulian, W. van Wamel, M.
factors are required for Staphylococcus aureus- Herrmann, S. M. Simon, G. Kaplan, G.
induced apoptosis in endothelial cells. Cell Micro- Peters, and A. L. Cheung. 2000. Staphylococcus
biol 7:1087–1097. aureus RN6390 replicates and induces apoptosis in
50. Herbert, S., P. Barry, and R. P. Novick. 2001. a pulmonary epithelial cell line. Infect. Immun.
Subinhibitory clindamycin differentially inhibits 68:5385–5392.
transcription of exoprotein genes in Staphylococcus 62. Kalkum, M., G. J. Lyon, and B. T. Chait.
aureus. Infect. Immun. 69:2996–3003. 2003. Detection of secreted peptides by using
51. Horsburgh, M. J., J. L. Aish, I. J. White, L. hypothesis-driven multistage mass spectrometry.
Shaw, J. K. Lithgow, and S. J. Foster. 2002. Proc. Natl.Acad. Sci. USA 100:2795–2800.
sigma(B) modulates virulence determinant 63. Karlsson-Kanth, A., K. Tegmark-Wisell, S.
expression and stress resistance: characterization of Arvidson, and J. Oscarsson. 2006. Natural
a functional rsbU strain derived from Staphylococcus human isolates of Staphylococcus aureus selected for
aureus 8325–4. J. Bacteriol. 184:5457–5467. high production of proteases and alpha-hemolysin
52. Huntzinger, E., S. Boisset, C. Saveanu, Y. are sigmaB deficient. Int. J. Med. Microbiol
Benito,T. Geissmann, A. Namane, G. Lina, J. 296:229–236.
Etienne, B. Ehresmann, C. Ehresmann, A. 64. Kavenoff, R., and B. Bowen. 1976. Elec-
Jacquier, F.Vandenesch, and P. Romby. 2005. tron microscopy of membrane-free folded chro-
mosomes from Escherichia coli. Chromosoma inhibition to the receptor-histidine kinase, AgrC.
59:89–101. Proc. Natl.Acad. Sci. USA 97:13330–13335.
65. Koenig, R. L., J. L. Ray, S. J. Maleki, M. S. 77. Lyon, G. J., J. S.Wright, A. Christopoulos, R.
Smeltzer, and B. K. Hurlburt. 2004. Staphylo- P. Novick, and T. W. Muir. 2002. Reversible
coccus aureus AgrA binding to the RNAIII-agr reg- and specific extracellular antagonism of receptor
ulatory region. J. Bacteriol. 186:7549–7555. histidine-kinase signaling. J. Biol. Chem. 277:
66. Kullik, I., P. Giachino, and T. Fuchs. 1998. 6247–6253.
Deletion of the alternative sigma factor sigmaB 78. Lyon, G. J., J. S.Wright,T.W. Muir, and R. P.
in Staphylococcus aureus reveals its function as a Novick. 2002. Key determinants of receptor acti-
global regulator of virulence genes. J. Bacteriol. vation in the agr autoinducing peptides of Staphylo-
180:4814–4820. coccus aureus. Biochemistry 41:10095–10104.
67. Labandeira-Rey, M., F. Couzon, S. Boisset, E. 79. Manna, A., and A. L. Cheung. 2001. Charac-
L. Brown, M. Bes,Y. Benito, E. M. Barbu,V. terization of sarR, a modulator of sar expression in
Vazquez, M. Hook, J. Etienne, F.Vandenesch, Staphylococcus aureus. Infect Immun 69:885–896.
and M. G. Bowden. 2007. Staphylococcus aureus 80. Manna,A. C., and A. L. Cheung. 2006. Expres-
Panton-Valentine leukocidin causes necrotizing sion of SarX, a negative regulator of agr and exo-
pneumonia. Science 315:1130–1133. protein synthesis, is activated by MgrA in
68. Lebeau, C., F.Vandenesch,T. Greenland, R. P. Staphylococcus aureus. J. Bacteriol. 188:4288–4299.
Novick, and J. Etienne. 1994. Coagulase expres- 81. Manna, A. C., and A. L. Cheung. 2003. sarU, a
sion in Staphylococcus aureus is positively and nega- sarA homolog, is repressed by SarT and regulates
tively modulated by an agr-dependent mechanism. virulence genes in Staphylococcus aureus. Infect.
J. Bacteriol. 176:5534–5536. Immun. 71:343–53.
69. Lenz, D. H., K. C. Mok, B. N. Lilley, R. V. 82. Manna, A. C., and A. L. Cheung. 2006.Tran-
Kulkarni, N. S. Wingreen, and B. L. Bassler. scriptional regulation of the agr locus and the iden-
2004.The small RNA chaperone Hfq and multiple tification of DNA binding residues of the global
small RNAs control quorum sensing in Vibrio har- regulatory protein SarR in Staphylococcus aureus.
veyi and Vibrio cholerae. Cell 118:69–82. Mol. Microbiol. 60:1289–1301.
70. Liang, X., C.Yu, J. Sun, H. Liu, C. Landwehr, 83. Manna, A. C., S. S. Ingavale, M. Maloney, W.
D. Holmes, and Y. Ji. 2006.Inactivation of a two- van Wamel, and A. L. Cheung. 2004.Identifica-
component signal transduction system, SaeRS, tion of sarV (SA2062), a new transcriptional regu-
eliminates adherence and attenuates virulence of lator, is repressed by SarA and MgrA (SA0641) and
Staphylococcus aureus. Infect. Immun. 74:4655–4665. involved in the regulation of autolysis in Staphylo-
71. Liang, X., L. Zheng, C. Landwehr, D. coccus aureus. J. Bacteriol. 186:5267–80.
Lunsford, D. Holmes, and Y. Ji. 2005. Global 84. Manna,A. C., and B. Ray. 2007.Regulation and
regulation of gene expression by ArlRS, a two- characterization of rot transcription in Staphylococ-
component signal transduction regulatory system cus aureus. Microbiology 153:1538–1545.
of Staphylococcus aureus. J. Bacteriol. 187:5486–5492. 85. Martin, P. K.,T. Li, D. Sun, D. P. Biek, and M.
72. Lina, G., S. Jarraud, G. Ji, T. Greenland, A. B. Schmid. 1999. Role in cell permeability of an
Pedraza, J. Etienne, R. P. Novick, and F. essential two-component system in Staphylococcus
Vandenesch. 1998.Transmembrane topology and aureus. J Bacteriol 181:3666–3673.
histidine protein kinase activity of AgrC, the agr 86. Mayville, P., G. Ji, R. Beavis, H. Yang, M.
signal receptor in Staphylococcus aureus. Mol. Micro- Goger, R. P. Novick, and T. W. Muir. 1999.
biol. 28:655–662. Structure-activity analysis of synthetic autoinduc-
73. Luong,T.T., P. M. Dunman, E. Murphy, S. J. ing thiolactone peptides from Staphylococcus aureus
Projan, and C.Y. Lee. 2006.Transcription profil- responsible for virulence. Proc. Natl.Acad. Sci. USA
ing of the mgrA regulon in Staphylococcus aureus. J. 96:1218–1223.
Bacteriol. 188:1899–1910. 87. McAleese, F. M., E. J.Walsh, M. Sieprawska, J.
74. Luong, T. T., and C.Y. Lee. 2006.The arl locus Potempa, and T. J. Foster. 2001. Loss of clump-
positively regulates Staphylococcus aureus type 5 cap- ing factor B fibrinogen binding activity by Staphy-
sule via an mgrA-dependent pathway. Microbiology lococcus aureus involves cessation of transcription,
152:3123–3131. shedding and cleavage by metalloprotease. J. Biol.
75. Luong,T.T., S.W. Newell, and C.Y. Lee. 2003. Chem. 276:29969–29978.
Mgr, a novel global regulator in Staphylococcus 88. McCallum, N., M. Bischoff, H. Maki, A.
aureus. J. Bacteriol. 185:3703–3710. Wada, and B. Berger-Bachi. 2004. TcaR, a
76. Lyon, G. J., P. Mayville, T. W. Muir, and R. P. putative MarR-like regulator of sarS expression. J.
Novick. 2000. Rational design of a global Bacteriol. 186:2966–2972.
inhibitor of the virulence response in Staphylococcus 89. McDowell, P., Z. Affas, C. Reynolds, M. T.
aureus, based in part on localization of the site of Holden, S. J.Wood, S. Saint,A. Cockayne, P. J.
Hill, C. E. Dodd, B.W. Bycroft,W. C. Chan, assimilation regulatory genes ntrB and ntrC. Proc.
and P. Williams. 2001. Structure, activity and Natl.Acad. Sci. USA 83:7850–7854.
evolution of the group I thiolactone peptide 101. Novick, R. P., and D. Jiang. 2003.The staphy-
quorum-sensing system of Staphylococcus aureus. lococcal saeRS system coordinates environmen-
Mol. Microbiol. 41:503–512. tal signals with agr quorum sensing. Microbiology
90. McNamara, P. J., K. C. Milligan-Monroe, 149:2709–2717.
S. Khalili, and R.A. Proctor. 2000. Identifica- 102. Novick, R. P., S. J. Projan, J. Kornblum, H. F.
tion, cloning, and initial characterization of rot, a Ross, G. Ji, B. Kreiswirth, F.Vandenesch, and
locus encoding a regulator of virulence factor S. Moghazeh. 1995.The agr P2 operon:an auto-
expression in Staphylococcus aureus. J. Bacteriol. catalytic sensory transduction system in Staphylo-
182:3197–3203. coccus aureus. Mol. Gen. Genet. 248:446–458.
91. McNamara, P. J., and R. A. Proctor. 2000. 103. Novick, R. P., H. F. Ross, S. J. Projan, J.
Staphylococcus aureus small colony variants, elec- Kornblum, B. Kreiswirth, and S. Moghazeh.
tron transport and persistent infections. Int. J. 1993. Synthesis of staphylococcal virulence fac-
Antimicrob.Agents. 14:117–122. tors is controlled by a regulatory RNA molecule.
92. Morfeldt, E. 1995. Activation of alpha-toxin EMBO J. 12:3967–3975.
translation in Staphylococcus aureus by the trans- 104. Ohlendorf, D. H., C. A. Earhart, D. T.
encoded antisense RNA, RNAIII. EMBO J. 14: Mitchell, G. M.Vath, M. Roggiani, J.V. Rago,
4569–4577. M. H. Kim, G.A. Bohach, and P. M. Schliev-
93. Morfeldt, E., L. Janzon, S. Arvidson, and S. ert. 1998. Structural biology of toxins associated
Lofdahl. 1988. Cloning of a chromosomal locus with TSS. Int. Congr. Symp. Series 229:89–91.
(exp) which regulates the expression of several 105. Oscarsson, J., K. Tegmark-Wisell, and S.
exoprotein genes in Staphylococcus aureus. Mol. Arvidson. 2006. Coordinated and differential
Gen. Genet. 211:435–440. control of aureolysin (aur) and serine protease
94. Morfeldt, E., D.Taylor,A. von Gabain, and S. (sspA) transcription in Staphylococcus aureus by
Arvidson. 1995.Activation of alpha-toxin trans- sarA, rot and agr (RNAIII). Int. J. Med. Microbiol.
lation in Staphylococcus aureus by the trans-encoded 296:365–380.
antisense RNA, RNAIII. EMBO J. 14:4569– 106. Otto, M., R. Sussmuth, C. Vuong, G. Jung,
4577. and F. Gotz. 1999.Inhibition of virulence factor
95. Morfeldt, E., K. Tegmark, and S. Arvidson. expression in Staphylococcus aureus by the Staphylo-
1996.Transcriptional control of the agr-depend- coccus epidermidis agr pheromone and derivatives.
ent virulence gene regulator, RNAIII, in Staphy- FEBS Lett. 450:257–262.
lococcus aureus. Mol. Microbiol. 21:1227–1237. 107. Palma, M., A. Bayer, L. I. Kupferwasser, T.
96. Morikawa, K., Y. Inose, H. Okamura, A. Joska, M. R.Yeaman, and A. Cheung. 2006.
Maruyama, H. Hayashi, K.Takeyasu, and T. Salicylic acid activates sigma factor B by rsbU-
Ohta. 2003. A new staphylococcal sigma factor dependent and -independent mechanisms. J.
in the conserved gene cassette: functional signifi- Bacteriol. 188:5896–5903.
cance and implication for the evolutionary 108. Papakyriacou, H., D. Vaz, A. Simor, M.
processes. Genes Cells 8:699–712. Louie, and M. J. McGavin. 2000. Molecular
97. Nair, S. P., M. Bischoff, M. M. Senn, and B. analysis of the accessory gene regulator (agr) locus
Berger-Bachi. 2003. The sigma B regulon and balance of virulence factor expression in epi-
influences internalization of Staphylococcus aureus demic methicillin-resistant Staphylococcus aureus.
by osteoblasts. Infect. Immun. 71:4167–4170. J. Infect. Dis. 181:990–1000.
98. Nakayama, J.,Y. Cao,T. Horii, S. Sakuda, A. 109. Peng, H.-L., R. P. Novick, B. Kreiswirth, J.
D. Akkermans, W. M. de Vos, and H. Naga- Kornblum, and P. Schlievert. 1988. Cloning,
sawa. 2001. Gelatinase biosynthesis-activating characterization and sequencing of an accessory
pheromone: a peptide lactone that mediates a gene regulator (agr) in Staphylococcus aureus J.
quorum sensing in Enterococcus faecalis. Mol. Micro- Bacteriol. 179:4365–4372.
biol. 41:145–154. 110. Pohlmann-Dietze, P., M. Ulrich, K. B. Kiser,
99. Nicholas, R. O., T. Li, D. McDevitt, A. G. Doring, J. C. Lee, J.-M. Fournier, K.
Marra, S. Sucoloski, P. L. Demarsh, and D. Botzenhart, and C.Wolz. 2000.Adherence of
R. Gentry. 1999. Isolation and characterization Staphylococcus aureus to endothelial cells: influence
of a sigB deletion mutant of Staphylococcus aureus. of capsular polysaccharide, global regulator agr,
Infect. Immun. 67:3667–3669. and bacterial growth phase. Infect. Immun.
100. Nixon, B. C., C. W. Ronson, and F. M. 68:4865–4871.
Ausubel. 1986.Two-component regulatory sys- 111. Pragman, A. A., Y. Ji, and P. M. Schlievert.
tems responsive to environmental stimuli share 2007. Repression of Staphylococcus aureus SrrAB
strongly conserved domains with the nitrogen using inducible antisense srrA alters growth and
virulence factor transcript levels. Biochemistry 123. Saenz, H. L.,V. Augsburger, C.Vuong, R.W.
46:314–3121. Jack, F. Gotz, and M. Otto. 2000. Inducible
112. Pragman, A. A., J. M. Yarwood, T. J. Tripp, expression and cellular location of AgrB,a protein
and P. M. Schlievert. 2004. Characterization involved in the maturation of the staphylococcal
of virulence factor regulation by SrrAB, a two- quorum-sensing pheromone. Arch. Microbiol.
component system in Staphylococcus aureus. J. Bac- 174:452–455.
teriol. 186:2430–2438. 124. Said-Salim, B., P. M. Dunman, F. M.
113. Qazi, S. N., E. Counil, J. Morrissey, C. E. McAleese, D. Macapagal, E. Murphy, P. J.
Rees,A. Cockayne, K.Winzer,W. C. Chan, P. McNamara, S. Arvidson, T. J. Foster, S. J.
Williams, and P. J. Hill. 2001. agr expression Projan, and B. N. Kreisworth. 2003. Global
precedes escape of internalized Staphylococcus regulation of Staphylococcus aureus genes by rot. J.
aureus from the host endosome. Infect. Immun. Bacteriol. 185:610–619.
69:7074–7082. 125. Sakoulas, G., G. M. Eliopoulos, R. C. J.
114. Qin, X., K.V. Singh, G. M.Weinstock, and B. Moellering, C. Wennersten, L. Venkatara-
E. Murray. 2001. Characterization of fsr, a regu- man, R. P. Novick, and H. S. Gold.
lator controlling expression of gelatinase and ser- 2002. Accessory gene regulator (agr) locus in
ine protease in Enterococcus faecalis OG1RF. J. geographically diverse Staphylococcus aureus iso-
Bacteriol. 183:3372–3382. lates with reduced susceptibility to vancomycin.
115. Qiu, R., W. Pei, L. Zhang, J. Lin, and G. Ji. Antimicrob.Agents Chemother. 46:1492–1502.
2005. Identification of the putative staphylococ- 126. Saravia-Otten, P., H. Muller, and S. Arvid-
cal AgrB catalytic residues involving the prote- son. 1997. Transcription of Staphylococcus aureus
olytic cleavage of AgrD to generate autoinducing fibronectin binding protein genes is negatively
peptide. J. Biol. Chem. 280:16695–16704. regulated by agr and an agr-independent mecha-
116. Rampone, H. 1996. In vivo expression of nism. J. Bacteriol. 179:5259–3673.
exoprotein synthesis with a Sae mutant of Staphy- 127. Schmidt, K. A. 2001. SarT, a repressor of alpha-
lococcus aureus. Canad J.Vet. Res. 60:237–240. hemolysin in Staphylococcus aureus. Infect. Immun.
117. Rechtin, T. M., A. F. Gillaspy, M. A. 69:4749–4758.
Schumacher, R. G. Brennan, M. S. 128. Schmidt, K. A., A. C. Manna, S. Gill, and A.
Smeltzer, and B. K. Hurlburt. 1999. Charac- L. Cheung. 2001. SarT, a repressor of alpha-
terization of the SarA virulence gene regulator of hemolysin in Staphylococcus aureus. Infect. Immun.
Staphylococcus aureus. Mol. Microbiol. 33:307–316. 69:4749–4758.
118. Recsei, P., B. Kreiswirth, M. O’Reilly, P. 129. Scott, J. M., N. Smirnova, and W. G. Halden-
Schlievert, A. Gruss, and R. Novick. 1986. wang. 1999.A Bacillus-specific factor is needed
Regulation of exoprotein gene expression by agr. to trigger the stress-activated phosphatase/kinase
Mol. Gen. Genet. 202:58–61. cascade of sigmaB induction. Biochem. Biophys.
119. Recsei, P., B. Kreiswirth, M. O’Reilly, P. Res. Commun. 257:106–110.
Schlievert,A. Gruss, and R. P. Novick. 1986. 130. Sheehan, B. J. 1992. The regulation of epider-
Regulation of exoprotein gene expression in molytic toxin A expression in Staphylococcus aureus.
Staphylococcus aureus by agr. Mol. Gen. Genet. Ph.D. dissertation. University of Dublin, Ireland.
202:58–61. 131. Shompole, S., K. T. Henon, L. E. Liou, K.
120. Regassa, L. B., J. L. Couch, and M. J. Betley. Dziewanowska, G. A. Bohach, and K. W.
1991. Steady-state staphylococcal enterotoxin Bayles. 2003.Biphasic intracellular expression of
type C mRNA is affected by a product of the Staphylococcus aureus virulence factors and evi-
accessory gene regulator (\fIagr\fP) and by glu- dence for Agr-mediated diffusion sensing. Mol.
cose. Infect. Immun. 59:955–962. Microbiol. 49:919–27.
121. Risoen, P. A., M. B. Brurberg,V. G. Eijsink, 132. Smeltzer, M. S., M. E. Hart, and J. J. Iandolo.
and I. F. Nes. 2000. Functional analysis of pro- 1993. Phenotypic characterization of xpr, a global
moters involved in quorum sensing-based regu- regulator of extracellular virulence factors in
lation of bacteriocin production in Lactobacillus. Staphylococcus aureus. Infect. Immun. 61:919–925.
Mol. Microbiol. 37:619–628. 133. Somerville, G. A., S. B. Beres, J. R. Fitzger-
122. Rogasch, K.,V. Ruhmling, J. Pane-Farre, D. ald, F. R. DeLeo, R. L. Cole, J. S. Hoff, and J.
Hoper, C.Weinberg, S. Fuchs, M. Schmudde, M. Musser. 2002. In vitro serial passage of
B. M. Broker, C. Wolz, M. Hecker, and Staphylococcus aureus: changes in physiology, viru-
S. Engelmann. 2006. Influence of the two- lence factor production, and agr nucleotide
component system SaeRS on global gene expres- sequence. J. Bacteriol. 184:1430–1437.
sion in two different Staphylococcus aureus strains. J. 134. Somerville, G. A., M. S. Chaussee, C. I.
Bacteriol. 188:7742–7758. Morgan, J. R. Fitzgerald, D.W. Dorward, L.
J. Reitzer, and J. M. Musser. 2002. Staphylococ- 145. Vuong, C., H. L. Saenz, F. Gotz, and M.
cus aureus aconitase inactivation unexpectedly Otto. 2000. Impact of the agr quorum-sensing
inhibits post-exponential-phase growth and system on adherence to polystyrene in Staphylo-
enhances stationary-phase survival. Infect. Immun. coccus aureus. J. Infect. Dis. 182:1688–1693.
70:6373–6382. 146. Watson, S. P., M. O. Clements, and S. J.
135. Sturme, M. H., J. Nakayama, D. Molenaar,Y. Foster. 1998. Characterization of the starvation-
Murakami, R. Kunugi, T. Fujii, E. E. survival response of Staphylococcus aureus. J. Bacte-
Vaughan, M. Kleerebezem, and W. M. de riol. 180:1750–1758.
Vos. 2005. An agr-like two-component regula- 147. Weinrick, B., P. M. Dunman, F. McAleese, E.
tory system in Lactobacillus plantarum is involved Murphy, S. J. Projan, Y. Fang, and R. P.
in production of a novel cyclic peptide and regu- Novick. 2004. Effect of mild acid on gene
lation of adherence. J. Bacteriol. 187:5224–5235. expression in Staphylococcus aureus. J. Bacteriol.
136. Tegmark, K., A. Karlsson, and S. Arvidson. 186:8407–8423.
2000. Identification and characterization of 148. Wesson, C. A., L. E. Liou, K. M.Todd, G. A.
SarH1, a new global regulator of virulence gene Bohach, W. R. Trumble, and K. W. Bayles.
expression in Staphylococcus aureus. Mol. Microbiol. 1998. Staphylococcus aureus Agr and Sar global reg-
37:398–409. ulators influence internalization and induction of
137. Tegmark, K., E. Morfeldt, and S. Arvidson. apoptosis. Infect. Immun. 66:5238–52343.
1998. Regulation of agr-dependent virulence 149. Wolz, C., D. McDevitt,T. J. Foster, and A. L.
genes in Staphylococcus aureus by RNAIII from Cheung. 1996. Influence of agr on fibrinogen
coagulase-negative staphylococci. J. Bacteriol. binding in Staphylococcus aureus Newman. Infect
180:3181–3186. Immun. 64:3142–7.
138. Throup, J. P., F. Zappacosta, R. D. Lunsford, 150. Wright III, J. S., R. Jin, and R. P. Novick.
R. S. Annan, S. A. Carr, J. T. Lonsdale, 2005.Transient interference with staphylococcal
A. P. Bryant, D. McDevitt, M. Rosenberg, quorum sensing blocks abscess formation. Proc.
and M. K. Burnham. 2001. The srhSR gene Natl.Acad. Sci. USA 102:1691–1696.
pair from Staphylococcus aureus: genomic and 151. Wright, J. S., III, K. E.Traber, R. Corrigan, S.
proteomic approaches to the identification and A. Benson, J. M. Musser, and R. P. Novick.
characterization of gene function. Biochemistry 2005.The agr radiation:an early event in the evolu-
40:10392–10401. tion of staphylococci. J. Bacteriol. 187:5585–5594.
139. Tortosa, P., L. Logsdon, B. Kraigher,Y. Itoh, 152. Wright, J. S., III, G. J. Lyon, E. A. George,T.
I. Mandic-Mulec, and D. Dubnau. 2001. W. Muir, and R. P. Novick. 2004.Hydrophobic
Specificity and genetic polymorphism of the interactions drive ligand-receptor recognition
Bacillus competence quorum-sensing system. J. for activation and inhibition of staphylococcal
Bacteriol. 183:451–460. quorum sensing. Proc. Natl. Acad. Sci.USA
140. Tremaine, M. T., D. K. Brockman, and 101:16168–16173.
M. J. Betley. 1993. Staphylococcal enterotoxin 153. Wright, J. S. I., K. E. Traber, R. Corrigan,
A gene (sea) expression is not affected by the S. A. Benson, J. M. Musser, and R. P.
accessory gene regulator (agr). Infect. Immun. 61: Novick. 2005.The agr radiation:an early event in
356–359. the evolution of staphylococci. J. Bacteriol. 187:
141. Tseng, C. W., and G. C. Stewart. 2005. Rot 5585–5594.
repression of enterotoxin B expression in Staphy- 154. Yarwood, J. M., D. J. Bartels, E. M. Volper,
lococcus aureus. J. Bacteriol. 187:5301–5309. and E. P. Greenberg. 2004. Quorum sensing
142. Vandenesch, F. 1991. A temporal signal, inde- in Staphylococcus aureus biofilms. J. Bacteriol.
pendent of agr, is required for hla but not spa tran- 186:1838–1850.
scription in Staphylococcus aureus. J. Bacteriol. 155. Yarwood, J. M., J. K. McCormick, and P. M.
173:6313–6320. Schlievert. 2001. Identification of a novel two-
143. Van Wamel,W. J., G. van Rossum, J.Verhoef, component regulatory system that acts in global
C. M. Vandenbroucke-Grauls, and A. C. regulation of virulence factors of Staphylococcus
Fluit. 1998. Cloning and characterization of an aureus. J. Bacteriol. 183:1113–1123.
accessory gene regulator (agr)-like locus from 156. Zhang, L., L. Gray, R. P. Novick, and G. Ji.
Staphylococcus epidermidis. FEMS Microbiol. Lett 2002.Transmembrane topology of AgrB,the pro-
163:1–9. tein i nvolved in the post-translational modifica-
144. Vojtov, N., H. F. Ross, and R. P. Novick. tion of AgrD in Staphylococcus aureus.J.Biol.Chem.
2002. Global repression of exotoxin synthesis by 277:34736–34742.
staphylococcal superantigens. Proc. Natl.Acad. Sci. 157. Zhang, L., and G. Ji. 2004. Identification of a
USA 99:10102–10107. staphylococcal AgrB segment(s) responsible for
group-specific processing of AgrD by gene swap- in Staphylococcus aureus. J. Biol. Chem. 279:
ping. J. Bacteriol. 186:6706–6713. 19448–19456.
158. Zhang, L., J. Lin, and G. Ji. 2004. Mem- 159. Zhang, S., and G. C. Stewart. 2000. Charac-
brane anchoring of the AgrD N-terminal terization of the promoter elements for the
amphipathic region is required for its process- staphylococcal enterotoxin D gene. J. Bacteriol.
ing to produce a quorum-sensing pheromone 182:2321–2325.
QUORUM SENSING IN THE
SOFT-ROT ERWINIAS
12
The soft-rotting erwinias, Erwinia carotovora mation to regulate gene expression accord-
subsp. carotovora and Erwinia carotovora subsp. ingly. A wide range of important processes,
atroseptica, are gram-negative, enteric plant in diverse bacterial species, is regulated by
pathogens that cause economically significant QS, including virulence, antibiotic produc-
crop losses. E. carotovora subsp. atroseptica has a tion, symbiosis, sporulation, and biofilm forma-
narrow host range, causing blackleg (stem rot) tion (62). In gram-negative bacteria, the most
and tuber rot of potatoes, whereas E. cartovora extensively studied QS systems are those
subsp. carotova has a wider host range, including using N-acylhomoserine lactone (AHL) signal
potato, carrot, and celery (54). The soft-rot molecules.The first such system to be described
erwinias are characterized by the production of was the Lux system of Vibrio fischeri, where the
a range of plant cell wall-degrading enzymes LuxI protein synthesizes the AHL signal, N-
(PCWDEs), resulting in general plant tissue (3-oxo-hexanoyl)-L-homoserine lactone (3-
maceration (soft-rot disease). As reviewed (40, oxo-C6-HSL). At high cell densities, when the
53, 54), the primary virulence mechanism of E. V. fischeri population is confined inside the light
carotova subspp. carotovara and atroseptica is the organ of the squid symbiotic host, a sufficient
coordinated production of high levels of concentration of 3-oxo-C6-HSL is reached
secreted PCWDEs, principally pectinases.The for it to bind to the transcriptional regulator,
most important PCWDEs are the multiple LuxR, which is then able to activate transcrip-
pectate lyases (Pels); other secreted PCWDEs tion of the lux operon, producing light.
include polygalacturonases (Peh) and endo- Many such AHL QS systems have now been
glucanase (cellulase, Cel). described, in which different LuxI homologues
Quorum sensing (QS) is a process of inter- synthesize various AHL signals and LuxR-type
cellular communication by which bacteria transcriptional regulatory proteins bind their
detect their population cell density through cognate signal at high cell densities and then
diffusible signal molecules and use this infor- alter gene expression appropriately (27, 62).
One of the first bacterial species for which
AHL QS was described was Erwinia carotovora
Department of Biochemistry, University of Cambridge,Tennis (2). Since then, QS has been well studied in the
Court Road, Cambridge CB2 1QW, United Kingdom. soft-rot erwinias, where, as described in this
185
186 ■ COULTHURST ET AL.
review, QS plays a key role in the regulation of HSL, and CarR binds 3-oxo-C6-HSL at high
secreted PCWDE production and hence in vir- cell density and activates expression of the
ulence. In certain strains, a well-defined AHL carA-H genes (32, 52, 61). Car production and
QS system also controls production of a - expression of the carA-H operon are absolutely
lactam antibiotic, carbapenem. While the QS dependent on QS, being undetectable in carI
system controlling carbapenem production is or carR mutants and being increased, and tem-
relatively simple and typical, that controlling porally advanced, when excess endogenous 3-
production of secreted virulence factors has oxo-C6-HSL is added at low cell density (33).
proven more complex and harder to unravel. This contrasts with many other QS-dependent
phenotypes, including secreted enzyme pro-
REGULATION OF CARBAPENEM
ANTIBIOTIC PRODUCTION duction in E. carotovora subsp. carotovora ATCC
IN E. CAROTOVORA SUBSP. 39048, where multiple other regulatory inputs
CAROTOVORA ATCC 39048 are also required, and hence growth phase
Production of the simple -lactam antibiotic, dependence is maintained even in the absence
carbapenem (1-carbapen-2-em-3-carboxylic of QS or upon the premature addition of AHL
acid,Car) by E.carotovora subsp.carotovora ATCC (48, 50, 63).
39048 is one of the earliest-identified and best- QS regulation of Car production in E. caro-
studied AHL QS-dependent phenotypes. Car tovora subsp. carotovora ATCC 39048 is now rea-
production by E. carotovora subsp. carotovora sonably well understood at the molecular level
ATCC 39048 was one of the first phenotypes and is depicted in Fig. 1. The gene encoding
found to be dependent on AHL QS outside of CarR, carR, is located just upstream of and tran-
the marine vibrios, providing the first indica- scribed in the same direction as carA-H,
tion that AHL QS was not simply a curiosity of whereas carI is genetically unlinked (33). Syn-
V.fischeri and related organisms,but in fact might thesis of 3-oxo-C6-HSL by CarI causes accu-
represent a widespread mechanism of bacterial mulation of the signal in the culture
regulation (2). Biosynthesis of Car and intrinsic supernatant throughout growth.At a threshold
resistance to the antibiotic are encoded by the concentration of signal (~0.1 g/ml), tran-
car biosynthetic operon, carABCDEFGH (34, scription of carA-H is initiated and Car produc-
35), depicted schematically in Fig. 1. The tion is observed, during the late-exponential
biosynthesis of Car has been reviewed in some and stationary phases of growth in laboratory
detail (13). Briefly, CarA, CarB, and CarC culture (2, 33). This cell-density-dependent
represent the core biosynthetic enzymes, being induction of Car expression is mediated via
necessary and sufficient to synthesize the car- CarR. CarR is a transcriptional activator that
bapenem from cellular precursors, and CarD specifically binds the carR-carA intergenic
and E are believed to be involved in precursor region, even in the absence of ligand. In the
synthesis.CarF and G together provide intrinsic presence of 3-oxo-C6-HSL, CarR multimer-
resistance to Car by a novel mechanism, yet izes (from a dimeric state in the absence of
to be elucidated,whereas the role of CarH is still ligand) and forms high-molecular-weight
unclear. complexes with its target DNA, hence activat-
Regulation of Car production in E. caro- ing carA-H transcription (61). CarR binds 3-
tovora subsp.carotovora ATCC 39048 represents a oxo-C6-HSL with a stoichiometry of two
classical, relatively simple, and well-defined molecules of ligand per dimer of CarR and a
example of AHL QS control.Production of Car dissociation constant of 1.8 M. Use of a series
is dependent on the LuxI family AHL synthase, of different AHLs confirmed that the affinity of
CarI (also known as ExpI in other strains of E. a particular AHL for binding CarR is directly
carotovora, see below), and the LuxR family proportional to its ability to facilitate DNA
transcriptional regulator, CarR. CarI produces binding by CarR (61). In addition, CarR has
the major AHL signal molecule, 3-oxo-C6- been shown to bind a series of nonhydrolyzable
FIGURE 1 Regulation of carbapenem antibiotic (Car) production by AHL QS in E. carotovora subsp. carotovora
ATCC 39048. The signaling molecule, 3-oxo-C6-HSL, is synthesized by CarI. At low cell densities (top), 3-oxo-C6-
HSL diffuses away from the cell and CarR is in a transcriptionally inactive state.The carA-H gene cluster is not tran-
scribed from the QS-dependent promoter (PQS) upstream of carA; hence, no antibiotic is produced, but the resistance
functions, encoded by carFG, are expressed from the internal promoter (Pint).At high cell densities (bottom), a high
concentration of 3-oxo-C6-HSL is achieved and CarR binds 3-oxo-C6-HSL, making it competent to activate tran-
scription of the carA-H genes from PQS and also to upregulate its own transcription.As a result, Car antibiotic is pro-
duced. The regulator Hor is also required for carA-H expression and other regulatory inputs are indicated: carI
expression is dependent on carbon source and possibly also downregulated in the presence of high 3-oxo-C6-HSL
levels; temperature affects hor transcription and probably also affects carI expression on a posttranscriptional level.
Arrows with “” indicate positive regulation, and flattened arrowhead indicates repression. Refer to text and refer-
ences 13 and 33 for details.
187
AHL analogues with varying efficiencies in tionally on CarI, since 3-oxo-C6-HSL levels,
vitro, correlating well with their relative abili- but not carI transcription, are reduced at 37C
ties to activate Car production in vivo (60).The (33). Anaerobiosis also eliminates Car produc-
CarR-3-oxo-C6-HSL-dependent promoter tion, but regulatory effects are hard to disentan-
upstream of carA has now been mapped, as has gle from the severe growth rate diminution
the promoter of carR (33). Expression of carR is manifested under this condition (33).
itself activated by CarR-3-oxo-C6-HSL, gen- From a biological perspective, why should
erating a positive feedback loop in response to Car production in E. carotovora subsp. carotovora
3-oxo-C6-HSL. On the other hand, expression be QS-dependent? Several complementary
of carI is reduced in the presence of excess 3- rationales have been proposed (3, 63). First, QS
oxo-C6-HSL, suggesting the existence of a also regulates production of PCWDEs (see
negative feedback mechanism to regulate levels below); simultaneous production of a broad-
of AHL (33). spectrum antibiotic may afford some protec-
In addition to the strong, QS-dependent tion of the resulting nutritional “windfall” from
promoter upstream of carA, we have also competitors. Second, for an antibiotic to be
described a second promoter in the car gene effective, it must be present at a sufficient con-
cluster. A weak, constitutive, QS-independent centration in the extracellular milieu; high local
internal promoter, within the carD gene, directs concentrations may be achievable by a dense
transcription of carEFGH (33). This promoter population of cells simultaneously producing
allows QS-independent expression of Car the antibiotic, whereas production by an iso-
resistance. Such constitutive resistance may be lated cell may simply be a waste of metabolic
important if some parts of the population energy. It also seems likely that the ability to
become quorate and start making antibiotic produce Car at high cell densities provides a
just ahead of others, or if the resistance mecha- selective advantage only in some situations.We
nism takes longer to implement than Car syn- have found that, whereas only a very few strains
thesis. It may also be relevant in the context of of E. carotovora subsp. carotovora naturally pro-
cryptic Car gene clusters (see below). duce Car,a significant proportion (16% of those
Although Car production is primarily regu- screened) contain cryptic car clusters. Overex-
lated by QS,it is also responsive to other regula- pression of CarR from E. carotovora subsp. caro-
tors and environmental cues. However, to date, tovora ATCC 39048 restored Car production to
these other regulatory inputs seem to be pre- these isolates, most of which still carried a
dominantly channeled through the QS system, detectable, but presumably defective or “silent,”
emphasizing the principal role of QS in Car carR gene (24). One of these cryptic strains, E.
regulation.Transcription of carI, and concomi- carotovora subsp.carotovora SCRI 193,was shown
tantly AHL levels and Car production,is modu- to be resistant to Car produced by E. carotovora
lated by carbon source, e.g., repressed in the subsp. carotovora ATCC 39048, demonstrating
presence of glycerol (33). Car production is that the presence of a cryptic cluster allows the
exquisitely sensitive to temperature; carA strain to survive in the presence of its Car-
expression is maximal at 34C, greatly reduced producing relatives (33). Moreover, the gain or
at 36C, and eliminated at 37C. The SlyA- loss of a functional CarR protein appears to
family protein, Hor, is the only other regulator provide a mutational “switch,” allowing E. caro-
identified to date, apart from CarIR, as being tovora subsp. carotovora to gain or lose the ability
required for carA-H transcription and Car pro- to produce Car under the appropriate selection
duction. Transcription of hor is itself tempera- pressure. It is perhaps noteworthy in this con-
ture sensitive, being greatly reduced at 37C, text that it has been suggested that CarR might
suggesting that some of the temperature effect represent a mutational hotspot (32).
on Car is mediated via Hor; however, tempera- Car production, directed by a homologous
ture is also suggested to impact posttranscrip- carABCDEFGH cluster, has also been
12. QUORUM SENSING IN THE SOFT-ROT ERWINIAS ■ 189
described in Serratia ATCC 39006 (13). Car although not via CarR (25, 32). It is now clear
production is also AHL QS dependent in this that AHL QS is central to the regulation of vir-
strain, but the mechanisms of QS control are ulence in Erwinia, as we will now describe.
not the same as in E. carotovora subsp. carotovora
ATCC 39048 and provide an interesting com- QUORUM SENSING REGULATION
parison. Similar to E. carotovora subsp. carotovora, OF VIRULENCE FACTOR
PRODUCTION IN E. CAROTOVORA:
a LuxR-family transcriptional regulator, CHARACTERIZATION OF THE
CarR39006, is encoded upstream of and tran- COMPONENTS OF THE CENTRAL
scribed in the same direction as carA-H. How- QUORUM-SENSING LOCUS
ever, genetic data suggest that CarR39006 is a The primary AHL signaling molecule pro-
ligand-independent transcriptional activator, duced by most strains of E. carotovora subsp.
the expression of which is itself controlled by atroseptica and carotovora is 3-oxo-C6-HSL,
the SmaIR QS system (50). SmaI is a LuxI- which at sufficient concentrations is able to
family AHL synthase that produces C4-HSL indirectly activate transcription of genes
and C6-HSL. SmaR is a LuxR-family tran- encoding the key virulence factors, PCWDEs,
scriptional regulator that,unusually,behaves as a and also secreted protease (6, 25, 41).The luxI
repressor in the absence of AHL; this repression homologue in E. carotovora subsp. carotovora, expI
is lifted in the presence of AHL. Genetic evi- (also known as carI, hslI ), encodes a 26-kDa
dence has led to the following model. At low protein that synthesizes 3-oxo-C6-HSL as its
cell densities, SmaR represses expression of major product (5, 41). An expI mutant strain is
carR39006 and carA-H. At high cell densities, severely impaired in virulence, but can be
SmaR binds AHL, relieving this repression and restored by the addition of culture supernatant
allowing the expression of CarR39006 and from a wild-type culture or the addition of syn-
hence the activation of carA-H transcription by thetic 3-oxo-C6-HSL (25, 41).The expI gene is
CarR39006 (50). It has been shown biochemi- expressed throughout growth and produces a
cally that SmaR binds to the carA promoter in constant supply of 3-oxo-C6-HSL, which dif-
the absence, but not the presence, of C4-HSL. fuses into the surrounding environment (1, 33).
However, SmaR repression of carR39006 and Transcription of expI is dependent on both the
other targets is likely to be indirect (21). It carbon source present in the culture medium
therefore appears that Car production has been and oxygen availability in the surrounding
successfully brought under the tight control environment. Grown with glycerol as the sole
of QS in both E. carotovora subsp. carotovora and carbon source, expression of expI is only half
Serratia, but the way in which QS control is that seen in cultures grown with glucose as a
imposed has been adjusted to fit the regulatory carbon source (33). Thus it appears that both
networks of each strain.This is consistent with environmental and nutritional queues regulate
our observation that QS systems and their sec- QS indirectly, by controlling transcription of
ondary metabolite gene cluster targets are the expI gene.
apparently easily moved and integrated into In V. fischeri, the luxR gene, encoding the
existing regulatory systems by horizontal gene cognate regulator of LuxI, is located adjacent to
transfer (15). the luxI gene. In several different Erwinia
Although a well-characterized example of species, a luxR homologue, expR, was identified
AHL QS regulation, carbapenem is only pro- adjacent to the expI gene (1, 17).This apparent
duced by relatively few Erwinia strains from a partner luxR homologue, expR, is transcribed
culture collection, whereas AHL QS appears to convergently with expI in both E. carotovora
be present ubiquitously, at least among these subspp. carotovora and atroseptica. However,
strains (24). Early work showed that production ExpR appears to have only a small or negligible
of secreted virulence factors is also dependent effect on virulence or extracellular enzyme
on AHL QS in E. carotovora subsp. carotovora, production in several different E. carotovora
subsp.carotovora and E.carotovora subsp.atroseptica In strains that are unable to synthesize their
strains (1, 16, 17, 49).A few studies do indicate cognate signaling ligand (expI mutants), high-
that ExpR may play a minor or modulatory level extracellular PCWDE production is lost
role in the QS regulatory network in some since VirR continues to repress expression of
strains (1, 17). In E. carotovora subsp. carotovora these virulence determinants, even at high cell
strain SCC 3193, overproduction of ExpR density (Fig. 2) (6). However, in strains carrying
(also called ExpR1) has been shown to cause a both an expI and a virR mutation, PCWDE
decrease in the production of extracellular production is restored to approximately wild-
PCWDEs and production of macerated rot tis- type levels (6, 16, 49). Proteomic and transcrip-
sue during potato infection, suggesting that tional data confirm that repression of secreted
ExpR might function as a weak repressor of enzyme production in an expI mutant back-
extracellular enzymes (1). However, mutants ground is alleviated when combined with a
carrying mutations in both expI and expR are virR mutation (6). Homologues of virR are
indistinguishable from strains carrying a muta- found in many different strains of E. carotovora,
tion in expI,and strains with a mutation in expR although not in the other soft-rot Erwinia,
alone are not affected significantly in virulence Erwinia chrysanthemi (6) (see below). In E. caro-
(1). ExpR from E. carotovora subsp. carotovora tovora subsp. carotovora strains 71 and SCC 3193,
SCC 3193 also retains the ability to act as an homologues of VirR have been shown to play
activator of a plasmid-based V. fischeri lux pro- a central role in their corresponding QS sys-
moter in a reconstituted Escherichia coli system; tems (16, 49). In E. carotovora subsp. carotovora
this activation is abolished by the addition of 3- strain 71,VirR has been shown to bind to the
oxo-C6-HSL (59). Similarly, in E. carotovora rsmA promoter (16). In addition, in a reconsti-
subsp. carotovora strain 71, ExpR acts as an acti- tuted E. coli system,VirR was shown to activate
vator of rsmA and thus has an indirect repressive production of an rsmA-lacZ fusion (16). RsmA
effect on secreted enzyme production (the role represses PCWDE synthesis by preventing the
of the Rsm regulatory system is discussed in translation of the cognate gene transcripts
more detail below). This activation of rsmA is (10) (see below). Thus, VirR appears to be
alleviated by the addition of exogenous 3-oxo- indirectly regulating virulence, at least in part,
C6-HSL (17). Overall, however, ExpR does not by acting through the Rsm system. However,
appear to act as the major LuxR-type regulator VirR is unlikely to be acting solely through the
in the QS system of E. carotovora. Rsm system. For example, in E. carotovora subsp.
The genome sequence of E. carotovora subsp. carotovora strain 71, the effect of a single virR
atroseptica strain SCRI 1043 revealed another mutation on the transcript levels of several vir-
luxR homologue in this organism: ECA1561 ulence genes did not appear to entirely overlap
(4). Bell et al. postulated that this new luxR with the effect of rsmA inactivation (6). Thus,
homologue, ECA1561, might be the “missing VirR appears to act as the central repressor of
link” in E. carotovora subsp. actroseptica QS con- virulence determinant production in both E.
trol of virulence (4). Subsequently, several carotovora subsp. atroseptica and E. carotovora
reports confirmed that this gene, ECA1561, subsp. carotovora, and this repression is alleviated
now called virR (virulence repressor, also by the cognate AHL signaling molecule.
known as expR2), encodes the central QS Although virR has been shown to act as the
repressor in multiple E. carotovora strains, albeit major regulator of QS in E.carotovora subsp.caro-
with slightly differing roles in different strains tovora and E. carotovora subsp. atroseptica, it can be
(6, 16, 49). assumed that there are reasons for maintaining
VirR functions phenotypically as a repressor the second luxR-type regulator, expR, although
of extracellular PCWDEs and other QS- these reasons remain to be fully clarified.Why
controlled accessory virulence factors at low have two different LuxR-type regulators but
cell density in several Erwinia species (6, 16, 49). only one cognate AHL species? Typically, pairs
FIGURE 2 Model for the regulation of virulence factor production by AHL QS in E. carotovora. At low cell
density (top), ExpI synthesizes the AHL signaling molecule, 3-oxo-C6-HSL, which diffuses into the environment.
VirR, the cognate LuxR homologue, is expressed constitutively and may directly repress expression of PCWDEs.
RsmA is activated by VirR and itself represses expression of rsmB and PCWDEs.At high cell density, high concen-
trations of 3-oxo-C6 HSL are achieved and the signal binds to VirR.This causes derepression and allows expression
of PCWDEs and rsmB.The untranslated RNA rsmB further sequesters RsmA in the cell, thus immediately affecting
PCWDE expression in a cell-density-dependent manner.
191
of LuxIR homologues exist together in organ- tovora subsp. carotovora SCC 3193, a mutation in
isms, producing and responding to cognate the expR gene causes little change in PCWDE
AHLs, for example, the RhlIR and LasIR sys- production or virulence, either on its own or in
tems of Pseudomonas aeruginosa. However, in E. combination with an expI mutation (1). Sur-
carotovora subsp. carotovora and E. carotovora subsp. prisingly, in a strain carrying both expI and virR
atroseptica, this relationship does not appear to mutations, Cel production is not fully restored.
exist,with one LuxI homologue (ExpI) but two Only in a strain carrying mutations in the expI,
(or three) LuxR homologues (VirR, ExpR expR, and virR genes is Cel production fully
[CarR]). One reason for having two LuxR restored to wild-type levels (49).Thus, unlike in
homologues might be to allow perception of E. carotovora subsp. atroseptica SCRI 1043 and E.
AHL molecules produced by other bacteria. In carotovora subsp. carotovora strains 71 and ATCC
E. carotovora subsp. carotovora strain 71, the two 39048, where a single mutation in virR is
LuxR homologues involved in regulating enough to relieve repression of PCWDE
PCWDE production appear to respond to two expression in an expI (AHL-non-producing)
different AHL species (16). Purified VirR is able mutant, E. carotovora subsp. carotovora SCC 3193
to bind to the rsmA promoter region in vivo, requires mutations in both expR and virR for full
and this complex is disrupted by the addition of restoration of PCWDE production when no
either 3-oxo-C6-HSL or 3-oxo-C8-HSL, AHL signaling molecule is present (6, 16, 49).
although this strain only produces a single Our current understanding of QS in E. caro-
detectable lactone species, 3-oxo-C6-HSL (16, tovora subsp. carotovora and E. carotovora subsp.
17). Hence, E. carotovora subsp. carotovora strain atroseptica does not explain how this complex
71 appears to be able to respond to multiple network of regulators evolved.It is clear that the
AHL signals in the surrounding environment. expR gene in several strains of E.carotovora subsp.
In E. carotovora subsp. carotovora SCC 3193, a dif- carotovora still forms an integral part of the QS
ferent primary AHL signal, 3-oxo-C8-HSL, is system. However, the fitness benefits of acquir-
produced by expI,but this strain also responds to ing a new,more important,luxR-type regulator,
other noncognate AHL signals in its immediate virR,at another location distinct from expI-expR
environment (49). It appears that the two on the E. carotovora subsp. atroseptica or carotovora
LuxR-type regulators in this strain, ExpR chromosome have yet to be determined. Per-
(ExpR1) and VirR (ExpR2), are each able to haps an ability to respond to a diverse array of
respond to different AHLs in the environment. AHL signaling molecules reflects a varied range
In an expI mutant, a second mutation in expR is of environments and bacterial communities
able to restore Cel production in conjunction encountered by this phytopathogen. It may
with the addition of many different AHL mole- allow Erwinia to respond to other organisms in
cules to counteract the repression of VirR, the near vicinity that are also using QS to regu-
including 3-oxo-C6-HSL, 3-oxo-C8-HSL, 3- late virulence, perhaps allowing strains of E.
oxo-C10-HSL, C6-HSL, and C7-HSL (49). In carotovora subsp. carotovora and E. carotovora subsp.
contrast, a virR mutation is only able to fully atroseptica to participate in mixed infections
restore Cel production to an expI mutant when involving diverse species.The classical rationale
supplemented with 3-oxo-C8-HSL or 3-oxo- for why Erwinia uses QS to control production
C10-HSL (49). Thus ExpR and VirR seem to of secreted virulence factors in a population-
be responding to different lactone species. dependent manner is as follows. At low cell
Furthermore, the example of E. carotovora numbers, production of PCWDEs will trigger
subsp.carotovora SCC 3193,where the two regu- plant defense responses without causing signifi-
lators appear to have overlapping roles,exempli- cant harm to the plant;QS prevents the elabora-
fies the point that the role and relative tion of these enzymes at low cell densities.
contribution of the two virulence LuxR When a high density of cells is reached at the
homologues can vary between strains.In E.caro- infection site and an infection is likely to
be successful, QS allows concerted massive (18).RsmA is a homologue of the E.coli regula-

induction of PCWDE expression throughout tory protein CsrA; the two proteins share 95%
the population, allowing plant defenses to be identity and are thus presumed to be function-
overwhelmed.This is likely to be an oversimpli- ally conserved. RsmA and CsrA both contain
fication, however, since non-PCWDE QS- predicted RNA-binding domains, which are
controlled host interaction and virulence likely to be involved with target transcript
factors, likely to be important early in infection, RNA stability (10).CsrA was originally discov-
have also been identified (see below) (55). In ered due to its role as a repressor of glycogen
addition, it must be emphasized that QS is only synthesis in E. coli (46). CsrA mediates this
one of many regulatory inputs into virulence repression by binding to the ribosome-binding
factor production in Erwinia. As reviewed (63), site of target transcripts and destabilizing them
multiple environmental cues are integrated into (29, 30). The partner molecule of CsrA, csrB
PCWDE production via regulators such as (rsmB in Erwinia), is a noncoding RNA that
KdgR, Hor, ExpA/S, and the Rsm system. sequesters CsrA and thus indirectly affects
Future work will be directed toward fully expression of those transcripts targeted by CsrA
understanding how the QS system is integrated (28).The Csr system in E.coli has been reviewed
into this complex regulatory network. extensively (45).
In conclusion, although research in E. caro- The Rsm system in E. carotovora subspecies,
tovora subspp. carotovora and atroseptica has been while different in its targets, appears to function
carried out on several different strains with sub- like the Csr system in E. coli. RsmA represses
tle differences in their QS systems, overall a extracellular enzymes by promoting transcript
common theme between all the differing QS degradation (10). Overproduction of RsmA
systems is apparent. Despite strain-dependent has also been shown to suppress expression of
differences in the AHL signaling molecules expI (18), and thus, indirectly, RsmA regulates
produced and the relative importance of a sec- both the components and terminal products of
ond LuxR-type regulator (ExpR), the basic E. the QS system in Erwinia. As in the E. coli Csr
carotovora QS system appears to utilize a repres- system, the small untranslated RNA, rsmB
sor of virulence determinants at low cell den- (originally called aepH), is thought to bind to
sity and relief of this repression by interaction RsmA and prevent it from binding to its target
with AHL at high cell densities. transcripts, thus indirectly mediating activation
of extracellular enzymes (31). rsmB RNA exists
THE Rsm (REGULATION OF in two forms, a larger 479-bp RNA (rsmB) and
SECONDARY METABOLITE) SYSTEM a 259-bp RNA (rsmB). The smaller RNA,
AND ITS ROLE IN QUORUM SENSING rsmB, is a processed form of the original longer
It is currently unclear exactly how the Erwinia rsmB transcript and it is rsmB RNA that is
QS components expI/expR/virR act to influ- responsible for the activation of virulence fac-
ence transcription of the ultimate target genes tors within the cell (31).This effect appears to
involved in virulence,such as PCWDE genes.It be mediated through the action of RsmA (9).
is not yet clear whether this regulation is always Another regulator, rsmC (hexY), has also been
direct or indirect and how it fits in with other identified.rsmC indirectly controls extracellular
regulatory networks within the cell. One such enzyme production by modulating levels of
network is the RsmAB system found in E. caro- both RsmA and rsmB (19). It appears that the
tovora subspp. carotovora and atroseptica. A mutant QS locus in Erwinia controls the Rsm system
defective in rsmA exhibits overproduction of through rsmA (10, 18) and that the Rsm system
extracellular enzymes in E. carotovora subsp. also affects the QS machinery by modulating
carotovora strain 71 (10). This hypervirulent expression of expI and thus AHL production
mutant is able to produce extracellular enzymes (10,18).However,as mentioned above,it is cur-
and cause disease by bypassing the QS system rently unclear whether QS regulation of down-
stream target genes is mediated entirely via the Serratia liquefaciens and the Xcp type II secretion
Rsm system or if some direct regulation of the system in P. aeruginosa (8, 44).
PCWDEs or other targets by the QS compo- The majority of the key secreted virulence
nents also exists. Most important, it remains to factors of E. carotovora and E. chrysanthemi,
be determined whether VirR binds directly to including multiple Pels, Peh, Cel, and Svx, are
any target promoters or whether it only acts secreted by a type II secretion system known as
through other regulators, including RsmA. the Out system (11, 23, 42). Type II secretion
systems are complex multiprotein assemblies
QUORUM SENSING MODULATION that span the periplasm and translocate their
OF SUBTLE AND ACCESSORY substrate proteins from the periplasm to the
VIRULENCE DETERMINANTS exterior of the cell, following Sec- (or Tat-)
As described in the previous section, the regula- dependent export to the periplasm (20).Type II
tion of the expression of the primary viru- secretion is important for virulence in a variety
lence determinants of E. carotovora, secreted of gram-negative bacteria (47).We have recently
PCWDEs, by AHL QS is well documented. observed that expression of the out genes in E.
However, in addition to these “brute force” vir- carotovora subsp. atroseptica is modulated by QS,
ulence factors, it is now clear that Erwinia also being reduced in an expI mutant (G. P. C. S.,
produces multiple “subtle” virulence factors unpublished observations). The reason for this
involved in plant-pathogen interactions (54, QS modulation is easy to rationalize: although
55).Interestingly,many of these latter factors are the Out system is expressed and either in use or
also under QS control, suggesting that the role primed for use in the absence of threshold AHL
of QS during the infection process may be more levels, once threshold AHL levels are reached
complicated than simply orchestrating a simul- and the massive induction of PCWDE produc-
taneous induction of PCWDE production at a tion is induced, expression of Out is concomi-
certain point during the infection. Examples of tantly increased to deal with the increase in
novel and/or subtle QS-dependent secreted substrates requiring secretion. Similarly, we have
virulence factors that have recently been identi- observed that expression of dsbA in E. carotovora
fied include Svx, a necrosis-inducing protein subsp. carotovora SCRI 193 is also modulated by
(Nip),and HrpN;several other secreted proteins QS (G.P.C.S.,unpublished observations).DsbA
(ECA0852 and ECA2220) have been identified is required for the introduction of disulfide
as QS dependent,making them good candidates bonds and correct folding of Pel and PehA in
for novel virulence factors (11, 39, 51). the periplasm (58), and hence can also be envis-
Thus, multiple secreted proteins have been aged to be required in larger quantities once
identified as key virulence determinants and expression of these enzymes is induced by QS.
exhibit QS-dependent expression. They can Although expression of the type III-secreted
also be seen as “terminal” virulence determi- protein,HrpN,is reported to be under AHL QS
nants: the final end product to which the plant control in E. carotovora subsp. carotovora and
is exposed. However, for a PCWDE or other atroseptica (37, 51), there is no evidence to date
secreted virulence factor to interact with the that type III secretion is QS dependent in
plant, other processes in addition to gene Erwinia; indeed, expression of the type III struc-
expression must be fulfilled. For example, the tural gene, hrcC, is QS independent (51).
relevant protein must be secreted from the bac-
terial cell. Hence protein secretion systems may OTHER EXAMPLES OF QUORUM
be considered as “accessory” virulence deter- SENSING IN ERWINIA
minants and could also be subject to QS regula- Although the majority of the work to date on
tion. It is noteworthy that QS-dependent QS in Erwinia has been performed in E. caro-
secretion systems have been reported in other tovora subsp. corotovora and atroseptica, AHL QS
bacteria, e.g., the Lip type I secretion system in has also been reported in other Erwinia species.
Apart from E. carotovora, the other major soft- described for several Italian strains of E.
rotting Erwinia is E. chrysanthemi (also known as amylovora; for one strain, production of AHL
Dickeya dadantii).E.chrysanthemi has been shown was observed in planta (57).Second,AHL activ-
to produce AHL signaling molecules, of which ity was detected in the culture supernatant of a
the major species is 3-oxo-C6-HSL, and to pos- Swiss strain of E. amylovora. In this strain, over-
sess convergently transcribed expI and expR expression of aiiA lactonase eliminated
genes, encoding LuxI- and LuxR-family pro- detectable AHL activity and considerably
teins, respectively. ExpI is required for the syn- reduced extracellular polysaccharide produc-
thesis of 3-oxo-C6-HSL (22, 38).The biological tion,oxidative stress tolerance,and virulence on
role of expIR in E. chrysanthemi is currently ill- apple leaves,indicating that AHL-QS is likely to
defined, since expI mutants show no defect in play an important role in E. amylovora virulence
virulence or in total secreted Pel activity (22, (36). Both reports describe the detection and
38). However, expression of two pel genes, pelA partial sequencing of pairs of convergent luxIR
and pelB, is decreased in the expI mutant, and homologues, named eamIR. However, the
ExpR binds upstream of their promoter impact of inactivation of these genes is yet to be
sequences in the presence of 3-oxo-C6-HSL, described. Another related plant pathogen,Pan-
suggesting that ExpR is an 3-oxo-C6-HSL- toea stewartii spp. stewartii (formerly known as
dependent activator of at least some pel genes Erwinia stewartii), the causative agent of Stew-
(22, 38). ExpR also specifically interacts with art’s wilt and leaf blight in maize, has a well-
the regulatory region of expI and its own gene, studied AHL QS system, which controls
expR, and activates expression of the pectinase extracellular polysaccharide production and
repressor, pecS. Conversely, PecS (and CRP) virulence. This system is discussed in detail in
repress expI expression, suggesting that expIR chapter 13.
may be linked into a complex regulatory net- In addition to AHL QS, another type of
work (43). A recent study has described, at the QS system has been described in gram-
molecular level, the repression of expR gene negative bacteria: LuxS-dependent signaling.
expression by ExpR in the absence of AHL, and As described elsewhere (56), LuxS produces an
its subsequent derepression in the presence of 3- extracellular activity known as Autoinducer-2,
oxo-C6-HSL (7). In the absence of AHL, ExpR which appears to be used as a signaling mole-
binds to expR promoter sequences and prevents cule by some, but certainly not all, bacteria that
access of RNA polymerase and initiation of produce it. Since LuxS also has a metabolic role
transcription. In the presence of 3-oxo-C6- (in the activated methyl cycle), it is not clear in
HSL, the ligand binds to ExpR, disrupts the many organisms whether the phenotypes
ExpR-DNA complex, and renders the expR resulting from luxS inactivation should be
promoter accessible to RNA polymerase, thus ascribed to loss of signaling or loss of metabolic
allowing expR expression to proceed (7). How- function.The presence of luxS and production
ever,other targets of ExpR,and hence QS,in E. of AI-2 has been described in E. carotovora
chrysanthemi remain to be fully defined. Finally, subspp. carotovora and atroseptica. Inactivation of
AHL QS has also been reported to control pro- luxS in two strains of E. carotovora subsp. caro-
duction of secreted enzymes and an unidenti- tovora resulted in a modest impact on PCWDE
fied antibiotic in another soft-rot Erwinia, E. production and virulence, whereas in E. caro-
carotovora subsp. betavasculorum (12). tovora subsp. atroseptica no impact on virulence
Erwinia amylovora is not a soft-rot Erwinia, was observed (14, 26).There is no evidence that
but causes the necrotic disease, fire blight, in these phenotypes are due to signaling rather
apple, pear and related plant species.There have than a metabolic role for LuxS, and so the
been two reports describing the existence of importance of luxS in the respective strains is
AHL QS in E. amylovora. First, production of a apparently very minor compared with that of
single AHL, most likely 3-oxo-C6-HSL, was AHL QS.
CONCLUDING REMARKS REFERENCES

AHL-mediated QS plays a key role in regulat- 1. Andersson, R. A., A. R. Eriksson, R. Heikin-
heimo,A. Mae, M. Pirhonen,V. Koiv, H. Hyyti-
ing the production of virulence factors, both
ainen, A. Tuikkala, and E. T. Palva. 2000.
“brute force” and “stealth,” and thus in viru- Quorum sensing in the plant pathogen Erwinia caro-
lence, in the soft-rotting plant pathogen E. carotovora subsp. carotovora: the role of expR(Ecc). Mol.
tovora. In some strains,AHL QS concomitantly Plant-Microbe Interact 13:384–393.
regulates the production of carbapenem antibi- 2. Bainton, N. J., P. Stead, S. R. Chhabra, B. W.
Bycroft, G. P. Salmond, G. S. Stewart, and P.
otic, perhaps for niche defense. While the QS
Williams. 1992. N-(3-oxohexanoyl)-L-homoser-
regulation of Car is simple, canonical, and well ine lactone regulates carbapenem antibiotic
characterized, QS regulation of virulence is production in Erwinia carotovora. Biochem. J.
more complex. The missing link in virulence 288:997–1004.
regulation, the major regulator VirR, has been 3. Barnard, A. M., and G. P. Salmond. 2006.
Quorum sensing in Erwinia species. Anal. Bioanal.
discovered, and it has been established that the
Chem. 387:415–423.
other LuxR homologue, ExpR, can play an 4. Bell, K. S., M. Sebaihia, L. Pritchard, M. T.
accessory role in some strains. Interestingly, Holden, L. J. Hyman, M. C. Holeva, N. R.
both VirR and ExpR (where determined) Thomson, S. D. Bentley, L. J. Churcher, K.
appear to act by an atypical derepression mech- Mungall, R. Atkin, N. Bason, K. Brooks, T.
Chillingworth, K. Clark, J. Doggett, A. Fraser,
anism,whereby they repress target gene expres-
Z. Hance, H. Hauser, K. Jagels, S. Moule, H.
sion in the absence of AHL, and this repression Norbertczak, D. Ormond, C. Price, M. A.
is relieved in the presence of AHL at high cell Quail, M. Sanders, D.Walker, S.Whitehead, G.
density. However, it remains to be determined P. Salmond, P. R. Birch, J. Parkhill, and I. K.
exactly howVirR impinges on downstream tar- Toth. 2004. Genome sequence of the enterobacte-
rial phytopathogen Erwinia carotovora subsp. atrosep-
get gene expression—is it solely via activation
tica and characterization of virulence factors.
of RsmA, via direct binding to target promot- Proc. Natl.Acad. Sci. USA 101:11105–11110.
ers,or via interaction with other regulatory sys- 5. Brader, G., S. Sjoblom, H. Hyytiainen, K.
tems? We speculate that it is most likely to Sims-Huopaniemi, and E. T. Palva. 2005.
involve some combination of all of these possi- Altering substrate chain length specificity of an
acylhomoserine lactone synthase in bacterial com-
bilities. One other important point that is
munication. J. Biol. Chem. 280:10403–10409.
becoming more apparent is that despite the 6. Burr, T., A. M. Barnard, M. J. Corbett, C.
common theme of VirR- (and ExpR-) medi- L. Pemberton, N. J. Simpson, and G. P.
ated repression, QS regulation of virulence Salmond. 2006. Identification of the central quo-
determinants is subject to much strain-to-strain rum sensing regulator of virulence in the enteric
phytopathogen, Erwinia carotovora: the VirR repres-
variation, including differences in molecules
sor. Mol. Microbiol. 59:113–125.
produced and detected, traits regulated, and rel- 7. Castang, S., S. Reverchon, P. Gouet, and W.
ative contributions of different components. Nasser. 2006. Direct evidence for the modula-
Several intriguing and challenging questions tion of the activity of the Erwinia chrysanthemi
about the role of QS in Erwinia remain, how- quorum-sensing regulator ExpR by acylho-
moserine lactone pheromone. J. Biol. Chem.
ever. These unresolved issues include exactly
281:29972–29987.
how the QS system is integrated with all the 8. Chapon-Herve, V., M. Akrim, A. Latifi, P.
other regulators and environmental inputs con- Williams, A. Lazdunski, and M. Bally. 1997.
trolling virulence, and what the true role of QS Regulation of the xcp secretion pathway by multiple
in Erwinia is during host interaction and disease quorum-sensing modulons in Pseudomonas aerugi-
nosa. Mol. Microbiol. 24:1169–1178.
progression in the plant.
9. Chatterjee, A., Y. Cui, and A. K. Chatterjee.
2002. RsmA and the quorum-sensing signal,
ACKNOWLEDGMENTS N-[3-oxohexanoyl]-L-homoserine lactone, control
We thank the BBSRC for financial support. R. E. M. is the levels of rsmB RNA in Erwinia carotovora subsp.
funded by the Gates Cambridge Trust through the Bill carotovora by affecting its stability. J. Bacteriol.
and Melinda Gates Foundation. 184:4089–4095.
10. Chatterjee,A.,Y. Cui,Y. Liu, C. K. Dumenyo, and harpin(Ecc) production and virulence by
and A. K. Chatterjee. 1995. Inactivation of rsmA modulating levels of regulatory RNA (rsmB) and
leads to overproduction of extracellular pectinases, RNA-binding protein (RsmA). J. Bacteriol.
cellulases, and proteases in Erwinia carotovora 181:6042–6052.
subsp. carotovora in the absence of the starvation/ 20. Filloux, A. 2004.The underlying mechanisms of
cell density-sensing signal, N-(3-oxohexanoyl)-L- type II protein secretion. Biochim. Biophys. Acta
homoserine lactone. Appl. Environ. Microbiol. 1694:163–179.
61:1959–1967. 21. Fineran, P. C., H. Slater, L. Everson, K.
11. Corbett, M., S. Virtue, K. Bell, P. Birch, T. Hughes, and G. P. Salmond. 2005. Biosynthesis
Burr, L. Hyman, K. Lilley, S. Poock, I. Toth, of tripyrrole and beta-lactam secondary metabo-
and G. Salmond. 2005. Identification of a new lites in Serratia: integration of quorum sensing
quorum-sensing-controlled virulence factor in with multiple new regulatory components in the
Erwinia carotovora subsp. atroseptica secreted via the control of prodigiosin and carbapenem antibiotic
type II targeting pathway. Mol. Plant-Microbe Inter- production. Mol. Microbiol. 56:1495–1517.
act. 18:334–342. 22. Ham, J. H.,Y. Cui, J. R.Alfano, P. Rodriguez-
12. Costa, J. M., and J. E. Loper. 1997. EcbI and Palenzuela, C. M. Rojas,A. K. Chatterjee, and
EcbR: homologs of LuxI and LuxR affecting A. Collmer. 2004.Analysis of Erwinia chrysanthemi
antibiotic and exoenzyme production by Erwinia EC16 pelE::uidA, pelL::uidA, and hrpN::uidA
carotovora subsp. betavasculorum. Can. J. Microbiol. mutants reveals strain-specific atypical regulation
43:1164–1171. of the Hrp type III secretion system. Mol. Plant-
13. Coulthurst, S. J., A. M. Barnard, and G. P. Microbe Interact. 17:184–194.
Salmond. 2005. Regulation and biosynthesis of 23. He, S.Y., M. Lindeberg, A. K. Chatterjee, and
carbapenem antibiotics in bacteria. Nat. Rev. Micro- A. Collmer. 1991. Cloned Erwinia chrysanthemi
biol. 3:295–306. out genes enable Escherichia coli to selectively
14. Coulthurst, S. J., K. S. Lilley, and G. P. secrete a diverse family of heterologous proteins to
Salmond. 2006. Genetic and proteomic analysis its milieu.Proc.Natl.Acad.Sci.USA 88:1079–1083.
of the role of luxS in the enteric phytopathogen, 24. Holden, M.T., S. J. McGowan, B.W. Bycroft,
Erwinia carotovora. Mol. Plant Pathol. 7:31–35. G. S. Stewart, P.Williams, and G. P. Salmond.
15. Coulthurst, S. J., N. R.Williamson,A. K. Har- 1998. Cryptic carbapenem antibiotic production
ris, D. R. Spring, and G. P. Salmond. 2006. genes are widespread in Erwinia carotovora: facile
Metabolic and regulatory engineering of Serratia trans activation by the carR transcriptional regula-
marcescens: mimicking phage-mediated horizontal tor. Microbiology 144:1495–1508.
acquisition of antibiotic biosynthesis and quorum- 25. Jones, S., B.Yu, N. J. Bainton, M. Birdsall, B.W.
sensing capacities. Microbiology 152:1899–1911. Bycroft, S. R. Chhabra,A. J. Cox, P. Golby, P. J.
16. Cui,Y.,A. Chatterjee, H. Hasegawa, and A. K. Reeves, S. Stephens, et al. 1993.The lux autoin-
Chatterjee. 2006. Erwinia carotovora subspecies ducer regulates the production of exoenzyme viru-
produce duplicate variants of ExpR, LuxR lence determinants in Erwinia carotovora and
homologs that activate rsmA transcription but dif- Pseudomonas aeruginosa. EMBO J. 12:2477–2482.
fer in their interactions with N-acylhomoserine 26. Laasik, E., L.Andresen, and A. Mae. 2006.Type
lactone signals. J. Bacteriol. 188:4715–4726. II quorum sensing regulates virulence in Erwinia
17. Cui,Y., A. Chatterjee, H. Hasegawa,V. Dixit, carotovora ssp. carotovora. FEMS Microbiol. Lett.
N. Leigh, and A. K. Chatterjee. 2005. ExpR, a 258:227–234.
LuxR homolog of Erwinia carotovora subsp. caro- 27. Lazdunski, A. M., I.Ventre, and J. N. Sturgis.
tovora, activates transcription of rsmA, which speci- 2004. Regulatory circuits and communication in
fies a global regulatory RNA-binding protein. J. gram-negative bacteria. Nat. Rev. Microbiol.
Bacteriol. 187:4792–4803. 2:581–592.
18. Cui,Y., A. Chatterjee,Y. Liu, C. K. Dumenyo, 28. Liu, M.Y., G. Gui, B. Wei, J. F. Preston III, L.
and A. K. Chatterjee. 1995. Identification of a Oakford, U. Yuksel, D. P. Giedroc, and T.
global repressor gene, rsmA, of Erwinia carotovora Romeo. 1997. The RNA molecule CsrB binds to
subsp. carotovora that controls extracellular enzymes, the global regulatory protein CsrA and antagonizes
N-(3-oxohexanoyl)-L-homoserine lactone, and its activity in Escherichia coli. J. Biol. Chem.
pathogenicity in soft-rotting Erwinia spp.J.Bacteriol. 272:17502–17510.
177:5108–5115. 29. Liu, M.Y., and T. Romeo. 1997.The global reg-
19. Cui,Y.,A. Mukherjee, C. K. Dumenyo,Y. Liu, ulator CsrA of Escherichia coli is a specific mRNA-
and A. K. Chatterjee. 1999. rsmC of the softrot- binding protein. J. Bacteriol. 179:4639–4642.
ting bacterium Erwinia carotovora subsp. caro- 30. Liu, M.Y., H.Yang, and T. Romeo. 1995.The
tovora negatively controls extracellular enzyme product of the pleiotropic Escherichia coli gene csrA
modulates glycogen biosynthesis via effects on 40. Perombelon, M. C. M. 2002. Potato diseases
mRNA stability. J. Bacteriol. 177:2663–2672. caused by soft rot erwinias: an overview of patho-
31. Liu,Y.,Y. Cui, A. Mukherjee, and A. K. Chat- genesis. Plant Pathol. 51:1–12.
terjee. 1998. Characterization of a novel RNA 41. Pirhonen, M., D. Flego, R. Heikinheimo, and
regulator of Erwinia carotovora ssp. carotovora that E.T. Palva. 1993.A small diffusible signal molecule
controls production of extracellular enzymes and is responsible for the global control of virulence
secondary metabolites. Mol. Microbiol. 29:219–234. and exoenzyme production in the plant pathogen
32. McGowan, S., M. Sebaihia, S. Jones, B.Yu, N. Erwinia carotovora. EMBO. J. 12:2467–2476.
Bainton, P. F. Chan, B. Bycroft, G. S. Stewart, 42. Reeves, P. J., D. Whitcombe, S. Wharam, M.
P. Williams, and G. P. Salmond. 1995. Car- Gibson, G. Allison, N. Bunce, R. Barallon, P.
bapenem antibiotic production in Erwinia caro- Douglas, V. Mulholland, S. Stevens, et al.
tovora is regulated by CarR, a homologue of the 1993. Molecular cloning and characterization of
LuxR transcriptional activator. Microbiology. 13 out genes from Erwinia carotovora subspecies caro-
141:541–550. tovora: genes encoding members of a general secre-
33. McGowan, S. J., A. M. Barnard, G. tion pathway (GSP) widespread in gram-negative
Bosgelmez, M. Sebaihia, N. J. Simpson, N. R. bacteria. Mol. Microbiol. 8:443–456.
Thomson, D. E.Todd, M.Welch, N.A.White- 43. Reverchon, S., M. L. Bouillant, G. Salmond,
head, and G. P. Salmond. 2005. Carbapenem and W. Nasser. 1998. Integration of the quorum-
antibiotic biosynthesis in Erwinia carotovora is regu- sensing system in the regulatory networks control-
lated by physiological and genetic factors modulat- ling virulence factor synthesis in Erwinia
ing the quorum sensing-dependent control chrysanthemi. Mol. Microbiol. 29:1407–1418.
pathway. Mol. Microbiol. 55:526–545. 44. Riedel, K., T. Ohnesorg, K. A. Krogfelt, T. S.
34. McGowan, S. J., M. Sebaihia, S. O’Leary, K. Hansen, K. Omori, M. Givskov, and L. Eberl.
R. Hardie, P. Williams, G. S. Stewart, B. W. 2001. N-acyl-L-homoserine lactone-mediated
Bycroft, and G. P. Salmond. 1997. Analysis of regulation of the lip secretion system in Serratia liq-
the carbapenem gene cluster of Erwinia carotovora: uefaciens MG1. J. Bacteriol. 183:1805–1809.
definition of the antibiotic biosynthetic genes and 45. Romeo, T. 1998. Global regulation by the small
evidence for a novel beta-lactam resistance mecha- RNA-binding protein CsrA and the non-
nism. Mol. Microbiol. 26:545–556. coding RNA molecule CsrB. Mol. Microbiol.
35. McGowan, S. J., M. Sebaihia, L. E. Porter, G. 29:1321–1330.
S. Stewart, P.Williams, B.W. Bycroft, and G. 46. Romeo, T., and M. Gong. 1993. Genetic and
P. Salmond. 1996. Analysis of bacterial car- physical mapping of the regulatory gene csrA on
bapenem antibiotic production genes reveals a the Escherichia coli K-12 chromosome. J. Bacteriol.
novel beta-lactam biosynthesis pathway. Mol. 175:5740–5741.
Microbiol. 22:415–426. 47. Sandkvist, M. 2001. Biology of type II secretion.
36. Molina, L., F. Rezzonico, G. Defago, and Mol. Microbiol. 40:271–283.
B. Duffy. 2005. Autoinduction in Erwinia 48. Schuster, M., C. P. Lostroh, T. Ogi, and E. P.
amylovora: evidence of an acyl-homoserine lactone Greenberg. 2003. Identification, timing, and sig-
signal in the fire blight pathogen. J. Bacteriol. nal specificity of Pseudomonas aeruginosa quorum-
187:3206–3213. controlled genes: a transcriptome analysis. J.
37. Mukherjee, A., Y. Cui, Y. Liu, and A. K. Bacteriol. 185:2066–2079.
Chatterjee. 1997. Molecular characterization and 49. Sjoblom, S., G. Brader, G. Koch, and E. T.
expression of the Erwinia carotovora hrpNEcc gene, Palva. 2006. Cooperation of two distinct ExpR
which encodes an elicitor of the hypersensitive regulators controls quorum sensing specificity and
reaction. Mol. Plant-Microbe Interact. 10:462–471. virulence in the plant pathogen Erwinia carotovora.
38. Nasser,W., M. L. Bouillant, G. Salmond, and Mol. Microbiol. 60:1474–1489.
S. Reverchon. 1998. Characterization of the 50. Slater, H., M. Crow, L. Everson, and G. P.
Erwinia chrysanthemi expI-expR locus directing the Salmond. 2003. Phosphate availability regulates
synthesis of two N-acyl-homoserine lactone signal biosynthesis of two antibiotics, prodigiosin and
molecules. Mol. Microbiol. 29:1391–1405. carbapenem, in Serratia via both quorum-sensing-
39. Pemberton, C. L., N.A.Whitehead, M. Sebai- dependent and -independent pathways. Mol.
hia, K. S. Bell, L. J. Hyman, S. J. Harris, A. J. Microbiol. 47:303–320.
Matlin, N. D. Robson, P. R. Birch, J. P. Carr, I. 51. Smadja, B., X. Latour, D. Faure, S. Chevalier,
K. Toth, and G. P. Salmond. 2005. Novel Y. Dessaux, and N. Orange. 2004. Involvement
quorum-sensing-controlled genes in Erwinia caro- of N-acylhomoserine lactones throughout plant
tovora subsp. carotovora: identification of a fungal infection by Erwinia carotovora subsp.atroseptica (Pec-
elicitor homologue in a soft-rotting bacterium. tobacterium atrosepticum). Mol. Plant-Microbe Interact.
Mol. Plant-Microbe Interact. 18:343–353. 17:1269–1278.
52. Swift, S., M. K. Winson, P. F. Chan, N. J. scription of genes encoding secreted proteins.
Bainton, M. Birdsall, P. J. Reeves, C. E. Rees, Microbiology 145:1945–1958.
S. R. Chhabra, P. J. Hill, J. P. Throup, et al. 59. von Bodman, S. B., J. K. Ball, M. A. Faini, C.
1993. A novel strategy for the isolation of luxI M. Herrera,T. D. Minogue, M. L. Urbanowski,
homologues:evidence for the widespread distribu- and A. M. Stevens. 2003. The quorum sensing
tion of a LuxR:LuxI superfamily in enteric bacte- negative regulators EsaR and ExpR(Ecc), homo-
ria. Mol. Microbiol. 10:511–520. logues within the LuxR family, retain the ability to
53. Thomson, N. R., J. D. Thomas, and G. P. function as activators of transcription. J. Bacteriol.
Salmond. 1999. Virulence determinants in 185:7001–7007.
the bacterial phytopathogen Erwinia. Methods 60. Welch, M., J. M. Dutton, F. G. Glansdorp, G.
Microbiol. 29:347–426. L.Thomas, D. S. Smith, S. J. Coulthurst,A. M.
54. Toth, I. K., K. S. Bell, M. C. Holeva, and P. R. Barnard, G. P. Salmond, and D. R. Spring.
J. Birch. 2003. Soft rot erwiniae: from genes to 2005. Structure-activity relationships of Erwinia
genomes. Mol. Plant Pathol. 4:17–30. carotovora quorum sensing signaling molecules.
55. Toth, I. K., and P. R. Birch. 2005. Rotting Bioorg. Med. Chem. Lett. 15:4235–4238.
softly and stealthily. Curr. Opin. Plant Biol. 61. Welch, M., D. E.Todd, N. A.Whitehead, S. J.
8:424–429. McGowan, B.W. Bycroft, and G. P. Salmond.
56. Vendeville, A., K. Winzer, K. Heurlier, 2000. N-acyl homoserine lactone binding to the
C. M. Tang, and K. R. Hardie. 2005. Making CarR receptor determines quorum-sensing speci-
‘sense’ of metabolism: autoinducer-2, LuxS ficity in Erwinia. EMBO J. 19:631–641.
and pathogenic bacteria. Nat. Rev. Microbiol. 62. Whitehead, N. A., A. M. Barnard, H. Slater,
3:383–396. N. J. Simpson, and G. P. Salmond. 2001.
57. Venturi, V., C. Venuti, G. Devescovi, C. Quorum-sensing in gram-negative bacteria.
Lucchese,A. Friscina, G. Degrassi, C.Aguilar, FEMS Microbiol. Rev. 25:365–404.
and U. Mazzucchi. 2004. The plant pathogen 63. Whitehead, N. A., J.T. Byers, P. Commander,
Erwinia amylovora produces acyl-homoserine M. J. Corbett, S. J. Coulthurst, L. Everson, A.
lactone signal molecules in vitro and in planta. K. Harris, C. L. Pemberton, N. J. Simpson, H.
FEMS Microbiol. Lett. 241:179–183. Slater, D. S. Smith, M.Welch, N.Williamson,
58. Vincent-Sealy, L. V., J. D. Thomas, P and G. P. Salmond. 2002.The regulation of viru-
Commander, and G. P. Salmond. 1999.Erwinia lence in phytopathogenic Erwinia species: quorum
carotovora DsbA mutants:evidence for a periplasmic- sensing, antibiotics and ecological considerations.
stress signal transduction system affecting tran- Antonie Van Leeuwenhoek 81:223–231.
ROLE OF QUORUM-SENSING REGULATION
IN PATHOGENESIS OF PANTOEA
STEWARTII SUBSP. STEWARTII
Susanne B. von Bodman, Aurelien L. Carlier, and Ann M. Stevens
13
THE PATHOGEN AND structural studies of the organism in culture and
DISEASE BIOLOGY in the infected maize tissues (4, 5). Subse-
Pantoea stewartii subsp. stewartii (Erwinia stewartii) quently, molecular approaches identified two
is an enterobacterial pathogen that causes Stew- major pathogenicity-linked gene systems,
art’s vascular wilt and leaf blight of sweet corn which contribute to two distinct phases of vir-
and maize (4).The bacterium also colonizes an ulence (7, 9, 23). First, an hrp (hypersensitive
insect vector, the corn flea beetle, Chaetocnema response and pathogenicity)-encoded type III
pulicaria, in which the bacterium survives harsh secretion system deploys disease-specific Wts
winter temperatures (30). The beetles emerge (for water soaking) effector proteins that induce
from hibernation in spring and feed on young electrolyte leakage and tissue water soaking
maize seedlings, thereby depositing the bacter- (19).This symptomology is characteristic of the
ial inoculum directly into the host tissue (30) early phase of infection when the bacteria still
(Fig. 1). How the bacterium interacts with the reside in the host apoplast (19). Subsequently,
beetle remains a largely open question because the bacteria preferentially colonize the xylem
of the difficulty in maintaining and studying of the host where they grow to high cell densi-
the beetle in a laboratory setting. Economic ties and produce an abundance of Stewartan
losses from Stewart’s wilt-infected sweet and capsular polysaccharides (CPSST, membrane-
seed corn can be significant and are directly associated) and exopolysaccharides (EPSST,
related to the winter survival of the beetle (30). cell-free) (7).This condition obstructs the free
A glimpse into the disease biology of P. stew- flow of xylem fluid, leading to seedling wilt and
artii comes from early biochemical and ultra- the chlorotic and necrotic parallel streaking
characteristic of Stewart’s wilt disease in mature
plants (30) (Fig. 1). EPSST is an acidic, high-
Susanne B. von Bodman Departments of Plant Science and
Molecular and Cell Biology, University of Connecticut, molecular-weight polymer of heptasaccharide
Storrs, Connecticut 06269-4163. Aurelien L. Carlier repeating units, composed of three glucose
Department of Plant Science, University of Connecticut, (Glc), three galactose (Gal), and one glucuronic
Storrs, Connecticut 06269-4163. Ann M. Stevens
Department of Biological Sciences,Virginia Tech, Blacksburg, acid (GlcA) (28) (Fig. 2). The structural genes
Virginia 24061. for EPSST synthesis are encoded by a second
201
202 ■ VON BODMAN ET AL.
FIGURE 1 Stewart’s wilt. (A) Typical parallel streaking symptoms due to

pathogen colonization of parallel longitudinal maize xylem. (B) Beetle vector feed-
ing on a maize leaf. Note the feeding scars (arrow).
major virulence locus,the wce gene cluster (pre- could be readily restored by the exogenous
viously cps, renamed [36]). It is a typical group 1 presentation of AHL signal (40). While these
polysaccharide gene system based on criteria data were expected, the mutagenesis of the
such as gene organization, genetic linkage to linked esaR gene presented two surprising
the chromosomal his locus, and regulation by results. First, expression of esaI and wild-type
the Rcs environmental sensing phosphorelay levels of AHL synthesis turned out to be EsaR
regulatory system (9). To fully appreciate the independent (40). Second, a single mutation in
functional role of the EsaI/EsaR quorum- esaR and, more important, an esaI/esaR double
sensing regulators, a basic functional under- mutant strain (ESIR) produced EPSST consti-
standing of the wce gene system is necessary. A tutively and was excessively mucoid. In con-
summary and schematic representation are pre- trast, the wild-type strain (DC283) produced
sented in Fig. 2. EPSST strictly as a function of cell density (41).
A simple explanation for these observations was
THE esaI/esaR QUORUM-SENSING that EsaR, directly or indirectly, repressed
SYSTEM OF P. STEWARTII DC283 EPSST synthesis at low cell density and that
Standard genetic procedures identified a 2.6-kb inducing levels of AHL promoted autoinducer-
DNA fragment, which when expressed in mediated derepression of EPSST synthesis.
Escherichia coli directed the synthesis of gener-
ous amounts of acyl homoserine lactone EsaR REPRESSES ITS OWN
(AHL), primarily N-(3-oxo-hexanoyl)-L- EXPRESSION
homoserine lactone (3-oxo-C6-HL), with The promoter region of esaR features a well-
minor amounts of N-(3-oxo-octanoyl)-L- conserved lux box-like element, the esaR box,
homoserine lactone (3-oxo-C8-HL) (40). The which spans a putative 10 hexameric 70
DNA sequence of this fragment revealed two promoter element (40). We focused initially
convergently oriented, slightly overlapping on this promoter to define the functional rol
open reading frames (Fig. 3). One, designated of EsaR as a transcription factor and DNA-
esaI, encoded a predicted LuxI homolog AHL binding protein. Genetic studies in an
synthase (24% identity), and the other, desig- Escherichia coli background showed that a
nated esaR, encoded a predicted LuxR PesaR::lacZ reporter was fully active in the
homolog regulatory protein (24% identity) absence of EsaR but was repressed by EsaR in a
(40). A chromosomal mutation of esaI led to dose-dependent manner. Moreover, the EsaR
parallel deficiencies in AHL synthesis, EPSST repressor activity was rapidly neutralized by
synthesis, and virulence (40). EPSST synthesis addition of exogenous AHL (27). Biochemical
13. ROLE OF QUORUM SENSING REGULATION IN PATHOGENESIS ■ 203
A wceG1 wza wzb wzc wceL wceB wceM wceN wceF wceJ wceK wzx1
wceG2 wceO wzx2
B C
WceG1 lip-PP-Gal transferase (I) ß-D-Glc (VII)
Wca OM channel for polymer export 1
Wzb phosphotyrosine phosphatase STEWARTAN
6
Wzc tyrosine protein kinase repeating units
-D-Gal (V)
WceL putative Wzy polymerase 1
WceN lip-PP-Gal-Glc (II)
WceM lip-PP-Gal-Glc-Gal (III) 4
WceB lip-PP-Gal-Glc-Gal-GlcA (IV) ß-D-GlcA (IV)
WceK lip-PP-Gal-Glc-Gal-GlcA-Gal (V) 1
WceJ putative Wzy ligase (III) (II) 4 (I)
WceF putative glycan hydrolase [ 1-ß-D-Gal-3 1-ß-D-Glc-6 1- -D-Gal-3 ]n
Wzx1 flippase 6
WceO putative glu transferase (VI & VII) 1
Wzx2 putative flippase
WceG2 alternate lip-PP-Gal transferase ß-D-Glc (VI)
FIGURE 2 Genes and proteins necessary for EPSST synthesis and its repeating unit structure. (A) Schematic
depiction of gene systems proven and predicted to contribute to Stewartan capsular/exopolysaccharide synthesis.
The primary wce gene system features 12 genes beginning with wceG at the 5′ end and ending with wzx1 at the
3′ end.The wceG1 gene is preceded by the primary RcsA/B-regulated promoter for induced expression of the sys-
tem ( ).Two internal promoters are located upstream of wza and wceB.These promoters appear to be constitutively
expressed ( ).Two putative attenuator loops ( o ) bracket wceL.The wceG2 gene, which appears to be constitu-
tively expressed, is functionally equivalent to the RcsA/B-regulated wceG1 (A. L. Carlier and S. B. von
Bodman, unpublished data).A separate two-gene system, encoding wceO and wzx2, is also RcsA/B regulated (A. L.
Carlier and S. B. von Bodman, unpublished data). (B). List of confirmed and predicted products of the CPSST/EPSST
biosynthetic genes.The roman numerals to the right of specific genes correspond to the residue designations of the
heptasaccharide structure in panel C.The hexose sugar linkages are as indicated. Polymerization between adjacent
repeating units involves a 1-3 linkage between Gal residues I and III.
confirmation for EsaR as a DNA-binding pro- mental treatment yielded SPR sensograms typi-
tein came from electrophoretic mobility shift cal of subsaturation analyte concentrations, sug-
assays and surface plasmon resonance (SPR) gesting that AHL reduces the fraction of
studies. These studies showed that purified, DNA-binding-competent EsaR protein in a
AHL-free EsaR specifically binds DNA frag- dose-dependent manner (27). Fluorescence
ments containing the esaR box palindrome, quenching studies showed a decrease in fluores-
although in electrophoretic mobility shift assays, cence intensity as a function of AHL concentra-
even excess levels of synthetic 3-oxo-C6-HL or tion, with maximal quenching requiring an
AHL extracts failed to induce EsaR/DNA approximate 1:1 molar EsaR:: AHL ratio (27).
complex dissociation (26, 27). This may be an Other evidence showed that EsaR can dimerize
artifact of the assay system or could hold the key in the absence of the AHL and therefore satis-
for the molecular understanding by which fied a general requirement for DNA-binding
EsaR responds and interacts with the AHL. proteins with specificity for palindromic DNA-
EsaR/DNA-binding studies by SPR con- binding sites (27, 32).Thus, EsaR showed func-
firmed the specificity of EsaR for the EsaR box tional attributes that were essentially the reverse
palindrome. More important, these assays from that of the LuxR and TraR functional par-
demonstrated that EsaR exposed to increasing adigms and other orthologous proteins that are
levels of synthetic 3-oxo-C6-HL during the AHL-dependent quorum-sensing activators.
protein/DNA association phase of the experi- Interestingly, those few members of the LuxR
FIGURE 3 Model of the converging EsaI/EsaR quorum-sensing and Rcs environmental

stimuli-sensing regulatory networks.The dominant quorum-sensing cell-cell signaling path-
way involves the EsaR transcription factor and cognate AHL signal synthase, EsaI. The two
genes are genetically linked, encoding two convergent, slightly overlapping open reading
frames.The esaI gene is constitutively expressed, leading to linear accumulation of AHL ( ).
EsaR represses its own gene expression and blocks the RcsA-dependent transcription of rcsA
under subthreshold AHL levels. The AHL-free form of EsaR binds DNA and represses and
activates target genes. DNA-binding and transcriptional activity of Apo-EsaR is neutralized by
inducing levels of AHL.Derepression leads to the RcsA/B-dependent activation of rcsA,which
represents a positive feedback regulatory loop. RcsA is a key component of the Rcs environ-
mental signal sensing ( ) phosphorelay system, which includes the outer membrane lipopro-
tein RcsF. RcsF perceives and transmits undefined environmental signals and membrane
pertubation stimuli to the inner membrane-bound RcsC sensor kinase with a receiver domain
that passes the phosphoryl group to RcsD (formerly YojN). RcsD is also an inner membrane-
bound phosphotransfer protein involved in transferring the phosphoryl group to the RcsB
response regulator. RcsB has multiple functions but, if complexed with the RcsA coactivator, is
dedicated for wce activation. RcsA/B-activated ( ) promoters include the primary pro-
moter of the wce gene system and the promoter upstream of wceO/wzx2. Their coordinate
activation leads to increased production of the heptasaccharide repeating units required
for synthesis of EPSST (■). The promoters (→) upstream of wceB and wceG2 are thought
to facilitate low-level constitutive repeat unit synthesis for CPSST (TTT), possible K-LPS
synthesis.
family that were first identified as repressors (3, ilar conformational changes also render AHL-
14) contain two unique regions: (i) an extended EsaR sensitive to proteolysis remains to be
linker region between the AHL-binding N-ter- established.
minal domain and C-terminal DNA-binding Finally, one can speculate about the role of a
domain and (ii) extra residues at the C terminus negative feedback regulatory mechanism at the
of the polypeptide (A.Thode, D. Donham, and level of esaR expression. First, synthesis of ele-
M. Churchill, personal communication).There vated levels of EsaR repressor under inducing
is good evidence that AHL serves as a scaffold levels seems paradoxical, although one may
for activator TraR folding and stabilization of envision a role in sequestering constitutively
the DNA-binding conformation and that the generated pools of AHL to raise the intrinsic
lack of AHL promotes the proteolytic degrada- inducer threshold. Second, increasing levels of
tion of nascent TraR protein (47).Whether sim- EsaR at higher cell densities may also be a
scheme for modulating target gene expression, green fluorescent protein retrieved three classes
which is a common feature of quorum-sensing of mutants based on EPS deficiency and dimin-
regulatory systems (10, 13, 20). ished fluorescence intensity after in trans expres-
sion of a plasmid-borne esaR (26). Specifically,
EsaR CAN ALSO FUNCTION multiple insertions localized to rcsA and none to
AS AN ACTIVATOR the other rcs (regulator of capsule synthesis) reg-
As described above, repressor activity during ulatory components (for an excellent review of
quorum sensing requires that EsaR exists as a the Rcs phosphorelay systems,see reference 25).
dimer and binds target promoters in the A number of insertions mapped to different
absence of AHL. Within these promoters, the genes within the wce gene cluster.One insertion
esaR box DNA-binding site is positioned to localized to a previously unknown glycosyl-
block the transcriptional activity of RNA poly- transferase gene, which has been designated
merase (6, 27). On the other hand, activation wceO (A. L. Carlier and S. B. von Bodman,
requires, in addition to dimerization and DNA manuscript in preparation).The multiple EsaR-
binding, appropriate surfaces with which the responsive insertions into rcsA strongly sug-
transcription factor may establish a productive gested that EsaR directly regulates rcsA, which
interaction with RNA polymerase (11, 12, 21, encodes the essential RcsA regulatory factor of
24, 45). The ability of EsaR to function as an the Rcs phosphorelay system known to operate
activator was examined using the LuxR- in P. stewartii (37). Subsequent genetic and bio-
activated class II-type lux operon promoter and chemical analyses of the P. stewartii rcsA pro-
LuxR-binding site (lux box) in an artificial moter region identified a semiconserved esaR
context (38). These studies confirmed that box positioned between two promoters: a con-
EsaR serves as an activator of transcription by stitutive, coding region-proximal promoter and
RNAP in this heterologous system, although a distal promoter located upstream of the esaR
with less efficiency than LuxR. The heterolo- box (6). The distal promoter features a well-
gous lux box,which differs from the esaR box at conserved RcsAB box indicative of RcsA/B-
5 of 20 positions, may reduce EsaR affinity for dependent gene activation. RcsA/B activation
lux box DNA. Additional unpublished studies of rcsA is common in E. coli and related enteric
have demonstrated that, like TraR, residues in bacteria (25, 43). That the esaR box is located
both the N- and C-terminal domains of EsaR downstream of the RcsA/B-regulated pro-
are essential for its activity as an activator (D. J. moter indicates that EsaR binding at this site
Schu, K. A. Penney, and A. M. Stevens, unpub- blocks transcript elongation from the distal pro-
lished data).The question of whether EsaR can moter with no appreciable effect on the down-
function as an activator in its native host, P. stew- stream constitutive promoter (6).
artii, is an active area of research. To date, two The EsaR-specific control of the RcsA-
promoters that are activated by EsaR in the dependent positive feedback loop at the level of
absence of AHL have been identified (D. J. rcsA transcription creates an important positive
Schu, A.L.Carlier,S.B.von Bodman,and A.M. feedback control mechanism characteristic of
Stevens, unpublished data). The physiological quorum-sensing networks, and is likely to
roles of the genes expressed from these promot- ensure a tightly controlled “on-off ” switching
ers with regard to the quorum-sensing response mechanism, particularly in a system that pro-
still have not been fully elucidated. duces AHL signal constitutively. In addition,
the EsaR/RcsA regulatory scheme integrates
EsaR-MEDIATED REPRESSION the Rcs environmental sensory pathway with
OF EPS SYNTHESIS the EsaR-controlled cell-cell sensory system,
Random mutagenesis of the hypermucoid presumably to guarantee that EPSST synthesis
strain ESIR (esaI/esaR) using a transposon occurs exclusively when both signal inputs
with a promoterless gene for expression of coincide. The convergence and integration of
these major signaling networks are significant related systems, is notably absent from the wce
considering that quorum-sensing control of gene cluster and the genome (46).
EPSST synthesis is obligatory for “normal” From published data and recent new infor-
biofilm development and virulence in P. stew- mation,a picture emerges that suggests a funda-
artii (see discussion below) (22, 39). mental role of the EsaR quorum-sensing
RcsA/B regulatory scheme as a switching
THE EsaR/RcsA REGULATORY mechanism between basal-level glycopolymer
EFFECT ON THE wce GENE CLUSTER synthesis and high-level EPSST synthesis. The
The wce gene cluster is a typical group 1 poly- evidence for this prediction is as follows. Under
saccharide gene system (9, 46). The functions AHL-inducing conditions, RcsA/B activates
attributed to individual genes are based on the transcription of the wce gene cluster from
partial functional analysis of individual wce the primary promoter located upstream of
genes, the related amylovoran ams gene system wceG1 (formerly cpsA), the first gene of the
of Erwinia amylovora (15, 16), and biochemically system. This promoter features a well-charac-
defined orthologs of related group 1 polysac- terized RcsAB-binding box,which is necessary
charide systems (46). These gene clusters, for RcsA/B-specific activation from this pro-
including wce, are basically bipartite, with moter (43,44).The wceG1 gene,which is nearly
the 5′ coding region separated from the 3′ cod- silent under noninducing conditions, expresses
ing region by at least one prominent stem- at fairly high levels upon RcsA/B activation.
loop transcriptional attenuator (Fig. 2 and 3) This gene is predicted to encode a priming
(34, 46). The 5′ region includes the wza, wzb, enzyme that transfers Gal-1-P from UDP-Gal
and wzc genes, whose gene products are dedi- to the lipid carrier to yield und-PP-Gal as a first
cated for the high-molecular-weight EPSST step in polymer repeat unit biosynthesis (see
polymerization and translocation (46). The Fig. 2) (15, 23, 46). Interestingly, a mutation in
downstream regions generally encode glycosyl- wceG1 has only a partial effect on EPSST syn-
transferases involved in glycopolymer repeat thesis, with the mutant strain exhibiting an
unit biosynthesis, and two inner membrane intermediary mucoid phenotype (9).The now
proteins, Wzy and Wzx. Wzx proteins translo- available P. stewartii genome (17) reveals an
cate specific undecaprenyl diphosphate (und- unlinked orthologous gene, which we desig-
PP)-linked repeating units from the cytoplasm nate wceG2. The two genes are 63% identical
across the inner membrane by a “flippase”-like and distinct from a third ortholog, wecA, also
mechanism.The Wzy polymerases assemble the present in P. stewartii, whose protein product
periplasmic und-PP-linked building blocks typically catalyzes the transfer of N-acetylhex-
into low-molecular-weight glycans, which can osamine-1-P to the lipid carrier as a first step in
serve as substrates for K-LPS, low-molecular- common antigen precursor synthesis (46). It is
weight CPS, and high-molecular-weight EPS possible that wceG2 satisfies the necessary levels
synthesis (23, 46). For the wce gene system, the und-PP-Gal generation under noninducing
assumption is that under noninducing condi- conditions and that expression of wceG1 associ-
tions internal promoters support low-level ated with the RcsA/B activated primary wce
constitutive expression, primarily of the 3′ promoter enhances the pool of available sub-
region genes, to give rise to a membrane- strate to accommodate the increased demand
bound CPSST (9, 23). CPSST is thought to be a during high-level EPSST synthesis.This predic-
low-molecular-weight version of EPSST com- tion awaits experimental verification.
posed of the same repeat units,although there is There is additional support for the hypothe-
little experimental evidence to verify this pre- sis that EsaR and RcsA/B control high-level
diction, nor is it known how CPSST is retained EPSST production. The wza/wzb/wzc genes
in the cell wall. A wzi gene, whose function is located adjacent to and downstream of wceG1
implicated in capsule surface attachment in are also strongly induced by RcsA/B (26, 37;
A.L.Carlier and S.B.von Bodman,unpublished notype is readily reversed by exogenous AHL.

data). As mentioned, these gene products are Thus, it is not surprising to find that the hyper-
dedicated for the high-molecular-weight EPS mucoid strain ESIR is virtually nonadherent.
polymerization and secretion (Fig. 2) (46).The In contrast, the wild-type strain exhibits low
primary wce promoter also features a “JUMP- but significant levels of adhesion at low cell
start/ops” element, which is implicated in the densities, with high-cell-density cultures
recruitment of the RfaH transcription elonga- behaving much like strain ESIR (22).
tion factor into the transcription complex (2, In terms of further in vitro biofilm develop-
33).This complex is capable of reading through ment, the wild-type strain, DC283, is capable of
intrinsic stem-loop attenuator(s) generally transitioning from an adhesion phase into a pro-
found in the region that separates the bipartite nounced microcolony developmental phase.
cps gene clusters (2, 33, 44, 46). Preliminary data Surface swarming motility, previously unrecog-
indicate the presence of a full-length transcript nized in P. stewartii, is believed to be involved in
under AHL-inducing conditions, which is microcolony formation (C. M. Herrera, M. D.
absent under noninducing conditions (A. L. Koutsoudis, and S. B. von Bodman, unpublished
Carlier and S. B. von Bodman, unpublished data). The microcolonies give rise to a mature
data).Finally,a separate RcsA/B-regulated gene biofilm characterized by towerlike three-
system encoding a predicted glycosyltrans- dimensional structures separated by region of
ferase, designated wceO, and a linked alternate voids (22). In contrast, the hyperadherent strain,
flippase gene, designated wzx2, appears to be ESN51, fails to form wild-type microcolonies
important for the synthesis of cell-free EPSST and instead forms numerous small bacterial
although the precise functional role of these aggregates, presumably as a result of clonal
two genes requires further investigation. growth, before colonizing the entire glass sur-
Together, these observations have led to the face as a compact bacterial mat.The hypermu-
working model depicted in Fig. 3. coid ESIR strain is capable of forming heavily
matrix-encased bacterial clumps, which in time
BIOLOGICAL SIGNIFICANCE OF EsaR develop into amorphous, nonadherent bacterial
QUORUM SENSING-CONTROLLED masses.
EPSST SYNTHESIS The in vitro biofilm behavior parallels a
The observation that P. stewartii expresses the fairly reasonable indication of how the bacteria
major EPSST virulence factor in a cell-density- behave in the xylem vessel (22). Specifically, the
dependent manner suggests a key role for the wild-type strain adheres to the xylem wall in
EsaI/EsaR quorum-sensing system in manag- specific places, with an apparent preference for
ing the transition between distinct phases of secondary wall structures (22). Adherent cells
bacterial/biofilm development, which may be develop into matrix-encased towerlike struc-
key to pathogen fitness during host coloniza- tures that grow toward the center of the xylem
tion.This is supported by the fact that repressed lumen, where they coalesce and eventually fill
and constitutive synthesis of the EPSST viru- the lumen of the xylem (22). In contrast, the
lence factor leads to loss or attenuated virulence hyperadherent, AHL-deficient strain, ESN51,
(22, 35, 39, 41). Hence, the wild-type and quo- forms highly compact biofilms devoid of
rum-sensing mutant strains were examined for EPS fibril material, while the hypermucoid
traits commonly associated with biofilm devel- strain, ESIR, develops less-densely-popu-
opment.These studies revealed that the AHL- lated, heavily matrix-enmeshed cellular clusters
deficient, EPSST-repressed mutant, ESN51, that collapse during drying procedures asso-
exhibits an unusually robust surface adhesion ciated with scanning electron microscopy
phenotype,as did mutants with insertions in the (SEM) (22). The aberrant biofilm characteris-
wce locus that are blocked for EPSST synthesis. tics of the quorum-sensing mutant strains sig-
In the case of strain ESN51, the adhesion phe- nificantly interfere with the pathogen’s ability
to effectively colonize the xylem and spread unpublished data). Further studies are necessary
within the plant host (22). to define the functional significance of the
Recently obtained SEM images of xylem rhlI/sdiA gene system in P. stewartii.
vessels colonized by the wild-type strain
DC283 and the AHL-deficient mutant ESN51 SUMMARY REMARKS
confirm the tendency by the bacteria to colo- P. stewartii features two potential quorum-
nize the secondary wall structures of the sensing regulatory systems.Only the EsaI/EsaR
xylem (Fig. 4). More interestingly, perhaps, is system has been extensively studied and charac-
the extensive network of gumlike strands asso- terized. In this system, EsaR dimerizes and
ciated with the wild-type strain DC283 that functions as a DNA-binding protein in an
appear to be largely bacterial in origin. The AHL-free state, capable of both repressing
structures seem to girdle the annular rings in an and activating target genes depending on the
almost parallel, semiorganized fashion, gener- position of the esaR box DNA-binding ele-
ally perpendicular to the grain of the annular ment within respective promoters. That at
rings (Fig. 4A).These strands also form a bridg- least two native EsaR-activated genes exist sug-
ing network across and between annular rings gests an even broader role for EsaR as a central
and individual bacterial cells (Fig. 4B). These switching mechanism between developmental
structures are notoriously absent in vessels phases. A major criterion for ApoEsaR DNA-
infected by strain ESN51 (EPSST repressed) binding activity is that it must occur under sub-
(Fig. 4D and 4E).The lack of excessive matrix threshold AHL conditions. Thus, EsaR may
in the ESN51 infections, however, exposes dif- correspondingly stimulate the expression of
ferent bacterial appendages and surface struc- developmentally keyed “low-cell-density”
tures involved in forming solid contacts traits, while repressing dedicated “high-cell-
between the bacterium and the xylem wall density” traits. A key area of future research
(Fig. 4D and 4E). While these images offer an (Stevens laboratory) is aimed at understanding
interesting first impression of how P. stewartii the structure/function criteria for ApoEsaR as
colonizes the xylem, they also reveal a complex a DNA-binding protein and the mechanisms
biology governed by EsaI/EsaR quorum- by which EsaR interacts with the AHL coin-
sensing regulation. ducer to neutralize the DNA-binding confor-
mation. Likewise, defining the physiological
A SECOND QUORUM-SENSING function of the native EsaR-activated genes
REGULATORY SYSTEM and identifying additional EsaR-regulated
P. stewartii features a second potential quorum- genes is an area of intense interest.
sensing regulatory gene system, making the That EsaR governs the regulation of the
quorum-sensing response more complex than RcsA auxiliary regulator in a scenario that leads
previously thought. The response regulator is to high-level expression of EPSST suggests that
homologous to sdiA with a genetic position a major role of the EsaR-mediated quorum-
next to the uvrY/uvrC gene system in the chro- sensing system is to govern the differential
mosome similar to E. coli and Salmonella (1). expression of surface glycopolymers. Definitive
However, P. stewartii also features a genetically experimental evidence about the composition
linked potential AHL signal synthase gene with and nature of the CPSST is largely lacking, as is
greatest homology to rhlI of Pseudomonas (8, 29, the role of other surface glycans produced by
31). Attempts to detect AHLs produced by the the organism. Sequence analysis of promoters
rhlI gene product expressed in native and het- associated with lipopolysaccharide biosynthetic
erologous chromosomal backgrounds were gene systems suggests that EsaR control of rcsA
negative by several bioassay reporter systems may have an effect on the expression of other
and liquid chromatography-tandem mass spec- glycopolymers in P. stewartii. Elucidating the
trometry (18;A. L. Carlier and M.A. Churchill, intricacies of these presumably complex regula-
FIGURE 4 P. stewartii colonization of xylem wall structures visualized by SEM. (A) An annular ring colonized by
the wild-type strain DC283.The black arrow indicates an extensive, gummy strandlike network that covers the
annular ring surfaces in heavily infected tissue. (B) These fibrils are less numerous in sparsely colonized areas
observed at higher magnification. (C) Annular ring excised from an uninfected xylem vessel. (D) The AHL, EPSST-
deficient strain, ESN51, also colonizes the annular rings. Note the absence of the gummy strands and the presence
of undefined appendages that mediate adhesion to the cell wall, as indicated by the black arrow (E).
tory systems and defining the nature, role, and AHL. This initial observation has led to the
timing of various glycopolymers will be development of an interesting biological system
another major area of future research interest that has provided key mechanistic insights into
(von Bodman laboratory).This focus becomes the AHL class of quorum sensing regulatory
even more compelling with the observations of systems.These includes the demonstration that
the extensive network of potential glycan- the LuxR homolog, EsaR, functions as a
based strands associated with the wild-type col- repressor in an AHL-independent manner and
onization of the maize xylem cell wall (Fig. 4). the first successful structure/function charac-
In addition, other surface localized functions terization of an AHL synthase (42).The latter is
will be characterized to identify components the specific focus of chapter 17.
that initiate the contact between the bacterium
and the xylem wall, as seen in infections with ACKNOWLEDGMENTS
the AHL mutant strain ESN51. We thank Carmen M. Herrera, Maria D. Koutsoudis,
Finally,it might be of interest to know that P. and Dimitrios Tsaltas in the von Bodman laboratory,
and Daniel J. Schu and Katherine A. Penney in the
stewartii was initially selected as an experimental Stevens laboratory for providing unpublished results
organism for quorum-sensing control because and helpful discussions during the development of this
of its capacity to synthesize high amounts of document.
This work was supported by the National Science 11. Egland, K. A., and E. P. Greenberg. 2001.
Foundation [(grant MCB-0211687 and MCB- Quorum sensing in Vibrio fischeri: analysis of the
0619104 [S. B.]); careerAward MCB-9875479 ([A. M. LuxR DNA binding region by alanine-scanning
S.]), United States Department of Agriculture National mutagenesis. J. Bacteriol. 183:382–386.
Research Initiative grant 2002-35319-12637 (S. v. B.), 12. Finney, A. H., R. J. Blick, K. Murakami, A.
Cooperative State Research Service, U.S. Depart- Ishihama, and A. M. Stevens. 2002. Role of the
ment of Agriculture under Project #CON00775 (S. v. C-terminal domain of the alpha subunit of RNA
B.) and National Institutes of Health grant GM066786 polymerase in LuxR-dependent transcriptional
(A. M. S.). activation of the lux operon during quorum sens-
ing. J. Bacteriol. 184:4520–4528.
REFERENCES 13. Fuqua, C., M. Burbea, and S. C.Winans. 1995.
1. Ahmer, B. M. M. 2004. Cell-to-cell signalling in Activity of the Agrobacterium Ti plasmid conjugal
Escherichia coli and Salmonella enterica. Mol. Micro- transfer regulator TraR is inhibited by the product
biol. 52:933–945. of the traM gene. J. Bacteriol. 177:1367–1373.
2. Artsimovitch, J., and R. Landick. 2002. The 14. Fuqua, C., S. C.Winans, and E. P. Greenberg.
transcriptional regulator RfaH stimulates RNA 1996. Census and consensus in bacterial ecosys-
chain synthesis after recruitment to elongation tems: the LuxR-LuxI family of quorum-sensing
complexes by the exposed nontemplate DNA transcriptional regulators. Annu. Rev. Microbiol.
strand. Cell 109:192–203. 50:727–751.
3. Barnard, A. M., and G. P. C. Salmond. 2006. 15. Geider, K. 2000. Exopolysaccharides of Erwinia
Quorum sensing in Erwinia species. Anal. Bioanal. amylovora:structure,biosynthesis,regulation,role in
Chem. 387:415–423. pathogenicity of amylovoran and levan. CAB
4. Bradshaw-Rouse, J. J., M. Whatley, D. International.
L. Coplin, A. Woods, L. Sequeira, and A. 16. Geider, K., Y. Zhang, H. Ullrich, and C.
Kelman. 1981.Agglutination of strains of Erwinia Langlotz. 1999. Expression of virulence
stewartii with a corn agglutinin: correlation factors secreted by Erwinia amylovora. Acta Hort.
with extracellular polysaccharide production 489:347–352.
and pathogenicity. Appl. Environ. Microbiol. 42: 17. Glasner, J. D., M. Rusch, P. Liss, G. R.
344–350. Plunket, E. L. Cabot, A. Darling, A. B. D., P.
5. Braun, E. J. 1982. Ultrastructural investigation Infield-Ham, M. C. Gislon, and N. T. Perna.
of resistant and susceptible maize inbreds 2006. ASAP: a resource for annotating, curating,
infected with Erwinia stewartii. Phytopathology comparing, and disseminating genomic data.
72:159–166. Nucleic Acids Res. 34 (Database issue):D41-D45.
6. Carlier, A. L., and S. B. von Bodman. 2006. 18. Gould, T. A., J. Herman, R. C. Murphy, and
The rcsA promoter of Pantoea stewartii subsp. stew- M. E. A. Churchill. 2006. Specificity of acyl-
artii features a low-level constitutive promoter and homoserine lactone synthases examined by mass
an EsaR quorum-sensing-regulated promoter. J. spectrometry. J. Bacteriol. 188:773–783.
Bacteriol. 188:4581–4584. 19. Ham, J. H., D. R. Majerczak, A. S. Arroyo-
7. Coplin, D. L., and D. Cook. 1990. Molecular Rodriguez, D. M. Mackey, and D. L. Coplin.
genetics of extracellular polysaccharide biosynthe- 2006.WtsE, an AvrE-family effector protein from
sis in vascular phytopathogenic bacteria. Mol. Pantoea stewartii subsp. stewartii, causes disease-asso-
Plant-Microbe Interact. 3:271–279. ciated cell death in corn and requires a chaperone
8. Davies, D. G., M. R. Parsek, J. P. Pearson, protein for stability. Mol. Plant-Microbe Interact.
B. H. Iglewski, J. W. Costerton, and E. P. 19:1092–1102.
Greenberg. 1998.The involvement of cell-to-cell 20. Hwang, I., A. J. Smyth, Z. Q. Luo, and S. K.
signals in the development of a bacterial biofilm. Farrand. 1999. Modulating quorum sensing by
Science 280:295–298. antiactivation:TraM interacts with TraR to inhibit
9. Dolph, P. J., D. R. Majerczak, and D. L. activation of Ti plasmid conjugal transfer genes.
Coplin. 1988. Characterization of a gene cluster Mol. Microbiol. 34:282–294.
for exopolysaccharide biosynthesis and virulence 21. Johnson, D. C., A. Ishihama, and A. M.
in Erwinia stewartii. J. Bacteriol. 170:865–871. Stevens. 2003. Involvement of region 4 of the
10. Dong, Y.-H., X.-F. Zhang, J.-L. Xu, A.-T. sigma70 subunit of RNA polymerase in transcrip-
Tan, and L.-H. Zhang. 2005. VqsM, a novel tional activation of the lux operon during quorum
AraC-type global regulator of quorum-sensing sensing. FEMS Microbiol. Lett. 228:193–201.
signalling and virulence in Pseudomonas aeruginosa. 22. Koutsoudis, M. D., D.Tsaltas,T. D. Minogue,
Mol. Microbiol. 58:552–564. and S. B. von Bodman. 2006. Quorum sensing
regulation by Pantoea stewartii subsp. stewartii, a crit- Escherichia coli K30 group 1 capsule biosynthesis
ical factor in bacterial adhesion, biofilm develop- (cps) gene cluster. Mol. Microbiol. 47:1045–1060.
ment and host colonization. Proc. Natl. Acad. Sci. 35. Ramey, B. E., M. Koutsoudis, S. B. von
USA. 103:5983–5988. Bodman, and C. Fuqua. 2004. Biofilm forma-
23. Leigh, J. A., and D. L. Coplin. 1992. tion in plant-microbe associations. Curr. Opin.
Exopolysaccharides in plant-bacterial interactions. Microbiol. 7:602–609.
Annu. Rev. Microbiol. 46:307–346. 36. Reeves, P. R., M. Hobbs, M. A. Valvano, M.
24. Luo, Z. Q., and S. K. Farrand. 1999. Signal- Skurnik, C.Whitfield, D. L. Coplin, N. Kido, J.
dependent DNA binding and functional domains Klena, D. Maskell, C. R. H. Raetz, and P. D.
of the quorum-sensing activator TraR as identified Rick. 1996.Bacterial polysaccharide synthesis and
by repressor activity. Proc. Natl. Acad. Sci. USA gene nomenclature. Trends Microbiol. 4:495–502.
96:9009–9014. 37. Torres-Cabassa, A., S. Gottesman, R. D.
25. Majdalani, N., and S. Gottesman. 2005. The Frederick, P. J. Dolph, and D. L. Coplin. 1987.
Rcs phosphorelay: a complex signal transduction Control of extracellular polysaccharide synthesis in
system. Annu. Rev. Microbiol. 59:379–405. Erwinia stewartii and Escherichia coli K-12:a common
26. Minogue, T. D., A. L. Carlier, M. D. regulatory function. J. Bacteriol. 169:4525–4531.
Koutsoudis, and S. B. von Bodman. 2005.The 38. von Bodman, S. B., J. K. Ball, M. A. Faini, C.
cell density-dependent expression of stewartan M. Herrera, T. D. Minogue, M. L.
exopolysaccharide in Pantoea stewartii ssp.stewartii is Urbanowski, and A. M. Stevens. 2003. The
a function of EsaR-mediated repression of the rcsA quorum sensing negative regulators EsaR and
gene. Mol. Microbiol. 56:189–203. ExpREcc, homologues within the LuxR family,
27. Minogue, T. D., M. Wehland-von Trebra, F. retain the ability to function as activators of tran-
Bernhard, and S. B. von Bodman. 2002. scription. J. Bacteriol. 185:7001–7007.
The autoregulatory role of EsaR, a quorum- 39. von Bodman, S. B., W. D. Bauer, and D. L.
sensing regulator in Pantoea stewartii ssp. stewartii: Coplin. 2003. Quorum sensing in plant-patho-
evidence for a repressor function. Mol. Microbiol. genic bacteria.Annu.Rev.Phytopathol. 41:455–482.
44:1625–1635. 40. von Bodman, S. B., and S. K. Farrand. 1995.
28. Nimtz, M., A. Mort, V. Wray, T. Domke, Y. Capsular polysaccharide biosynthesis and patho-
Zhang, D. L. Coplin, and K. Geider. 1996. genicity in Erwinia stewartii require induction by an
Structure of stewartan, the capsular exopolysac- N-acylhomoserine lactone autoinducer. J. Bacte-
charide from the corn pathogen Erwinia stewartii. riol. 177:5000–5008.
Carbohydr. Res. 288:189–201. 41. von Bodman, S. B., D. R. Majerczak, and D.
29. Ochsner, U. A., and J. Reiser. 1995. Autoin- L. Coplin. 1998. A negative regulator mediates
ducer-mediated regulation of rhamnolipid biosur- quorum-sensing control of exopolysaccharide
factant synthesis in Pseudomonas aeruginosa. Proc. production in Pantoea stewartii subsp. stewartii. Proc.
Natl.Acad. Sci. USA 92:6424–6428. Natl.Acad. Sci. USA 95:7687–7692.
30. Pataky, J. K. 2003. Stewart’s wilt of corn. In M. L. 42. Watson,W.T.,T. D. Minogue, D. L.Val, S. Beck
Elliott (ed.), APSnet Feature Story. American Phy- von Bodman, and M. E. A. Churchill. 2002.
topathological Society, St. Paul, MN. Structural basis and specificity of acyl-homoserine
31. Pesci, E. C., and B. H. Iglewski. 1997. The lactone signal production in bacterial quorum
chain of command in Pseudomonas quorum sens- sensing. Mol. Cell 9:685–694.
ing. Trends Microbiol. 5:132–135. 43. Wehland, M., and F. Bernhard. 2000. The
32. Qin, Y., Z. Q. Luo, A. J. Smyth, P. Gao, S. RcsAB box. Characterization of a new operator
Beck von Bodman, and S. K. Farrand. essential for the regulation of exopolysaccharide
2000. Quorum-sensing signal binding results in biosynthesis in enteric bacteria. J. Biol. Chem.
dimerization of TraR and its release from 275:7013–7020.
membranes into the cytoplasm. EMBO J. 19: 44. Wehland, M., C. Kiecker, D. L. Coplin, O.
5212–5221. Kelm, W. Saenger, and F. Bernhard. 1999.
33. Rahn, A., J. Drummelsmith, and C. Identification of an RcsA/RcsB recognition motif
Whitfield. 1999. Conserved organization in the in the promoters of exopolysaccharide biosyn-
cps gene clusters for expression of Escherichia coli thetic operons from Erwinia amylovora and Pantoea
group 1 K antigens: relationship to the colanic acid stewartii subspecies stewartii. J. Biol. Chem. 274:
biosynthesis locus and the cps genes from Klebsiella 3300–3307.
pneumoniae. J. Bacteriol. 181:2307–2313. 45. White, C. E., and S. C.Winans. 2005. Identifi-
34. Rahn, A., and C. Whitfield. 2003. Transcrip- cation of amino acid residues of the Agrobacterium
tional organization and regulation of the tumefaciens quorum-sensing regulator TraR that are
critical for positive control of transcription. Mol. 47. Zhu, J., and S. C. Winans. 2001.The quorum-
Microbiol. 55:1473–1486. sensing transcriptional regulator TraR requires its
46. Whitfield, C. 2006. Biosynthesis and assembly of cognate signaling ligand for protein folding, pro-
capsular polysaccharides in Escherichia coli. Annu. tease resistance, and dimerization. Proc. Natl.Acad.
Rev. Biochem. 75:39–68. Sci. USA 98:1507–1512.
CELL-TO-CELL COMMUNICATION IN
RHIZOBIA: QUORUM SENSING AND
PLANT SIGNALING
J. Allan Downie and Juan E. González
14
Communication is a key aspect of the lifestyle their chances of being at the appropriate sites
of legume-nodulating rhizobia.The communi- for infection to occur (77, 100). Rhizobia
cation with legumes and the consequent infec- form populations on root surfaces where they
tion of roots and nodule formation enable the communicate via N-acyl-homoserine lactone
bacteria to enter an ecological niche, within (AHL)-based signaling systems (19, 47, 131).
which they can grow rapidly (108 to 109 bacte- This probably enhances occupation of such
ria in a nodule) without competition from niches and plays other more direct roles in sym-
other bacteria (47, 94). Even though the major- biosis and stimulates genetic exchange among
ity of bacteria in legume nodules do not survive the bacterial populations in the rhizosphere. In
nodule senescence, even a 0.1% survival rate this review, we briefly describe the nodulation
would lead to a great increase in bacterial num- process and some of the quorum-sensing regu-
bers.Therefore, rhizobial growth in the rhizo- latory systems that rhizobia use to monitor
sphere and nodulation signaling are very their population density.
important characteristics for a given rhizobial
strain to succeed in infecting a legume root. RHIZOBIUM-LEGUME SIGNALING
There appears to be a general signaling mecha- Rhizobial Perception
nism by which rhizobia generate Nod factors of Legume Signals
to initiate legume infection and nodule mor- Rhizobia recognize and metabolize many
phogenesis (88), although some unusual compounds in root exudates, which are a
legumes can be nodulated by strains of rhizobia potentially rich source of nutrients to which the
that do not make Nod factors (43). However, bacteria are attracted by chemotaxis (82). How-
there are probably many mechanisms by which ever, flavonoids and isoflavonoids (and some
rhizobia can optimize their growth and attach- other compounds such as stachydrine and
ment to root hairs so that they can enhance trigonelline) are specifically recognized by rhi-
zobia and these induce the expression of the
J. Allan Downie John Innes Centre, Colney Lane, Norwich bacterial nod genes (47, 106). Several of the
NR4 7UH, United Kingdom. Juan E. González
Department of Molecular and Cell Biology, University of nod genes are essential for nodulation (27, 112)
Texas at Dallas, Richardson,Texas 75083-0688. and are usually regulated by NodD proteins,
213
214 ■ DOWNIE AND GONZÁLEZ
transcriptional regulators belonging to the modifications determine a wide range of Nod

LysR family of regulators (53). Multimeric factors that can be made by different rhizobia.
forms of NodD bind to extended conserved In addition to Nod factors, some rhizobia
promoter regions of 45 to 50 nt upstream of nod secrete proteins that can influence legume
gene operons (31, 33), and flavonoids or nodulation. One of these proteins (NodO),
isoflavonoids activate transcription by binding which influences the range of legumes nodu-
to NodD (15, 33).Transcription by each NodD lated by Rhizobium spp., is secreted by a type I
protein is induced by different flavonoids, so, for secretion system (32) and is thought to enhance
example, the rhizobia that nodulate peas and infection by forming cation-selective pores in
vetch induce nod genes in response to flavonoids root-hair membranes (116, 127). Proteins such
such as eriodictyol and naringenin, whereas the as NolA and NolB secreted via a type III secre-
bacteria that nodulate soybeans induce nod tion system can affect the formation of
genes in response to isoflavonoids such as genis- appendages such as pili and a rhamnose-rich
tein (106).The rhizobia that have multiple nodD polysaccharide, which can influence nodula-
genes have the potential to recognize a wider tion specificity (22, 72, 73, 109), while other
variety of plant-made signals (94). proteins, such as NopL, can be delivered into
plant cells where they can interfere with plant
Rhizobial Nod-Factor Signals defense reactions (4).
Inducing Nodulation
Several of the induced nod gene products are Legume Recognition of
involved in the synthesis and secretion of Nod Rhizobial Signals
factors, which have a backbone of usually four Nod factors are perceived by specific plant
or five -1, 4-linked N-acetyl glucosamine receptors that are transmembrane kinases with
residues and a long acyl chain replacing the N- extracellular domains that are rather like LysM
acetyl group on the terminal nonreducing chitin-binding motifs (70).The receptors bind
sugar. As reviewed previously (27, 47, 80, 94, Nodfactors with high specificity, because 1012
112), NodM produces N-acetyl glucosamine, to 1013 M Nod factor can induce responses
NodC polymerizes UDP-N-acetyl glu- such as deformation of root hairs and oscilla-
cosamine into the sugar backbone, NodB tions in intracellular calcium (88). Localized
removes the acetyl group from the terminal changes to the microtubule cytoskeleton cause
nonreducing sugar, and then NodA attaches a the root hairs to bend back on themselves
long chain acyl group to the free amine.Specific (108), entrapping rhizobia in the crook of the
modifications to this basic lipo-oligosaccharide curl. As the bacteria grow in this niche, the
structure are catalyzed by the various different increased localized levels of Nod factor may
nod gene products that are found in diverse rhi- then promote infection thread growth (83).To
zobial strains. Thus, for example, NodX and this end,the direction of growth of the root hair
NodL can add O-linked acetyl groups to the tip is reversed,producing a tunnel-like structure
terminal reducing and nonreducing residues, that grows within the plant cell and is rather
respectively; NodE and NodF can synthesize similar to an intercellular space (9).The bacteria
novel acyl groups that can be attached by grow at the tip of this tunnel,forming a column
NodA;NodH,NodP,and NodQ can synthesize of bacteria only a couple of cells wide (38, 39).
and attach a sulfate group to the reducing sugar; The net effect of the limited growth at the tip is
NodS can attach an N-linked methyl group to that, even if two different bacteria are included
the acylated sugar; and NodU can add a car- at the initiation of an infection thread, only one
bamoyl group.There are also sugar transferases tends to be selected as the tip of the infection
that can add different sugars such as arabinose thread grows (38).
and fucose (and modified derivatives) to the The biosynthesis of different rhizobial sur-
backbone. Different combinations of these face polysaccharides has also been implicated in
14. QUORUM SENSING AND PLANT SIGNALING IN RHIZOBIA ■ 215
the initiation and elongation of infection In parallel with the infection, Nod factors
threads. Most rhizobia produce a variety of induce nodule morphogenesis by activating cell
polysaccharides (110), and it was hypothesized division in root cortical cells (88).The infection
that they might play roles in bacteria-plant thread grows down through successive layers of
interactions, such as being at least partially root cells until it meets the growing meristem,
responsible for the host specificity of various where the infection threads then branch into
rhizobia.Arguably, the best-characterized sym- the meristematic cells (84).At this point the cell
biotically important exopolysaccharides are wall of the infection thread is no longer made
those in Sinorhizobium meliloti (34, 110). S. and the bacteria are budded off, surrounded by
meliloti is capable of synthesizing two different a plant-made membrane that was derived from
exopolysaccharides, succinoglycan and EPS II. the plasma membrane (9).The bacteria and the
The synthesis of at least one of these membrane differentiate into symbiosomes,
exopolysaccharides is absolutely required for which are effectively organelles containing dif-
the development of normal nitrogen-fixing ferentiated bacteria called bacteroids, and there
nodules (45). Both succinoglycan and EPS II are hundreds of symbiosomes per plant cell. In
are secreted in two major fractions, high- and some legumes the bacteria undergo several
low-molecular-weight polymers. Mutants that rounds of endoreduplication, becoming large
are unable to produce the low-molecular- and polyploid; this increase in ploidy depends
weight fraction of either polymer form empty on the legume rather than the rhizobial strain
nodules that lack bacteria and bacteroids (48, (81).The bacteroids dedicate their metabolism
61, 124) and are similar to the nodules elicited to the reduction of N2 to NH3, switching off
by treatment of alfalfa roots with purified Nod the major enzyme of ammonia assimilation
factor (120). Root-hair curling is delayed and (glutamine synthetase) and relying on amino-
normal infection threads are not seen in the acid exchange with the plant cells for their own
curled root hairs; infection threads are detected nitrogen (63). The plant supplies dicarboxylic
upon sectioning, but these abort within the acids such as malate as a carbon source for the
peripheral cells of the developing nodule (16, bacteroids (98). The plant cytoplasm contains
90).The nodules elicited by mutants unable to relatively high levels of plant-made leghemo-
make either exopolysaccharide appear to be globin.This oxygen-binding protein buffers the
arrested at an intermediate state of nodule free oxygen at a low level, while maintaining a
development, and only 2 of 17 nodule-specific high oxygen flux to the bacteroids (2).Thus,the
plant proteins (nodulins) identified in nodules nodule provides a perfect environment for the
containing wild-type bacteria were expressed oxygen-sensitive nitrogenase reductase and
in these plants (85). A subset of plant defense nitrogenase enzymes in the bacteroids, thereby
responses appears to be induced (86), suggest- producing one of the most efficient nitrogen-
ing a possible role for exopolysaccharides in fixing systems found in nature (28). Ammonia
modulating the plant defenses. released by the symbiosomes is assimilated by
Exopolysaccharides also play an important glutamine synthetase in the plant cell and, fol-
role in the symbiotic role of Rhizobium legumi- lowing transamination, is translocated out of
nosarum.Both biovars (viciae and trifolii) produce the nodule in the form of asparagine or ureides
the same conserved octasaccharide repeating such as allantoin (123).
unit of EPS (87), but minor strain variations
have been reported (1, 11). Defects in the syn- CONJUGAL TRANSFER IN RHIZOBIA
thesis of EPS have pleiotropic effects that Conjugation is common among Rhizobiaceae,
include a dramatic increase in the synthesis and and there are very strong selection pressures to
secretion of cyclic -1, 2-glucans (6, 59) and optimize growth in the rhizosphere and nodu-
result in empty nodules with clear symptoms of lation competitiveness (8, 19, 131). This selec-
plant defense reactions (130). tion has even enabled transfer of nodulation
and nitrogen-fixation genes from rhizobia ally in the traI-trb operon (Fig. 1) and so is
(which are in the -division of the proteobac- induced with positive feedback by TraI-made
teria) to totally unrelated bacteria (similar AHLs, resulting in high levels of induction of
to Burkholderia and Ralstonia spp.) in the - the tra/trb operons as the bacterial population
division of the proteobacteria (14).Several gen- density increases (30, 96).TraM is an antiactiva-
era of bacteria can nodulate legumes (105), tor of TraR, forming a stable oligomeric com-
and horizontal transfer of nodulation and plex probably containing two TraR and two
nitrogen-fixation genes occurs by conjugation. TraM dimers (125) titring out TraR (36, 95),
There is good evidence that genes on so-called thereby attenuating gene induction, particularly
“cryptic” plasmids can enhance bacterial at low population densities (69). The traI-trb
growth in the rhizosphere and hence nodule operon is transcribed divergently from the
competitiveness (3, 23, 79, 89). However, it is repABC operon (Fig. 1), which is also enhanced
not only plasmid transfer that plays an impor- in expression due to TraR activation of an
tant role; one of the most compelling examples inducible alternative promoter upstream of
of gene transfer resulted from the mobilization repA, resulting in enhanced plasmid copy num-
of a chromosomal symbiosis island. Seven ber during conjugation (10).
years after inoculation of a strain of Mesorhizo- As illustrated (Fig. 1), similar traI, traR, and
bium loti into an area devoid of naturalized M. traM genes are conserved on pRL1JI, the
loti, over 80% of the bacteria isolated from Lotus pSYM of R. leguminosarum bv. viciae 248 (21,
corniculatus nodules were genetically different 131); p42a from Rhizobium etli CFN42 (121);
from the introduced strain but had acquired pNGR234a, the pSYM of NGR234 (14, 35);
the chromosomal symbiosis gene region of and pRme41a of S. meliloti Rm41 (47, 75), and
the inoculant strain (113). in general terms, the regulation of plasmid
The regulation of gene transfer by quorum- transfer by these genes is similar to that seen in
sensing regulation is common (but not ubiqui- agrobacteria. Genome sequencing has revealed
tous) among rhizobia. Although we focus here somewhat different gene arrangements (Fig. 1)
on quorum-sensing regulation, it should be on the transmissible plasmids pRL7JI and
noted that there are other mechanisms of pRL8JI in R. leguminosarum 3841 (133) and the
inducing conjugation, including a novel repres- chromosomal island of M.loti strain R7A (115),
sion of plasmid conjugation by a repressor all of which also seem likely to have a TraR-
encoded by rctA and the induction of transfer mediated induction of conjugation.
via an unusual traA-encoded cis-acting relaxase The best-studied plasmid transfer in R. legu-
that catalyzes DNA-strand-specific cleavage at minosarum is with pRL1JI, which is transferred
the nic site of oriT (92, 93). at very high rates (54).The conjugal transfer of
The paradigm for quorum-sensing regula- pRL1JI is regulated via TraI-made AHLs acti-
tion of conjugation was first established for plas- vating TraR to induce plasmid transfer operons
mid transfer between agrobacteria (129).Three (Fig. 2).TraI from pRL1JI in R. leguminosarum
regulatory genes that are conserved are traI, determines the production of several AHLs, the
traR, and traM. TraI-makes N-oxo-octanoyl- most abundant being 3-O-C8-HSL and C8-
homoserine lactone (3-O-C8-HSL), and the HSL. Both of these AHLs induce the traI-
induction of traI is under the control of TraR. trbBCDEJKLFGHI operon and so traI is posi-
The traR genes are induced by plant-made tively autoregulated (21). Adjacent to traR is
opines. Once traR is induced, TraR binds the traM, which attenuates the activity of TraR on
TraI-made 3-O-C8-HSL, stabilizing the forma- the promoter of the traI-trb operon (21). A
tion of TraR dimers (126,136) that bind to “tra- major difference compared with Agrobacterium
box” sequences in the promoters of the strains is the regulation of traR. Adjacent to traR
traItrbBCDEJKLFGHI, traAFBH, and traCDG on pRL1JI is bisR (Fig. 1), encoding another
operons (30, 37, 139, 140).The traI gene is usu- LuxR-type regulator (131), which induces the
FIGURE 1 Arrangement of plasmid transfer and replication genes in different plasmids

in the Rhizobiacea.The characterized genes involved in plasmid replication (repABC), con-
jugation (trbBCDEJKLFGHI, traAFBH, traCDG), and regulation of conjugation (traI,
traR, traM, bisR) are shown.The traR and traM genes on pRL8JI are in the opposite orien-
tation to those in the other strains. The traABFH and traCDG genes have not been
sequenced on pRL1JI and pRme41 but are probably present.The trbBCDEJKLFGHI and
traG genes are not present on pRL7JI.
traR promoter in response to CinI-made 3- ficient to totally block 3-OH-C14:1-HSL pro-

OH-C14:1-HSL (21). However, BisR also duction, and so some transfer can occur even in
represses the expression of cinI on the chromo- the absence of 3-OH-C14:1-HSL production by
some (62, 131), and so when pRL1JI is present, potential recipient strains (131). As in Agrobac-
R. leguminosarum strains produce very little 3- terium tumefaciens, when the traI-trb operon is
OH-C14:1-HSL. This dual induction and induced, the divergently transcribed plasmid
repression by BisR lead to a mechanism of replication repABC operon (Fig. 2) is coin-
recipient-induced plasmid transfer (Fig. 2). duced under TraR and AHL control (78).This
Potential recipient strains of R. leguminosarum coordinated regulation appears to be mediated
lacking pRL1JI (and bisR) produce CinI-made via conserved tra-box promoter sites upstream
3-OH-C14:1-HSL, which can activate BisR in of both repA and traI.
nearby donor strains carrying pRL1JI. This A similar gene arrangement to that on
induces traR and the concomitant population- pRL1JI is seen on the (nonsymbiosis) plasmid
density-dependent induction of the tra/trb p42a in R. etli CFN42 (Fig. 1), and the bisR-like
operons meditated via TraI-made AHLs activat- gene adjacent to traR on that plasmid is also
ing TraR (21). In pRL1JI-containing strains, required for traR induction (121), suggesting a
induction of traR expression does not normally similar mechanism of recipient-induced induc-
occur, because of the BisR repression of cinI tion. However, although there is a traM gene on
expression and the consequent lack of 3-OH- p42a, there is no evidence showing that its
C14:1-HSL production. However, if the popula- product can act as an antiactivator of TraR,pos-
tion density of a pRL1JI-containing strain gets sibly explaining the apparent constitutive con-
to be very high, the repression by BisR is insuf- jugation observed with this strain.
FIGURE 2 Model for recipient-induced transfer of plasmid pRL1JI from R. leguminosarum bv. viciae.
BisR is activated by CinI-made 3-OH-C14:1-HSL from potential recipients (i), resulting in the induction
of traR expression.The traR induction of traI (ii) and the resulting TraI-made AHLs induce a positive feed-
back loop (iii),causing high-level expression of the traI-trb operon;the plasmid-replication repABC operon
is induced in parallel with the traI-trb operon. Premature accumulation of TraR is avoided by the expres-
sion of the TraR anti-activator TraM (iv). BisR, produced when pRL1JI is present, represses cinI expression
(v), thereby preventing 3-OH-C14:1-HSL activation of TraR in pRL1JI-containing strains.The repression
of cinI and induction of traR by BisR enable the pRL1JI donor to respond to CinI-made 3-OH-C14:1-
HSL from potential recipients. Figure adapted from Danino et al. (21) with permission from Blackwell
Publishing.
Recipient-induced plasmid transfer is The mechanisms of induction of the traR

unlikely in R. leguminosarum bv. viciae strain genes on pNGR234a and pRme41a are not
3841, because bisR is not found in the genome known (14, 47, 75). TraI-made AHLs can be
sequence (http://www.sanger.ac.uk/Projects/ detected in culture supernatants of NGR234
R_leguminosarum/).There are six plasmids in (14) and S.meliloti strain Rm41 (75),but in both
strain 3841, accounting for about one-third of strains the frequency of conjugal transfer is low
the total genome of 7.75 Mb (133).The sym- (around 1 transconjugant/107 to 109 donors)
biosis plasmid pRL10JI appears to be nontrans- (14, 19, 47). In NGR234 this low frequency
missible, but pRL7JI and pRL8JI are (55). of conjugation is unaffected by constitu-
pRL8JI has traR located between the trbBCDE- tive expression of traR, and so induction of
JLFGHI and the traAFBH operons (Fig. 1) plasmid transfer may need to be activated
(http://www.sanger.ac.uk/Projects/R_legu by some unidentified factor or unusual growth
minosarum/).There is no traI gene on pRL8JI, conditions (14). However, although TraR
but there is a traI-like gene adjacent to traAFBH induced some tra operons in response to 3-O-
and traDC on pRL7JI (Fig. 1). Probably AHLs C8-HSL in pNGR234a, the traAFB operon
determined by traI on pRL7JI activate TraR was not significantly expressed, even though
(encoded on pRL8JI) to induce transfer of both there appeared to be a conserved tra box
plasmids. However, there is as yet no proof that (14). This lack of traAFB induction could
traI and traR are required for conjugation of explain the very low levels of transfer of
pRL7JI and pRL8JI. pNGR234a.
Conjugal transfer of the symbiosis island of of the diversity of genes regulated by quorum
M. loti is different from plasmid transfer because sensing.Therefore, in this review we deal sepa-
it requires the excision and integration as well rately with what is known of the quorumsens-
as transfer of the excised island (around 500 kb) ing regulatory systems of those rhizobia that
(114). The integration requires intS, which have been worked on most intensively.
encodes a phage P4-type integrase (115) that
promotes integration into the phetRNA gene R. leguminosarum
in a manner that reconstructs the phetRNA There are three biovars of R. leguminosarum:
gene. Excision of the symbiosis island requires strains of biovar viciae nodulate peas, lentils, and
both intS and a recombination directionality field (Faba) bean; strains of biovar trifolii nodu-
factor encoded by rdfS. In exponential cultures, late clovers; strains of biovar phaseoli nodulate
excision occurred at a low frequency (in varieties of Phaseolus bean. Genes encoding
around 0.05% of the cells), but increased about four different types of AHL synthase/regulator
20-fold in stationary-phase cultures (101). One systems have been described in different strains
traR-like and two traI-like genes are located on (131), and these are cinI/cinR, rhiI/rhiR,
the symbiosis island, and when the cloned traR raiI/raiR, and traI/traR. The proteomes of a
and traI2 genes were introduced into wild-type wild-type strain (UPM791) of R.leguminosarum
M. loti, the AHL levels were greatly increased bv. viciae and a derivative lacking AHLs due to
and the symbiosis island was excised in 100% of expression of an AHL lactonase were com-
the cells.This strongly suggests that the excision pared, showing that about 1.7% of the resolved
is regulated by quorum sensing via a TraI and proteins were altered more than twofold.About
TraR quorum-sensing system and that intS and half increased in levels, suggesting the possibil-
rdfS are likely to be induced along with other ity that there might be some repression of gene
genes required for conjugal transfer (101). expression (12).
However, the details of the regulation and
AHLs produced have still to be described.
cinI AND cinR
cinI was so named because CinI makes a
Quorum-Sensing Systems long-chain AHL (3-OH-C14:1-HSL) that had
in Different Rhizobia initially been identified because it had bacteri-
Individual rhizobial strains can contain up to ocin-like properties. Working independently,
four different LuxI-type AHL synthases and one group (107) isolated what was thought to
associated regulators plus several other LuxR- be a bacteriocin called small, while others (50)
type regulators lacking dedicated AHL syn- isolated an AHL inducer of the rhiABC genes,
thases. Table 1 summarizes work on rhizobial noticing that the AHL inhibited the growth of
quorum-sensing systems. Each strain that has some (“small bacteriocin”-sensitive) strains of
been characterized seems to have a different R. leguminosarum bv. viciae; structural analyses
complement of quorum-sensing regulatory revealed that the two groups had identified the
systems (19, 47, 131). On the basis of gene sim- same AHL. Subsequently, cinI was shown to
ilarities, it is also clear that there has been determine the synthesis of 3-OH-C14:1-HSL
horizontal transfer of these genes, sometimes and to be induced in response to 3-OH-C14:1-
between rhizobia that are relatively distantly HSL by the product of the adjacent gene cinR
related (137).Therefore, there is no single para- (62), which probably binds to a domain 60 nt
digm for quorum-sensing regulation in rhizo- upstream of cinI (78).The expression of nearby
bia and only recently,with the availability of the genes was also decreased in cinI and cinR
genome sequences of different rhizobia (40, 49, mutants (62).The apparent induction of the rhi-
56, 57, 133) and the associated analyses of tran- ABC genes by 3-OH-C14:1-HSL was due to an
script levels (52),we are getting a clearer picture indirect effect, because cinI and cinR are at the
TABLE 1 Rhizobial quorum-sensing systems
Rhizobial strain Gene (location) Signal Phenotypes regulated Reference(s)
R. leguminosarum
bv. viciae cinR/cinI (chromosome) 3-OH-C14:1-HSL Growth inhibition 62
rhiR/rhiI (pRLIJI) C6-HSL, C7-HSL, C8-HSL Nodulation efficiency 17, 102
traR/traI (pRLIJI) 3-O-C8-HSL, C8-HSL Plasmid transfer 21, 131
excR (chromosome) Unknown Unknown 133
bv. phaseoli raiR/raiI (nonsymbiotic plasmid) 3-OH-C8-HSL, C8-HSL Unknown 132
R. etli
CNPAF512 cinR/cinI (chromosome) 3-OH-(slc)-HSL Nitrogen fixation, symbiosome 18
development, growth inhibition
raiR/raiI (chromosome) Short-chain AHLs Nitrogen fixation, growth 18, 103
inhibition
CNP42 traR/traI (p42a) 3-O-C8-HSL, 3-OH-C8-HSL Plasmid transfer 121
S. meliloti
Rm1021 sinR/sinI (chromosome) 3-O-C14-HSL, C16:1-HSL, 3- EPS II production, swarming 41, 47, 75, 76, 93, 117
O-C16:1-HSL, 3-O-C16-
HSL, C18-HSL, C12-HSL
expR (chromosome) C16:1-HSL EPS II production, swarming 41, 74, 91
Rm41 traR/traI (pRm41a) 3-O-C8-HSL Plasmid transfer 75
RU10/406 visN/visR (chromosome) Unknown effector Motility (flagellar regulon: fli, mot, 111
fla, and che genes)
Rhizobium sp.
NGR234 traR/traI (pNGR234a) 3-oxo-C8-HSL Plasmid transfer 51
Unknown genes (chromosome) Other AHLs Growth inhibition
M. loti R7A traR/traI2, traI2 Not defined Symbiosis island transfer 101
M. itanshanense mrtR/mrtI Not defined Legume nodulation 138
M. huakii Not defined Biofilms and legume nodulation 42, 128
B. japonicum
USDA110 Unknown Bradyoxetin nod gene control 64, 66
USA 10/290 and Unknown Several AHLs Unknown 7, 97
B. elkanii
top of a quorum-sensing hierarchy that is and located between the nodulation (nod) and
required for the normal induction of the rhiI- nitrogen-fixation (nif) genes in R.leguminosarum
rhiR quorum-sensing system (62, 131), as well bv. viciae (24). RhiR regulates rhiI and the rhi-
as the raiI-raiR (131) and traI-traR (21) encoded ABC operon in response to RhiI-made C6-
quorum-sensing systems. HSL, C7-HSL, and C8-HSL (102). Although
The growth inhibition by 3-OH-C14:1-HSL mutations of the rhi genes alone do not affect
(131) is related to high levels of activated TraR, nodulation, mutations in rhiR or rhiA decreased
because growth inhibition requires traI (or TraI- nodulation in strains that are already decreased
made AHLs) in addition to 3-OH-C14:1-HSL for nodulation due to the absence of some of
induction of traR (21, 131). One protein the nod genes (17).RhiA is highly expressed and
strongly upregulated during growth inhibition present in all biovar viciae strains that nodulate
is the translation factor Ef-Ts (131). A similar peas and vetch but absent from even very
growth inhibition has been observed when traR closely related biovars trifolii and phaseoli that
from Rhizobium sp. NGR234 is expressed at nodulate clovers and Phaseolus beans (24).This,
artificially high levels (14); however, the growth together with the localization in the symbiosis
inhibition cannot be due to the induction of cluster and nodulation phenotype, suggests that
the identified tra/trb genes because the the rhiABC genes may allow R. leguminosarum
inhibitory effect is seen when traR is induced bv. viciae to interact better with the roots of a
strongly in Rhizobium and Agrobacterium strains specific group of legumes as the rhizobial popu-
lacking other tra/trb genes (51, 131). lation density increases.
The cinI and cinR genes in R. leguminosarum
strains clearly play a role in plasmid transfer, but raiI AND raiR
beyond that their physiological role remains RaiR induces raiI in response to the RaiI-made
unclear. Mutation of cinI or cinR does not affect 3-OH-C8-HSL and C8-HSL, but other genes
growth, nodulation, or nitrogen fixation on regulated by RaiR have not been identified
peas and vetch, although mutations in the (131).These genes were identified in R. legumi-
orthologous genes in R. etli caused slower nosarum bv. phaseoli strain 8002 and are absent
growth and decreased symbiotic nitrogen fixa- from the genome of the sequenced strain of R.
tion in Phaseolus bean (18), while mutations in leguminosarum bv. viciae 3841 (http://www.
the orthologous genes (mrtI and mrtR) in sanger.ac.uk/Projects/R_leguminosarum/).
Mesorhizobium tianshanense blocked nodule for- Mutations in cinI or cinR decrease raiI expres-
mation on its host Glycyrrhiza uralensis (137). sion and RaiI-made AHLs (131), but the link
It remains to be determined whether the between the induction of cinI and raiI has not
expression of similar or different groups of been described.
genes is influenced by cinI and cinR in these dif-
ferent strains. 3-OH-C14:1-HSL can influence traI AND traR
adaptation to starvation stress, because R. legu- As described above, strains of R. leguminosarum
minosarum bv. phaseoli entering stationary phase can have several plasmids, some of which carry
at a low population density survived much less traI and traR genes that induce plasmid transfer.
well than those entering stationary phase at a One of the consequences of the strong induc-
high population density, and added 3-OH- tion of traI is the production of relatively high
C14:1-HSL significantly enhanced long-term levels of AHLs (21), which can influence the
viability to bacteria entering stationary phase at other quorum-sensing regulatory systems.
low population density (119). Thus,TraI-made AHLs can enhance the expres-
sion of rhiI (102) and raiI (131) by activating
rhiI AND rhiR their respective regulators RhiR and RaiR.
The rhi (rhizosphere-expressed) genes were Conversely, TraI-made AHLs can compete
first identified because they are highly expressed with low levels of 3-OH-C14:1-HSL for the
induction of traR by BisR (21).The picture that quorum-sensing system are required for nor-
emerges is that each of the different quorum- mal symbiotic development (18). It is difficult
sensing regulatory systems can have an impact in this system to distinguish between a direct
on the regulation of the others, forming an effect of quorum-sensing regulation on sym-
interactive network with cinI and cinR at the biosome development and the possibility of an
top of the network. indirect effect resulting from the relatively poor
growth of the mutants.
R. etli R. etli CNAPF512 swarms on agar plates,
Quorum-sensing regulation has been analyzed and this is blocked by mutations in cinI or cinR
in detail in two different strains of R. etli called (19). The swarming is enhanced by added 3-
CNPAF512 and CFN42. The genome of OH-(slc)-HSL and by cloned cinI. Surprisingly,
CFN42 has been sequenced (49), but most of swarming in a cinR mutant was restored by
the quorum-sensing work on this strain has cloned cinI, implying a direct role for AHLs in
focused on the transfer of plasmid p42a (121). swarming, rather than activation of gene
cinI and cinR are on the chromosome and raiI expression via cinR (19). Added AHLs could
and raiR on plasmid p42f (49), but no other also induce limited extrusions to the smooth
work on these genes from CFN42 has been borders of colonies formed by a cinR mutant,
described. However, as described below, the and the AHLs could decrease the surface ten-
orthologous genes in CNPAF512 have been sion when added to water droplets. Further-
worked on extensively (20). more, the viscosity of the slime produced by
swarming colonies was reduced following the
cinI AND cinR addition of 3-OH-(slc)-HSL. On the basis of
These genes in R. etli CNPAF512 appear to be these observations, it was concluded that the
orthologous to cinI and cinR from R. legumi- AHLs act as biosurfactants that can affect
nosarum strains (18); CinI produces saturated swarming by R. etli (19).
long-chain AHLs (3-OH-C14-HSL), referred
to as 3-OH-(slc)-HSL, and cinI induction by raiI AND raiR
CinR can be activated by C12-HSL, C14-HSL, These genes were first described in R. etli
and most strongly by 3-OH-(slc)-HSL. Muta- CPNAF512, and RaiR induces the expression
tions in cinI and cinR delayed and decreased the of raiI in response to AHLs in a population-
growth rate of R. etli (18); because such altered density-dependent manner (103). Different
growth was not observed in R. leguminosarum AHLs made by RaiI in R.etli were detected,but
cinI or cinR mutants, this suggests that different they were not definitively identified, although
sets of genes are induced via this quorum- apparently raiI in R. etli is induced by added 3-
sensing regulon in these two Rhizobium species. OH-C8-HSL (19),as seen with raiI in R.legumi-
The cinI and cinR mutants of R. etli CNPAF512 nosarum. Mutation of raiI in R. etli CPNAF512
formed nodules, but symbiotic nitrogen fixa- caused an increased number of nodules and
tion was decreased by about 60 to 70%, and this increased nitrogenase activity on nodulated
correlated with abnormal development of sym- bean roots, although this did not lead to
biosomes. The cinI-mutant bacteroids were increased nitrogen fixation. Surprisingly, muta-
individually packed within the symbiosome tion of raiR had no effect on nodulation, lead-
membrane, whereas the wild type had multiple ing to the suggestion that raiI-made AHLs may
bacteroids within the symbiosomal membrane. suppress nodulation in this system (103).
Because AHLs [including one comigrating
with 3-OH-(slc)-HSL] could be isolated from Rhizobium sp. strain NGR234
nodules and cinI expression could be detected This strain is phylogenetically very close to S.
in infection threads using a gene fusion, it was meliloti (105) but has been widely studied
concluded that genes regulated via the cinI-cinR because it has an unusually wide host range
and can nodulate over 100 legume species quence lead to the overall conjugal deficiency
(99). Sequencing of the symbiosis plasmid (C. Fuqua, personal communication).
(pNGR234a) of this strain revealed the pres- An additional feature of the NGR234
ence of traI and traR genes associated with quorum-sensing system is that expression of
plasmid transfer genes (35), and as described TraR increases production of the other AHLs
above, their action has been analyzed in some produced by NGR234 and, in the presence of
detail (14). Work by He et al. showed that 3-O-C8-HSL, resulted in growth inhibition, an
the NGR234 tra system shares striking similar- observation reminiscent of the bacteriocin-like
ities to that of A. tumefaciens (51).The traI gene activity in R. leguminosarum bv. viciae (50).
product is responsible for the synthesis of an Although the mechanism of the growth inhibi-
AHL likely to be 3-oxo-C8-HSL. This AHL tion is unknown, data indicated that it requires
and TraR are responsible for the autoregulation one or more genes that are not located on
of traI and activation of conjugal transfer pNGR234a, in addition to traI, traR, and traM
genes. Furthermore, TraR activity is inhibited (51).The pNGR234a TraR also seems to con-
by TraM. Interestingly, a traI mutant still pro- trol several genes that are not localized to the
duces a compound that comigrates with the 3- plasmid (C. Fuqua, personal communication).
oxo-C8-HSL and one or more long-chain
AHLs. It is therefore likely that one or more S. meliloti
additional AHL synthases may be present. Since Quorum sensing in S. meliloti has been studied
synthesis of these putative AHLs can be in some detail in two different strains, Rm41
detected in a strain lacking pNGR234a, the and Rm1021. At least two quorum-sensing
synthase gene(s) must reside elsewhere in the systems have been identified (41, 75, 76).
genome. Recently a homologue of the S. One of them, the tra system, is present in a
meliloti sinI gene was cloned and shown to be Rm41-specific plasmid. The second quorum-
involved in the synthesis of various AHLs sensing system,the sin system,is chromosomally
(C. Fuqua, personal communication). Despite located in all strains examined so far.
the obvious similarities with the A. tumefaciens
quorum-sensing system, plasmid transfer traI AND traR
in NGR234 seems to be hobbled. pNGR234a The S. meliloti tra system, named for its homol-
transfer occurs at a frequency of only 109 (51). ogy to the tra systems in A. tumefaciens, resides
One possible explanation for this extremely on a plasmid called pRme41a (75).This plasmid
low transfer frequency is that an environmental is found in the commonly used S. meliloti strain,
signal analogous to A. tumefaciens opines may Rm41, but it is also present in many natural
be required for full activity. This would be in Sinorhizobium isolates. Interestingly, the other
contrast to the transfer frequency (102) seen commonly used wild-type S. meliloti strain,
for R. leguminosarum and R. etli, which appar- Rm1021, appears to have lost this plasmid
ently do not require an extra signal to induce along with other traits (see below).At least three
plasmid transfer. regulatory genes (traR, traI, and traM), in addi-
Another possible explanation for the low tion to genes with homology to the trb operon,
transfer frequency of pNGR234a is that the have been identified in pRme41a (75). TraI is
traAFB operon does not seem to be signifi- an AHL synthase responsible for the production
cantly expressed. In addition, the trbE coding of at least three different AHLs, 3-O-C8-HSL,
sequence is split into two separate reading 3-OH-C8-HSL, and C8-HSL (75). The pres-
frames, trbE1 and trbE2. Despite the observa- ence of the transcriptional activator TraR is
tion that each of these two open reading frames necessary for full synthase activity. In a manner
has a putative start codon and ribosome- analogous to the tra system in A. tumefaciens, the
binding site, the possibility remains that these product of the traM gene was shown to nega-
do not function in conjugation and as a conse- tively regulate TraR activity, ensuring that the
tra system is active only at high population den- bined with a sinI mutation, is no longer capable
sities.The S. meliloti Rm41 tra system also con- of forming nitrogen-fixing nodules (74).
trols conjugal plasmid transfer, as in other More recently, the Sin quorum-sensing sys-
organisms, by mediating the transfer of tem, in conjunction with the ExpR regulator,
pRme41a (75). Disruption of traR, traI, or the has been shown to play a role in other impor-
trb operon abolishes plasmid transfer, while distant cellular processes (52). Genomewide
ruption of the traM gene results in a 100-fold expression analysis of mutants lacking compo-
increase in transfer. Interestingly, as in nents of this quorum-sensing system showed
NGR234, the conjugal transfer frequency is that over 200 genes are controlled by ExpR
low (about 107) (75). either in the presence or absence of SinI-made
AHLs. For example, biosynthesis of succinogly-
sinI AND sinR can,the second exopolysaccharide produced by
The second quorum-sensing system in S. S. meliloti, is also positively regulated by
meliloti, the Sin system, is present in all strains of ExpR/SinI. Moreover, production of the sym-
Sinorhizobium analyzed so far. It consists of the biotically active low-molecular-weight fraction
sinI autoinducer synthase and its regulator, of this polymer is strongly controlled by this
sinR. Together they are responsible for the pro- quorum-sensing system (46). On the other
duction of several novel AHLs, ranging in size hand, bacterial motility is negatively regulated
from C12-HSL to C18-HSL (41, 76, 117). Dis- by high population density via ExpR and SinI.
ruption of either sinI or sinR results in a delay in In the presence of these two components, fla-
the appearance of nitrogen-fixing nodules, as gellar and chemotaxis genes are repressed when
well as an overall decrease in the number of cultures achieve stationary phase (Hoang,
nodules, suggesting a role for the SinI-made Gurich, and González, unpublished data).
AHLs in establishing a successful symbiosis In addition to ExpR, the S. meliloti genome
with Medicago sativa (41, 76). It appears that the contains LuxR homologues that appear to play
SinI AHLs work in conjunction with various a role in gene expression. Preliminary evidence
orphan LuxR-like activators present in the S. shows that the gene products of Smc04032
meliloti genome. For example, the presence of Smc00658, Smc00877, and Smc00878 play a
both the LuxR homologue ExpR and the role in nitrogen utilization, amino acid trans-
SinI-made AHLs is required for synthesis of port, and membrane stability, among others (A.
EPS II, one of the symbiotically important Patankar and J. E. González, unpublished data).
exopolysaccharides produced by S.meliloti (74).
The Rm1021 strain, which normally does not BRADYRHIZOBIUM JAPONICUM
produce EPS II, has an insertion sequence that A non-AHL molecule has been shown to
disrupts the expR gene (91), which seems to be mediate a quorum-sensing response in B. japon-
present and active in most strains of S. meliloti icum. Early studies showed repression of the nod
analyzed so far (74, 91). ExpR is a positive reg- genes at high population densities, suggesting
ulator of the exp genes, responsible for EPS II quorum-sensing control. This population-
biosynthesis (45). In a sinI mutant, expression of density-dependent control appeared to be
several of the exp genes is abolished, and this mediated by an extracellular signal molecule
deficiency can be fully complemented by the termed cell density factor (CDF) (65, 67, 68,
addition of either crude AHL extracts or syn- 134). The chemical structure of CDF was
thetic C16:1-HSL (74). Therefore, it seems that shown to be 2-(4-{[4-(3-aminooxetan-2-yl)
the sinRI locus controls EPS II production, phenyl]-(imino)methyl]phenyl}oxetan-3-
either directly or indirectly, via ExpR. Regula- ylamine,also designated as bradyoxetin (64–66).
tion of EPS II production by sinRI was shown Bradyoxetin has strong structural similarity to
to be important for nodule invasion, since a a variety of antibiotics and to siderophores.
strain that exclusively produces EPS II, com- Interestingly, synthesis of bradyoxetin is iron
regulated, with maximal production under Bassler and coworkers (5) suggest that quorum
iron-depleted conditions (64, 66). Loh et al. sensing modulates both intra- and inter-species
identified a response regulator, nwsB, which is cell-cell communications. It is interesting to
part of a two-component system and is required speculate that evolution might have allowed the
for detection of CDF (64). NwsB responds to a development of this type of chemical commu-
rise in population and the corresponding nication to ultimately increase cell survivability
increase in CDF levels by inducing nolA, an by coordinating interactions among potential
activator of the nodD2 regulator (64). The bacterial competitors and their plant hosts.
NodD2 represses the nod genes at high popula- Molecular cross-talk between bacteria and
tion densities in the presence of flavonoids (64, eukaryotes has been described for a variety of
68). In planta studies showed that both nolA and symbiotic or pathogenic relationships (29, 112,
nwsB were required for the repression of the nod 122). Recent studies have revealed that eukary-
genes in plants (64), suggesting that the brady- otes are capable of interfering with bacterial
oxetin-mediated quorum sensing in B. japon- communication by the production of molecu-
icum played a role in the early signaling events of lar signals that interact with the bacterial
the symbiotic process. Bradyoxetin activity was quorum-sensing system (58, 71, 118, 135), and
found in extracts of all -proteobacteria tested the evidence for legume perception of, and
(65,66).This suggests that compounds similar to interference with, rhizobial quorum sensing has
bradyoxetin may play an important role, not been reviewed recently (104). Quorum-sens-
only in rhizobial symbiosis, but also in other ing-interfering compounds have been intensely
plant- and animal-bacterial interactions. investigated for their potential as microbial con-
More recently, production of AHLs by dif- trol agents.For example,Pisum sativum (pea) and
ferent strains of Bradyrhizobium spp. has been other higher plants have been shown to pro-
reported (7,97),but it remains to be seen if they duce AHL-mimic compounds that interfere
play a role in gene regulation or in the symbi- with the quorum-sensing-regulated behavior
otic interaction. of several reporter strains (118).Another exam-
ple of a plant signal affecting quorum sensing in
MESORHIZOBIUM SPP. an associated bacterium is that of the Australian
In addition to the regulation of conjugal transfer red alga, Delisea pulchra, which interferes with
of the M. loti symbiosis island described above, the swarming motility of Serratia liquefaciens. D.
other species of Mesorhizobium, including M. pulchra produces halogenated furanones that are
huakii and M. tianshanense, produce AHLs (42, structurally similar to the AHL signals produced
128, 137).The mrtR and mrtI genes in M. tian- by S. liquefaciens. The furanones successfully
shanense (137) appear to be orthologous to cinR inhibit swarming motility in S. liquefaciens (44).
and cinI from R. leguminosarum bv. viciae and are It was demonstrated that the halogenated
required for nodulation and for optimal adher- furanones modulate LuxR activity through
ence to root hairs.AHL production by M.huakii accelerated degradation of the transcriptional
was also required for nodulation (42), and AHLs activator, rather than by blocking or displacing
also appear to affect biofilm formation (128). the binding of the AHL signal (71).
Another potential way to interfere with
THE SOIL ECOLOGY OF quorum sensing is through the degradation or
QUORUM SENSING inactivation of the AHL signal molecules. A
A wide variety of soil- and plant-associated bac- strain of Variovorax paradoxus was isolated from
teria produce AHLs (13). It has been suggested soil based on its ability to utilize AHLs as the
that, in the soil, various chemical signals serve as sole source of energy and nitrogen, an activity
a bacterial Esperanto, helping microorganisms that could disrupt the signaling process of other
interact or avoid each other in their mission to bacteria sharing the same environment (60).
bond with their plant hosts. Elegant studies by Zhang and colleagues isolated a Bacillus sp.
strain capable of inactivating AHLs. They behavior of Rhizobium leguminosarum. J. Bacteriol.

cloned from this species a gene encoding an 175:750–757.
7. Brelles-Marino, G., and E. J. Bedmar. 2001.
AHL-lactonase capable of inactivating AHLs by Detection, purification and characterisation
hydrolyzing the lactone bond (26). Further- of quorum-sensing signal molecules in plant-
more, transgenic plants expressing the AHL- associated bacteria. J. Biotechnol. 91:197–209.
lactonase activity were found to be more 8. Brencic, A., and S. C.Winans. 2005. Detection
resistant to Erwinia carotovora, a plant pathogen of and response to signals involved in host-microbe
interactions by plant-associated bacteria. Microbiol.
that requires AHLs for expression of the genes Mol. Biol. Rev. 69:155–194.
necessary for pathogenicity (25). 9. Brewin, N. J. 2004. Plant cell wall remodelling in
The study of quorum sensing in the symbi- the Rhizobium-legume symbiosis. Crit. Rev. Plant
otic nitrogen-fixing rhizobia has revealed new Sci. 23:293–316.
paradigms and novel signal molecules for this 10. Burn, J., L. Rossen, and A. W. B. Johnston.
1987.Four classes of mutations in the nodD gene of
type of gene regulation. Further studies are Rhizobium leguminosarum biovar viciae that affect its
bound to expose new mechanisms utilized by ability to autoregulate and or activate other nod
the plant hosts’ eukaryotic hosts to interfere genes in the presence of flavonoid Inducers. Genes
with the quorum-sensing behavior of associ- Dev. 1:456–464.
ated bacteria. 11. Canter Cremers, H. C., K. Stevens, B. J.
Lugtenberg, C. A. Wijffelman, M. Batley, J.
W. Redmond, M. W. Breedveld, and L. P.
ACKNOWLEDGMENTS
Zevenhuizen. 1991. Unusual structure of the
We thank Clay Fuqua and Craig McAnulla for helpful exopolysaccharide of Rhizobium leguminosarum bv.
comments and input during the preparation of this viciae strain 248. Carbohydrate Res. 218:185–200.
manuscript. The work in the authors’ laboratories 12. Cantero, L., J. M. Palacios, T. Ruiz-Argueso,
is supported by the BBSRC (via a grant in aid and and J. Imperial. 2006.Proteomic analysis of quo-
grant P19980 to J. A. D.) and the National Science rum sensing in Rhizobium leguminosarum biovar
Foundation and the National Institutes of Health viciae UPM791. Proteomics 6: S97–S106.
(grants MCB-9733532 and 1RO1GM069925 to 13. Cha, C., P. Gao, Y. C. Chen, P. D. Shaw, and
J. E. G.). S. K. Farrand. 1998. Production of acyl-
homoserine lactone quorum-sensing signals by
REFERENCES gram-negative plant-associated bacteria. Mol.
1. Amemura, A., T. Harada, M. Abe, and S. Plant-Microbe Interact. 11:1119–1129.
Higashi. 1983. Structural studies of the acidic 14. Chen, W. M., L. Moulin, C. Bontemps, P.
polysaccharide from Rhizobium trifolii 4S. Carbohy- Vandamme, G. Bena, and C. Boivin-Masson.
drate Res. 115:165–174. 2003.Legume symbiotic nitrogen fixation by beta-
2. Appleby, C. A. 1984. Leghemoglobin and Rhizo- proteobacteria is widespread in nature. J. Bacteriol.
bium respiration. Ann. Rev. Plant Physiol. Plant Mol. 185:7266–7272.
Biol. 35:443–478. 15. Chen, X. C., J. Feng, B. H. Hou, F. Q. Li, Q. Li,
3. Baldani, J. I., R.W.Weaver, M. F. Hynes, and B. and G. F. Hong. 2005. Modulating DNA bend-
D. Eardly. 1992. Utilization of carbon substrates, ing affects NodD-mediated transcriptional control
electrophoretic enzyme patterns, and symbiotic in Rhizobium leguminosarum. Nucleic Acids Res.
performance of plasmid-cured clover rhizobia. 33:2540–2548.
Appl. Environ. Microbiol. 58:2308–2314. 16. Cheng, H. P., and G. C.Walker. 1998. Succino-
4. Bartsev, A.V.,W. J. Deakin, N. M. Boukli, C. B. glycan is required for initiation and elongation of
McAlvin, G. Stacey, P. Malnoe, W. J. infection threads during nodulation of alfalfa by
Broughton, and C. Staehelin. 2004. NopL, an Rhizobium meliloti. J. Bacteriol. 180:5183–5191.
effector protein of Rhizobium sp NGR234, thwarts 17. Cubo, M.T., A. Economou, G. Murphy, A.W.
activation of plant defense reactions. Plant Physiol. Johnston, and J. A. Downie. 1992. Molecular
134:871–879. characterization and regulation of the rhizosphere-
5. Bassler, B. L. 2002. Small talk. Cell-to-cell com- expressed genes rhiABCR that can influence nodu-
munication in bacteria. Cell 109:421–424. lation by Rhizobium leguminosarum biovar viciae. J.
6. Breedveld, M. W., H. C. Cremers, M. Batley, Bacteriol. 174:4026–4035.
M. A. Posthumus, L. P. Zevenhuizen, C. A. 18. Daniels, R., D. E. De Vos, J. Desair, G. Raed-
Wijffelman, and A. J. Zehnder. 1993. Poly- schelders, E. Luyten, V. Rosemeyer, C.
saccharide synthesis in relation to nodulation Verreth, E. Schoeters, J.Vanderleyden, and J.
Michiels. 2002.The cin quorum sensing locus of Kluwer Academic Publishers, Dordrecht, The
Rhizobium etli CNPAF512 affects growth and Netherlands.
symbiotic nitrogen fixation. J. Biol. Chem. 31. Feng, J., Q. Li, H. L. Hu, X. C. Chen, and G. F.
277:462–468. Hong. 2003. Inactivation of the nod box distal
19. Daniels, R., S. Reynaert, H. Hoekstra, C. half-site allows tetrameric NodD to activate nodA
Verreth, J. Janssens, K. Braeken, M. Fauvart, transcription in an inducer-independent manner.
S. Beullens, C. Heusdens, I. Lambrichts, D. E. Nucleic Acids Res. 31:3143–3156.
De Vos, J. Vanderleyden, J. Vermant, and J. 32. Finnie, C., N. M. Hartley, K. C. Findlay, and J.
Michiels. 2006. Quorum signal molecules as bio- A. Downie. 1997. The Rhizobium leguminosarum
surfactants affecting swarming in Rhizobium etli. prsDE genes are required for secretion of several
Proc. Natl.Acad. Sci. USA 103:14965–14970. proteins, some of which influence nodulation,
20. Daniels, R., J.Vanderleyden, and J. Michiels. symbiotic nitrogen fixation and exopolysaccharide
2004.Quorum sensing and swarming migration in modification. Mol. Microbiol. 25:135–146.
bacteria. FEMS Microbiol. Rev. 28:261–289. 33. Fisher, R. F., and S. R. Long. 1993. Interactions
21. Danino,V. E., A. Wilkinson, A. Edwards, and of NodD at the nod box: NodD binds to two dis-
J. A. Downie. 2003. Recipient-induced transfer tinct sites on the same face of the helix and induces
of the symbiotic plasmid pRL1JI in Rhizobium a bend in the DNA. J. Mol. Biol. 233:336–348.
leguminosarum bv. viciae is regulated by a quorum- 34. Fraysse, N., F. Couderc, and V. Poinsot. 2003.
sensing relay. Mol. Microbiol. 50:511–525. Surface polysaccharide involvement in establishing
22. Deakin, W. J., C. Marie, M. M. Saad, H. B. the Rhizobium-legume symbiosis. Eur. J. Biochem.
Krishnan, and W. J. Broughton. 2005. NopA is 270:1365–1380.
associated with cell surface appendages produced 35. Freiberg, C., R. Fellay, A. Bairoch, W. J.
by the type III secretion system of Rhizobium sp Broughton, A. Rosenthal, and X. Perret.
strain NGR234. Mol. Plant-Microbe Interact. 1997.Molecular basis of symbiosis between Rhizo-
18:499–507. bium and legumes. Nature 387:394–401.
23. Denison, R. F., and E. T. Kiers. 2004. Lifestyle 36. Fuqua, C., M. Burbea, and S. C.Winans. 1995.
alternatives for rhizobia: mutualism, parasitism, and Activity of the Agrobacterium Ti plasmid conjugal
forgoing symbiosis. FEMS Microbiol. Lett. transfer regulator TraR is inhibited by the product
237:187–193. of the traM gene. J. Bacteriol. 177:1367–1373.
24. Dibb, N. J., J. A. Downie, and N. J. Brewin. 37. Fuqua, C., M. R. Parsek, and E. P. Greenberg.
1984. Identification of a rhizosphere protein 2001. Regulation of gene expression by cell-to-
encoded by the symbiotic plasmid of Rhizobium cell communication: acyl-homoserine lactone
leguminosarum. J. Bacteriol. 158:621–627. quorum sensing. Annu. Rev. Genet. 35:439–468.
25. Dong, Y. H., L. H. Wang, J. L. Xu, H. B. 38. Gage, D. J. 2002. Analysis of infection thread
Zhang, X. F. Zhang, and L. H. Zhang. 2001. development using Gfp- and DsRed-expressing
Quenching quorum-sensing-dependent bacterial Sinorhizobium meliloti. J. Bacteriol. 184:7042–7046.
infection by an N-acyl homoserine lactonase. 39. Gage, D. J. 2004. Infection and invasion of roots
Nature 411:813–817. by symbiotic, nitrogen-fixing rhizobia during
26. Dong, Y. H., J. L. Xu, X. Z. Li, and L. H. nodulation of temperate legumes. Microbiol. Mol.
Zhang. 2000.AiiA, an enzyme that inactivates the Biol. Rev. 68:280–300.
acylhomoserine lactone quorum-sensing signal 40. Galibert, F.,T. M. Finan, S. R. Long,A. Pühler,
and attenuates the virulence of Erwinia carotovora. P. Abola, F. Ampe, F. Barloy-Hubler, M. J.
Proc. Natl.Acad. Sci. USA 97:3526–3531. Barnett, A. Becker, P. Boistard, G. Bothe, M.
27. Downie, J.A. 1998. Functions of rhizobial nodu- Boutry, L. Bowser, J. Buhrmester, E. Cadieu,
lation genes, p. 387–402. In H. P. Spaink, A. Kon- D. Capela, P. Chain, A. Cowie, R.W. Davis, S.
dorosi, and P. J. J. Hooykaas (ed.), The Rhizobiaceae. Dreano, N. A. Federspiel, R. F. Fisher, S.
Kluwer, Dordrecht,The Netherlands. Gloux, T. Godrie, A. Goffeau, B. Golding, J.
28. Downie, J.A. 2005. Legume haemoglobins: sym- Gouzy, M. Gurjal, I. Hernandez-Lucas, A.
biotic nitrogen fixation needs bloody nodules. Hong, L. Huizar, R. W. Hyman, T. Jones, D.
Curr. Biol. 15:R196–R198. Kahn, M. L. Kahn, S. Kalman, D. H. Keating,
29. Dunn,A. K., and J. Handelsman. 2002.Toward E. Kiss, C. Komp, V. Lelaure, D. Masuy, C.
an understanding of microbial communities Palm, M. C. Peck,T. M. Pohl, D. Portetelle, B.
through analysis of communication networks. Purnelle, U. Ramsperger, R. Surzycki, P.The-
Antonie Van Leeuwenhoek. 81:565–574. bault, M.Vandenbol, F. J.Vorholter, S.Weidner,
30. Farrand, S. K. 1998. Conjugal plasmids and D. H. Wells, K. Wong, K. C.Yeh, and J. Batut.
their transfer, p. 199–233. In H. P. Spaink, A. Kon- 2001.The composite genome of the legume sym-
dorosi, and P. J. J. Hooykaas (ed.), The Rhizobiaceae. biont Sinorhizobium meliloti. Science 293:668–672.
41. Gao, M., H. Chen, A. Eberhard, M. R. 51. He, X. S.,W. Chang, D. L. Pierce, L. O. Seib, J.
Gronquist, J. B. Robinson, B. G. Rolfe, and W. Wagner, and C. Fuqua. 2003. Quorum sensing
D. Bauer. 2005. sinI- and expR-dependent quo- in Rhizobium sp strain NGR234 regulates conjugal
rum sensing in Sinorhizobium meliloti. J. Bacteriol. transfer (tra) gene expression and influences
187:7931–7944. growth rate. J. Bacteriol. 185:809–822.
42. Gao,Y. J., Z.T. Zhong, K. J. Sun, H.Wang, and 52. Hoang, H. H., A. Becker, and J. E. González.
J. Zhu. 2006. The quorum-sensing system in a 2004.The LuxR homolog ExpR, in combination
plant bacterium Mesorhizobium huakuii affects with the Sin quorum sensing system,plays a central
growth rate and symbiotic nodulation. Plant Soil role in Sinorhizobium meliloti gene expression. J.
286:53–60. Bacteriol. 186:5460–5472.
43. Giraud, E., L. Moulin, D.Vallenet,V. Barbe, E. 53. John, M., H. Rohrig, J. Schmidt, U.Wieneke,
Cytryn, J. C. Avarre, M. Jaubert, D. Simon, F. and J. Schell. 1993. Rhizobium NodB protein
Cartieaux, Y. Prin, G. Bena, L. Hannibal, J. involved in nodulation signal synthesis is a chi-
Fardoux, M. Kojadinovic, L.Vuillet, A. Lajus, tooligosaccharide deacetylase. Proc. Natl.Acad. Sci.
S. Cruveiller, Z. Rouy, S. Mangenot, B. USA 90:625–629.
Segurens, C. Dossat, W. L. Franck, W. S. 54. Johnston,A.W., J. L. Beynon,A.V. Buchanan-
Chang, E. Saunders, D. Bruce, P. Richardson, Wollaston, S. M. Setchell, P. R. Hirsch, and J.
P. Normand, B. Dreyfus, D. Pignol, G. Stacey, E. Beringer. 1978. High frequency transfer of
D. Emerich, A. Vermeglio, C. Medigue, and nodulating ability between strains and species of
M. Sadowsky. 2007. Legumes symbioses: absence Rhizobium. Nature 276:634–636.
of nod genes in photosynthetic Bradyrhizobia. Sci- 55. Johnston, A. W. B., G. Hombrecher, N. J.
ence 316:1307–1312. Brewin, and M. C. Cooper. 1982.Two transmis-
44. Givskov, M., R. de Nys, M. Manefield, L. sible plasmids in Rhizobium leguminosarum strain
Gram, R. Maximilien, L. Eberl, S. Molin, P.D. 300. J. Gen. Microbiol. 128:85–93.
Steinberg, and S. Kjelleberg. 1996. Eukaryotic 56. Kaneko,T.,Y. Nakamura, S. Sato, E.Asamizu,
interference with homoserine lactone-mediated T. Kato, S. Sasamoto, A. Watanabe, K.
prokaryotic signalling. J. Bacteriol. 178:6618–6622. Idesawa, A. Ishikawa, K. Kawashima, T.
45. Glazebrook, J., and G. C.Walker. 1989.A novel Kimura,Y. Kishida, C. Kiyokawa, M. Kohara,
exopolysaccharide can function in place of the M. Matsumoto, A. Matsuno,Y. Mochizuki, S.
Calcofluor-binding exopolysaccharide in nodula- Nakayama, N. Nakazaki, S. Shimpo, M.
tion of alfalfa by Rhizobium meliloti. Cell 56: Sugimoto, C. Takeuchi, M. Yamada, and S.
661–672. Tabata. 2000. Complete genome structure of the
46. Glenn, S. A., N. Gurich, M. A. Feeney, and J. nitrogen-fixing symbiotic bacterium Mesorhizo-
E. González. 2007.The ExpR/Sin quorum sens- bium loti. DNA Res. 7:331–338.
ing system controls succinoglycan production in 57. Kaneko,T.,Y. Nakamura, S. Sato, K. Minami-
Sinorhizobium meliloti. J. Bacteriol. 189:7077–7088. sawa,T. Uchiumi, S. Sasamoto, A.Watanabe,
47. González, J. E., and M. M. Marketon. 2003. K. Idesawa, M. Iriguchi, K. Kawashima, M.
Quorum sensing in nitrogen-fixing rhizobia. Kohara, M. Matsumoto, S. Shimpo, H.
Microbiol. Mol. Biol. Rev. 67:574–592. Tsuruoka,T.Wada, M.Yamada, and S.Tabata.
48. González, J. E., B. L. Reuhs, and G. C.Walker. 2002. Complete genomic sequence of nitrogen-
1996. Low molecular weight EPS II of Rhizobium fixing symbiotic bacterium Bradyrhizobium japon-
meliloti allows nodule invasion in Medicago sativa. icum USDA110. DNA Res. 9:189–197.
Proc. Natl.Acad. Sci. USA 93:8636–8641. 58. Keshavan, N. D., P. K. Chowdhary, D. C.
49. González, V., R. I. Santamaria, P. Bustos, I. Haines, and J. E. González. 2005. L-Canavanine
Hernandez-González, A. Medrano-Soto, G. made by Medicago sativa interferes with quorum
Moreno-Hagelsieb, S. C. Janga, M. A. sensing in Sinorhizobium meliloti. J. Bacteriol.
Ramirez, V. Jimenez-Jacinto, J. Collado- 187:8427–8436.
Vides, and G. Davila. 2006.The partitioned Rhi- 59. Laus, M. C., T. J. Logman, A. A.Van Brussel,
zobium etli genome: genetic and metabolic R. W. Carlson, P. Azadi, M.Y. Gao, and J. W.
redundancy in seven interacting replicons. Proc. Kijne. 2004. Involvement of exo5 in production
Natl.Acad. Sci. USA 103:3834–3839. of surface polysaccharides in Rhizobium legumi-
50. Gray, K. M., J. P. Pearson, J. A. Downie, B. E. nosarum and its role in nodulation of Vicia sativa
Boboye, and E. P. Greenberg. 1996. Cell-to- subsp. nigra. J. Bacteriol. 186:6617–6625.
cell signaling in the symbiotic nitrogen-fixing bac- 60. Leadbetter, J. R., and E. P. Greenberg. 2000.
terium Rhizobium leguminosarum: autoinduction of Metabolism of acyl-homoserine lactone quorum-
a stationary phase and rhizosphere-expressed sensing signals by Variovorax paradoxus. J. Bacteriol.
genes. J. Bacteriol. 178:372–376. 182:6921–6926.
61. Leigh, J. A., E. R. Signer, and G. C. Walker. rhamnose-rich polysaccharides. Mol. Plant-Microbe
1985. Exopolysaccharide-deficient mutants of Interact. 17:958–966.
Rhizobium meliloti that form ineffective nodules. 73. Marie, C.,W. J. Deakin,V.Viprey, J. Kopcinska,
Proc. Natl.Acad. Sci. USA 82:6231–6235. W. Golinowski, H. B. Krishnan, X. Perret, and
62. Lithgow, J. K., A.Wilkinson, A. Hardman, B. W. J. Broughton. 2003. Characterization of
Rodelas, F. Wisniewski-Dye, P. Williams, and Nops, nodulation outer proteins, secreted via the
J.A. Downie. 2000.The regulatory locus cinRI in type III secretion system of NGR234. Mol. Plant-
Rhizobium leguminosarum controls a network of Microbe Interact. 16:743–751.
quorum-sensing loci. Mol. Microbiol. 37:81–97. 74. Marketon, M. M., S. A. Glenn, A. Eberhard,
63. Lodwig, E. M., A. H. Hosie, A. Bourdes, K. and J. E. González. 2003. Quorum sensing con-
Findlay, D. Allaway, R. Karunakaran, J. A. trols exopolysaccharide production in Sinorhizo-
Downie, and P. S. Poole. 2003. Amino-acid bium meliloti. J. Bacteriol. 185:325–331.
cycling drives nitrogen fixation in the legume- 75. Marketon, M. M., and J. E. González.
Rhizobium symbiosis. Nature 422:722–726. 2002. Identification of two quorum-sensing
64. Loh, J., D. P. Lohar, B. Andersen, and G. systems in Sinorhizobium meliloti. J. Bacteriol.
Stacey. 2002. A two-component regulator medi- 184:3466–3475.
ates population-density-dependent expression of 76. Marketon, M. M., M. R. Gronquist, A. Eber-
the Bradyrhizobium japonicum nodulation genes. J. hard, and J. E. González. 2002. Characteriza-
Bacteriol. 184:1759–1766. tion of the Sinorhizobium meliloti sinR/sinI locus
65. Loh, J., E. A. Pierson, L. S. Pierson, 3rd, G. and the production of novel N-acyl homoserine
Stacey, and A. Chatterjee. 2002. Quorum sens- lactones. J. Bacteriol. 184:5686–5695.
ing in plant-associated bacteria. Curr. Opin. Plant 77. Matthysse,A. G., and J.W. Kijne. 1998.Attach-
Biol. 5:285–290. ment of Rhizobiaceae to plant cells, p. 135–249. In
66. Loh, J., and G. Stacey. 2003. Nodulation gene H. P. Spaink, A. Kondorosi, and P. J. J. Hooykaas
regulation in Bradyrhizobium japonicum: a unique (ed.), The Rhizobiaceae. Kluwer Academic, Dor-
integration of global regulatory circuits. Appl. Env- drecht,The Netherlands.
iron. Microbiol. 69:10–17. 78. McAnulla, C., A. Edwards, M. Sanchez-
67. Loh, J.T., and G. Stacey. 2001. Feedback regula- Contreras, R. G. Sawers, and J. A. Downie.
tion of the Bradyrhizobium japonicum nodulation 2007. Quorum-sensing-regulated transcriptional
genes. Mol. Microbiol. 41:1357–1364. initiation of plasmid transfer and replication genes
68. Loh, J. T., J. P. Yuen-Tsai, M. G. Stacey, D. in Rhizobium leguminosarum biovar viciae. Microbiol-
Lohar,A.Welborn, and G. Stacey. 2001. Popu- ogy 153:2074–2082.
lation density-dependent regulation of the 79. Mercado-Blanco, J., and N. Toro. 1996. Plas-
Bradyrhizobium japonicum nodulation genes. Mol. mids in rhizobia: the role of nonsymbiotic plas-
Microbiol. 42:37–46. mids. Mol. Plant-Microbe Interact. 9:535–545.
69. Luo, Z. Q., Y. Qin, and S. K. Farrand. 2000. 80. Mergaert, P., M. Ferro, W. Dhaeze, M. Van-
The antiactivator TraM interferes with the Montagu, M. Holsters, and J. C. Prome. 1997.
autoinducer-dependent binding of TraR to Nod factors of Azorhizobium caulinodans strain
DNA by interacting with the C-terminal region of ORS571 can be glycosylated with an arabinosyl
the quorum-sensing activator. J. Biol. Chem. group, a fucosyl group, or both. Mol. Plant-Microbe
275:7713–7722. Interact. 10:683–687.
70. Madsen, E. B., L. H. Madsen, S. Radutoiu, M. 81. Mergaert, P., T. Uchiumi, B. Alunni, G.
Olbryt, M. Rakwalska, K. Szczyglowski, S. Evanno,A. Cheron, O. Catrice,A. E. Mausset,
Sato, T. Kaneko, S. Tabata, N. Sandal, and J. F. Barloy-Hubler, F. Galibert, A. Kondorosi,
Stougaard. 2003. A receptor kinase gene of the and E. Kondorosi. 2006. Eukaryotic control on
LysM type is involved in legume perception of rhi- bacterial cell cycle and differentiation in the Rhi-
zobial signals. Nature 425:637–640. zobium-legume symbiosis. Proc. Natl. Acad. Sci.
71. Manefield, M.,T. B. Rasmussen, M. Henzter, USA 103:5230–5235.
J. B.Andersen, P. Steinberg, S. Kjelleberg, and 82. Miller, L. D., C. K.Yost, M. F. Hynes, and G.
M. Givskov. 2002. Halogenated furanones inhibit Alexandre. 2006. The major chemotaxis gene
quorum sensing through accelerated LuxR cluster of Rhizobium leguminosarum bv. viciae is
turnover. Microbiology 148:1119–1127. essential for competitive nodulation. Mol. Micro-
72. Marie, C., W. J. Deakin, T. Ojanen-Reuhs, biol. 63:348–362.
E. Diallo, B. Reuhs, W. J. Broughton, and X. 83. Miwa, H., J. Sun, G. E. Oldroyd, and J. A.
Perret. 2004. TtsI, a key regulator of Rhizobium Downie. 2006.Analysis of calcium spiking using a
species, NGR234 is required for type III- cameleon calcium sensor reveals that nodulation
dependent protein secretion and synthesis of gene expression is regulated by calcium spike
number and the developmental status of the cell. 95. Piper, K. R., and S. K. Farrand. 2000. Quo-
Plant J. 48:883–894. rum sensing but not autoinduction of Ti plasmid
84. Monahan-Giovanelli, H., C. A. Pinedo, and conjugal transfer requires control by the opine
D. J. Gage. 2006.Architecture of infection thread regulon and the antiactivator TraM. J. Bacteriol.
networks in developing root nodules induced by 182:1080–1088.
the symbiotic bacterium Sinorhizobium meliloti on 96. Piper, K. R., S. B. von Bodman, I. Hwang,
Medicago truncatula. Plant Physiol. 140:661–670. and S. K. Farrand. 1999. Hierarchical gene
85. Morrison, N., and D. P. S.Verma. 1987.A block regulatory systems arising from fortuitous gene
in endocytosis of Rhizobium allows cellular differ- associations: controlling quorum sensing by the
entiation in nodules but affects the expression of opine regulon in Agrobacterium. Mol. Microbiol.
some peribacteroid membrane nodulins. Plant 32:1077–1089.
Mol. Biol. 9:185–196. 97. Pongsilp, N., E. W. Triplett, and M. J. Sad-
86. Niehaus, K., D. Kapp, and A. M. Pühler. 1993. owsky. 2005. Detection of homoserine lactone-
Plant defense and delayed infection of alfalfa like quorum sensing molecules in Bradyrhizobium
pseudonodules induced by an exopolysaccharide strains. Curr. Microbiol. 51:250–254.
(EPS I)-deficient Rhizobium meliloti. Planta 98. Poole, P., and D. Allaway. 2000. Carbon and
190:415–425. nitrogen metabolism in Rhizobium. Adv. Microb.
87. O’Neill, M. A., A. G. Darvill, and P. Alber- Physiol. 43:117–163.
sheim. 1991. The degree of esterification and 99. Pueppke, S. G., and W. J. Broughton. 1999.
points of substitution by O-acetyl and O-(3- Rhizobium sp. strain NGR234 and R. fredii
hydroxybutanoyl) groups in the acidic extracellular USDA257 share exceptionally broad, nested
polysaccharides secreted by Rhizobium legumi- host ranges. Mol. Plant-Microbe Interact. 12:
nosarum biovars viciae, trifolii, and phaseoli are not 293–318.
related to host range.J.Biol.Chem. 266:9549–9555. 100. Ramey, B. E., M. Koutsoudis, S. B. von Bod-
88. Oldroyd, G. E., and J. A. Downie. 2004. Cal- man, and C. Fuqua. 2004. Biofilm formation
cium, kinases and nodulation signalling in in plant-microbe associations. Curr. Opin. Micro-
legumes. Nat. Rev. Mol. Cell Biol. 5:566–576. biol. 7:602–609.
89. Oresnik, I. J., L. A. Pacarynuk, S. A. P. 101. Ramsay, J. P., J.T. Sullivan, G. S. Stuart, I. L.
O’Brien, C. K. Yost, and M. F. Hynes. 1998. Lamont, and C.W. Ronson. 2006.Excision and
Plasmid-encoded catabolic genes in Rhizobium transfer of the Mesorhizobium loti R7A symbiosis
leguminosarum bv. trifolii: evidence for a plant- island requires an integrase IntS, a novel recombi-
inducible rhamnose locus involved in competition nation directionality factor RdfS, and a putative
for nodulation. Mol. Plant-Microbe Interact. relaxase RlxS. Mol. Microbiol. 62:723–734.
11:1175–1185. 102. Rodelas, B., J. K. Lithgow, F. Wisniewski-
90. Pellock, B. J., H. P. Cheng, and G. C.Walker. Dye, A. Hardman, A. Wilkinson, A.
2000. Alfalfa root nodule invasion efficiency is Economou, P. Williams, and J. A. Downie.
dependent on Sinorhizobium meliloti polysaccha- 1999. Analysis of quorum-sensing-dependent
rides. J. Bacteriol. 182:4310–4318. control of rhizosphere-expressed (rhi) genes in
91. Pellock, B. J., M.Teplitski, R. P. Boinay,W. D. Rhizobium leguminosarum bv. viciae. J. Bacteriol.
Bauer, and G. C. Walker. 2002. A LuxR 181:3816–3823.
homolog controls production of symbiotically 103. Rosemeyer, V., J. Michiels, C. Verreth, and
active extracellular polysaccharide II by Sinorhizo- J. Vanderleyden. 1998. luxI- and luxR-
bium meliloti. J. Bacteriol. 184:5067–5076. homologous genes of Rhizobium etli CNPAF512
92. Perez-Mendoza, D., M. Lucas, S. Munoz, J.A. contribute to synthesis of autoinducer molecules
Herrera-Cervera, J. Olivares, F. de la Cruz, and nodulation of Phaseolus vulgaris. J. Bacteriol.
and J. Sanjuan. 2006.The relaxase of the Rhizo- 180:815–821.
bium etli symbiotic plasmid shows nic site cis-acting 104. Sanchez-Contreras, M., W. D. Bauer, M.
preference. J. Bacteriol. 188:7488–7499. Gao, J. B. Robinson, and J.A. Downie. 2007.
93. Perez-Mendoza, D., E. Sepulveda, V. Pando, Quorum-sensing regulation in rhizobia and its
S. Munoz, J. Nogales, J. Olivares, M. J. Soto, J. role in symbiotic interactions with legumes. Phil.
A. Herrera-Cervera, D. Romero, S. Brom, Trans. R. Soc. B 362:1149–1163.
and J. Sanjuan. 2005. Identification of the rctA 105. Sawada, H., L. D. Kuykendall, and J. M.
gene, which is required for repression of conjuga- Young. 2003. Changing concepts in the system-
tive transfer of rhizobial symbiotic megaplasmids.J. atics of bacterial nitrogen-fixing legume sym-
Bacteriol. 187:7341–7350. bionts. J. Gen.Appl. Microbiol. 49:155–179.
94. Perret, X., C. Staehelin, and W. J. Broughton. 106. Schlaman, H. R. M., D.A. Phillips, and K. E.
2000. Molecular basis of symbiotic promiscuity. 1998. Genetic organisation and transcriptional
Microbiol. Mol. Biol. Rev. 64:180–201. regulation of rhizobial nodulation genes,
p. 361–386. In H. P. Spaink,A. Kondorosi, and P. J. 117. Teplitski, M.,A. Eberhard, M. R. Gronquist,

J. Hooykaas (ed.), The Rhizobiaceae. Kluwer Acad- M. Gao, J. B. Robinson, and W. D. Bauer.
emic Publishers, Dordrecht,The Netherlands. 2003. Chemical identification of N-acyl
107. Schripsema, J., K. E. E. deRudder,T. B. van- homoserine lactone quorum-sensing signals pro-
Vliet, P. P. Lankhorst, E. deVroom, J. W. duced by Sinorhizobium meliloti strains in defined
Kijne, and A.A. N. vanBrussel. 1996. Bacteri- medium. Arch. Microbiol. 180:494–497.
ocin small of Rhizobium leguminosarium belongs 118. Teplitski, M., J. B. Robinson, and W. D.
to the class of N-acyl-1-homoserine lactone Bauer. 2000. Plants secrete substances that
molecules,known as autoinducers and as quorum mimic bacterial N-acyl homoserine lactone sig-
sensing co-transcription factors. J. Bacteriol. nal activities and affect population density-
178:366–371. dependent behaviors in associated bacteria. Mol.
108. Sielberer, B. J., A. C. J. Timmers, and A. M. Plant–Microbe Interact. 13:637–648.
C. Emons. 2005. Nod factors alter the micro- 119. Thorne, S. H., and H. D.Williams. 1999. Cell
tubule cytoskeleton in Medicago truncatula root density-dependent starvation survival of Rhizo-
hairs to allow root hair reorientation. Mol. Plant- bium leguminosarum bv. phaseoli: identification of
Microbe Interact. 18:1195–1204. the role of an N-acyl homoserine lactone in
109. Skorpil, P., M. M. Saad, N. M. Boukli, H. adaptation to stationary-phase survival. J. Bacte-
Kobayashi, F. Ares-Orpel, W. J. Broughton, riol. 181:981–990.
and W. J. Deakin. 2005.NopP,a phosphorylated 120. Truchet, G., P. Roche, P. Lerouge, J.Vasse, S.
effector of Rhizobium sp strain NGR234, is a Camut, F. Debilly, J. C. Prome, and J.
major determinant of nodulation of the tropical Denarie. 1991. Sulfated lipo-oligosaccharide
legumes Flemingia congesta and Tephrosia vogelii. signals of Rhizobium meliloti elicit root nodule
Mol. Microbiol. 57:1304–1317. organogenesis in alfalfa. Nature 351:670–673.
110. Skorupska,A., M. Janczarek, M. Marczak,A. 121. Tun-Garrido, C., P. Bustos,V. Gonzalez, and
Mazur, and J. Krol. 2006. Rhizobial S. Brom. 2003.Conjugative transfer of p42a from
exopolysaccharides: genetic control and symbi- Rhizobium etli CFN42,which is required for mobi-
otic functions. Microb. Cell Fact. 5:7. lization of the symbiotic plasmid, is regulated by
111. Sourjik, V., P. Muschler, B. Scharf, and R. quorum sensing. J. Bacteriol. 185:1681–1692.
Schmitt. 2000.VisN and VisR are global regula- 122. Tzfira, T., and V. Citovsky. 2002. Partners-in-
tors of chemotaxis, flagellar, and motility genes in infection: host proteins involved in the transfor-
Sinorhizobium (Rhizobium) meliloti. J. Bacteriol. mation of plant cells by Agrobacterium.Trends Cell
182:782–788. Biol. 12:121–129.
112. Spaink, H. P. 2000. Root nodulation and infec- 123. Udvardi, M. K., and D.A. Day. 1997. Metabo-
tion factors produced by rhizobial bacteria.Annu. lite transport across symbiotic membranes of
Rev. Microbiol. 54:257–288. legume nodules. Annu. Rev. Plant Physiol. Plant
113. Sullivan, J. T., H. N. Patrick, W. L. Lowther, Mol. Biol. 48:493–523.
D. B. Scott, and C.W. Ronson. 1995.Nodulat- 124. Urzainqui,A., and G. C.Walker. 1992.Exoge-
ing strains of Rhizobium loti arise through nous suppression of the symbiotic deficiencies of
chromosomal symbiotic gene transfer in the Rhizobium meliloti exo mutants. J. Bacteriol.
environment. Proc. Natl. Acad. Sci. USA 92: 174:3403–3406.
8985–8989. 125. Vannini,A., C.Volpari, and S. Di Marco. 2004.
114. Sullivan, J.T., and C.W. Ronson. 1998.Evolu- Crystal structure of the quorum-sensing protein
tion of rhizobia by acquisition of a 500-kb TraM and its interaction with the transcriptional
symbiosis island that integrates into a phe-t RNA regulator TraR. J. Biol. Chem. 279:24291–24296.
gene Proc. Natl.Acad. Sci. USA 95:5145–5149. 126. Vannini, A., C. Volpari, C. Gargioli, E.
115. Sullivan, J. T., J. R. Trzebiatowski, R. W. Muraglia, R. Cortese, R. De Francesco, P.
Cruickshank, J. Gouzy, S. D. Brown, R. M. Neddermann, and S. D. Marco. 2002. The
Elliot, D. J. Fleetwood, N. G. McCallum, U. crystal structure of the quorum sensing protein
Rossbach, G. S. Stuart, J. E. Weaver, R. J. TraR bound to its autoinducer and target DNA.
Webby, F. J. de Bruijn, and C. W. Ronson. EMBO J. 21:4393–4401.
2002. Comparative sequence analysis of the sym- 127. Walker, S.A., and J.A. Downie. 2000.Entry of
biosis island of Mesorhizobium loti strain R7A. J. Rhizobium leguminosarum bv. viciae into root hairs
Bacteriol. 184:3086–3095. requires minimal Nod factor specificity, but sub-
116. Sutton, J. M., E. J. Lea, and J. A. Downie. sequent infection thread growth requires nodO or
1994. The nodulation-signaling protein NodO nodE. Mol. Plant-Microbe Interact. 13:754–762.
from Rhizobium leguminosarum biovar viciae forms 128. Wang, H., Z.T. G. Zhong,T. Cai, S. P. Li, and
ion channels in membranes. Proc. Natl. Acad. Sci. J. Zhu. 2004. Heterologous overexpression
USA 91:9990–9994. of quorum-sensing regulators to study cell-
density-dependent phenotypes in a symbiotic and accessory components. Genome Biol. 7:R34.

plant bacterium Mesorhizobium huakuii. Arch. [Online.] http://genomebiology.com/2006/7/
Microbiol. 182:520–525. 4/R34.
129. White, C. E., and S. C.Winans. 2007.Cell-cell 134. Yuen, J. P.-Y., and G. Stacey. 1996. Inhibition
communication in the plant pathogen Agrobac- of nod gene expression in Bradyrhizobium japon-
terium tumefaciens. Phil. Trans. R. Soc. Lond. B icum by organic acids. Mol. Plant-Microbe Interact.
362:1135–1148. 9:424–428.
130. Wielbo, J., A. Mazur, J. Krol, M. Marczak, J. 135. Zhang, L. H. 2003. Quorum quenching
Kutkowska, and A. Skorupska. 2004. Com- and proactive host defense. Trends Plant Sci.
plexity of phenotypes and symbiotic behaviour 8:238–244.
of Rhizobium leguminosarum biovar trifolii 136. Zhang, R. G., T. Pappas, J. L. Brace, P. C.
exopolysaccharide mutants. Arch. Microbiol. Miller, T. Oulmassov, J. M. Molyneaux, J. C.
182:331–336. Anderson, J. K. Bashkin, S. C.Winans, and A.
131. Wilkinson,A.,V. Danino, F.Wisniewski-Dye, Joachimiak. 2002. Structure of a bacterial
J. K. Lithgow, and J.A. Downie. 2002.N-acyl- quorum-sensing transcription factor complexed
homoserine lactone inhibition of rhizobial with pheromone and DNA.Nature 417:971–974.
growth is mediated by two quorum-sensing 137. Zheng, H., Z. Zhong, X. Lai,W. X. Chen, S.
genes that regulate plasmid transfer. J. Bacteriol. Li, and J. Zhu. 2006. A LuxR/LuxI-type
184:4510–4519. quorum-sensing system in a plant bacterium,
132. Wisniewski-Dye, F., J. Jones, S. R. Chhabra, Mesorhizobium tianshanense, controls symbiotic
and J.A. Downie. 2002. raiIR genes are part of a nodulation. J. Bacteriol. 188:1943–1949.
quorum-sensing network controlled by cinI and 138. Zheng, H. M., Z. T. Zhong, X. Lai, W. X.
cinR in Rhizobium leguminosarum. J. Bacteriol. Chen, S. P. Li, and J. Zhu. 2006. A LuxR/
184:1597–1606. LuxI-type quorum-sensing system in a plant bac-
133. Young, J. P. W., L. C. Crossman, A. W. B. terium, Mesorhizobium tianshanense, controls sym-
Johnston, N. R.Thomson, Z. F. Ghazoui, K. biotic nodulation. J. Bacteriol. 188:1943–1949.
H. Hull, M. Wexler, A. R. J. Curson, J. D. 139. Zhu, J., and S. C. Winans. 1999. Autoinducer
Todd, P. S. Poole, T. H. Mauchline, A. K. binding by the quorum-sensing regulator TraR
East, M. A. Quail, C. Churcher, C. Arrow- increases affinity for target promoters in vitro and
smith, I. Cherevach, T. Chillingworth, K. decreases TraR turnover rates in whole cells. Proc.
Clarke, A. Cronin, P. Davis, A. Fraser, Z. Natl.Acad. Sci. USA 96:4832–4837.
Hance, H. Hauser, K. Jagels, S. Moule, K. 140. Zhu, J., and S. C.Winans. 2001.The quorum-
Mungall, H. Norbertczak, E. Rabbinow- sensing transcriptional regulator TraR requires its
itsch, M. Sanders, M. Simmonds, S. cognate signaling ligand for protein folding, pro-
Whitehead, and J. Parkhill. 2006.The genome tease resistance, and dimerization. Proc. Natl.Acad.
of Rhizobium leguminosarum has recognizable core Sci. USA 98:1507–1512.
QUORUM SIGNALING AND SYMBIOSIS IN
THE MARINE LUMINOUS BACTERIUM
VIBRIO FISCHERI
E.V. Stabb,A. Schaefer, J.L. Bose, and E. G. Ruby
15
Recent studies of the biology of the marine AHLs and sense the presence and abundance of
bacterium Vibrio fischeri have focused primarily related bacteria by the accumulation of these
on two related characteristics of this species: its autoinducers in their environment. As a result,
signature capacity to produce bioluminescence the quorum-induced cells are able to differen-
and its ability to symbiotically colonize the tially regulate genes whose products convey a
light-emitting organs of certain species of selective advantage only at a high bacterial den-
squids and fishes (65, 77). Not surprisingly, sity (30). AHL signaling was initially identified
these environment-specific activities have been because it regulated the expression of the lux
shown to be under the control of a series of gene clusters in V. fischeri and Vibrio harveyi (64)
cellular signal and response systems that and encoded bacterial luciferase and other pro-
include two-component phosphorelays (59, teins required for bioluminescence (20).Later it
101, 105), chemical receptors (6, 94), and quo- was recognized that quorum sensing controls
rum sensing (30, 96). additional gene products in these and other
Acyl-homoserine lactone (AHL)-based microbial species (see references 8, 42, and 56).
quorum sensing was first described in V. fischeri This signal-response system is both biologically
(64) in the early 1970s and has subsequently and genetically flexible and has evolved in dif-
been found to be present in at least 70 species ferent bacteria to regulate a variety of different
of other gram-negative bacteria, including a genes, even within a single genus (60). Genetic
number of important pathogens (39, 99). In an investigations have revealed that several species
idealized and simplified model of the quorum- encode two or more AHL systems, often with
sensing process (sometimes called “autoinduc- (i) one controlling the expression of the other,
tion”), cells continuously emit species-specific and/or (ii) each operating independently but
providing parallel inputs into a circuit that
E. V. Stabb and J. L. Bose Department of Microbiology, appears to function like a coincidence detector
University of Georgia,Athens, Georgia 30602. A. Schaefer (30, 62). The study of quorum sensing over
Department of Microbiology, University of Washington, the last 3 decades has been responsible for a
Seattle, Washington 98195. E. G. Ruby Department of
Medical Microbiology and Immunology, University of fundamental change in the way we think about
Wisconsin-Madison, Madison,Wisconsin 53706 bacterial behavior (39), yet the single most
233
234 ■ STABB ET AL.
intriguing characteristic of this form of com- have in their mind’s eye. The seminal paper
munication may be that it is deeply embedded describing autoinduction in luminous bacteria
in the way bacteria have evolved specific rela- used a strain designated MAV (63).When MAV
tionships, either beneficial or pathogenic, with was lost, V. fischeri MJ1 became the workhorse
eukaryotic hosts. strain in the biochemical and molecular dissec-
Recognition of the ubiquity and impor- tion of luminescence regulation, starting with
tance of bacterial symbioses in animals and the isolation and chemical identification of the
plants has led to an increasing interest in how first AHL signal (22), and continuing into the
bacterium-bacterium and bacterium-host sig- cloning, sequencing, and functional characteri-
naling maintains specificity and modulates zation of the MJ1 lux region (24, 25).Although
functionality in beneficial associations.The list strain ATCC 7744 has also been used (19),
of microbial symbioses in which AHL quorum sequence differences between the genes of the
sensing has been identified continues to grow ATCC 7744 and MJ1 lux operons are minimal
(33, 54, 87, 96, 97, 102) and has led to the devel- (18), and while these strains behave similarly
opment of a number of natural experimental with respect to luminescence and autoinduc-
systems for studying the mechanisms of this tion,ATCC 7744 often forms distinct colonies
cell-cell signaling process and its biological less readily. Thus, MJ1 became a useful type
consequences under conditions that exist in the strain, and most of our fundamental knowledge
real world (95). of the mechanisms underlying quorum sensing
In this chapter we review the role of quorum in V. fischeri was discovered in MJ1 or inferred
sensing in V.fischeri,focusing on recent develop- from experiments using the MJ1 lux genes
ments in our understanding of the genetics and cloned in Escherichia coli.
physiology of cell-cell signaling by populations
of these bacteria, both in culture and in their Development of an Environmentally
light-organ symbioses. Particular emphasis is Relevant Model System of Quorum
placed on outlining the regulatory factors and Signaling
pathways by which quorum sensing coordi- Unfortunately, MJ1 and other bright strains
nates the biological activities of this biolumi- posed an experimental limitation: they could
nescent microbe. not be studied under the natural conditions in
which V.fischeri produces light.Specifically,MJ1
MJ1 AND ES114:THE DIFFERENCE was isolated from the light-emitting organ of
A STRAIN MAKES the Japanese pinecone fish, Monocentris japonica,
The last 15 years have brought an easily over- where V. fischeri is highly bioluminescent (79);
looked shift in the major strains being used to however, these animals have yet to be induced
study quorum sensing in V. fischeri, and impor- to reproduce in the laboratory, making them an
tant differences between these isolates bear intractable model to manipulate and study the
emphasizing. Quorum sensing was discovered initiation of symbiosis.As interest shifted to the
because it was used by several marine bacteria implications of quorum sensing in the environ-
to control the induction of the luxICDABEG ment, particularly during growth in a host, the
operon, which is responsible for generating inability to reconstitute a natural symbiosis
bioluminescence, an easily measured pheno- with MJ1 became problematic. For example,
type. Earlier work typically utilized very bright although addition of autoinducers was discov-
isolates of V.fischeri that formed visibly lumines- ered to stimulate transcription of non-lux genes
cent colonies on plates and intensely light- in MJ1 (8), there was no way to determine
emitting broth cultures. This is, after all, the which, if any, of these genes were important for
phenotype that drew interest to V. fischeri and this strain’s interactions with M. japonica.
quorum sensing in the first place, and it is the A breakthrough in V. fischeri research came
image of this bacterium that most readers will when this bacterium was described as the light-
15. QUORUM SIGNALING AND SYMBIOSIS IN V. FISCHERI ■ 235
organ symbiont of the Hawaiian bobtailed cence is fully induced, ES114 cells produce
squid, Euprymna scolopes (100).This invertebrate only 1/10 the luminescence of cultured MJ1
had previously been reared in the laboratory cells.This puzzling property of poor light out-
(1), and when its juveniles hatched, they were put (i.e., cell-specific luminescence level) in
aposymbiotic (i.e., free of symbionts) and culture is conserved among virtually all V. fis-
acquired V. fischeri from the surrounding seawa- cheri isolates from E. scolopes (49, 50), and the
ter (100). Early work showed that V. fischeri reason for the prevalence of this trait among the
strains native to E. scolopes were especially well symbionts of this host, but not others (27, 68), is
adapted to this host and that nonnative strains not well understood. Importantly, it is not due
such as MJ1 did not colonize juvenile squid to a large-scale genetic difference: the orienta-
well; thus, MJ1 was not an appropriate strain for tion and function of the lux genes, which are
studying the symbiosis between V.fischeri and E. described below and underpin the quorum-
scolopes (77).Instead,V.fischeri ES114,which was sensing and bioluminescence functions of V. fis-
isolated from an E. scolopes light organ (3), was cheri, are similar in ES114 and MJ1 (38). Other
chosen for the subsequent array of ongoing components of the V. fischeri quorum-sensing
symbiotic studies. system are also found in both strains.
Although MJ1 continues to be used produc- The relatively dim luminescence of ES114
tively for biochemical studies of lux-encoded in culture has been attributed to a low produc-
proteins, ES114 has become the wild-type tion of the AHL autoinducer N-3-oxo-hexa-
strain of choice for most researchers investigat- noyl-homoserine lactone (3-O-C6-HSL) (3),
ing quorum sensing and/or symbiosis. For this and a convincing argument can be made for
reason, ES114 was the first V. fischeri strain to this assertion. For example, in culture ES114
have its genome sequenced (80), and as a result, makes only 0.01% of the 3-O-C6-HSL that
commercially produced microarrays are based MJ1 does (Table 1). Moreover, even when the
on the gene content of ES114 and most mutant lux operon of ES114 is fully induced with
analyses are now done in ES114. Furthermore, exogenous 3-O-C6-HSL, de novo production
many of the recent advances in understanding of this autoinducer is far lower than what is
quorum sensing in V. fischeri, as well as all of the observed in MJ1 (38). On the other hand, a low
symbiotic studies described below, have used level of 3-O-C6-HSL production is a some-
ES114 as the wild-type and/or parent strain. what unsatisfying and circular explanation for
For these reasons, the following sections of this low luminescence, considering that the 3-O-
chapter focus primarily on V. fischeri ES114; C6-HSL autoinducer synthase gene, luxI, is
however, before doing so, it is relevant to sum- cotranscribed with the luxCDABEG genes
marize certain important similarities and differ- that encode the enzymes responsible for biolu-
ences between strains ES114 and MJ1. minescence.Thus, one might reasonably assert
that low 3-O-C6-HSL production does not
A Comparison of Quorum Signaling cause low luminescence in ES114 so much as
between V. fischeri Strains both phenomena are the result of some other
The most obvious difference between MJ1 and mechanism that attenuates expression of the
ES114 is that the latter is not visibly lumines- luxICDABEG operon (at least in culture).
cent in culture.While the specific luminescence Interestingly, 3-O-C6-HSL levels are consider-
(i.e., light emission per cell) of ES114 increases ably higher in the light organ of colonized E.
2 to 3 orders of magnitude when it colonizes scolopes than they are in cultures of ES114
the E. scolopes light organ (3), the luminescence grown outside the host (4), and engineered
of cultured ES114 cells, even at very high cell overexpression of ES114 luxI and the other
density, typically can be detected only using a lux genes within cells of ES114 leads to a high
luminometer. Interestingly, even in the symbio- level of luminescence (38). These results indi-
sis, where cells are packed densely and lumines- cate that strain ES114 has the metabolic means
TABLE 1 Output of HSL autoinducers by V. fischeri strains ES114 and MJ1

3-oxo-C6-HSL (nM) C8-HSL (nM)
Culture optical density
ES114 MJ1 ES114 MJ1
0.5 0.01 2 35 11
1.5 0.1 2,100 1,100 180
2.8 0.2 7,400 ND 1,700
to produce higher levels of both luminescence larly, two other notable and documented differ-
and 3-O-C6-HSL than it normally does and ences between lux regulation in ES114 and
suggest that an important regulatory difference MJ1 are that glucose (29, 94) and iron (41)
between MJ1 and ES114—located outside the repress luminescence in strain MJ1, but these
lux operon—exerts an effect on this locus. physiological factors do not affect strain ES114
One candidate regulator known to affect in culture (3).
bioluminescence in V. fischeri is ainS (32), which Given the dramatic differences between
produces a second autoinducer, N-octanoyl- ES114 and MJ1, it is clear that in V. fischeri the
homoserine lactone (C8-HSL), as discussed in processes of quorum sensing and lux regulation
greater detail below. Production of C8-HSL are strain dependent. Thus, when considering
initiates stimulation of the lux operon at mod- individual studies, readers should be careful to
erate cell densities; if ES114 had relatively low note which of these model strains is being used.
expression of AinS and low C8-HSL output, The differences between gene regulation in
this deficiency could potentially explain why
ES114 is so much dimmer than MJ1. However,
not only is C8-HSL production unimpaired in
ES114, it is even greater than the output of this
autoinducer by MJ1 (Table 1).
Most likely, then, the large difference in the
levels of 3-O-C6-HSL production and lumi-
nescence seen between cultures of ES114 and
MJ1 is due to external regulatory influences on
the autoinducer synthase genes, and such regu-
lation may be multifactorial. In this regard, it is
worth noting that bioluminescence is affected
by environmental conditions very differently in
these two strains. One example of this differ-
ence is how the expression of luminescence in
ES114 and MJ1 varies with culture aeration
(Fig. 1). In the experiment presented, aliquots
from each culture were vigorously shaken (so
that the cell’s luciferase was fully oxygenated)
immediately before their specific light emission
was measured in a luminometer. With ES114,
the more highly aerated the culture, the greater
the maximal expression of luminescence (Fig. FIGURE 1 Specific luminescence (luminescence
1A); in contrast, while differences in aeration per A595) of ES114 (A) or MJ1 (B) grown at 24C in
250-ml flasks, shaken at 200 rpm, in 50 (diamonds), 100
affected the culture density (A595) at which MJ1 (squares), or 200 (triangles) ml of SWTO, a rich nutri-
cells induced luminescence, they did not affect ent medium (7, 66). Bacterial cell density was measured
maximal luminescence output (Fig. 1B). Simi- by absorbance of the culture at 595 nm (A595).
ES114 and MJ1 raise interesting questions and oxygen to a long-chain fatty acid, flavin
regarding the evolution of V. fischeri and provide mononuleotide, H2O, and light. The luxC,
the grist for potentially enlightening compara- luxD, and luxE genes encode a fatty acid reduc-
tive studies; however, given our increased use of tase complex that resynthesizes the aldehyde
ES114 and the fact that researchers can utilize substrate, and luxG is involved in flavin
this strain in a reconstituted symbiosis, we focus mononucleotide biosynthesis. In V. fischeri the
the rest of our consideration on this strain. expression of these lux genes is regulated in a
cell-density-dependent fashion through bind-
A MECHANISTIC MODEL OF THE ing of the LuxI-synthesized AHL signal, 3-O-
THREE QUORUM-SIGNALING C6-HSL,to the transcriptional activator protein
SYSTEMS OF V. FISCHERI LuxR, encoded by luxR. The LuxR-AHL
The last 5 years have seen a considerable complex then binds to the luxICDABEG pro-
advance in our understanding of quorum sens- moter and induces the transcription of this
ing within the genus Vibrio and, specifically, V. operon (85). Thus, the transcription of luxI
fischeri. Figure 2 presents a summary of some of leads to the production of an AHL signal
our current knowledge concerning the cir- (autoinducer) that results in a further increase
cuitry of signaling in V. fischeri.Three quorum- in its transcription (Fig. 2). Homologs of this
sensing signals have been described in V. fischeri LuxI-LuxR system are present in dozens of
ES114. As mentioned above, two of them are bacterial species, controlling a wide diversity of
AHLs: (i) C8-HSL, synthesized by AinS (32, activities, including luminescence, motility,
48), and (ii) 3-O-C6-HSL, synthesized by LuxI biofilm formation, pigmentation, antibiotic
(20, 22).The third signal is produced by V. fis- production, and virulence (16, 30, 60).
cheri LuxS, a homolog of the V. harveyi protein
that synthesizes a furanosyl borate diester prod- Systems 2 and 3:The C8-HSL
uct called autoinducer-2 (AI-2) (9).These three and AI-2 Signals
signal synthases are all active in V. fischeri (55, As first described in V. harveyi (99), parallel (or
57), and their products accumulate in the envi- “hybrid”) inputs are used by V. fischeri to sense
ronment around the cells.Together these signals the accumulation of both C8-HSL and AI-2.
comprise three systems that coordinately regu- The C8-HSL synthesized by AinS presumably
late specific genes in a cell-density-dependent functions with its cognate receptor, termed
manner (Fig. 2). AinR, activating it as a transcriptional regulator
(32). The ainR gene is located immediately
System 1:The 3-O-C6-HSL Signal downstream of ainS (Fig. 2), and the two genes
The first level of quorum signaling recognized are transcriptionally linked (unpublished data).
in V. fischeri was that controlled by the canonical The arrangement of these genes is analogous to
LuxI-LuxR system.This system is embedded in that of their homologs in V. harveyi, luxM and
the lux operon of both V. fischeri (24) and Vibrio luxN. Biochemical and genetic studies of AHL
salmonicida (67) and has been the subject of ex- signaling in V. harveyi have shown that, in the
tensive physiological, biochemical, and genetic presence of an inducing concentration of
studies (20, 30). Briefly, the enzymes catalyzing the LuxM-synthesized AHL, the receptor,
the bacterial luminescence reaction are LuxN,participates in a phosphorelay cascade by
encoded by the last six genes of the luxICD- stimulating the relative dephosphorylation of
ABEG operon,which in V.fischeri is adjacent to, LuxU (Fig. 2).
but divergently transcribed from, the luxR The LuxS system of V. fischeri also appears
gene.The luxA and luxB genes encode, respec- homologous to that described in V. harveyi,
tively, the and subunits of luciferase, an although the roles and activities of several of
enzyme that catalyzes the conversion of a long- the components have not been directly
chain aldehyde,reduced flavin mononucleotide demonstrated in V. fischeri.As in other bacteria,
FIGURE 2 Model of quorum sensing in V. fischeri ES114. Each gene is indicated by a labeled open arrow with the open reading frame designation from the V. fis-
cheri genome database provided underneath in parentheses (e.g., luxI is designated VFA924).The structures of three autoinducer molecules are presented underneath
their respective synthases, and a “?” by AI-2 indicates that its structure is inferred but has not been identified in V. fischeri. Interactions between autoinducers, proteins,
genes, and small RNAs (designated “sRNAs”) are indicated, and dotted lines around sRNA or protein components indicate that these have been identified in the
genome database but not functionally confirmed in ES114 experimentally.
the V. fischeri LuxS is responsible for AI-2 syn- In the V. fischeri core circuit, the C8-HSL sig-
thesis, and the integration of this system into nal appears to have a considerably stronger
quorum sensing is illustrated by the observation influence on the expression of bioluminescence
that a luxS mutant of V. fischeri is dimmer than than does AI-2. For example, a luxS deletion
wild type (55).LuxS can give rise to structurally mutant, which lacks AI-2 production, still
distinct versions of AI-2 (58).To date, the spe- achieves 70% of the level of bioluminescence
cific structure of AI-2 produced by V. fischeri produced by wild type (55). However, the
has not been determined; however, it seems importance of luxS is probably context depend-
likely that this species’ bioactive AI-2 autoin- ent; i.e., AI-2 may have a significant effect on
ducer is also a furanosyl borate diester. Also, sensing and regulation of bioluminescence in
by analogy to work done in V. harveyi, and on certain environments. As one might predict
the basis of genomic analysis of V. fischeri, we (Fig. 2), the fold-decrease in luminescence asso-
predict that AI-2 is perceived by LuxQ via ciated with a luxS mutation is greater in an ainS
LuxP and that the resulting signal is transduced mutant background (55), and conditions that
through the relative dephosphorylation of favor AI-2 accumulation over that of C8-HSL
LuxU (Fig. 2). may change the extent of the regulatory contri-
bution of each of these systems. As discussed
Systems 2 and 3: Convergence below, environmentally responsive regulators
on a Core Circuitry in Vibrio can influence the rate of autoinducer signal pro-
Quorum Sensing duction and accumulation, so it would not be
As illustrated in Fig.2,the V.fischeri C8-HSL and surprising to find that environmental condi-
AI-2 signaling systems converge at LuxU. In tions influence the relative importance of the
our current model, the presence of either three quorum-signaling systems (Fig. 2).
autoinducer leads to the relative dephosphory- In conjunction with the AI-2-activated cir-
lation of LuxU, which in turn leads to a more cuit, the result of the parallel AHL-signaling
dephosphorylated, inactive condition for the pathway is to dephosphorylate the transcrip-
sigma-54-dependent regulator LuxO (Fig. 2). tional regulator LuxO (52, 99). In V. harveyi, this
As mentioned above with respect to transduc- inactivation reduces the concentration of sev-
tion of the C8-HSL signal, the genes encoding eral regulatory sRNAs whose presence inhibits
components of this pathway have been identi- the translation of the V. harveyi master regulator,
fied in the V. fischeri genome, but only the roles LuxR (a homolog of V. fischeri LitR [Table 2]),
of LuxO, sigma-54, and LitR have been tested resulting in an increase in expression of not
experimentally. Importantly, as predicted from only the lux operon but also a number of other
the current model (Fig. 2), the influence of luxS recently discovered target genes (62).The com-
on bioluminescence is dependent on luxO (55). ponents of this “core” phosphorelay system are
Recently, another signaling input to LuxO has encoded in the genomes of many if not all Vib-
been discovered: the global regulator CsrA is rio species (61) (Table 2).While this core signal-
implicated in quorum sensing by Vibrio cholerae, ing circuitry is best described in the V. harveyi
feeding information into LuxO that affects this paradigm (99), it is already clear that there are
protein’s output (51). All sequenced Vibrio both qualitative and quantitative differences
species, including V. fischeri, have homologs of among the phosphorelay cascades found in dif-
CsrA, as well as small RNAs (sRNAs) that reg- ferent species (15,56,106).For instance,in V.fis-
ulate it, encoded by several csrB genes (47). At cheri, LuxO may regulate the synthesis of as few
present there is no direct evidence that V. fischeri as one sRNA (personal communication) (Fig.
quorum sensing is affected by the CsrA system; 2), rather than the four described for V. harveyi
however, a mutation in GacA, which regulates (51, 99). Similarly, while Vibrio anguillarum has
CsrA in other species (89), suppresses induction all the typical core circuit components (Table
of V. fischeri luminescence (101). 2), it regulates LuxO quite differently; e.g.,
TABLE 2 Presence of homologous quorum-sensing systems in some Vibrio speciesa

Ortholog in the following Vibrio species:
System
Vf Vp Vv Vc Vh Va
System 1 LuxI — — — — Vanl
LuxR — — — — VanR
System 2 AinS LuxM — — LuxM VanM
AinR LuxN — — LuxN VanN
System 3 LuxS LuxS LuxS LuxS LuxS VanS
LuxP LuxP LuxP LuxP LuxP VanP
LuxQ LuxQ LuxQ LuxQ LuxQ VanQ
System 4 — CqsA — CqsA CqsA ND
— CqsS — CqsS CqsS ND
Core circuit LuxU LuxU LuxU LuxU LuxU VanU
(for systems
2, 3, and/or 4)
LuxO LuxO LuxO LuxO LuxO VanO
LitR OpaR SmcR HapR LuxRb VanT
CsrA CsrA CsrA CsrA ND ND
a
Vf, V. fischeri;Vp, V. parahaemolyticus;Vv, V. vulnificus;Vh, V. harveyi;Va, V. anguillarum. Systems 1 to 4 are listed in the order in which they
were discovered; the absence of a recognizable ortholog in the published genome is indicated by a dash (—). Core components constitute a
signaling pathway shared by systems 2 and 3. Bold type indicates the signal synthases. ND, not determined (no genome sequence available).
b
The V. harveyi LuxR is not a homolog of the V. fischeri LuxR.
phosphorelay signaling apparently can proceed studies indicate that the output of this sensor
independently of LuxU and is subject to feed- feeds into the core circuitry at the level of
back by the LitR homolog VanT (14, 17). LuxU.CqsS signaling has also been described in
Finally, there is evidence that, in addition to V. cholerae (51), where it appears to play a more
luminescence, the AinS/LuxO pathway regu- dominant role than in V. harveyi (53, 99). Exam-
lates genes affecting general cellular functions; ination of the genome of V. fischeri ES114 pro-
i.e.,in contrast to a luxI mutant (92),the growth vides no evidence of a CqsS-CqsR system in
yield of an ainS mutant is only 75% of the level this organism (Table 2).
of wild-type cells (57).These differences may be
mediated through the AinS/LuxO-dependent Evidence for Sequential AHL
induction of a unique V. fischeri sigma factor Signaling in V. fischeri
(56) and have not been described in mutants of In strain ES114, levels of C8-HSL quickly
other Vibrio species. become saturating in culture.While exogenous
addition of this autoinducer can complement
System 4: a Parallel Input in Some an ainS mutant’s luminescence defect, it has no
Signaling Pathways effect on either a wild-type or luxI mutant (57),
In the preceding sections we described the an observation that is consistent with a sequen-
pathways by which two AHLs and the AI-2 tial effect of the two AHLs (Fig. 2). Adding 3-
signal work both in parallel and sequentially to O-C6-HSL to an ainS luxI double mutant
regulate the induction of luminescence and recovers luminescence only to the level charac-
other activities in V. fischeri (Fig. 2). In a subset teristic of an ainS mutant, indicating that C8-
of Vibrio species (Table 2), another signaling HSL signaling must be normal before cells can
system feeds into the central parallel circuitry respond to even a high concentration of 3-O-
along with systems 2 and 3 (99). In V. harveyi C-HSL.Thus, the impact of C8-HSL signaling
the protein CqsS apparently synthesizes an as (at least on ES114 lux gene expression) is evi-
yet unidentified signal that interacts with dent at cell concentrations occurring in culture
CqsR,a membrane-bound sensor (43).Genetic and continues to be important at the higher
densities more characteristic of the light organ remains more than a 100-fold dimmer (56).
environment. In contrast, the inducing effect of During the early stages of growth, this differ-
the LuxI signal is apparent only at the very high ence between the ainS and litR mutants may be
bacterial concentrations in the symbiosis. A due to C8-HSL combining with LuxR to
similar mechanism for sequentially linking two directly pre-induce the lux operon (including
AHL systems has been well documented in luxI), thereby “jump-starting” the powerful 3-
Pseudomonas aeruginosa (chapter 9), and results O-C6-HSL/LuxR transcriptional effects.
in a highly complex and interacting network of However, late in growth, a luxO “on” mutation,
signaling (82). which is unable to induce the repressing
sRNAs, completely complements ainS (57).
Multiple Roles of C8-HSL This result indicates that the ainS luminescence
in Quorum Sensing defect is expressed primarily through LuxO.
The C8-HSL autoinducer in V.fischeri functions The nature of this LitR-independent LuxO
through more than one mechanism (Fig. 2), regulation remains unknown, but it may work
although this signal apparently has other activi- through another, physiological, level of regula-
ties that cannot yet be integrated into the cur- tion of the lux operon (see below).
rent model. Besides signaling through AinR
and the “conserved core circuitry” described QUORUM SIGNALING IN A
LIGHT-ORGAN SYMBIOSIS
above, C8-HSL can interact with LuxR
directly. When 3-O-C6-HSL is limiting, C8- The V. fischeri-E. scolopes
HSL can serve as a weak activator of LuxR (23, Light-Organ Association
81) (Fig. 2); however, when 3-O-C6-HSL is Even correcting for our disproportionate focus
abundant, C8-HSL can actually inhibit 3-O- on pathogenic bacteria, almost all bacterial
C6-HSL signaling (48; unpublished data). This species recognized to use AHL quorum sensing
result can be interpreted as reflecting the differ- are host associated (39).This trend suggests that
ential capacity of these two AHLs to bind and for many microorganisms the biologically sig-
activate LuxR. C8-HSL signaling may work nificant function of quorum signaling may be
through another mechanism as well.While the fully revealed only by studying the colonization
ainS mutant of V. fischeri ES114 produces very of their natural host.The beneficial association
little luminescence, the addition of 3-O-C6- between V. fischeri and its squid host is an exam-
HSL does not induce luminescence to a wild- ple of just such a biologically relevant system.
type level (57), suggesting that activation by Symbiotic colonization of the light organ of E.
C8-HSL must precede or prepare the way for scolopes proceeds through a series of well-
subsequent induction by 3-O-C6-HSL. described stages (72, 87) involving develop-
From the current model (Fig. 2), one might mental adaptation and accommodation by both
predict that C8-HSL accomplishes this prepara- partners (96). Briefly, free-living V. fischeri cells
tory activity by enhancing LitR levels and that present in the ambient seawater attach to a
this enhancement must be critical for LuxR host-derived mucus matrix produced by
accumulation and full induction by 3-O-C6- epithelial appendages located immediately out-
HSL. However, it is difficult to reconcile such a side of pores that lie on the surface of the nas-
model with all the available data: notably, the cent light organ of a juvenile squid (69, 70).
important phenotypic differences between ainS These pores lead to six epithelium-lined crypt
and litR mutants.For example,an ainS mutant is spaces deep within the organ. The aggregated
dark in culture, whereas a litR mutant is only bacteria migrate along the mucus, through the
delayed in its induction of luminescence (26, pores and into the crypts, where they rapidly
57).Moreover,addition of 3-O-C6-HSL stimu- proliferate, colonizing the light organ with
lates luminescence in a litR mutant just as well as approximately 106 bacteria (72, 76). Once it
it does in wild type, while an ainS mutant reaches a critical cell density, the symbiont
population induces luminescence that is ulti- The Importance of Cell-Cell

mately used by mature squid in their nocturnal Signaling in Symbiosis
behavior (44). At dawn every morning At bacterial concentrations below 109 cells/ml
throughout its life, the squid expels 90 to 95% (i.e., the maximum cell density normally
of the symbiont population, and the remaining achieved in broth cultures), C8-HSL dominates
cells repopulate the organ within a few hours, the control of lux expression in V. fischeri ES114
creating a fresh bacterial culture for the coming (57). In contrast, in the light organ, where sym-
night (37, 71).The resulting cyclic daily rhythm biont concentrations are well over 109 cells/ml,
plays an integral part in the biology of this both signals are required for maximum luci-
dynamic association (5, 49). ferase synthesis and luminescence, although the
The ability to colonize the squid light organ ability to synthesize 3-O-C6-HSL clearly plays
is remarkably specific: only V. fischeri and, to a the greatest role (57).Bioassays of 3-O-C6-HSL
much lesser extent, the very closely related Vib- in the adult light organ have indicated a con-
rio. logei (27) can successfully initiate an associa- centration of approximately 100 nM, signifi-
tion. Studies of mutants of V. fischeri have shown cantly above that necessary to fully induce
that a number of bacterial products (e.g., cap- luminescence in culture (4). Because a func-
sule [96, 105]) and behaviors (e.g., motility and tional luxI gene is required for normal levels of
luminescence [36, 59, 92]) play important roles light emission in the symbiosis, and mutants
in the symbiont’s ability to initiate and persist in defective in luminescence fail to persist, 3-O-
the association. In all cases studied to date, these C6 signaling is an important colonization factor
colonization factors are under signal-induced (57, 92).Thus, although the LuxI signal has little
regulation involving two-component phos- effect on luminescence in culture, it is critical
phorelays and/or quorum sensing (31, 96). for symbiotic performance.
The symbiotic roles of the three V. fischeri While a luxI mutant initiates the symbiosis
quorum-sensing systems (Fig. 2) have been normally,an ainS mutant is delayed in coloniza-
examined during the initiation and stabilization tion, suggesting that C8-HSL signaling plays a
stages of the juvenile light-organ association. role during the first few hours of the symbiotic
The two AHL quorum-sensing signaling sys- interaction, as the bacteria form aggregates and
tems of V. fischeri contribute to colonization of enter the light organ (73).Thus, the sequential
the light organ in distinct ways. Mutation of the nature of the pathway of quorum signaling,
ainS (but not the luxI) signal synthase delays ini- proceeding from AinS to LuxI, is not only
tiation of the symbiosis, while mutation of luxI found in culture but is similarly evident in the
results in a much more severe luminescence natural process of colonization (57).ainS signal-
defect in the symbiosis (57). As described ing is also required for normal persistence in the
below, loss of either of these AHL signals leads symbiosis (57); however, this requirement
to an inability of V. fischeri to persist normally in appears to extend beyond the role of C8-HSL
the light organ beyond the first 24 h. In con- in inducing a normal level of luminescence.
trast, the role of LuxS signaling in colonization Specifically, an ainS strain engineered to have
is minor and can only be observed as a small a lower specific luminescence (i.e., less light
decrease in an ainS luxS double mutant back- emitted per cell) than the ainS mutant never-
ground relative to the much larger ainS defect. theless persists normally (93). This and other
While there is apparently little role for LuxS in observations suggest that some additional, non-
the monospecific light-organ symbiosis, it luminescence activity that is regulated by ainS
remains possible that AI-2 signaling may serve through the LuxO-dependent pathway con-
V. fischeri in its other ecological niches (such as tributes to symbiotic persistence in the light
the enteric tracts of marine organisms [78]) organ. It is not surprising, then, that other ainS-
where it resides in multispecies communities regulated cellular functions should exist, owing
with other AI-2 signaling species (77). to the presence of a number of LuxO-regulated
phenotypes reported for different Vibrio species rently, studies are under way to determine
(52, 91, 107). Microarray analysis of the ainS whether quorum signaling induces symbiotic V.
regulon has presented a series of candidates fischeri cells to induce the release of lipid A
required for symbiotic persistence, as well as for and/or tracheal cytotoxin,either during the ini-
the pathway of their regulation downstream of tiation of colonization or during the daily cycle
LuxO (56). of expulsion and regrowth that characterizes
the persistent association.
Signaling between V. fischeri While such indirect signaling of host devel-
and Its Host opment may result from V.fischeri quorum sens-
As described above, the initiation of light-organ ing, examples of a more direct action of
colonization brings about a specific series of bacterial AHLs on host tissue are appearing in
genetic and behavioral modifications in the pathogenic models (reviewed in reference 84),
bacterial symbionts, including several quorum- and this mode of signaling may be a factor in
induced activities. Similarly, V. fischeri coloniza- the light-organ symbiosis as well. Not only is 3-
tion of a juvenile squid triggers a program of O-C6-HSL present in the squid light-organ tis-
biochemical and morphological development sues at near-micromolar concentrations, but
that transforms the nascent light organ into a this AHL has also been shown to freely diffuse
mature symbiotic structure (72, 96). Recent into those tissues (4). Current investigations by
studies have begun to reveal the signals that trig- M. McFall-Ngai and her collaborators are
ger host development and, to date, three bacter- aimed at determining whether this or other
ial signal compounds, all of them modified cell quorum-sensing signals of the V. fischeri sym-
envelope components, have been identified. bionts are also sensed, and responded to, by the
The first signal is the lipid A component of V.fis- host, leading to changes in gene expression in
cheri lipopolysaccharide, a potent morphogen the light-organ epithelium. In pathogenic asso-
inducing apoptosis of the mucus-secreting ciations, one response of animal tissue is the
appendages that facilitate the bacterial attach- inactivation of bacteria-produced AHLs (10),
ment and colonization (28, 46). In addition, possibly playing a role in the host’s innate
peptidoglycan (PGN) fragments shed by V. fis- defenses. Such an activity in the light organ,
cheri cells induce the secretion of mucus by the while as yet unobserved, could modulate the
light organ (69), possibly by a signaling pathway response of both the host and the symbionts to
using the host’s PGN receptor proteins (34). bacteria-produced quorum-sensing signals.
Finally, a tetrapeptide fragment of PGN, first
described as the tracheal cytotoxin of Bordetella POSITIVE FEEDBACK AND
pertussis (13) and the PGN-derived cytotoxin of ENVIRONMENTAL CONTROL
Neisseria gonorrhoeae (11, 75), has been shown to OF SIGNAL SYNTHESIS
induce trafficking of hemocytes into the light In V. fischeri, as in many other bacteria, the rate
organ (45).Tracheal cytotoxin works synergisti- of autoinducer accumulation is not constant
cally with lipid A, inducing the full program of but rather is tied to certain environmental con-
tissue regression as the organ matures (46). ditions.Throughout this volume it is clear that
These three V. fischeri-derived signals, required environmentally responsive regulators often
for normal symbiotic development in the light- control the rate of autoinducer production, and
organ association, have previously been shown also may modulate the rate of signal degrada-
to function as virulence determinants in several tion by the autoinducer-producing bacterium
pathogenic bacterial infections (12). Thus, the (e.g., chapter 24). Moreover, several autoin-
discovery of the remarkably different roles these ducer systems control themselves in a positive
compounds play in microbe-host interactions feedback manner (40, 74, 83) and, as a conse-
emphasizes the importance of context in the quence, regulatory inputs from environmental
evolution of signal-response pathways. Cur- stimuli can be greatly amplified.
Both the luxI and ainS genes of V. fischeri are functions as a repressor by interfering with
autostimulatory; i.e., an accumulation of 3-O- LuxR-mediated activation. Interestingly, the
C6-HSL and C8-HSL results in an increased ArcA/ArcB system does not repress lumines-
transcription of their synthase genes, luxI and cence during colonization of juvenile E.
ainS, respectively (55). The positive-feedback scolopes, and an arcA mutant achieves the same
autoregulation of 3-O-C6-HSL is readily illus- level of brightness in culture as wild-type cells
trated in Fig. 2, which depicts how this signal do in the light organ (Bose et al., submitted).
combines with LuxR to stimulate expression of Thus, redox state appears to be a key environ-
LuxI, the 3-O-C6-HSL synthase. Multiple lines mental factor determining expression of luxI-
of evidence (55) suggest that C8-HSL stimulates CDABEG, and regulation by the ArcA/ArcB
ainS transcription through the AinR,LuxU,and system could account for the differences in
LuxO pathway, leading to an increased level of bioluminescence and autoinducer output
LitR (Fig. 2).Although AI-2 does not appear to observed in culture and in host tissues.Another
affect transcription of luxS (55), this signal is intriguing and as yet untested implication of
predicted to increase LitR abundance and these findings is that one V. fischeri cell perceiv-
therefore AinS and C8-HSL output. As ing a more oxidized environment through its
described above, this output leads to positive- ArcA/ArcB system might transmit this infor-
feedback regulation through core components mation to neighboring cells in the form of an
common to the AI-2 signaling pathway.Thus, if increased production of 3-O-C6-HSL.
environmentally responsive regulators modu- It seems likely that environmentally respon-
late transcription of luxI, ainS, or possibly luxS, sive regulators also modulate the C8-HSL and
such control can be amplified greatly by the AI-2 signals in V. fischeri, although data in this
positive-feedback activity of the autoinducers. regard are scant. Interestingly, there is a large
In V.fischeri,environmental factors play a role inverted repeat located near the 5′ end of the
in modulating autoinducer production, biolu- ainR gene of V. fischeri that is absent in its
minescence, and other coregulated functions, homolog (e.g., luxN) in other Vibrio species.
and autoinducer concentrations reflect both The presence of this element could indicate a
the ambient conditions and cell density, not novel regulatory mechanism controlling the
simply the latter. That cell density is not the receptor for C8-HSL, and because of the
only determinant governing bioluminescence autostimulatory nature of this signal, such con-
is illustrated by the observation that in colonies trol might influence C8-HSL accumulation as
on agar medium and in the E. scolopes light well. Similarly, environmental control of genes
organ, bacteria are packed to similarly high encoding several other intermediates in the C8-
densities, yet only in the latter environment are HSL signaling pathway, such as sigma-54 (103)
they highly bioluminescent. Therefore, some or LitR (26),could connect conditional regula-
environmental difference between the light tors with C8-HSL accumulation. There is also
organ and a nutrient-rich agar medium helps precedence for control of AI-2 accumulation
determine the level of luxICDABEG (and by the modulation of expression of luxS or the
other gene) expression. related enzyme pfs (2, 98).This level of regula-
We recently discovered that the redox- tion might be active in V. fischeri as well,
responsive ArcA/ArcB two-component regu- although V. fischeri apparently lacks the pathway
latory system, which is activated in response to for AI-2 destruction mediated by LsrF, LsrG,
more reduced conditions, mediates repression and LsrK in E. coli and Salmonella (90, 104).The
of the luxI promoter (J. L. Bose, C. S. Rosen- possibility that the C8-HSL or AI-2 signaling
berg, and E.V. Stabb, submitted for publication). systems are modulated merits further investiga-
The phosphorylated response regulator, ArcA- tion, as we may find that in certain environ-
P, binds the luxI promoter proximal to the ments these signals are even more important
LuxR-binding site, and we propose that it than is currently appreciated.
FUTURE DIRECTIONS We expect that bioinformatic and computa-

Recent advances in V. fischeri research, notably tional approaches will help build and test such
the completed genome sequence of strain models and reveal new pathways in this regula-
ES114 (80) and the development of molecular tory web. In the past, components of the cur-
genetics (21, 88), have enabled us to draw a rent model (Fig. 2) have been tested for their
fairly detailed model of how three autoinducer influence on bioluminescence, autoinducer
systems in this bacterium control the regulation synthesis, and/or transcription from the luxI or
of bioluminescence (Fig. 2); however, several luxR promoters; in some cases, their depend-
aspects of this model await experimental con- ence on other components of this regulatory
firmation. In particular, the details of the C8- pathway has been tested by mutational analyses.
HSL and AI-2 pathways draw extensively from Such approaches will continue to be useful,
work performed on homologous components but microarray analyses also now offer the
in V. harveyi, and genetic and biochemical tests opportunity to simultaneously and quantita-
of the roles for AinR, LuxP, LuxQ, Hfq, and tively assay regulatory effects on transcripts
sRNAs in V. fischeri await completion. Because for multiple components of the pathway,
these components are conserved in several Vib- potentially uncovering how a single regulator
rio species (Table 2), we expect that they will connects to the regulatory network at multiple
function the same in V. fischeri as in these other nodes. Such analyses combined with computa-
systems. Nonetheless, careful examination of tional approaches will prove a powerful way to
these components is likely to reveal differences model the many connections and their relative
unique to quorum sensing in V. fischeri and is importance in this complex system.
necessary to build a complete understanding of The sequencing of the first V.fischeri genome
quorum sensing in this model system. Such has led to the construction and application of
knowledge will be critical for elucidating glass-slide and Affymetrix chip microarrays,
exactly where and how environmentally opening up the opportunity to map out the
responsive regulators feed into the quorum- transcriptional circuitry defining a quorum-
sensing circuitry illustrated in Fig. 2. sensing pathway in this bacterium (56). As a
Unraveling the complex series of interac- result, we can expect soon to see what other
tions controlling autoinducer synthesis and genes and functions are part of the AHL-
bioluminescence in V. fischeri presents a chal- induced regulon, as well as the ways the several
lenge that is likely to grow exponentially as quorum-sensing regulons in V. fischeri interact
more regulatory components are uncovered. with each other and with other modulating cir-
For example, environmentally responsive regu- cuitry, as described in P. aeruginosa (82).
lators may feed into this circuitry at any of a Recently a second V. fischeri isolate, strain MJ11
number of sites and in some cases may exert obtained from a M. japonica light organ, has had
control over more than one component of the its genome sequenced, assembled, and anno-
pathway. Importantly, as each new regulator is tated (unpublished data).The availability of the
discovered, we are presented not only with the MJ11 sequence will open the door to exciting
question of how it connects to the current comparative studies of quorum-sensing net-
model (Fig. 2), but we must also consider works in V. fischeri, with an eye to discovering
whether other regulators affect it and how. the evolutionary flexibility of these signaling
Such questions may best be addressed using pathways as they adapt to the specific environ-
mathematical models and integrated analysis to ments presented by different hosts.
understand the dynamics of quorum-sensing
regulatory networks (35, 86), allowing us to REFERENCES
place quantitative values on the contributions 1. Arnold, J. M., C. Singley, and L. Williams-
to effector output of the individual pathways in Arnold. 1972. Embryonic development and
the network (Fig. 2). post-hatch survival of the sepiolid squid Euprymna
scolopes under laboratory conditions. Veliger LuxR, regulates serine, metalloprotease, pigment,
14:361–364. and biofilm production in Vibrio anguillarum.J.Bac-
2. Beeston, A. L., and M. G. Surette. 2002. pfs- teriol. 184:1617–1629.
dependent regulation of autoinducer-2 produc- 16. Croxatto, A., J. Lauritz, C. Chen, and D. L.
tion in Salmonella enterica serovar Typhimurium. Milton. 2007. Vibrio anguillarum colonization of
J. Bacteriol. 184:3450–3456. rainbow trout integument requires a DNA locus
3. Boettcher, K. J., and E. G. Ruby. 1990. involved in exopolysaccharide transport and
Depressed light emission by symbiotic Vibrio fis- biosynthesis. Environ. Microbiol. 9:370–382.
cheri of the sepiolid squid Euprymna scolopes.J.Bacte- 17. Croxatto, A., J. Pride, A. Hardman, P.
riol. 172:3701–3706. Williams, M. Camara, and D. L. Milton. 2004.
4. Boettcher, K. J., and E. G. Ruby. 1995. Detec- A distinctive dual-channel quorum-sensing system
tion and quantification of Vibrio fischeri autoin- operates in Vibrio anguillarum. Mol. Microbiol.
ducer from symbiotic squid light organs.J.Bacteriol. 52:1677–1689.
177:1053–1058. 18. Devine, J. H., C. Countryman, and T. O. Bald-
5. Boettcher, K. J., E. G. Ruby, and M. J. McFall- win. 1988. Nucleotide sequence of the luxR
Ngai. 1996. Bioluminescence in the symbiotic and luxI genes and structure of the primary regula-
squid Euprymna scolopes is controlled by a daily bio- tory region of the lux operon of Vibrio fischeri
logical rhythm. J. Comp. Physiol. 179:65–73. ATCC7744. Biochemistry 27:837–842.
6. Bose, J. L., U. Kim,W. Batrtkowski, R. P. Gun- 19. Devine, J. H., G. S. Shadel, and T. O. Baldwin.
salus, A. M. Overley, N. L. Lyell, K. L.Visick, 1989. Identification of the operator of the lux reg-
and E.V. Stabb. 2007. Bioluminescence in Vibrio ulon from Vibrio fischeri ATCC 7744. Proc. Natl.
fischeri is controlled by the redox-responsive regu- Acad. Sci. USA 86:5688–5692.
lator ArcA. Mol. Microbiol. 65:538–553. 20. Dunlap, P.V. 1999. Quorum regulation of lumi-
7. Reference deleted. nescence in Vibrio fischeri. J. Mol. Microbiol. Biotech-
8. Callahan, S. M., and P.V. Dunlap. 2000. LuxR- nol. 1:5–12.
and acyl-homoserine-lactone-controlled non-lux 21. Dunn, A. K., D. S. Millikan, D. M. Adin, J.
genes define a quorum-sensing regulon in Vibrio L. Bose, and E. V. Stabb. 2006. New rfp- and
fischeri. J. Bacteriol. 182:2811–2822. pES213-derived tools for analyzing symbiotic
9. Chen, X., S. Schauder, N. Potier,A.Van Dors- Vibrio fischeri reveal patterns of infection and
selaer, I. Pelczer, B. L. Bassler, and F. M. lux expression in situ. Appl. Environ. Microbiol.
Hughson. 2002. Structural identification of a 72:802–810.
bacterial quorum-sensing signal containing boron. 22. Eberhard,A.,A. L. Burlingame, C. Eberhard,
Nature 415:545–549. G. L. Kenyon, K. H. Nealson, and, and N. J.
10. Chun, C. K., E.A. Ozer, M. J.Welsh, J. Zabner, Oppenheimer. 1981. Structural identification of
and E. P. Greenberg. 2004. Inactivation of a autoinducer of Photobacterium fischeri luciferase.
Pseudomonas aeruginosa quorum-sensing signal by Biochemistry 20:2444–2449.
human airway epithelia. Proc. Natl. Acad. Sci. USA 23. Egland, K. A., and E. P. Greenberg. 2000.
101:3587–3590. Conversion of the Vibrio fischeri transcrip-
11. Cloud, K. A., and J. P. Dillard. 2002. A lytic tional activator, LuxR, to a repressor. J. Bacteriol.
transglycosylase of Neisseria gonorrhoeae is involved 182:805–811.
in peptidoglycan-derived cytotoxin production. 24. Engebrecht, J., K. Nealson, and M. Silver-
Infect. Immun. 70:2752–2757. man. 1983. Bacterial bioluminescence: isolation
12. Cloud-Hansen, K. A., S. B. Peterson, E. V. and genetic analysis of functions from Vibrio fis-
Stabb,W. E. Goldman, M. J. McFall-Ngai, and cheri. Cell 32:773–781.
J. Handelsman. 2006. Breaching the great wall: 25. Engebrecht, J., and M. Silverman. 1984. Iden-
peptidoglycan and microbial interactions.Nat.Rev. tification of genes and gene products necessary for
Microbiol. 4:710–716. bacterial bioluminescence. Proc. Natl. Acad. Sci.
13. Cookson, B. T., A. N. Tyler, and W. E. Gold- USA 81:4154–4158.
man. 1989. Primary structure of the peptidogly- 26. Fidopiastis, P. M., C. M. Miyamoto, M. G.
can-derived tracheal cytotoxin of Bordetella Jobling, E. A. Meighen, and E. G. Ruby. 2002.
pertussis. Biochemistry 28:1744–1749. LitR, a new transcriptional activator in Vibrio fis-
14. Croxatto, A. 2006. VanT, a central regulator of cheri, regulates luminescence and symbiotic light
quorum sensing signalling in Vibrio anguillarum. organ colonization. Mol. Microbiol. 45:131–143.
Umea University, Umea, Sweden. 27. Fidopiastis, P. M., S. von Boletzky, and E. G.
15. Croxatto,A.,V. J. Chalker, J. Lauritz, J. Jass,A. Ruby. 1998.A new niche for Vibrio logei, the pre-
Hardman, P.Williams, M. Camara, and D. L. dominant light organ symbiont of squids in the
Milton. 2002.VanT, a homologue of Vibrio harveyi genus Sepiola. J. Bacteriol. 180:59–64.
28. Foster, J. S., M. A. Apicella, and M. J. McFall- 43. Henke, J. M., and B. L. Bassler. 2004. Three
Ngai. 2000. Vibrio fischeri lipopolysaccharide parallel quorum-sensing systems regulate
induces developmental apoptosis, but not com- gene expression in Vibrio harveyi. J. Bacteriol.
plete morphogenesis, of the Euprymna scolopes 186:6902–6914.
symbiotic light organ. Dev. Biol. 226:242–254. 44. Jones, B. W., and M. K. Nishiguchi. 2004.
29. Friedrich, W. F., and E. P. Greenberg. 1983. Counterillumination in the Hawaiian bobtail
Glucose repression of luminescence and luciferase squid, Euprymna scolopes Berry (Mollusca:-
in Vibrio fischeri.Arch. Microbiol. 134:87–91. Cephalopoda). Mar. Biol. 144:1151–1155.
30. Fuqua, C., M. R. Parsek, and E. P. Greenberg. 45. Koropatnick, T., J. R. Kimbell, and M. J.
2001. Regulation of gene expression by cell-to- McFall-Ngai. 2007. Responses of host hemo-
cell communication: acyl-homoserine lactone cytes during initiation of the squid-Vibrio symbio-
quorum sensing. Annu. Rev. Genet. 35:439–468. sis. Biol. Bull. 212:29–39.
31. Geszvain, K., and K. L.Visick. 2006. Roles of 46. Koropatnick,T.A., J.T. Engle, M.A.Apicella,
bacterial regulators in the symbiosis between Vibrio E.V. Stabb,W. E. Goldman, and M. J. McFall-
fischeri and Euprymna scolopes. Prog. Mol. Subcell. Ngai. 2004. Microbial factor-mediated develop-
Biol. 41:277–290. ment in a host-bacterial mutualism. Science
32. Gilson, L.,A. Kuo, and P.V. Dunlap. 1995.AinS 306:1186–1188.
and a new family of autoinducer synthesis pro- 47. Kulkarni, P. R., X. Cui, J. W. Williams, A. M.
teins. J. Bacteriol. 177:6946–6951. Stevens, and R.V. Kulkarni. 2006. Prediction of
33. Gonzalez, J. E., and M. M. Marketon. 2003. CsrA-regulating small RNAs in bacteria and their
Quorum sensing in nitrogen-fixing rhizobia. experimental verification in Vibrio fischeri. Nucleic
Microbiol. Mol. Biol. Rev. 67:574–592. Acids Res. 34:3361–3369.
34. Goodson, M. S., M. Kojadinovic, J.V.Troll,T. 48. Kuo, A., S. M. Callahan, and P. V. Dunlap.
E. Scheetz, T. L. Casavant, M. B. Soares, and 1996. Modulation of luminescence operon
M. J. McFall-Ngai. 2005. Identifying compo- expression by N-octanoyl-L-homoserine lactone
nents of the NF-kappaB pathway in the beneficial in ainS mutants of Vibrio fischeri. J. Bacteriol.
Euprymna scolopes-Vibrio fischeri light organ sym- 178:971–976.
biosis. Appl. Environ. Microbiol. 71:6934–6946. 49. Lee, K.-H., and, and E. G. Ruby. 1994.Effect of
35. Goryachev, A. B., D. J.Toh, K. B.Wee,T. Lee, the squid host on the abundance and distribution
H. B. Zhang, and L. H. Zhang. 2005.Transition of symbiotic Vibrio fischeri in nature. Appl. Environ.
to quorum sensing in an Agrobacterium population: Microbiol. 60:1565–1571.
a stochastic model. PLoS Comput Biol 1:e37. 50. Lee, K.-H., and E. G. Ruby. 1992. Detection of
36. Graf, J., P. V. Dunlap, and E. G. Ruby. 1994. the light organ symbiont, Vibrio fischeri, in Hawai-
Effect of transposon-induced motility mutations ian seawater by using lux gene probes. Appl. Envi-
on colonization of the host light organ by Vibrio ron. Microbiol. 58:942–947.
fischeri. J. Bacteriol. 176:6986–6991. 51. Lenz, D. H., M. B. Miller, J. Zhu, R.V. Kulka-
37. Graf, J., and E. G. Ruby. 1998. Host- rni, and B. L. Bassler. 2005. CsrA and three
derived amino acids support the proliferation redundant small RNAs regulate quorum sensing
of symbiotic bacteria. Proc. Natl. Acad. Sci. USA in Vibrio cholerae. Mol. Microbiol. 58:1186–1202.
95:1818–1822. 55. Lilley, B. N., and B. L. Bassler. 2000. Regula-
38. Gray, K. M., and E. P. Greenberg. 1992. Physi- tion of quorum sensing in Vibrio harveyi by LuxO
cal and functional maps of the luminescence gene and sigma-54. Mol. Microbiol. 36:940–954.
cluster in an autoinducer-deficient Vibrio fischeri 53. Liu, Z., F. R. Stirling, and J. Zhu. 2007.
strain isolated from a squid light organ. J. Bacteriol. Temporal quorum-sensing induction regulates
174:4384–4390. Vibrio cholerae biofilm architecture. Infect. Immun.
39. Greenberg, E. P. 1997. Quorum sensing in 75:122–126.
gram-negative bacteria. ASM News 63:371–377. 54. Loh, J., E.A. Pierson, L. S. Pierson, G. Stacey,
40. Hao, G., and T. J. Burr. 2006. Regulation of and A. Chatterjee. 2002. Quorum sensing in
long-chain N-acyl-homoserine lactones in plant-associated bacteria. Curr. Opin. Plant Biol.
Agrobacterium vitis. J. Bacteriol. 188:2173–2183. 5:285–290.
41. Haygood, M. G., and K. H. Nealson. 1985. 55. Lupp, C., and E. G. Ruby. 2004. Vibrio fischeri
Mechanisms of iron regulation of luminescence in LuxS and AinS: comparative study of two signal
Vibrio fischeri. J. Bacteriol. 162:209–216. synthases. J. Bacteriol. 186:3873–3881.
42. Henke, J. M., and B. L. Bassler. 2004. Quo- 56. Lupp, C., and E. G. Ruby. 2005. Vibrio fischeri
rum sensing regulates type III secretion in Vibrio uses two quorum-sensing systems for the regula-
harveyi and Vibrio parahaemolyticus. J. Bacteriol. tion of early and late colonization factors. J. Bacte-
186:3794–3805. riol. 187:3620–3629.
57. Lupp, C., M. Urbanowski, E. P. Greenberg, Hastings,L.J.Krick,and P.E.Stanley (ed.),Biolumi-

and E. G. Ruby. 2003.The Vibrio fischeri quorum- nescence and Chemiluminescence: Molecular Reporting
sensing systems ain and lux sequentially induce with Photons.Wiley and Sons, New York, NY.
luminescence gene expression and are important 69. Nyholm, S.V., B. Deplancke, H. R. Gaskins,
for persistence in the squid host. Mol. Microbiol. M. A. Apicella, and M. J. McFall-Ngai. 2002.
50:319–331. Roles of Vibrio fischeri and nonsymbiotic bacteria
58. Miller, S.T., K. B. Xavier, S. R. Campagna, M. in the dynamics of mucus secretion during sym-
E.Taga, M. F. Semmelhack, B. L. Bassler, and biont colonization of the Euprymna scolopes light
F. M. Hughson. 2004. Salmonella typhimurium organ. Appl. Environ. Microbiol. 68:5113–5122.
recognizes a chemically distinct form of the 70. Nyholm, S. V., and M. J. McFall-Ngai. 2003.
bacterial quorum-sensing signal AI-2. Mol. Cell. Dominance of Vibrio fischeri in secreted mucus
15:677–687. outside the light organ of Euprymna scolopes: the
59. Millikan, D. S., and E. G. Ruby. 2003. FlrA, first site of symbiont specificity. Appl. Environ.
a sigma54-dependent transcriptional activator Microbiol. 69:3932–3937.
in Vibrio fischeri, is required for motility and 71. Nyholm, S. V., and M. J. McFall-Ngai. 1998.
symbiotic light-organ colonization. J. Bacteriol. Sampling the light-organ microenvironment of
185:3547–3557. Euprymna scolopes: description of a population of
60. Milton, D. L. 2006. Quorum sensing in vibrios: host cells in association with the bacterial sym-
complexity for diversification. Int. J. Med. Microbiol. biont Vibrio fischeri. Biol. Bull. 195:89–97.
296:61–71. 72. Nyholm, S. V., and M. J. McFall-Ngai. 2004.
61. Miyamoto, C. M., P. V. Dunlap, E. G. Ruby, The winnowing: establishing the squid-Vibrio
and E. A. Meighen. 2003. LuxO controls luxR symbiosis. Nat. Rev. Microbiol. 2:632–642.
expression in Vibrio harveyi: evidence for a com- 73. Nyholm, S.V., E.V. Stabb, E. G. Ruby, and M.
mon regulatory mechanism in Vibrio. Mol. Micro- J. McFall-Ngai. 2000. Establishment of an ani-
biol. 48:537–548. mal-bacterial association: recruiting symbiotic
62. Mok, K. C., N. S.Wingreen, and B. L. Bassler. vibrios from the environment. Proc. Natl.Acad. Sci.
2003. Vibrio harveyi quorum sensing: a coincidence USA 97:10231–10235.
detector for two autoinducers controls gene 74. Pestova, E.V., L. S. Havarstein, and D.A. Mor-
expression. EMBO J. 22:870–881. rison. 1996. Regulation of competence for
63. Nealson, K., and A. Markovitz. 1970. Mutant genetic transformation in Streptococcus pneumoniae
analysis and enzyme subunit complementation in by an auto-induced peptide pheromone and a
bacterial bioluminescence in Photobacterium fis- two-component regulatory system. Mol. Microbiol.
cheri. J. Bacteriol. 104:300–312. 21:853–862.
64. Nealson, K. H. 1999. Early observations defining 75. Rosenthal, R. S. 1979. Release of soluble pepti-
quorum-dependent gene expression, p. 277–289. doglycan from growing gonococci: hexaminidase
In G. M. Dunny and S. C. Winans (ed.), Cell-Cell and amidase activities. Infect. Immun. 24:869–878.
Signaling in Bacteria. ASM Press,Washington, DC. 76. Ruby, E. G., and L. M.Asato. 1993.Growth and
65. Nealson, K. H., and, and J.W. Hastings. 1991. flagellation of Vibrio fischeri during initiation of the
The luminous bacteria, p. 1332–1345. In A. sepiolid squid light organ symbiosis. Arch. Micro-
Balows, H. G.Truper, M. Dworkin,W. Harder, and biol. 159:160–167.
K. H. Schleifer (ed.), The Prokaryotes, a Handbook 77. Ruby, E. G., and K. H. Lee. 1998. The Vibrio
on the Biology of Bacteria: Ecophysiology, Isolation, fischeri-Euprymna scolopes light organ association:
Identification,Applications, 2nd ed. Springer, Berlin, current ecological paradigms. Appl. Environ. Micro-
Germany. biol. 64:805–812.
66. Nealson, K. H., and J.W. Hastings. 1977. Low 78. Ruby, E. G., and J. G. Morin. 1979. Luminous
oxygen is optimal for luciferase synthesis in some enteric bacteria of marine fishes in a study of their
bacteria. Ecological implications. Arch. Microbiol. distribution, density and dispersion. Appl. Environ.
112:9–16. Microbiol. 38:406–411.
67. Nelson, E. J., P. M. Fidopiastis, and E. G. 79. Ruby, E. G., and K. H. Nealson. 1976. Symbi-
Ruby. 2007.A novel lux operon in the cryptically otic association of Photobacterium fischeri with the
bioluminescent fish pathogen Vibrio salmonicida is marine luminous fish Monocentris japonica; a model
associated with virulence. Appl. Environ. Microbiol. of symbiosis based on bacterial studies. Biol. Bull.
73:1825–1833. 151:574–586.
68. Nishiguchi, M. K., E. G. Ruby, and M. J. 80. Ruby, E. G., M. Urbanowski, J. Campbell, A.
McFall-Ngai. 1997. Phenotypic biolumines- Dunn, M. Faini, R. Gunsalus, P. Lostroh, C.
cence as an indicator of competitive dominance in Lupp, J. McCann, D. Millikan, A. Schaefer, E.
the Euprymna-Vibrio symbiosis, p. 123–126. In J.W. Stabb, A. Stevens, K.Visick, C. Whistler, and
E. P. Greenberg. 2005. Complete genome tion and biofilm formation in Vibrio cholerae.
sequence of Vibrio fischeri: a symbiotic bacterium Infect. Immun. 71:2571–2576.
with pathogenic congeners.Proc Natl Acad Sci USA 92. Visick, K. L., J. Foster, J. Doino, M. McFall-
102:3004–3009. Ngai, and E. G. Ruby. 2000. Vibrio fischeri lux
81. Schaefer, A. L., D. L.Val, B. L. Hanzelka, J. E. genes play an important role in colonization and
Cronan, Jr., and E. P. Greenberg. 1996. Gener- development of the host light organ. J. Bacteriol.
ation of cell-to-cell signals in quorum sensing: acyl 182:4578–4586.
homoserine lactone synthase activity of a purified 93. Visick, K. L., and M. J. McFall-Ngai. 2000.
Vibrio fischeri LuxI protein. Proc. Natl. Acad. Sci. An exclusive contract: specificity in the Vibrio
USA 93:9505–9509. fischeri-Euprymna scolopes partnership. J. Bacteriol.
82. Schuster, M., and E. P. Greenberg. 2006. A 182:1779–1787.
network of networks: quorum-sensing gene regu- 94. Visick, K. L., T. M. O’Shea, A. H. Klein, K.
lation in Pseudomonas aeruginosa. Int. J. Med. Micro- Geszvain, and A. J. Wolfe. 2007. The sugar
biol. 296:73–81. phosphotransferase system (PTS) of Vibrio fischeri
83. Seed, P. C., L. Passador, and B. H. Iglewski. inhibits both motility and bioluminescence. J.
1995. Pseudomonas aeruginosa lasI gene by LasR and Bacteriol. 189:2571–2574.
the Pseudomonas autoinducer PAI: an autoinduc- 95. Visick, K. L., and E. G. Ruby. 1999.The emer-
tion regulatory hierarchy.J.Bacteriol. 177:654–659. gent properties of quorum sensing: consequences
84. Shiner, E. K., K. P. Rumbaugh, and S. C. to bacteria of autoinducer signaling in their natu-
Williams. 2005. Interkingdom signaling: deci- ral environment, p. 333–352. In G. M. Dunny and
phering the language of acyl homoserine lactones. S. C.Winans (ed.), Cell-Cell Signaling in Bacteria.
FEMS Microbiol. Rev. 29:935–947. ASM Press,Washington, DC.
85. Sitnikov, D. M., J. B. Schineller, and T. O. 96. Visick, K. L., and E. G. Ruby. 2006. Vibrio fis-
Baldwin. 1995.Transcriptional regulation of bio- cheri and its host: it takes two to tango. Curr. Opin.
luminescence genes from Vibrio fischeri. Mol. Micro- Microbiol. 9:632–638.
biol. 17:801–812. 97. Walters, M., and V. Sperandio. 2006. Quorum
86. Sonnleitner, E., M. Schuster, T. Sorger- sensing in Escherichia coli and Salmonella. Int. J.
Domenigg, E. P. Greenberg, and U. Blasi. Med. Microbiol. 296:125–131.
2006. Hfq-dependent alterations of the tran- 98. Wang, L., J. Li, J. C. March, J. J. Valdes,
scriptome profile and effects on quorum and W. E. Bentley. 2005. luxS-dependent
sensing in Pseudomonas aeruginosa. Mol. Microbiol. gene regulation in Escherichia coli K-12 revealed
59:1542–1558. by genomic expression profiling. J. Bacteriol.
87. Stabb, E. V. 2006. The Vibrio fischeri-Euprymna 187:8350–8360.
scolopes light organ symbiosis, p. 204–218. In F. L. 99. Waters, C. M., and B. L. Bassler. 2005. Quo-
Thompson, B.Austin, and J. Swings (ed.), The Biol- rum sensing: cell-to-cell communication in bac-
ogy of Vibrios. ASM Press,Washington, DC. teria. Annu. Rev. Cell. Dev. Biol. 21:319–346.
88. Stabb, E.V., K. L.Visick, D. S. Millikan, A. A. 100. Wei, S. L., and R. E. Young. 1989. Develop-
Corcoran, L. Gilson, S. V. Nyholm, M. ment of symbiotic bacterial luminescence in a
J. McFall-Ngai, and E. G. Ruby. 2001. The nearshore cephalopod, Euprymna scolopes. Mar.
Vibrio fischeri-Euprymna scolopes symbiosis: a model Biol. 103:541–546.
marine animal-bacteria interaction, p. 269–277. In 101. Whistler, C. A., and E. G. Ruby. 2003.
N. Saxena (ed.), Recent Aadvances in Marine Sciences GacA regulates symbiotic colonization traits
and Technology, 2000. Pacon International, Hon- of Vibrio fischeri and facilitates a beneficial
olulu, HI. association with an animal host. J. Bacteriol.
89. Suzuki, K., X. Wang, T. Weilbacher, A. K. 185:7202–7212.
Pernestig, O. Melefors, D. Georgellis, P. Bab- 102. Wisniewski-Dye, F., and J. A. Downie. 2002.
itzke, and T. Romeo. 2002. Regulatory circuitry Quorum-sensing in Rhizobium.Antonie, Leeuwen-
of the CsrA/CsrB and BarA/UvrY systems of hoek. 81:397–407.
Escherichia coli. J. Bacteriol. 184:5130–5140. 103. Wolfe, A. J., D. S. Millikan, J. M. Campbell,
90. Taga, M. E., J. L. Semmelhack, and B. L. and K. L. Visick. 2004. Vibrio fischeri sigma54
Bassler. 2001.The LuxS-dependent autoinducer controls motility, biofilm formation, lumines-
AI-2 controls the expression of an ABC trans- cence, and colonization. Appl. Environ. Microbiol.
porter that functions in AI-2 uptake in Salmonella 70:2520–2524.
typhimurium. Mol Microbiol. 42:777–793. 104. Xavier, K. B., and B. L. Bassler. 2005.Regula-
91. Vance, R. E., J. Zhu, and J. J. Mekalanos. 2003. tion of uptake and processing of the quorum-
A constitutively active variant of the quorum- sensing autoinducer AI-2 in Escherichia coli. J.
sensing regulator LuxO affects protease produc Bacteriol. 187:238–248.
105. Yip, E. S., K. Geszvain, C. R. Deloney- colonization in Vibrio cholerae. Dev. Cell
Marino, and K. L.Visick. 2006.The symbiosis 5:647–656.
regulator RscS controls the syp gene locus, 107. Zhu, J., M. B. Miller, R. E.Vance, M. Dziej-
biofilm formation and symbiotic aggregation by man, B. L. Bassler, and J. J. Mekalanos. 2002.
Vibrio fischeri. Mol. Microbiol. 62:1586–1600. Quorum-sensing regulators control virulence
106. Zhu, J., and J. J. Mekalanos. 2003. gene expression in Vibrio cholerae. Proc. Natl.Acad.
Quorum sensing-dependent biofilms enhance Sci. USA 99:3129–3134.
ACYLATED HOMOSERINE
LACTONE SIGNALING IN
MARINE BACTERIAL SYSTEMS
Elisha M. Cicirelli, Holly Williamson, Karen Tait, and Clay Fuqua
16
The oceans are well recognized as the primary current understanding of AHL signaling
source of biological productivity on the planet. in marine bacterial systems outside of the
As with so many important processes, our well-described Vibrio fischeri and Vibrio harveyi
understanding of bacterial communication sys- models.
tems can also be traced back to roots in marine
systems. One of the first and best examples of GENERAL AHL-BASED
bacterial communication was that of Vibrio QUORUM-SENSING MODELS
species that direct symbiotic light production The fundamental model for AHL signaling was
when associated with the light organs of fish. originally described for luminescence (lux)
Studies on the regulation of light production gene regulation in V. fischeri, but the under-
during symbiotic associations with fishes and standing of this model and its consequences has
squids led to the concept of bacterial autoin- been greatly augmented by work in several
duction, to chemical characterization of acy- other systems (described in greater detail in
lated homoserine lactone (AHL) autoinducers, chapters 9,12,13, and 18). In its simplest form,
and molecular definition of the paradigmatic the core regulatory elements consist of an
LuxI-LuxR quorum-sensing system (see chap- enzyme that synthesizes the diffusible signal
ters 15 and 20). Studies on the marine vibrios molecule and a transcription factor that
have continued to play pivotal roles in deter- responds to the same molecule,activating target
mining the modes and mechanisms of bacterial functions in response to accumulation of the
communication systems. Despite the intense signal.The details of specific regulatory systems
scrutiny received by the marine vibrios, there in different bacteria embellish on this basic
has been, until recently, a relative paucity of model but tend to adhere to its basic tenets.
information on signaling mechanisms in other AHLs are typically synthesized by the activ-
marine bacteria. In this chapter, we review the ity of proteins similar to LuxI from V. fischeri,
utilizing S-adenosylmethionine (SAM) and
Elisha M. Cicirelli and Clay Fuqua Department of Biology, fatty acyl precursors conjugated to the acyl car-
Indiana University, Bloomington, Indiana 47405. Holly
Williamson and Karen Tait Plymouth Marine Laboratory, rier protein (ACP) (24).The homoserine moi-
Prospect Place, Plymouth, United Kingdom, PL1 3DH. ety of the AHL is largely derived from the
251
252 ■ CICIRELLI ET AL.
methionine in SAM, and this is conjugated to systems activate their target genes, although
the acyl chain donated by acyl-ACP. The length several have been found to function via AHL-
and reduction state of the acyl-ACP conjugate dependent derepression (54). Often, but not
determine the eventual chemistry of the AHL always, expression of the AHL synthase is itself
product.The specificity of each LuxI-type pro- activated, resulting in a positive feedback loop
tein for its acyl-ACP substrate(s) leads to the that can foster a uniformity in response and
diversity of AHL signals that are now recog- also desensitizes the system to subsequent
nized (26, 35).At least one additional family of fluctuations in AHL concentration.
proteins, those similar to LuxM from V. harveyi,
can also direct AHL synthesis, and indirect evi- SALIENT PARAMETERS FOR
dence suggests there may be even greater diver- AHL QUORUM SENSING IN
sity in AHL synthases (4, 29). Even so, the THE MARINE ENVIRONMENT
majority of bona fide AHL synthases that have The distribution of microbes in the marine
been identified are LuxI homologues. environment does not seem, at first approxima-
AHLs ranging from short to medium acyl tion, to lend itself to the dense microbial popu-
chain lengths (C4 to C8) are thought to transit lations typically required to activate AHL
the bacterial envelope via passive diffusion. quorum-sensing systems, such as those reached
There is evidence that longer-acyl-chain-length by V. fischeri in host animal light organs (5).The
AHLs (C10 to C20) partition into lipid bilayers animals that harbor these symbionts promote
and thus require membrane transport systems to the growth of dense bacterial monocultures that
effectively cross the bacterial envelope (23). At stand in contrast to the generally low microbial
low population density or in environments with densities in bulk seawater (approximately 106
rapid mixing or flow, each cell produces AHL at cells ml of seawater1). On a smaller scale, how-
a basal level that exits the cell, progressing down ever, marine microbial densities can be quite
its diffusion gradient.The resulting residual lev- high, such as when they colonize specific sites
els of AHL are insufficient to drive interactions on particulate organic matter (27). Even at
with AHL receptors. The relative AHL levels larger scales, microbial population explosions
increase as a function of local population density can occur in response to large releases of organic
or due to alterations in the diffusive environ- material, such as during algal blooms, resulting
ment. LuxR-type transcriptional regulators are in far greater than average densities (41). Finally,
the most common AHL receptors and are local- the vast abundance of plant and animal life in
ized to the cytoplasm (73). In contrast, for sev- the ocean provides a wide range of host-specific
eral of the marine vibrios, two-component niches in which different microbes can reach
sensor kinases can mediate response to AHLs, high local densities.It is therefore not surprising
resulting in transcriptional control via a multi- that further investigation has revealed frequent
step regulatory cascade (3, 4, 32). By either and pervasive AHL production among diverse
mechanism, AHL accumulation is monitored, marine bacteria (27, 79).
and above a certain threshold concentration, in The chemistry of the marine environment
many cases thought to represent a microbial imposes certain constraints onto the process of
quorum, the AHL response ensues.The precise AHL quorum sensing. AHLs can be highly
AHL concentration required to activate each base-labile,particularly those with 3-oxo groups
system can vary widely and appears to be cali- and those with very short acyl side chains (12).
brated for the specific microbe and the environ- The average pH of oceanic surface waters is 8.2
mental context of the regulatory response.The (69). Many marine bacteria produce and
target(s) of the AHL response is typically tran- employ AHLs, suggesting that these signal mol-
scription of specific sets of genes, varying in ecules, despite their pH lability, can function
number and complexity depending on the spe- effectively under these conditions. At a pH of
cific bacterial species. Many AHL-responsive 8.2 the half-life of a 3-oxo-AHL will be fairly
16. AHL SIGNALING IN MARINE BACTERIAL SYSTEMS ■ 253
short, effectively reducing the “calling distance” the well-studied model systems.AHLs are thus
of this signal molecule (25). Tait et al. investi- far restricted to the proteobacteria, and one
gated the stability of AHLs in seawater at differ- would predict that AHL regulation would be
ent temperatures and found a marked increase most significant for environments in which
in hydrolysis with increasing temperature, with these microbes are indigenous. Given the per-
most hydrolysis occurring above 22C (75). If vasiveness of the proteobacteria, however (71),
we, however, consider the micrometer scale at this may be less of a limitation than those
which the bacterial communities in the ocean imposed by the chemical and physical condi-
typically function, perhaps AHLs in marine sys- tions specific to each zone of the ocean.
tems are strictly employed over short calling dis-
tances. It is interesting to consider that under PROTEOBACTERIAL DIVERSITY
certain conditions, AHLs could function as IN MARINE ECOSYSTEMS
effective pH indicators,similar to their proposed AHL production has thus far only been docu-
alternative utilization as diffusion sensors (67). mented in proteobacterial groups. Representa-
Recent global climate trends have been low- tives of these taxa, however, constitute one of
ering the average pH of the oceans through the most numerous and functionally diverse
partitioning of carbon emissions into seawater, classes of microorganisms, and they are parti-
driving oceanic carbonate pools to carbonic cularly abundant in marine environments.
acid (60). Many important oceanic processes, Metagenomic surveys of marine surface waters
such as calcification in the exoskeletons of cer- suggest that 63% of the microbial ribosomal
tain highly numerous zooplankton, are thought RNA sequences are proteobacterial,dominated
to be threatened by this oceanic acidification. by the alpha and gamma subgroups (71). Deep-
It remains unclear how the oceanic biota, and sea hydrothermal vent communities are like-
the global processes they drive, will respond to wise rich in proteobacteria, in this case
this change in ambient pH. This drop in pH dominated by those of the delta subgroup (37).
may also affect the half-life of AHL signals in Some of these bacteria numerically dominate
the environment, increasing their calling dis- their environment such as with marine alpha-
tance. It is possible that this change will inter- proteobacteria of the SAR11 clade, which can
fere with or otherwise alter microbe-microbe comprise from one-third to one-half of the bac-
and microbe-host interactions. The conse- terial biomass in marine surface waters (58).
quences could conceivably include emergence Members of the Roseobacter clade of alpha-
of new disease agents and degeneration of proteobacteria are also very prevalent in marine
key existing syntrophic interactions. On the environments and are estimated to comprise 20
other end of the spectrum,it is also possible that to 30% of the 16S rRNA gene sequences in the
given the time scale of this acidification, the photic zone (10, 78). Our understanding of the
microbial populations will simply adapt to complex and dynamic relationships that exist
maintain a dynamic balance similar to those among marine bacteria and their environment
under current conditions. is at a rudimentary level.Although many marine
The oceans also present high salinity and proteobacteria are free-living, others are found
extreme conditions, including high pressure associated with eukaryotic organisms such as
(benthic), low and high temperature (hydro- alga, sponges, dinoflagellates, and corals (10, 57).
thermal vents), and intense UV irradiation
(pelagic surface waters).All of these conditions AHL PRODUCTION AMONG MARINE
can significantly influence the dynamics of sig- PROTEOBACTERIAL GROUPS
nal production, stability, and perception. It is There are greater than 50 different proteobac-
clear that some oceanic zones are highly con- terial species recognized to produce and
ducive to the deployment of AHLs as diffusible respond to AHLs, predominantly from terres-
signal molecules, in a manner consistent with trial and aquatic environments (43, 81). Studies
using mutants of luminescent marine vibrios of the AHL strains did not form biofilms
that respond to exogenous AHLs suggested that or rosettes, the correlation with AHL synthesis
signal production might be broadly distributed was imperfect. A Phaeobacter isolate strongly
in the marine environment (2, 28).These stud- produced AHLs and also formed biofilms
ies focused predominantly on Vibrio species and rosettes. This Phaeobacter strain, originally
and related gamma-proteobacteria. More isolated from a turbot larval rearing unit,
recent survey efforts have revealed similar was shown to produce 3-hydroxy-decanoyl
widespread AHL production by additional homoserine lactone (3-OH-C10-HSL) (9).
marine proteobacterial taxa. In these studies, Bacteria that were isolated from a variety of
AHL-responsive biosensor strains have been sites including the North Sea,the German Wad-
commonly employed to detect synthesis of the den Sea, and the hypersaline Ekho Lake in
AHLs. These AHL biosensors are bacterial Antarctica were assayed for AHL production
derivatives that do not produce their own AHLs and phylogenetically ribotyped using their 16S-
but respond to exogenous signals via a LuxR- rRNA sequences (79).The vast majority of the
type protein, which activates expression of an isolates analyzed were proteobacteria of the
easily measurable activity such as -galactosi- alpha (64%) and gamma (14%) subgroups, and
dase, green fluorescent protein (GFP), biolumi- the remaining isolates were composed of bac-
nescence, or pigment production. teroidetes, actinobacteria, and firmicutes. Two
Gram et al. reasoned that one likely site in different,GFP-based AHL biosensor strains,one
the marine environment for high-density optimally responsive to 3-oxo-C12-HSL and
microbial colonization is on suspended parti- the other specific for short-chain AHLs, were
culate organic carbon, or marine snow (27). employed to screen for AHL production. In a
This study screened bacterial strains isolated useful variation on the AHL bioassays, Amber-
from marine snow for AHL production using lite XAD-16 resin was added to each culture for
biosensors derived from Agrobacterium tumefa- several hours before extraction in order to bind
ciens and Chromobacterium violaceum, responsive and stabilize AHLs that were produced over
to medium-chain-length and fully reduced, time. Use of this resin resulted in up to 50-fold
short-chain-length compounds, respectively yield improvements in extracted activity over
(15).Four out of 22 proteobacterial isolates acti- analysis of culture supernatants alone.
vated the A. tumefaciens reporter. None of these The Amberlite extracts were used in both
strains activated the C. violaceum biosensor or an bioassays and gas chromatography-mass spec-
additional system based on LuxR from V.fischeri. trometry (GC-MS) analysis.One or both of the
Of the four AHL strains, three belonged to the biosensors were induced by greater than
large Roseobacter clade, a marine subgroup of twofold by 41 of the 102 tested isolates (79).
the family Rhodobacteraceae within the alpha- Strikingly, 59% of the alpha-proteobacterial
proteobacteria (Fig. 1). In this case, these isolates isolates significantly activated one or both
were most similar to species of Sulfitobacter and biosensor systems. Additionally, two gamma-
Roseobacter (10).The additional AHL strain was proteobacteria, related to Glaciecoloa polaris and
a gamma-proteobacteria,likely to be a species of Pseudoalteromonas atlantica, were also judged to
Marinobacter (27). Four more Roseobacter iso- produce AHLs. As with the previous studies,
lates were also analyzed, and two of these members of the Roseobacter group dominated
induced a response in the A. tumefaciens biosen- the AHL alpha-proteobacteria. All 10 Roseo-
sor. In a separate study, 14 cultivated roseobac- varius isolates activated the long-chain reporter.
ters were screened for multicellular activities Bacteria most similar to Dinoroseobacter shibae,
such as surface attachment, rosette formation, Jannaschia helgolandensis, and Roseovarius mucosus
antibiotic synthesis and sensitivity, and AHL consistently activated either one or both
production (8). Five of these cultivated strains reporters. However, among other members of
were found to produce AHLs. Although many the Roseobacter group, AHL production var-
FIGURE 1 AHL production in the Roseobacter lineage. Phylogram of bacteria within the Roseobacter group
adapted with permission from Buchan et al. (10). Robust phylogenetic lineages are indicated with filled ovals at
branch nodes and vertical black lines.The numbers of clone and isolate sequences representing each cluster are pro-
vided in brackets.The tree was constructed using a neighbor-joining method.The bar represents Jukes-Cantor evo-
lutionary distances. Bootstrap values of greater than 50% are shown at branch nodes (100 iterations). AHL
indicates that representatives of these isolates and subgroups have been reported to produce AHL activities.
255
ied considerably. As an example, Roseobacter prising variation in the types and abundance of
litoralis strongly induced the long-chain biosen- signals produced. Many appeared to produce
sor; however, its close relatives, Roseobacter deni- long-chain, nonpolar AHLs, similar to the
trificans and Staleya guttiformis, did not activate free-living roseobacters (79). Interestingly, sam-
either system (79). ples from the bulk seawater at the collection site
In parallel with these bioassays, GC-MS identified AHL bacteria, not roseobacters, but
analysis of organically extracted, Amberlite- rather, marine species of Erythrobacter, a distinct
resin-enriched material was performed for group of alpha-proteobacteria (55). A small
each isolate. Of the 41 isolates determined number of SR-type roseobacters were isolated
to be AHL with the biosensors, 22 yielded from the bulk seawater but were not positive
AHLs that were identifiable by GC-MS (79). for AHL production.
A mixture of AHLs with acyl chains ranging Collectively,these survey studies suggest that
from 14 to 18 carbons, many with unsaturated despite the original identification of AHL quo-
bonds, were the most prevalent among the rum sensing in the gamma-proteobacterial
compounds identified. Some shorter-chain species of Vibrio, it is in fact bacteria from the
AHLs, most notably octanoyl derivatives (C8- Roseobacter group that may be the numeri-
HSL), were also observed. Several phylogenetically dominant AHL producers in marine sys-
cally distinct bacteria produced identical AHL tems. Even with these dominant cultivatable
profiles, and conversely, some closely related groups, the breadth of diversity for the marine
strains manifested distinctly different patterns proteobacterial taxa that synthesize AHLs is
of AHL synthesis. roughly equivalent to that observed for terres-
Marine animal, plant, and protozoan hosts trial and freshwater systems (15, 22). It is also
can foster dense populations of commensal very important to note that all of these studies
bacteria. Several of the Roseobacter-type bac- focused, for obvious practical reasons, on culti-
teria discussed above were isolated in associa- vatable bacteria. It is possible, and perhaps even
tion with dinoflagellate hosts (8). One study likely, that a significant fraction of uncultivated
specifically evaluated AHL production in the proteobacteria also produce AHL signals in
tissue samples from marine invertebrates and marine systems. Sensitive AHL biosensor sys-
marine plants (76). The signal molecules were tems were employed in almost all of these stud-
detected in the tissues from a large fraction of ies, and these have well-established biases for
sponges (77%) and corals (50%). Further analy- their cognate AHLs (61, 72, 83). Consequently,
sis of several different sponge taxa revealed bacteria that test negative for AHL biosynthesis
widespread AHL production, and subsequently, with these biosensors may in fact be synthesiz-
AHL-producing bacteria were isolated. One of ing AHLs that are outside the effective range of
these was a species of Vibrio that synthesized N- the biosensors employed. Direct chemical
hexanoyl-homoserine lactone (C6-HSL) and analyses, such as those employed by Wagner-
also N-(3-oxo)-hexanoyl-homoserine lactone Dobler et al. (79), do not have these intrinsic
(3-oxo-C6-HSL). Another isolate was identi- biases but are currently at least 10-fold less sen-
fied as an alpha-proteobacterium, a member of sitive than the best biosensors.Finally,AHL syn-
the Silicibacter-Ruegeria (SR) subgroup of the thesis is often strongly regulated by other
Roseobacter clade (Fig. 1). Recent compre- environmental conditions, and these activating
hensive bacterial isolation from two species conditions may not be recapitulated in standard
of shallow reef sponges, and screens for AHL laboratory culture.The surveys discussed above
production, found the SR-type roseobacters to should therefore be considered as conservative
be the dominant, cultivatable AHL producers, estimates of AHL production capacity among
along with a small number of Vibrio species cultivatable marine bacteria and should be inte-
(55). For the roseobacters isolated from these grated with emerging genomic information on
sponges, 100% produced AHLs, but with sur- marine bacteria.
AHL SIGNALING IN SILICIBACTER gral role in sulfur cycling in the marine envi-
SPECIES AND RELATED ISOLATES ronment and through this can have large-scale
As described above, the Roseobacteria are one impacts on oceanic processes such as cloud
of the dominant microbial groups in the ocean, formation (36). The DSS-3 sequence revealed
and several subgroups are also the most com- two pairs of LuxIR-type regulators, designated
mon AHL signal producers. Of these AHL- SilI1/SilR1 and SilI2/SilR2 (Table 1 and Fig.
producing microbes, Silicibacter pomeroyi DSS-3 2). Genes encoding both AHL synthases, silI1
is a free-living representative and was the first and silI2, direct the synthesis of AHLs when
roseobacter to have its genome completely they are expressed in Escherichia coli. Expression
sequenced (57).S.pomeroyi DSS-3 plays an inte- of silI1 results in an AHL profile that closely
TABLE 1 LuxI-LuxR homologues in the Roseobacteria

Local gene
Species/isolate LuxI homologuesa LuxR homologuesb organizationc
Silicibacter pomeroyi DSS-3 SPO2287 (SilI1/SsaI) Linked, SPO2286 (SilR1/SsaR) A
(and SR sponge isolates) SPO0372 (SilI2/SsbI) Linked, SPO0371 (SilR2/SsbR) B
Silicibacter sp. strain TM1040 No homologues TM1040_1212,TM1040_3102 NA
Jannaschia sp. strain CCS1 Jann_0620 Linked, Jann_0619 D
Sulfitobacter sp. strain EE-36 EE36_01635, (RhlL-type) Unlinked I
Sulfitobacter sp. strain NAS-14.1 NAS141_00695, (RhlL-type) Unlinked I
NAS141_01136 Linked, NAS141_01141 F
Roseovarius nubinhibens ISM ISM_03755 (RhlL-type) Unlinked J
Roseovarius sp. strain 217 ROS217_18267 (SilI1-type) Linked, ROS217_18272 A
(SilR1-type)
ROS217_01410 (RhlL-type) Linked, ROS217_01415 H
Roseovarius sp. strain HTCC 2601 No homologues No homologues NA
Dinoroseobacter shibae DFL12 DshiDraft_1388 (SilI2-type) Linked, DshiDraft_1389 B
(SilR2-type)
DshiDraft_2384 (Jann_0620- Linked, DshiDraft_2383 D
type) (Jann_0619-type)
Rhodobacteriales bacterium RB2654_09014 (SilI2-type) Linked, RB2654_09024 C
HTCC 2654 (SilR2-type)
RB2654_20048 (RhlL-type) Linked, RB2654_20053 G
Saggitula stellata E-37 SSE37_11164 (SilI2-type) Linked, SSE37_11169 B
(SilR2-type)
Roseobacter denitrificans Och114 RD1_1639 (SilI2-type) Linked, RD1_1638 (SilR2-type) B
Roseobacter sp. strain MED193 MED193_10423 (SilI2-type) Linked, MED193_10428 (SilR2- B
type)
MED193_08053 (RhlL-type) Unlinked K
Roseobacter sp. strain CCS2 RCCS2_02078 Multiple linked R genes, E
RCCS2_02083, RCCS2_02088
Loktanella vestfoldensis SKA53 SKA53_05839 (RCCS2-type) Multiple linked R genes (like E
RCCS2)
SKA53_05835, SKA53_05840
Oceanicola granulosus HTCC 2516 OG2516_07121 (RCCS2-type) Multiple linked R genes (like E
RCCS2)
OG2516_07116, OG2516_07111
Oceanicola batensis HTCC 2597 No homologues Imperfect homologue, NA
OB2597_03302
a End-to-end homologues of LuxI AHL synthase. Gene/protein code for Roseobase (www.roseobase.org) access is provided.
Parenthetical information gives general subclass of LuxI-type protein.

b End-to-end homologues of LuxR transcription regulator. Protein code for Roseobase access is provided. Parenthetical information
gives general subclass of LuxR-type protein. Indicated if the corresponding gene is genetically linked or unlinked to a luxI homologue.
c Local genetic organization.The letter code corresponds to Figure 2.
FIGURE 2 Genetic organization and context of roseobacterial LuxI-LuxR systems. LuxI
homologues are dark gray arrows, LuxR homologues are light gray arrows, and all flanking genes
are black arrows. In most cases, labels above genes indicate a reference homologue or, in the cases of
certain sequences, the genetic ID from Roseobase (www.roseobase.org). Maps are drawn roughly
to scale.Abbreviations: CoA, crotonyl CoA reductase; DBP, DNA-binding protein; HK, histidine
kinase; HK/RR, hybrid histidine kinase/response regulator; HK/PAS, histidine kinase with PAS
domain; Hyp, hypothetical protein; RND, resistance, nodulation, cell division multidrug efflux
pump homologue; RR, response regulator; SigB, sigma B.
258
resembles what is synthesized by the S. pomeroyi sequences reveals that AHL-based quorum
parental strain, whereas expression of silI2 pro- sensing is widely distributed, pervasive, and
duces AHLs that appear to be distinct from complex within this diverse group. Convincing
those produced by laboratory cultures of DSS- LuxI homologues are present in 14 roseobac-
3 (57). Interestingly, Silicibacter sp. TM1040, a ters, with 6 genomes encoding a pair of discrete
related subspecies associated with dinoflagel- homologues, totaling 19 LuxI-type proteins
lates, does not synthesize AHLs nor does its (Table 1). In general, all of these LuxI-type pro-
genome sequence encode any LuxI- or LuxM- teins are more closely related to other LuxI-
type AHL synthase proteins (8). type proteins from within the Roseobacteria
Several closely related members of the SR than to homologues outside this group. In 15
subgroup within the Roseobacter clade and out of 19 cases, the LuxI homologue is physi-
affiliated with S. pomeroyi DSS-3 were isolated cally linked to one or more luxR-type tran-
from marine sponges and found to be AHL scription factor genes (Fig. 2 and Table 1).There
producers (55). Although all of these bacteria are several recurring types of LuxI homologues
synthesized AHLs, several classes could be dis- within this set of genome sequences. These
tinguished by the spectrum of AHLs produced. orthologues are identifiable not only by pri-
Two luxRI regulatory genes were isolated from mary sequence homology but also often by the
a representative of these SR sponge isolates genes that flank them (Fig. 2).The SilI2/SsbI-
using a genetic screen (E. M. Cicirelli, N. type AHL synthases and SilR2/SsbR-type tran-
Mohamed, J. Herman, M. E.A. Churchill, R.T. scription factors are encoded within 6 of the 14
Hill, and C. Fuqua, unpublished data). These genomes (Fig. 2B). Only two genomes, S.
genes were designated ssaI/R and ssbI/R and pomeroyi DSS-3 itself and Roseovarius sp. 217,
share high identity and synteny with silI1/ harbor a SilI1/SsaI-SilR1/SsaR regulatory pair.
R1 and silI2/R2, respectively (Fig. 2). Ortho- There are three genomes (we have designated as
logues of these genes were also amplified RCCS2-type) in which two separate luxR-type
by PCR from other SR bacteria associated genes are immediately linked to an AHL syn-
with sponges. Expression of the ssaI/R and thase gene (e.g., RCCS2_02078), which is itself
ssbI/R systems in E.coli resulted in AHL synthe- distinct in primary sequence from SilI1 and
sis, consistent with the findings from DSS-3. SilI2 (Fig. 2E). Two genomes, Jannaschia sp.
Although null mutants in the ssaIR and CCS1 and D. shibae DFL12, also have their own
ssbIR genes are affected for synthesis of a distinct luxR-luxI locus (defined as Jann_0620-
significant number of proteins as assessed by type, Fig. 2D), although D. shibae also encodes a
two-dimensional gel electrophoresis,the nature SilR2-type system (Table 1 and Fig. 2B). Five
of these target functions is still under investiga- genomes encode a LuxI-type protein that has
tion. Preliminary findings suggest that the ssaIR been annotated as RhlL (Fig. 2G to 2K).
system may influence motility of the sponge Although the primary sequence is conserved
symbionts (Cicirelli et al., unpublished). among this group,the genetic context and loca-
tion are highly variable,with three clusters lack-
LuxI-LuxR HOMOLOGUES ing a linked luxR homologue and four different
IN THE ROSEOBACTERIA flanking gene configurations.Three roseobacter
Genomic sequencing has been performed genomes, Silicibacter sp. TM1040, Oceanocola
on a growing number of marine bacteria batensis,and Roseovarius sp.HTCC 2601,encode
through support from the Moore Foundation no LuxI-type proteins. In all three cases, other
(http://www.moore.org/marine-micro.aspx). species of these same genera do encode LuxI-
As a subset of this wider effort, 17 genome type proteins.
sequences, representing 10 genera, are publicly An attempt to map the pattern of AHL sig-
available from within the Roseobacter group naling systems onto the phylogeny of the
(www.roseobase.org).Probing these 17 genome Roseobacteria does not reveal a readily inter-
preted geneology for these systems among this also increasingly implicated in human intestinal
group (56). Closely related species, such as Sul- and extraintestinal infections. Antibiotic treat-
fitobacter sp.EE-36 and Sulfitobacter sp.NAS14.1, ments are currently in use for these infections,
can share highly conserved systems such as the but the development of antibiotic-resistant
RhlL orphan AHL synthase homologue linked populations has led to an increased interest in
to a TetR-type repressor (Fig. 2I). Another alternative methods of control (19). Many plant
example is the conserved LuxI homologue and animal pathogens regulate virulence deter-
paired with two different luxR-type genes (Fig. minants via quorum sensing. Studies of the
2E). In this case, three highly divergent and quorum sensing pathways in V. anguillarum, A.
phylogenetically separated genera, Roseobacteria hydrophila, and A. salmonicida have been stimu-
sp. RCCS2, Loktanella vestfoldensis SKA53, and lated by an interest in the correlation between
Oceanicola granulosus HTCC 2516, share a signaling and pathogenesis.
highly conserved LuxI, two conserved LuxR
proteins, and an identical genomic context Aeromonas spp.
(Table 1 and Fig. 2E). In contrast, O. batensis Aeromonas spp. possess a range of virulence
HTCC 2597, highly similar to O. granulosus determinants, including -hemolysin, glycero-
HTCC 2516, completely lacks this regulatory phospholipid-cholesterol acyltransferase, lipase,
locus (56). Perhaps most striking, Silicibacter sp. and proteases. Expression of several of these
TM1040 entirely lacks both the SilI1/R1 and excreted products in laboratory culture is
the SilI2/R2 loci,identified in S.pomeroyi DSS- associated with late exponential and stationary-
3 (Table 1).The processes of gene duplication, phase growth. Several protease activities are
acquisition, and loss among the Roseobacteria regulated by the LuxRI homologues AhyRI
are readily apparent in these comparisons.The and AsaRI and their cognate AHLs in A.
complexity and diversity of the roseobacteral hydrophila and A. salmonicida, respectively (74).
AHL signaling systems rival that of the terres- Mutations to the ahyI and ahyR genes in A.
trial alpha-proteobacteria in the Rhizobiaceae salmonicida virtually abolished serine and metal-
family (30). Determining the functionalities of loprotease activities, and in the ahyI mutant, this
these pervasive signaling systems in the was rescued by exogenous butryl homoserine
Roseobacteria should provide important lactone (C4-HSL) (74). A. hydrophila primarily
insights into the evolution of signaling in produces C4-HSL via the AHL synthase AhyI.
marine microbes and the selective pressures to On the basis of several lines of evidence, the
which they respond. ahyR gene product is reported to function
as both a negative and positive regulator of
QUORUM-SENSING SYSTEMS ahyI, controlling C4-HSL synthesis in a growth
IN FISH PATHOGENS phase-dependent manner (40). A. hydrophila
In marine systems gamma-proteobacteria of forms thick, architecturally complex biofilms
the genera Vibrio and Aeromonas are common interspersed with microcolonies, a process
AHL producers and are fairly well studied. that is under quorum-sensing control. Biofilms
In addition to better characterized human of the ahyI mutant failed to form microcolonies,
pathogens, several Aeromonas spp. and Vibrio whereas those of the ahyR mutant occupied
spp. cause diseases of fish. Vibrio anguillarum is a notably greater proportion of the surface area
the cause of hemorrhagic septicemia in fish and (44). This suggests that AhyR acts to limit
has a massive impact on the aquaculture indus- growth on surfaces and that C4-HSL control
try (51). Similarly, Aeromonas salmonicida is the of this activity promotes formation of micro-
causative agent of furunculosis in salmonoid colonies. It is unclear why these mutants
fish, whereas Aeromonas hydrophila is responsible are effectively identical for virulence factor syn-
for motile aeromonad septicemia.These agents thesis but so distinct in regards to their biofilm
are significant problems in aquaculture and are phenotypes.
V. anguillarum L-homoserine lactone (3-oxo- C10-HSL) (53).

Within many Vibrio species, including V. Two additional shorter-chain AHLs, N-
harveyi, V. cholerae, V. vulnificus, and V. fischeri, hexanoyl-L-homoserine lactone (C6-HSL) and
quorum sensing is regulated by multiple, over- N-(3-hydroxyhexanoyl)-L-homoserine lac-
lapping signaling systems (51). The quorum- tone (3-OH-C6-HSL), are synthesized by the
sensing system of V. anguillarum is similar to VanM protein, a homologue of LuxM from V.
those described in the other Vibrio spp., in that harveyi (52). Response to these signals requires
homologues to quorum-sensing regulators in the two-component sensor kinase VanM, again
V. harveyi appear to account for most of these homologous to the LuxN pathway of V. harveyi.
functions. However, certain aspects of the V. Analysis of laboratory cultures revealed 3-oxo-
anguillarum system differ significantly from the C10-HSL to be the predominant AHL, whereas
V. harveyi paradigm (18). 3-OH-C6-HSL was dominant in the AHL
Three quorum-sensing circuits have been profiles of extracts from fish infected with V.
described for V. anguillarum, and a fourth was anguillarum (11).
recently suggested (Fig. 3) (51).VanI and VanR A third signaling circuit involves VanS,
are homologues of LuxIR and function which is thought to direct synthesis of the
through their cognate AHL N-3-oxodecanoyl- autoinducer-2 (AI-2)-type signal (Fig. 3). The
FIGURE 3 Model for quorum-sensing regulation in V. anguillarum. Depicts current understanding of quorum-
sensing mechanisms in V. anguillarum relative to the cell exterior and interior.
signal is presumptively a furanosyl borate (17). Interestingly, the VanS enzyme is required
diester, 3A-methyl-5,6-dihydro-furo[2,3-D] for virulence in a model infection of brine
[1,3,2] dioxaborole-2,2,6,6A-tetranol (18). shrimp, and this is rescued by addition of AI-2,
This is perceived through interaction with the whereas the vanM gene is dispensable in this
periplasmic binding protein VanP and the model system (19). This is a curious observa-
membrane-associated sensor kinase VanQ, tion, given that the current model for the VanM
which initiates a phosphorelay cascade through and VanS pathways converges at VanU to con-
a histidine-containing phosphotransfer (Hpt) trol target functions,and suggests that these two
protein,VanU. The AI-2/VanPQ pathway and pathways can function somewhat independ-
the 3-OH-C6-HSL/VanN pathway converge ently. AHLs are implicated in virulence, how-
on the VanU histidine-containing phospho- ever, by the observation that a furanone-type
transfer protein, subsequently directing phos- inhibitor (designated C30), known to impact
photransfer to theVanO response regulator (Fig. perception of AHLs by LuxR-type proteins,
3). Phospho-VanO is likely to be a transcrip- reduced vibriosis-linked mortality in rainbow
tional activator that indirectly represses VanT, a trout (64). The role of the CqsA signal in V.
homologue of the V. harveyi LuxR transcription anguillarum is not yet clear.
activator (18).In V.harveyi theVanO homologue Although the V. anguillarum quorum-sensing
LuxO activates expression of several small quo- systems show similarities to those from the
rum sensing regulatory RNAs that control other species of Vibrio, there are also several dif-
translation of the mRNA encoding the LuxR ferences. Most often, at low population densi-
transcription activator (42). It is not yet clear if ties, homologues of VanO completely repress
VanT is similarly controlled in V. anguillarum. the VanT homologue. In V. anguillarum, VanO
The fourth predicted signaling circuit, based diminishes but does not completely repress
on the V. anguillarum genome sequence, is simi- vanT expression.VanT is thought to repress its
lar to the CqsA/S circuit found in V. cholerae own expression but also to activate the vanOU
(32). In this system the gene cqsA is required for promoter, an unusual regulatory pattern among
the production of an autoinducer molecule the vibrios (18). The overall effect of the
known as CAI-1 (not yet chemically defined) VanM/N and VanS/PQ cascades appears to be
and cqsS for its detection (Fig. 3). Preliminary through inhibition of VanT activity.This com-
evidence suggests that CAI-1 may be involved plex upstream pathway allows multiple layers of
in communication between species of Vibrio, regulation to integrate at VanT, modulating the
indicated by the response of both V. harveyi and control of its regulon (31).
V. cholerae to the same CAI-1 preparation.The
role of this system may therefore be interspecies ULVA ZOOSPORES UTILIZE
communication, although there is as yet no AHL SIGNAL MOLECULES
evidence to support this for V. anguillarum (32). FOR SURFACE SELECTION
The relationships between the Van I/R,Van Most studies of quorum sensing have examined
M/N,Van S/PQ, and CqsA signaling pathways cell-cell communication within single bacterial
are not yet understood.Inactivation of the AHL species, although a growing number of studies
synthase vanM abolishes production of all three have reported signaling between different bac-
AHLs, whereas a vanI mutation abolishes pro- terial species (1, 63, 70). An emerging area
duction of only 3-oxo-C10-HSL (52). This of bacterial signaling is communication
suggests a hierarchy of VanM/N upstream of between prokaryotes and eukaryotes, including
VanI/R.The genes regulated by VanR are not microbes, fungi, plants, and mammals via quo-
known,butVanT,the target of theVanM/N sys- rum-sensing signals (16, 48, 77). The marine
tem, positively regulates virulence-related phe- environment provides ample opportunities for
notypes such as biofilm formation, extracellular such interactions, and consequently several
protease activity, and siderophore production such cross-kingdom signaling examples have
been documented in marine systems.A partic- to influence Ulva zoospores during their
ularly compelling version of cross-kingdom surface-selection (Fig. 4), including negative
signaling is found for the interactions between phototaxis, surface chemistry and wettability,
marine bacteria residing in biofilms and unicel- surface topography (13, 14), and also the
lular zoospores of the green macroalgae Ulva. presence of a bacterial biofilm (38). Image
Ulva is the most common macroalga con- analysis revealed a positive correlation between
tributing to biofouling of man-made surfaces zoospore settlement and the presence of a
throughout the world, mainly due to the vast bacterial biofilm. Ulva zoospores preferen-
quantities of motile zoospores (or zygotes from tially settled on top of bacteria, suggesting a
the fusion of gametes) produced by the mature direct interaction between the bacteria and
plant.Once released from the thallus,the swim- zoospores and providing evidence that attach-
ming zoospores select sites for attachment by ment is not a random process (38). This work
“sensing” a surface, and this is followed by tem- led to the discovery that bacterial quorum-
porary adhesion (Fig.4).If conditions are unsat- sensing molecules, specifically AHLs, are
isfactory, the zoospore detaches and continues involved in zoospore settlement.
to search for an optimum location. Under As with biofilms of natural bacterial
appropriate conditions, the zoospore excretes a assemblages, biofilms of wild-type (WT) AHL-
glycoprotein adhesive to form a permanent producing V. anguillarum also showed a positive
attachment, differentiates, and grows to form a correlation between the numbers of zoospores
mature plant (14). attaching and bacterial density (39).However,if
Many studies have shown the presence of a AHL production was inactivated either by
bacterial biofilm to be an important factor in expressing the recombinant Bacillus lactonase
the settlement and subsequent metamorphosis coding gene aiiA (20) or through mutation of
of invertebrate larvae, and this is also true of the AHL biosynthetic genes, the attraction
Ulva zoospores (38). Several factors are thought of the zoospores was abolished (39, 75). This
FIGURE 4 Settlement of Ulva zoospores. Processes and signals that attract Ulva zoospores to surfaces in the
marine environment, including AHL signaling.
strongly suggested that Ulva zoospores cient mutant harboring the AHL reporter sys-
responded to AHLs produced by V. anguillarum tem revealed no differences in zoospore settle-
but did not exclude the possibility that there ment patterns between the biofilms and clean
were other unidentified phenotypes dependent cover-glass controls, and only background GFP
on AHL production that may have affected activity. Zoospores, however, preferentially set-
zoospore settlement.To address this, biofilms of tled on top of WT microcolonies, specifically
E.coli expressing the recombinant V.anguillarum areas associated with GFP induction, indicating
AHL synthase genes VanI (producing 3-oxo- elevated local AHL production.Biofilms treated
C10-HSL) and VanM (producing C6-HSL and with UV or antibiotic treatment were reduced
3-OH-C6-HSL) were tested for the ability to in their ability to attract the Ulva zoospores,
attract zoospores. These biofilms significantly consistent with AHL production rather than
enhanced zoospore settlement (39). Finally, the surface topography playing the dominant role.
attraction of zoospores to the synthetic AHLs Addition of synthetic AHL to the seawater
3-oxo-C10-HSL, C6-HSL, and 3-OH-C6-HSL in which the biofilms were immersed abolished
was tested. Incorporation of all three AHLs into the attraction of Ulva zoospores to the surface,
agarose films enhanced settlement (75). AHLs suggesting that the biofilm was the source of a
with a range of different chain lengths and lev- chemoattractant (80). Recent studies have
els of substitution, both synthetic and those shown that although AHLs act as a chemoat-
produced by biofilms, enhanced settlement tractant for Ulva zoospores, the mechanism
with the exception of C4-HSL (75). This through which they act is not through chemo-
response is specific to functional AHL mole- taxis but rather due to an effect on chemokine-
cules, as opening the lactone ring of the AHL sis. Zoospores rapidly accumulate around an
molecule by increasing pH destroys the signal- AHL point source, indicating AHLs are strong
ing properties of the molecule and fails to chemoattractants for zoospores. Examination
attract zoospores (39).AHLs with open lactone of zoospore swimming,however,revealed com-
rings are also nonfunctional in bacterial pletely random swimming in the proximity of
quorum-sensing signaling (82). the AHL point source (80). Surprisingly, it was
The number of Ulva zoospores attaching was the speed of swimming that was affected by the
positively correlated with the density of bacte- presence of AHL, which caused an immediate
ria within the biofilms for both mixed assem- decrease. Similar decreases in swimming speed
blages of natural marine bacteria (38) and single were observed in the presence of V. anguillarum
species cultures (39). Statistical analysis revealed biofilms, but biofilms of a QS-deficient mutant
that the zoospores directly attach to bacterial had only a weak effect. The rapid decrease in
cells rather than to the surrounding substratum. swimming speed provides an attractive mecha-
Zoospores are known to preferentially settle nism for zoospore accumulation around an
within crevices and micrometer-scale imper- AHL point source. If the swimming velocity of
fections in steel surfaces (14). It was therefore a fast, randomly moving population is reduced
plausible that some of the interactions between upon entering a zone of elevated AHL concen-
zoospores and bacterial biofilms were due to tration, they will tend to accumulate at this
variations in surface topography and not AHL point. Only zoospores close to the surface were
production. Tait et al. (75) combined the decreased in swimming speed, suggesting that
approach of Joint et al. (38) with direct detec- there may be a requirement for surface contact,
tion of AHL production within biofilms of WT which would ensure that a zoospore does not
and AHL-defective V. anguillarum. A gfp-based completely cease swimming in the absence of a
AHL reporter system was introduced into the substratum for settlement.
WT and quorum-sensing mutant,and zoospore The benefit of AHL responsiveness in
settlement was monitored with biofilms of zoospores has not been investigated but is pre-
these derivatives. The quorum-sensing-defi- sumably used by the alga to determine an
appropriate surface, prepopulated by a biofilm sampling of milky seas revealed V. harveyi as the
composed of diverse microbes. AHLs are pro- source of bioluminescence (41). In general,
duced by proteobacteria,and thus the zoospores algal blooms are known to support the growth
would be specifically attracted to biofilms of large numbers of bacteria, particularly as
containing AHL members of this group.Given the bloom begins to break down, resulting in
the high proteobacterial abundance in ocean massive nutrient release available for bacterial
surface layers,it seems likely that such a require- consumption.
ment would be readily met by most biofilm The occurrence of milky seas is certainly
communities. Symbiotic interactions of bacte- intriguing but is logistically difficult to study.
ria and algae are known to occur, and in some Reports of milky seas are infrequent, the
cases normal development of the alga requires majority made by shipping operatives within
the microbial symbiont (49, 62). It is plausible the northwest Indian Ocean during the sum-
that the swimming zoospores are targeting mer southwest monsoon season (34).The prob-
bacteria that influence their morphological ability of a scientific team being in the right
development. place at the right time to sample a milky sea is
very low. However, a recent report used remote
MILKY SEAS: QUORUM SENSING sensing by satellites to detect a vast luminescent
IN OPEN OCEANS? area (≈15,400 km2) of the northwest Indian
Reported for centuries by mariners, “milky Ocean (Fig. 5) (50). Luminescence lasted for
sea” is used to describe the nocturnal phenom- 3 days and was confirmed on the first night by a
enon where sea surfaces emit an intense, uni- ship-based account. This satellite observation
form, and prolonged glow extending over vast was consistent with a bacterial origin for these
areas.The glow is undoubtedly due to biolumi- events, although by no means unequivocal
nescence, but there has been disagreement proof.Whether this type of phenomenon rep-
over its source. Many marine bioluminescent resents a massive bacterial quorum or some
organisms emit light in a brief (millisecond to other means of luminescent induction remains
second) burst or flash, such as with the dinofla- to be determined.
gellates that “sparkle” when mechanically stim-
ulated. In addition, sampling of milky seas failed INHIBITION OF AHL QUORUM
to identify dinoflagellates as the causative agent SENSING BY MACROALGAE
(41). Milky seas are commonly reported in Many marine organisms remain relatively free
calm waters, and dinoflagellate blooms, even in from fouling. For marine invertebrates such as
turbulent seas,fail to produce constant,uninter- sponges, this may be largely due to the produc-
rupted light. tion of deterrents by their endogenous micro-
Bioluminescent marine bacteria are the biota. Other marine organisms also synthesize
other likely candidates that produce light for their own deterrents to discourage colonization
several hours or even days without mechanical of their surfaces. The red macroalga Delisea
stimulation, consistent with the observations of pulchra is thought to control biofilm formation
milky seas.Most of these bacteria,however,reg- on its surface by production of halogenated
ulate light production via quorum sensing, and furanones (46). These halogenated furanones
typical bacterial densities in open waters (≈106 have been shown to inhibit AHL-dependent
cells ml1) would be insufficient to represent a gene expression in Serratia liquefaciens, Erwinia
quorum for the bioluminescent vibrios, calcu- carotovora, and Pseudomonas aeruginosa (33, 47,
lated to require approximately 100-fold greater 66). Concentrations of halogenated furanones
density (6, 59). However, bacteria can accumu- at the surface of the alga appear to be in quanti-
late to relatively high density during algal ties sufficient to inhibit AHL signaling in colo-
blooms, and local colonization of algal micro- nizing bacterial populations and, through this
colonies can also increase this density. Indeed, activity and perhaps others, inhibit biofilm for-
FIGURE 5 Milky seas off the coast of Africa. Satellite imagery of a milky sea. Study areas (Top) corresponding
to unfiltered (A–C) and filtered (D–F) images over three successive days: (A and D) Jan. 25, 1995, 1836 GMT; (B
and E) Jan. 26, 1995, 1804 GMT; and (C and F) Jan. 27, 1995, 1725 GMT.Arrowheads in F indicate low signal-to-
noise ratio artifacts. Shown in D are the track of a ship (dashed line) and positions at time of first sighting on the
horizon (point a) and exit from the glowing waters (point b), based on details of the ship report. Reprinted from
reference 50. Copyright 2007 National Academy of Sciences, USA.
mation (21).The relatively low ratios of gram- In addition to activation or deactivation of

negative to gram-positive bacteria on marine the DNA-binding capacity of LuxR-type pro-
algal surfaces, despite the dominance of teins,AHLs are also thought to modulate cellu-
gram-negative bacteria in bulk seawater, may lar concentrations of these proteins by altering
be a consequence of this inhibition of quorum their susceptibility to proteolytic degradation
sensing. (84). The halogenated furanones decrease the
residence time of the LuxR proteins that they CONCLUSIONS AND

inhibit but do not have activity consistent with a FUTURE PROSPECTS
competitive inhibition for AHL binding (46). The marine environment is a source of abun-
Synthetic derivatives of the natural halogenated dant materials and resources across the globe.
furanones are effective quorum-sensing inhibi- The oceans are sources of diverse and complex
tion (QSI) compounds that effectively reduce quorum-sensing signal molecules produced by
biofilm formation of certain microbes such as the many of their endogenous proteobacteria.
during lung infections by P. aeruginosa in mouse The original quorum-sensing mechanism was
models (33). An interesting potential applica- defined for V.fischeri and related marine vibrios,
tion of QSI using halogenated furanones for but much more recent work has discovered
mariculture was demonstrated by inhibiting the AHL signals and/or LuxI-LuxR-type regula-
pathogenesis of V. harveyi on the black tiger tors in a wide diversity of marine bacteria.
prawn Paneus monodon (45). The furanones Although the gamma-proteobacteria com-
clearly are multifunctional inhibitors as they monly utilize these systems, it is in fact the
also reduce growth, swarming, and biofilm for- alpha-proteobacteria that appear to represent
mation of gram-positive bacteria but use a dif- the most complex and pervasive AHL signaling
ferent mechanism (68). mechanisms in the marine environment. The
Other macroalgae may also control colo- target functions of AHL quorum sensing in
nization of their surfaces by interfering with non-vibrio marine bacteria are just now being
AHL signaling in bacteria. A number of sea- identified and promise to provide great insights
weeds produce haloperoxidases that catalyze into the utility of this signaling mechanism in
the oxidation of bromide and/or chlorine with the marine environment.The quorum-sensing
hydrogen peroxide to produce the bacterioci- mechanism may be escalated to a strikingly
dal compounds hypobromous acid (HOBr) or larger dimension in the ocean, such as in the
hyperchlorous acid (HOCl). HOBr and HOCl milky seas phenomenon. As in terrestrial sys-
are highly toxic to a wide variety of microor- tems, host associations, including symbiotic and
ganisms and, as such, are used to control bio- pathogenic mechanisms, are likely to be a com-
fouling in industrial and potable water systems. mon function under the influence of AHLs.
Specifically,AHLs carrying the 3-oxo group are Examples of intricate cross-kingdom signaling,
rapidly inactivated by exposure to HOCl and exemplified by the stimulation of Ulva-
HOBr, whereas AHLs without the 3-oxo sub- zoospore settlement by AHLs, are probably
stitution retained their activity (7).The brown much more likely than we would have initially
alga Laminaria digitata produces bromoperoxi- anticipated. Future efforts to elucidate inter-
dase, which can also catalyze the reaction of bacterial and interkingdom signaling in the
bromide to HOBr in the presence of H2O2,and marine environment will continue to yield
incubation of 3-oxo-C6-HSL with L. digitata important insights into this process and the
led to degradation of this compound. Marine many ecological contexts in which it is
organisms are an excellent source for bioactive employed. In addition, important applications
compounds,and this clearly includes those with are certain to result, including the prevention
QSI activity, as well as QS mimics.A rapid and of biofouling, the health and productivity of
reliable screening method for QSI compounds oceanic sources of food, and the discovery of
has recently been developed and applied to a novel pharmaceuticals for human therapeutics.
variety of natural samples (65). Given the Leading from the groundswell of studies ini-
wide prevalence of AHL quorum sensing in tiated from work on the luminescent marine
marine bacteria, screening organisms that vibrios, through the ongoing examination of
consistently interact with these microbes is AHL signaling in terrestrial and aquatic bacte-
virtually certain to yield novel inhibitory and ria that followed, our understanding of AHL
stimulatory compounds. signaling in the marine environment has now
begun to build momentum. Propelled by the 10. Buchan,A., J. M. Gonzalez, and M.A. Moran.
amazing rate and diversity of microbial genome 2005. Overview of the marine Roseobacter lineage.
Appl. Environ. Microbiol. 71:5665–5677.
sequencing, new high-throughput biological 11. Buchholtz, C., K. F. Nielsen, D. L. Milton, J. L.
and chemical analyses, and our increasing Larsen, and L. Gram. 2006. Profiling of acylated
recognition of the impact of microbial activity homoserine lactones of Vibrio anguillarum in vitro
in the oceans, it seems certain that the wave of and in vivo: influence of growth conditions and
interest and activity on AHL-based signaling in serotype. Syst.Appl. Microbiol. 29:433–445.
12. Byers, J. T., C. Lucas, G. P. Salmond, and M.
marine systems has yet to crest and will in fact Welch. 2002. Nonenzymatic turnover of an
continue to grow over the coming years. Erwinia carotovora quorum-sensing signaling mole-
cule. J. Bacteriol. 184:1163–1171.
13. Callow, M. E., J. A. Callow, L. K. Ista, S. E.
REFERENCES Coleman, A. C. Nolasco, and G. P. Lopez.
1. An, D., T. Danhorn, C. Fuqua, and M. 2000. Use of self-assembled monolayers of
R. Parsek. 2006. Quorum sensing and motility different wettabilities to study surface selection
mediate interactions between Pseudomonas and primary adhesion processes of green algal
aeruginosa and Agrobacterium tumefaciens in (Enteromorpha) zoospores. Appl. Environ. Microbiol.
biofilm cocultures. Proc. Natl. Acad. Sci. USA 66:3249–3254.
103:3828–3833. 14. Callow, M. E., A. R. Jennings, A. B. Brennan,
2. Bassler, B. L., E. P. Greenberg, and A. M. C. E. Seegert, A. Gibson, L. Wilson, A.
Stevens. 1997.Cross-species induction of lumines- Feinberg, R. Baney, and J. A. Callow.
cence in the quorum-sensing bacterium Vibrio 2002. Microtopographic cues for settlement of
harveyi. J. Bacteriol. 179:4043–4045. zoospores of the green fouling alga Enteromorpha.
3. Bassler, B. L., M.Wright, R. E. Showalter, and Biofouling 18:237–245.
M. R. Silverman. 1993. Intercellular signalling 15. Cha, C., P. Gao, Y.-C. Chen, P. D. Shaw,
in Vibrio harveyi: sequence and function of genes and S. K. Farrand. 1998. Production of acyl-
regulating expression of luminescence. Mol. Micro- homoserine lactone quorum-sensing signals by
biol. 9:773–786. gram-negative plant-associated bacteria. Mol.
4. Bassler, B. L., M.Wright, and M. R. Silverman. Plant-Microbe Interact. 11:1119–1129.
1994. Multiple signalling systems controlling 16. Chen, H., and G. R. Fink. 2006. Feedback
expression of luminescence in Vibrio harveyi: control of morphogenesis in fungi by aromatic
sequence and function of genes encoding a second alcohols. Genes Dev. 20:1150–1161. (First pub-
sensory pathway. Mol. Microbiol. 13:273–286. lished 17 April 2006; doi 10.1101/gad.1411806)
5. Boettcher, K. J., and E. G. Ruby. 1995. Detec- 17. Croxatto,A.,V. J. Chalker, J. Lauritz, J. Jass,A.
tion and quantification of Vibrio fischeri autoinducer Hardman, P.Williams, M. Camara, and D. L.
from the symbiotic squid light organs. J. Bacteriol. Milton. 2002.VanT, a homologue of Vibrio harveyi
177:1053–1058. LuxR, regulates serine, metalloprotease, pigment,
6. Booth, C. R., and K. H. Nealson. 1975. Light- and biofilm production in Vibrio anguillarum.J.Bac-
emission by luminous bacteria in open ocean. Bio- teriol. 184:1617–1629.
phys. J. 15:A56–A56. 18. Croxatto, A., J. Pride, A. Hardman, P.
7. Borchardt, S. A., E. J. Allain, J. J. Michells, Williams, M. Camara, and D. L. Milton. 2004.
G. W. Stearns, R. F. Kelly, and W. F. McCoy. A distinctive dual-channel quorum-sensing system
2001. Reaction of acylated homoserine lactone operates in Vibrio anguillarum. Mol. Microbiol.
bacterial signaling molecules with oxidized 52:1677–1689.
halogen antimicrobials. Appl. Env. Microbiol. 19. Defoirdt, T., P. Bossier, P. Sorgeloos, and W.
67:3174–3179. Verstraete. 2005.The impact of mutations in the
8. Bruhn, J. B., L. Gram, and R. Belas. 2007. Pro- quorum sensing systems of Aeromonas hydrophila,
duction of antibacterial compounds and biofilm Vibrio anguillarum and Vibrio harveyi on their viru-
formation by Roseobacter species are influenced by lence towards gnotobiotically cultured Artemia
culture conditions. Appl. Environ. Microbiol. franciscana. Environ. Microbiol. 7:1239–1247.
73:442–450. 20. Dong, Y.-H., J.-L. Xu, X.-Z. Li, and L.-H.
9. Bruhn, J. B., K. F. Nielsen, M. Hjelm, M. Zhang. 2000.AiiA, an enzyme that inactivates the
Hansen, J. Bresciani, S. Schulz, and L. Gram. acylhomoserine lactone quorum-sensing signal
2005. Ecology, inhibitory activity, and morphogen- and attenuates virulence of Erwinia carotovora. Proc.
esis of a marine antagonistic bacterium belonging Natl.Acad. Sci. USA 97:3526–3531.
to the Roseobacter clade. Appl. Environ. Microbiol. 21. Dworjanyn, S. A., R. De Nys, and P. D.
71:7263–7270. Steinberg. 1999. Localisation and surface quan-
tification of secondary metabolites in the red alga 33. Hentzer, M., H.Wu, J. B.Andersen, K. Riedel,
Delisea pulchra. Mar. Biol. 133:727–736. T. B. Rasmussen, N. Bagge, N. Kumar, M. A.
22. Elasri, M., S. Delorme, P. Lemanceau, Schembri, Z. Song, P. Kristoffersen, M.
G. Stewart, B. Laue, E. Glickmann, P. M. Manefield, J.W. Costerton, S. Molin, L. Eberl,
Oger, and Y. Dessaux. 2001. Acyl-homoserine P. Steinberg, S. Kjelleberg, N. Hoiby, and M.
lactone production is more common among Givskov. 2003.Attenuation of Pseudomonas aerugi-
plantassociated Pseudomonas spp. than among soil- nosa virulence by quorum sensing inhibitors.
borne Pseudomonas spp. Appl. Environ. Microbiol. EMBO J. 22:3803–3815.
67:1198–1209. 34. Herring, P. J., and M.Watson. 1993. Milky seas:
23. Evans, K., L. Passador, R. Srikumar, E. a bioluminescent puzzle. Mar. Observer 63:22–30.
Tsang, J. Nezezon, and K. Poole. 1998. Influ- 35. Hoang, T. T., S. A. Sullivan, J. K. Cusick, and
ence of the MexAB-OprM multidrug efflux sys- H. P. Schweizer. 2002. Beta-ketoacyl acyl carrier
tem on quorum sensing in Pseudomonas aeruginosa. protein reductase (FabG) activity of the fatty acid
J. Bacteriol. 180:5443–5447. biosynthetic pathway is a determining factor of 3-
24. Fuqua, C., and A. Eberhard. 1999. Signal oxo-homoserine lactone acyl chain lengths. Micro-
generation in autoinduction systems: synthesis biology 148:3849–3856.
of acylated homoserine lactones by LuxI-type 36. Howard, E. C., J. R. Henriksen,A. Buchan, C.
proteins, p. 211–230. In G. M. Dunny and S. C. R. Reisch, H. Burgmann, R.Welsh,W.Ye, J. M.
Winans (ed.), Cell-Cell Signaling in Bacteria. ASM Gonzalez, K. Mace, S. B. Joye, R. P. Kiene,W.
Press,Washington, DC. B.Whitman, and M. A. Moran. 2006. Bacterial
25. Gantner, S., M. Schmid, C. Dürr, R. taxa that limit sulfur flux from the ocean. Science.
Schuhegger, A. Steidle, P. Hutzler, C. 314:649–652.
Langebartels, L. Eberl, A. Hartmann, and F. 37. Huber, J. A., D. A. Butterfield, and J. A.
B. Dazzo. 2006. In situ quantitation of the spatial Baross. 2003. Bacterial diversity in a subseafloor
scale of calling distances and population density- habitat following a deep-sea volcanic eruption.
independent N-acylhomoserine lactone-mediated FEMS Microbiol. Ecol. 43:393–409.
communication by rhizobacteria colonized on 38. Joint, I., M. E. Callow, J. A. Callow, and K. R.
plant roots. FEMS Microbiol. Ecol. 56:188–194. Clarke. 2000. The attachment of Enteromorpha
26. Gould, T. A., H. P. Schweizer, and M. E. zoospores to a bacterial biofilm assemblage. Bio-
Churchill. 2004. Structure of the Pseudomonas fouling 16:151–158.
aeruginosa acyl-homoserinelactone synthase LasI. 39. Joint, I., K.Tait, M. E. Callow, J.A. Callow, D.
Mol. Microbiol. 53:1135–1146. Milton, P. Williams, and M. Camara. 2002.
27. Gram, L., H. P. Grossart, A. Schlingloff, and Cell-to-cell communication across the prokary-
T. Kiorboe. 2002. Possible quorum sensing in ote-eukaryote boundary. Science 298:1207.
marine snow bacteria: production of acylated 40. Kirke, D. F., S. Swift, M. J. Lynch, and P.
homoserine lactones by Roseobacter strains isolated Williams. 2004.The Aeromonas hydrophila LuxR
from marine snow. Appl. Environ. Microbiol. homologue AhyR regulates the N-acyl homoser-
68:4111–4116. ine lactone synthase,AhyI positively and negatively
28. Greenberg, E. P., J. W. Hastings, and S. in a growth phase-dependent manner. FEMS
Ulitzur. 1979. Induction of luciferase synthesis in Microbiol. Lett. 241:109–117.
Beneckea harveyi by other marine bacteria. Arch. 41. Lapota, D., C. Galt, J. R. Losee, H. D. Huddell,
Microbiol. 120:87–91. J. K. Orzech, and K. H. Nealson. 1988. Obser-
29. Hanzelka, B. L., M. R. Parsek, D. L.Val, P.V. vations and measurements of planktonic biolumi-
Dunlap, J. E. J. Cronan, and E. P. Greenberg. nescence in and around a milky sea J.Exp.Mar.Biol.
1999.Acylhomoserine lactone synthase activity of Ecol. 119:55–81.
the Vibrio fischeri AinS protein. J. Bacteriol. 42. Lenz, D. H., K. C. Mok, B. N. Lilley, R. V.
181:5766–5770. Kulkarni, N. S. Wingreen, and B. L. Bassler.
30. He, X., and C. Fuqua. 2006. Rhizosphere com- 2004.The small RNA chaperone Hfq and multiple
munication: quorum sensing by the rhizobia. J. small RNAs control quorum sensing in Vibrio har-
Microbiol. Biotechnol. 16:1661–1677. veyi and Vibrio cholerae. Cell 118:69–82.
31. Hellingwerf, K. J. 2004. A network of net- 43. Lerat, E., and N. A. Moran. 2004. The evolu-
workers:report of the Euresco conference on ‘Bac- tionary history of quorum-sensing systems in
terial Neural Networks’ held at San Feliu (Spain) bacteria. Mol. Biol. Evol. 21:903–913. (First
from 8 to 14 May 2004. Mol. Microbiol. 54:2–13. published 10 March 2004; doi 10.1371/
32. Henke, J. M., and B. L. Bassler. 2004. journal.pbio.0030130)
Three parallel quorum-sensing systems regulate 44. Lynch, M. J., S. Swift, D. F. Kirke, C.W. Keevil,
gene expression in Vibrio harveyi. J. Bacteriol. C. E. Dodd, and P. Williams. 2002.The regu-
186:6902–6914. lation of biofilm development by quorum sensing
in Aeromonas hydrophila. Environ. Microbiol. ing regulator in Pantoea stewartii subsp.stewartii:evi-

4:18–28. dence for a repressor function. Mol. Microbiol.
45. Manefield, M., L. Harris, S. A. Rice, R. De 44:1625–1635.
Nys, and S. Kjelleberg. 2000. Inhibition of 55. Mohamed, N. M., E. M. Cicirelli, J. Kan, F.
luminescence and virulence in the black tiger Chen, C. Fuqua, and R.T. Hill. 2008. Diversity
prawn (Penaeus monodon) pathogen Vibrio harveyi and quorum sensing signal production of pro-
by intercellular signal antagonists. Appl. Environ. teobacteria associated with marine sponges. Envi-
Microbiol. 66:2079–2084. ron. Microbiol. 10:75–86.
46. Manefield, M.,T. B. Rasmussen, M. Hentzer, 56. Moran, M. A., R. Belas, M. A. Schell, J. M.
J. B.Andersen, P. Steinberg, S. Kjelleberg, and Gonzalez, F. Sun, S. Sun, B. J. Binder, J.
M. Givskov. 2002. Halogenated furanones inhibit Edmonds,W.Ye, B. Orcutt, E. C. Howard, C.
quorum sensing through accelerated LuxR Meile, W. Palefsky, A. Goesmann, Q. Ren, I.
turnover. Microbiology 148:1119–1127. Paulsen, L. E. Ulrich, L. S. Thompson, E.
47. Manefield, M., M.Welch, M. Givskov, G. P. C. Saunders, and A. Buchan. 2007. Ecological
Salmond, and S. Kjelleberg. 2001.Halogenated genomics of marine roseobacters. Appl. Environ.
furanones from the red alga, Delisea pulchra, Microbiol. 25:25.
inhibit carbapenem antibiotic synthesis and 57. Moran, M.A.,A. Buchan, J. M. Gonzalez, J. F.
exoenzyme virulence factor production in the Heidelberg,W. B.Whitman, R. P. Kiene, J. R.
phytopathogen Erwinia carotovora. FEMS Microbiol. Henriksen, G. M. King, R. Belas, C. Fuqua, L.
Lett. 205:131–138. Brinkac, M. Lewis, S. Johri, B.Weaver, G. Pai,
48. Mathesius, U., S. Mulders, M. Gao, M. J.A. Eisen, E. Rahe,W. M. Sheldon,W.Ye,T. R.
Teplitski, G. Caetano-Anolles, B. G. Rolfe, Miller, J. Carlton, D. A. Rasko, I. T. Paulsen,
and W. D. Bauer. 2003. Extensive and specific Q. Ren, S. C. Daugherty, R. T. Deboy, R. J.
responses of a eukaryote to bacterial quorum- Dodson, A. S. Durkin, R. Madupu, W. C.
sensing signals. Proc. Natl. Acad. Sci. USA Nelson, S. A. Sullivan, M. J. Rosovitz, D.
100:1444–1449. H. Haft, J. Selengut, and N. Ward. 2004.
49. Matsuo, Y., M. Suzuki, H. Kasai, Y. Shizuri, Genome sequence of Silicibacter pomeroyi reveals
and S. Harayama. 2003. Isolation and phyloge- adaptations to the marine environment. Nature
netic characterization of bacteria capable of 432:910–913.
inducing differentiation in the green alga 58. Morris, R. M., M. S. Rappe, S.A. Connon, K.
Monostroma oxyspermum. Environ. Microbiol. L.Vergin,W. A. Siebold, C. A. Carlson, and S.
5:25–35. J. Giovannoni. 2002. SAR11 clade dominates
50. Miller, S. D., S. H. D. Haddock, C. D. Elvidge, ocean surface bacterioplankton communities.
and T. F. Lee. 2005. Detection of a biolumines- Nature 420:806–810.
cent milky sea from space.Proc.Natl.Acad.Sci.USA 59. Nealson, K. H., T. Platt, and J. W. Hastings.
102:14181–14184. 1970. Cellular control of the synthesis and activity
51. Milton, D. L. 2006. Quorum sensing in vibrios: of the bacterial luminescent system. J. Bacteriol.
complexity for diversification. Int. J. Med. Microbiol. 104:313–322.
296:61–71. (First published 17 February 2006; 60. Orr, J. C.,V. J. Fabry, O. Aumont, L. Bopp, S.
doi 10.1016/j.ijmm.2006.01.044) C. Doney, R. A. Feely, A. Gnanadesikan, N.
52. Milton, D. L., V. J. Chalker, D. Kirke, A. Gruber, A. Ishida, F. Joos, R. M. Key, K.
Hardman, M. Camara, and P. Williams. Lindsay, E. Maier-Reimer, R. Matear, P.
2001. The LuxM homologue VanM from Vibrio Monfray, A. Mouchet, R. G. Najjar, G. K.
anguillarum directs the synthesis of N-(3- Plattner, K. B. Rodgers, C. L. Sabine, J. L.
hydroxyhexanoyl) homoserine lactone and Sarmiento, R. Schlitzer, R. D. Slater, I. J.
N-hexanoylhomoserine lactone. J. Bacteriol. Totterdell, M. F. Weirig,Y.Yamanaka, and A.
183:3537–3547. Yool. 2005. Anthropogenic ocean acidification
53. Milton, D. L., A. Hardman, M. Camara, S. R. over the twenty-first century and its impact on cal-
Chhabra, B. W. Bycroft, G. S. A. B. Stewart, cifying organisms. Nature 437:681–686.
and P.Williams. 1997. Quorum sensing in Vibrio 61. Passador, L., K. D. Tucker, K. R. Guertin,
anguillarum: characterization of the vanI/vanR M. P. Journet,A. S. Kende, and B. H. Iglewski.
locus and identification of the autoinducer N-(3- 1996. Functional analysis of the Pseudomonas aerug-
oxodecanoyl)-L-homoserine lactone. J. Bacteriol. inosa autoinducer PAI. J. Bacteriol. 178:5995–6000.
179:3004–3012. 62. Provasoli, L., and I. J. Pintner. 1980. Bacteria
54. Minogue, T. D., M. Wehland-von Trebra, F. induced polymorphism in an axenic laboratory
Bernhard, and S. Beck von Bodman. 2002. strain of Ulva lactuca (Chlorophyceae). J. Phycol.
The autoregulatory role of EsaR, a quorum sens- 16:196–201.
63. Qazi, S., B. Middleton, S. H. Muharram, G. Tamayo-Castillo, M. R. Ferrari, R. L.

A. Cockayne, P. Hill, P. O’Shea, S. R. Strausberg, K. Nealson, R. Friedman, M. Fra-
Chhabra, M. Camara, and P. Williams. 2006. zier, and J. C.Venter. 2007.The Sorcerer II global
N-acylhomoserine lactones antagonize virulence ocean sampling expedition: northwest Atlantic
gene expression and quorum sensing in Staphylo- through eastern tropical Pacific. PLoS Biol. 5:e77.
coccus aureus. Infect Immun. 74:910–919. 72. Schaefer, A. L., B. L. Hanzelka, A. Eberhard,
64. Rasch, M., C. Buch, B. Austin, W. J. and E. P. Greenberg. 1996. Quorum-sensing in
Slierendrecht, K. S. Ekmann, J. L. Larsen, C. Vibrio fischeri: probing autoinducer-LuxR interac-
Johansen, K. Riedel, L. Eberl, M. Givskov, and tions with autoinducer analogs. J. Bacteriol.
L. Gram. 2004.An inhibitor of bacterial quorum 178:2897–2901.
sensing reduces mortalities caused by vibriosis in 73. Stevens, A. M., and E. P. Greenberg. 1999.
rainbow trout (Oncorhynchus mykiss, Walbaum). Transcriptional activation by LuxR, p. 231–242. In
Syst.Appl. Microbiol. 27:350–359. G. M. Dunny and S. C.Winans (ed.), Cell-Cell Sig-
65. Rasmussen, T. B., T. Bjarnsholt, M. E. naling in Bacteria. ASM Press,Washington, DC.
Skindersoe, M. Hentzer, P. Kristoffersen, M. 74. Swift, S.,A.V. Karlyshev, L. Fish, E. L. Durant,
Kote, J. Nielsen, L. Eberl, and M. Givskov. M. Winson, S. R. Chhabra, P. Williams, S.
2005. Screening for quorum-sensing inhibitors Macintyre, and G. S. A. B. Stewart. 1997.
(QSI) by use of a novel genetic system, the QSI Quorum-sensing in Aeromonas hydrophila and
selector. J. Bacteriol. 187:1799–1814. Aeromonas salmonicida: identification of the LuxRI
66. Rasmussen, T. B., M. Manefield, J. B. Ander- homologs AhyRI and AsaRI and their cognate N-
sen, L. Eberl, U. Anthoni, C. Christophersen, acylhomoserine lactone molecules. J. Bacteriol.
P. Steinberg, S. Kjelleberg, and M. Givskov. 179:5271–5281.
2000.How Delisea pulchra furanones affect quorum 75. Tait, K., I. Joint, M. Daykin, D. L. Milton, P.
sensing and swarming motility in Serratia liquefa- Williams, and M. Camara. 2005. Disruption of
ciens MG1. Microbiology 146:3237–3244. quorum sensing in seawater abolishes attraction of
67. Redfield, R. J. 2002. Is quorum sensing a side zoospores of the green alga Ulva to bacterial
effect of diffusion sensing? Trends Microbiol. biofilms. Environ. Microbiol. 7:229–240.
10:365–370. 76. Taylor, M. W., P. J. Schupp, H. J. Baillie, T. S.
68. Ren, D. C., L. A. Bedzyk, P. Setlow, D. F. Charlton, R. de Nys, S. Kjelleberg, and P. D.
England, S. Kjelleberg, S. M. Thomas, R. W. Steinberg. 2004. Evidence for acyl homoserine
Ye, and T. K. Wood. 2004. Differential gene lactone signal production in bacteria associated
expression to investigate the effect of (5Z)-4- with marine sponges. Appl. Environ. Microbiol.
bromo-5-(bromomethylene)-3-butyl-2(5H)- 70:4387–4389.
furanone on Bacillus subtilis. Appl. Environ. 77. Telford, G., D. Wheeler, P. Williams, P. T.
Microbiol. 70:4941–4949. Tomkins, P. Appleby, H. Sewell, G. S. A. B.
69. Riebesell, U., I. Zondervan, B. Rost, P. D. Stewart, B. W. Bycroft, and D. I. Pritchard.
Tortell, R. E. Zeebe, and F. M. Morel. 2000. 1998. The Pseudomonas aeruginosa quorum-
Reduced calcification of marine plankton in sensing signal molecule N-(3-oxododecanoyl)-l-
response to increased atmospheric CO2. Nature homoserine lactone has immunomodulatory
407:364–367. activity. Infect. Immun. 66:36–42.
70. Riedel, K., M. Hentzer, O. Geisenberger, B. 78. Wagner-Dobler, I., and H. Biebl. 2006. Envi-
Huber, A. Steidle, H. Wu, N. Hoiby, M. ronmental biology of the marine Roseobacter line-
Givskov, S. Molin, and L. Eberl. 2001. N- age. Annu. Rev. Microbiol. 60:255–280.
acylhomoserine-lactone-mediated communica- 79. Wagner-Dobler, I., V. Thiel, L. Eberl, M.
tion between Pseudomonas aeruginosa and Burk- Allgaier, A. Bodor, S. Meyer, S. Ebner, A.
holderia cepacia in mixed biofilms. Microbiology Hennig, R. Pukall, and S. Schulz. 2005. Dis-
147:3249–3262. covery of complex mixtures of novel long-
71. Rusch, D. B., A. L. Halpern, G. Sutton, K. B. chain quorum sensing signals in free-living and
Heidelberg, S.Williamson, S.Yooseph, D.Wu, host-associated marine alphaproteobacteria.
J. A. Eisen, J. M. Hoffman, K. Remington, K. Chembiochem. 6:2195–2206.
Beeson, B.Tran, H. Smith, H. Baden-Tillson, 80. Wheeler, G. L., K. Tait, A. Taylor, C. Brown-
C. Stewart, J. Thorpe, J. Freeman, C. lee, and I. Joint. 2006.Acyl-homoserine lactones
Andrews-Pfannkoch, J. E. Venter, K. Li, S. modulate the settlement rate of zoospores of the
Kravitz, J. F. Heidelberg, T. Utterback, Y. H. marine alga Ulva intestinalis via a novel chemoki-
Rogers, L. I. Falcon, V. Souza, G. Bonilla- netic mechanism. Plant Cell Environ. 29:608–618.
Rosso, L. E. Eguiarte, D. M. Karl, S. Sathyen- 81. Whitehead, N.A.,A. M. L. Barnard, H. Slater,
dranath,T. Platt, E. Bermingham,V. Gallardo, N. J. L. Simpson, and G. P. C. Salmond. 2001.
Quorum-sensing in gram-negative bacteria. 83. Zhu, J., J.W. Beaber, M. I. More, C. Fuqua,A.
FEMS Microbiol. Rev. 25:365–404. Eberhard, and S. C. Winans. 1998. Analogs of
82. Yates, E. A., B. Philipp, C. Buckley, S. the autoinducer 3-oxooctanoyl-homoserine
Atkinson, S. R. Chhabra, R. E. Sockett, lactone strongly inhibit activity of the TraR pro-
M. Goldner, Y. Dessaux, M. Camara, H. tein of Agrobacterium tumefaciens. J. Bacteriol.
Smith, and P. Williams. 2002. N-acylhomoser- 180:5398–5405.
ine lactones undergo lactonolysis in a pH-, 84. Zhu, J., and S. C. Winans. 1999. Autoinducer
temperature-,and acyl chain length-dependent binding by the quorum-sensing regulator TraR
manner during growth of Yersinia pseudotu- increases affinity for target promoters in vitro and
berculosis and Pseudomonas aeruginosa. Infect. Immun. decreases TraR turnover rates in whole cells. Proc.
70: 5635–5646. Natl.Acad. Sci. USA 96:4832–4837.
PRODUCTION,
DETECTION,AND
QUENCHING OF
CHEMICAL SIGNALS
III
ACYL-HOMOSERINE LACTONE
BIOSYNTHESIS: STRUCTURE
AND MECHANISM
Mair E. A. Churchill and Jake P. Herman
17
Bacteria communicate using chemical signals to posed of a homoserine lactone (HSL) ring with
sense cell density and activate differentiation to an acyl chain (Fig. 1). The acyl-chain length
diverse lifestyles by a process known as quorum typically varies from C4 to C18 (23, 47) and may
sensing.The first quorum-sensing system to be be modified, most commonly by a 3-oxo sub-
discovered was the Lux system of Vibrio fischeri stituent or, in some cases, a 3-hydroxy sub-
(19, 53), which regulates light production in the stituent or a degree of unsaturation. The
light organ of deep sea fish and squid via an lipid-soluble AHLs diffuse in and out of the cell
acyl-homoserine lactone (AHL) signaling mol- by passive as well as active transport (39). At a
ecule (reviewed in reference 25). Decades of given threshold cell number or bacterial “quo-
study into quorum sensing of many gram-neg- rum,” the AHL reaches a sufficiently high con-
ative bacteria have shown that the downstream centration and is recognized by the binding of
consequences of this intercellular signaling the receptor protein, which is the second com-
mechanism include a variety of complex cellu- ponent of the system.These include a family of
lar behavioral processes, such as biolumines- sensor kinases related to LuxN (22) and a wide
cence, sporulation, secretion of virulence variety of “R proteins,” such as LuxR or LasR,
factors, pigment and drug production, and reg- which are AHL-responsive transcriptional reg-
ulation of bacterial virulence (15;chapters 9,10, ulators (reviewed in chapter 9 and references
and 12–16). More recently, the molecular basis 25 and 71).The binding of the AHL by some R
for quorum sensing mediated by the AHL signal proteins initiates the activation and repression
in gram-negative bacteria has become much of target genes,and in some cases, AHL binding
clearer (reviewed in references 24, 79, and 82). leads to target gene derepression (50, 64, 66, 77,
Three elements are central to quorum sens- 78). Finally, the AHLs are synthesized by
ing mediated by AHLs. First, the AHL is com- enzymes known as the AHL synthases (23, 62).
Mair E. A. Churchill Department of Pharmacology, Program
in Biomolecular Structure,The University of Colorado at AHL SYNTHASE FAMILIES
Denver and Health Sciences Center,Aurora, Colorado 80045. Production of AHLs has been demonstrated in
Jake P. Herman Department of Pharmacology,The University
of Colorado at Denver Health Sciences Center, Aurora, more than 70 species, and three different types
Colorado 80045. of enzymes are now known to synthesize AHLs
© 2008 ASM Press, Washington, DC
275
276 ■ CHURCHILL AND HERMAN
FIGURE 1 Chemical structure of AHLs. AHLs found in gram-negative bacteria vary

by substitution at the C-3 position (R1) and the length and unsaturation of R2. Shown
also are the structures of a subset of the AHLs that are produced by the AHL synthases EsaI
from P. stewartii (or LuxI from V. fischeri) and LasI from P. aeruginosa.
in vivo, including enzymes related to LuxI, species, the AinS and VanM enzymes were
HdtS, and LuxM (reviewed in reference 79). found to produce AHLs (26,48). TheVanM and
The AHL synthases typified by the LuxI AinS enzymes each have sequence similarity to
enzymes, the archetype of the class, were first LuxL in their N-terminal regions and similarity
discovered in the lux operon in V. fischeri (21). to LuxM in the C-terminal region, and appear
Hence,these enzymes are frequently referred to to be a composite of both LuxM and LuxL.
as I proteins and have been assigned a name Unlike LuxL and LuxM,VanM or AinS is each
based on the species of origin or on the pheno- capable of synthesizing AHLs by itself.Although
type or systems that they regulate; for example, these enzymes do not appear to be related to
LasI was named for its role in the induction of any other enzyme family, enzymatic studies
the elastase virulence factors in Pseudomonas conducted in vitro show that they use SAM and
aeruginosa (56). The LuxI-type AHL synthases acyl-ACP as well as acyl-coenzyme A (acyl-
catalyze the formation of the AHL from the CoA) as substrates (31).These substrate require-
substrates S-adenosyl-L-methionine (SAM) ments are similar to those of the LuxI-type AHL
and acyl-acyl carrier protein (acyl-ACP) (30) synthases. However, little more is known about
(Fig. 2A). Enzymatic synthesis of AHLs using their mechanism of AHL synthesis.
purified substrates for TraI (from Agrobacterium The HdtS enzyme was first identified in
tumefaciens) verified that both SAM and acyl- Pseudomonas fluorescens F113, where it was
ACP are substrates for AHL synthesis in vitro shown to produce N-(3-OH-7-cis-tetrade-
(51). This finding was confirmed in vivo with cenoyl)-HSL, as well as C6-HSL and C10-HSL
studies of LuxI (75). as a recombinant protein exogenously expressed
The LuxM family of enzymes has been in Escherichia coli (42).An HdtS homologue has
shown to make AHLs in only a few species of - also been suggested to be responsible for the
proteobacteria. luxM was identified as a gene in synthesis of C6-HSL, C8-HSL, and C10-HSL in
Vibrio harveyi that together with luxL was the Nitrosomonas europaea strain Schmidt (8),
required for production of specific AHLs (2). although other N. europaea strains appear to
However, in V. fischeri and other related Vibrio carry a distant LuxI-type AHL synthase gene in
17. ACYL-HOMOSERINE LACTONE BIOSYNTHESIS ■ 277
FIGURE 2 Schematic diagrams of the reactions performed by AHL synthases and structural homologues.
(A) AHL synthesis reaction. AHL synthases catalyze the formation of AHLs from SAM and acyl-ACP by acyla-
tion of SAM and lactonization of the methionine moiety to give, in addition to the AHL, holo-ACP and
5′-methylthioadenosine products (55, 75). (B) The GNATs catalyze acetylation of lysine or other primary amines
using acetyl-CoA as a substrate (12). (C) The N-acyl tyrosine synthases catalyze the acylation of amino acids using
acyl-ACP as a substrate (76).
the genomes. Interestingly, HdtS is a member is likely to be some interest as to the specific role
of the lysophosphatidic acid acyltransferase fam- of this enzyme for each function in vivo.
ily (42) and appears to be required for correct
acylation of lysophosphatidic acid in P.fluorescens AHL SYNTHASES OF THE LuxI TYPE
(13). HdtS is capable of using both acyl-CoA By far, the most widespread and best under-
and acyl-ACP as substrates for acylation of stood of the AHL synthases are the LuxI-type
lysophosphatidic acid, and its homologues are enzymes. These enzymes are composed of a
ubiquitous among bacteria and eukaryotes (68). single domain of approximately 205 amino acid
The enzymatic mechanism of AHL synthesis residues in length. Identifiable homologues
remains uncharacterized. Furthermore, it is have been found in more than 150 species of
unknown which features of the HdtS enzymes , , and classes of proteobacteria. There is
are required for AHL production, but there are great diversity among the AHL synthases in
homologues in organisms that do not produce their production of AHLs (Table 1), and they
AHLs. Therefore, this bacterial enzyme family cluster into several ill-defined groups based on
appears to have two functions: acylation of sequence and phylogenic relationships (44) and
lysophosphatidic acid and AHL synthesis.There likely structural relatedness.The sequence con-
TABLE 1 AHL signals and AHL synthases

AHL synthase Species Signal Function Reference
RhlI Pseudomonas aeruginosa C4-HSL Opportunistic pathogen 15
CviI Chromobacterium violaceum C6-HSL Antibiotic producer 10
CepI Burkholderia cepacia C8-HSL Soil bacterium 45
BmaI1 Burkholderia mallei C8/C10-HSL Glanders causative agent 74
CerI Rhodobacter sphaeroides 7-cis-C14-HSL Photosynthetic bacterium 58
SinI Sinorhizobium melioti 9-cis-C16-HSL Nitrogen-fixing symbiont 47
C18-HSL
EsaI Pantoea sterwartii 3-oxo-C6-HSL Plant pathogen 4
LuxI Vibrio fischeri 3-oxo-C6-HSL Marine symbiont 19
YspI Yersinia pestis 3-oxo-C6/C8-HSL Plague causative agent 41
TraI Agrobacterium tumefaciens 3-oxo-C8-HSL Crown gall causative agent 85
VanI Vibrio anguillarum 3-oxo-C10-HSL Fish pathogen 49
LasI Pseudomonas aeruginosa 3-oxo-C12-HSL Opportunistic pathogen 14
SinI Sinorhizobium meliloti 3-oxo-C14-HSL Nitrogen-fixing symbiont 47
3-oxo-9-cis-C16-HSL
Xenorhabdus nematophilus 3-hydroxy-C4-HSL Nematode symbiont 18
PhzI Pseudomonas fluorescens 3-hydroxy-C6/C8-HSL Phenazine producer 40
CinI Rhizobium leguminosarum 3-hydroxy-7-cis-C14-HSL Nitrogen-fixing symbiont 46
servation between any two members of the terminal region of the enzyme is less conserved
AHL synthase family is typically more than overall and appears to be less important for
20% identity/40% similarity (80). In fact, the P. activity of RhlI compared to LuxI for reasons
aeruginosa enzyme LasI, a member of the class that are still not clear (32,54).However,we now
of proteobacteria, is more similar in sequence know that this region is involved in recognition
to enzymes from the and classes (≥30% of the most variable part of the acyl-ACP sub-
identity/≥50% similarity) than to EsaI, which is strate, the acyl chain, which explains the degree
also from the class. Despite these differences, of variability in this region. Thus, the entire
there is a single sequence signature that defines LuxI-type AHL synthase family will likely share
the LuxI-type AHL synthase family.These con- a similar structure and mechanism of AHL syn-
served sequence blocks and conserved amino thesis, but what will differ is the manner by
acid residues are apparent in an alignment of which they recognize the varying substrate
four selected AHL synthases shown in Fig. 3A. acyl-ACP.
The N-terminal region (amino acid residues
1 to 100) of the LuxI-type AHL synthase is the STRUCTURAL ANALYSIS OF
most conserved part of the enzyme.There are LuxI-TYPE AHL SYNTHASES
eight invariant residues, including Arg24, X-ray crystallographic structural analyses of
Phe28,Trp34,Asp45,Asp48,Arg68, Glu97, and two LuxI-type AHL synthases form the basis of
Arg100 (using the numbering based on the the current molecular understanding of AHL
Pantoea stewartii enzyme EsaI [Fig. 3A]), in this synthesis. The structure of the EsaI enzyme
region of the enzyme that are crucial for enzy- from P. stewartii was determined from the native
matic activity. Of these, Arg24, Asp45, Asp48, sequence (80, 81). The structure of the LasI
Arg68, Glu97, Ser99, and Arg100 were shown enzyme from P. aeruginosa was determined from
to be essential for AHL synthesis based on an active form that had been engineered to
mutagenesis analysis of LuxI, RhlI, and EsaI improve solubility and crystallization properties
enzymes (32, 54, 80). The high degree of (28, 29). EsaI produces a 3-oxo-C6-HSL, and
sequence conservation in this region suggested LasI produces a significantly longer AHL, 3-
a role in the binding of the conserved substrate oxo-C12-HSL, which both contribute to the
SAM and in catalysis (32). In contrast, the C- quorum-sensing regulation of pathogenicity in
their respective organisms (3, 56). EsaI (Fig. 3B) interactions between N-terminal helices and
and LasI (Fig. 3C) are divergent AHL synthases, the front surface of the -sheet but may also
with 28% identity/44% similarity.Their struc- perform a catalytic function.
tures show remarkably similar core elements, The LasI structure (Fig. 3C) suggested a role
but there are some notable differences in the N- for the remaining conserved AHL synthase sig-
terminal region of the enzyme. nature residues. The loop between 2 and 3
The AHL synthase is composed of a mixed that contains the conserved residues Trp34 and
–– sandwich with a V-shaped cleft and a Phe28 (EsaI numbering) positions them
deep cavity/tunnel (Fig. 3B and C). The pre- directly in front of the active site.These residues
dominantly antiparallel -sheet has a promi- may be conserved for the purpose of binding
nent V-shaped cleft (V cleft in the figure) SAM, perhaps by helping position the SAM
between the parallel -strands 4 and 5.A - amine nitrogen for N-acylation at the C-1
bulge between residues 99 and 100 in 4 (EsaI position of acyl-ACP.The aromatic side chains
numbering) also appears to be a unique and would allow a stacked-ring sandwich type of
conserved feature of the AHL synthases (28, binding with the adenine of SAM, as is com-
80).There is a structurally conserved core,commonly observed in SAM-binding enzymes.The
posed of 74 residues within the beta-strands N-terminal region of EsaI is in a more open
1, 2, 3, 4, 5, 6, and 8, and -helices conformation compared to LasI, but the con-
3 and 4, which has a root mean square devi- served residues of EsaI must be brought into
ation between LasI and EsaI of less than 1 Å proximity of the active site for catalysis to
(28). Helices 6 and 7 are also conserved but occur. That this region of the AHL synthase
occupy slightly different positions in the struc- varies greatly in structure between EsaI and
tures. Portions of the relatively unstructured LasI may be attributed to crystal packing forces,
loop between strands 3 and 4 are also con- but it also indicates the degree of potential flex-
served. Interestingly, only three of the con- ibility in this region of the enzyme and suggests
served AHL synthase signature residues,Arg68, that conformational changes take place when
Glu97, and Arg100, are found in this struc- the substrates bind to the AHL synthase in the
turally conserved core (Fig. 3B).The V-shaped course of the reaction.
cleft is the active site of the enzyme, and its base
consists of a hydrophobic cavity that is the site THE AHL SYNTHASE ACTIVE SITE
of acyl-chain binding (6, 28, 80). To date, there are no structures of acyl-ACP
Adjacent to the V-shaped cleft is an electro- bound to any enzyme, and thus the binding
static cluster in the core of the N-terminal site of the acyl chain, the phosphopantethiene
region of EsaI and LasI (Fig. 3B and C). This moiety, and protein core of ACP within the
cluster contains a complex network of ion pairs AHL synthase had to be interpreted from struc-
formed by the side chains from six of the eight tural analysis of other enzymes and the loca-
conserved residues in the AHL synthase family tions of the conserved residues. Fortunately, the
(Fig. 3A and C). Mutagenesis analysis of LuxI, structure of the highly conserved core domain
RhlI,and EsaI enzymes demonstrated that most of the AHL synthases is shared with other
of the residues in this cluster, Arg24, Asp45, enzymes, such as the GCN5-related histone N-
Asp48, Arg68, Glu97, Ser99, and Arg100 (in acetyltransferases (GNATs) (12, 80) and amino
EsaI numbering), are essential for AHL synthe- acid acyltransferases (76; reviewed in reference
sis (32, 54, 80). Ser99, which is conserved as 11).In both the acyl-ACP and acetyl-CoA sub-
either serine or threonine in all known LuxI- strates used by these enzymes,the terminal thiol
like AHL synthases, lies at the center of this of phosphopantetheine forms a thioester bond
cluster and interacts directly with Arg68 and a to either a variable length acyl chain or an
bridging water molecule bound to Glu97.This acetyl group. Holo- and acyl-ACP carry phos-
cluster appears to contribute to the stability of phopantetheine via a phosphodiester bond to
the hydroxyl oxygen atom of Ser36. In acetyltransferase Esa1 (not to be confused with the
CoA, however, phosphopantetheine forms a AHL synthase EsaI),which proceeds via a thioa-
pyrophosphate linkage to the 5′ phosphate of cyl covalent intermediate (83). No cysteine
adenosine 3′5′-diphosphate. The common residues were seen in the AHL synthase active
phosphopantetheine portion of acetyl-CoA site (80) or found to be important for catalysis
in the GCN5-acetyl-CoA complex (61) occu- (32, 54), which ruled out the long-standing
pies the central catalytic cleft that is highly hypothesis that AHL synthesis would proceed
conserved structurally with EsaI and LasI (Fig. through a covalent thio-acyl-enzyme interme-
3B and C). This close structural relationship diate through an active-site cysteine (51).
of the AHL synthases with the GNATs and Cyclization of the methionine moiety of
amino acid acyltransferases has provided great SAM gives rise to the lactone portion of the
insight into how the AHL synthases may bind AHL through a lactonization reaction. There
phosphopantetheine and carry out acylation are multiple chemical mechanisms by which
reactions. lactonization could take place. Tipton and
coworkers examined the mechanism of lac-
ENZYMATIC MECHANISM tonization of SAM using deuterium incorpora-
OF AHL SYNTHESIS tion with the RhlI enzyme (59) (Fig. 4C). In
The reaction mechanism of AHL synthesis has their study,no deuterons were taken up into the
two main steps: acylation and laconization (Fig. product when the reaction was conducted in
2A). The model of the acyl-phosphopanteth- deuterated buffer. This observation ruled out
eine of acyl-ACP in the active sites of EsaI and indirect lactonization mechanisms, such as
LasI places the acyl chain C-1 carbonyl oxygen elimination, which would proceed through an
within hydrogen bonding distance of the back- intermediate such as N-butyrylvinylglycine.
bone amides of residues 100 and 101 that form Therefore, the lactonization reaction most
an unusual -bulge in theV-shaped cleft (Fig.4). likely proceeds through direct nucleophilic
Therefore, the current proposed mechanism of attack on the C- of SAM by the carboxylate
N-acylation by AHL synthase is similar to that of oxygen via an SN2 chemical reaction (59). For
the GNATs, which involves proton abstraction the SN2 reaction to occur (illustrated in Fig.
from the substrate amine to be acylated by an 4C), the methionine moiety must be in a near
activated water molecule (33, 73) (Fig. 4B).This cyclic conformation (20). It is currently
water molecule acts as a base to abstract a proton unknown which residues contribute to this
from the amine of SAM, making it a better specific conformation. In addition, the available
nucleophile for direct attack on the C-1 atom of data do not distinguish between the alternative
the acyl chain. In support of this, the acyl O-1 reaction pathways: acylation followed by lac-
carbonyl oxygen that will form an oxyanion tonization of acyl-SAM, SAM lactonization
during the acylation reaction could be stabilized prior to acylation, or where acylation and lac-
by the -bulge, as described earlier. Further- tonization are concerted reaction.
more, as seen in all of the GNAT and AHL syn-
thase structures,there is a water molecule bound INTRINSIC SPECIFICITY OF AHL
to a conserved glutamic acid residue (Glu97 in PRODUCTION BY AHL SYNTHASES
EsaI). There is a second possible N-acylation A wide variety of AHL signals are produced in
mechanism that is used by the histone acetyl- nature (Table 1).They vary greatly in acyl-chain
FIGURE 3 Structure of the AHL synthases EsaI and LasI. (A) Sequence and topology diagram of the AHL syn-
thases LasI and EsaI.The gray shaded regions are the most conserved sequence blocks within the AHL synthase fam-
ily. The eight conserved residues of the AHL synthase family are highlighted in black, and those that are similar
among AHL synthases are boxed in white. Below the sequences are shown the alpha-helices and beta-strands
observed in the structures. (B and C) Ribbon diagrams depict the backbone structures of EsaI (B) and LasI (C).The
conserved residues are drawn in dark gray. The active site V-shaped clefts are indicated by arrows.
length (67), from the C4-HSL produced by P. binding pocket is an enclosed cavity, composed
aeruginosa RhlI (57), up to C18-HSLs produced of a shell of residues directly surrounding the
by Sinorhizobium meliloti SinI (47) (Table 1). cavity, including Ser98, Met126, Thr140,
AHLs produced by different bacterial species Val142, Met146, and Leu176. Numerous other
also vary in the degree of oxidation at the AHL residues within the protein core that contact
C-3 position.The preference for unsubstituted, these cavity residues but not the hexanoyl chain
3-oxo-, or 3-hydroxy-acyl-ACPs is thought to also direct the size and shape of the cavity
be due to the intrinsic selectivity of the AHL through hydrophobic packing (Fig. 5C) (80).
synthase for a particular subset of a pool of The importance of these two layers of residues
available acyl-ACP substrates. For example, the was demonstrated by Palva and coworkers, who
AHL synthase,LasI,produces predominantly 3- identified mutations in the gene encoding
oxo-C12-HSL, whereas RhlI produces an ExpI(SCC1) of Erwinia carotovora that increased
unsubstituted, C4-HSL from the same cellular the acyl-chain lengths of the AHLs produced
pool of acyl-ACPs. in vivo (6). These mutations occurred at one
The preference of AHL synthases for 3-oxo- position that lines the acyl-chain binding
substituted acyl-ACP substrates appears to be pocket and one that is in the next layer of
due to hydrogen-bonding interactions between residues that would not directly contact the
the C-3 carbonyl of 3-oxo-hexanoyl-ACP and acyl-chain (80).Thus, acyl-chain length restric-
residues in the acyl-chain binding site. In EsaI tion plays a role in production of AHLs with
and LasI, two hydrogen bonds are predicted to relatively short acyl chains.
form between the enzyme and the 3-oxo group The LasI structure provided a less clear
of acyl-ACP. One of these occurs from a threo- explanation for the selectivity of AHL synthases
nine O1, at amino acid position 140 in EsaI for acyl-ACPs with long acyl chains.The LasI
(Fig. 4A), and the other is from the main-chain acyl-chain binding pocket is actually an elon-
carbonyl group of residue 141. For all enzymes gated tunnel through the enzyme that is
that are known to produce 3-oxo-AHLs, the formed by hydrophobic residues at similar posi-
residue at this position is a threonine (Fig. 3A). tions in the enzyme as those in the EsaI pocket
Substitution of this threonine to alanine or (Fig. 5D) (28).Although the same residue posi-
other amino acids in EsaI and LasI, respectively, tions are involved in formation of both acyl-
increased the amount of the unsubstituted AHL chain binding sites, the different sizes of the
produced relative to 3-oxo-AHL, an indication hydrophobic side chains and slight changes in
of some lost degree of specificity for the oxo- the orientations of 6 and 8 contribute to the
group (Fig. 5A and B) (27). Notably, the structure of the closed pocket in EsaI versus the
enzymes that make 3-hydroxy-AHLs or unsub- tunnel observed in LasI. The residues in EsaI
stituted AHLs have either a glycine or alanine, that occlude the pocket are larger than those in
or occasionally a serine, at that position (40). the same position in LasI, which limits the acyl
The structural studies of EsaI provided an chain size to a C6 acyl-chain. Interestingly for
explanation for the selectivity of AHL synthases LasI, the tunnel places no steric restriction on
for acyl-ACPs with short acyl-chain lengths. the acyl-chain length of acyl-ACPs that could
Sequence analysis of the AHL synthase family bind to LasI. However, there is still a rather nar-
has failed to reveal a robust correlation between row distribution of AHLs produced by LasI in
sequence composition and acyl-chain length. vivo, which raises the question of how LasI
However, to accommodate and create a prefer- restricts the population of AHLs produced to
ence for acyl-ACPs of different acyl-chain those with the appropriate length of acyl chain.
lengths, sequence and tertiary structure differ-
ences must occur in different AHL synthases. MODULATION OF AHL LEVELS
For example, EsaI produces 3-oxo-C6-HSL, AND SPECIFICITY
which has a relatively short acyl chain. The Synthesis of AHLs appears to be modulated
structure of EsaI revealed that the acyl-chain both quantitatively and qualitatively. Multiple
FIGURE 4 Schematic diagram of the proposed AHL synthesis mechanism. (A) The interactions predicted to
form between acyl-phosphopantetheine bound in the active site of EsaI are depicted as dotted lines for hydrogen
bonds and curves for other types of interactions. (B) The proposed mechanism of acylation involves a direct nucle-
ophilic attack by the amine of SAM on the C-1 position of the acyl chain. (C) The mechanism of lactonization is a
direct nucleophilic attack on the C- position of SAM by the carboxylate oxygen atom,which will produce the lac-
tone and release the product 5′-methylthioadenosine.
AHL synthases and AHLs are produced at dif- (16). On top of this, the expression levels of the
ferent times in a variety of organisms (20). In P. AHL synthases are regulated in complex but as
aeruginosa, 3-oxo-C12-HSL is produced first, yet incompletely understood ways (reviewed
and as quorum sensing progresses, the primary in chapters 9,10,12-16,18, and 20 and refer-
AHL becomes C4-HSL, the product of RhlI ences 7 and 64). For example, in some organ-
FIGURE 5 Specificity of AHL synthases. (A) Comparison of AHLs produced by EsaI and the EsaI T140A
mutant, as analyzed by mass spectrometry (27). (B) Comparison of AHLs produced by LasI and LasI-substitution
mutants at the equivalent position T142 (27). (C) The acyl-chain cavity/tunnel is shown as a surface representation
that is visible though the protein ribbon diagram, shown in gray for both EsaI (C) (80) and LasI (D) (28).
isms,such as P.stewartii,the AHL synthase gene is have differential activity as substrates for the
constitutively expressed (3, 5). However, in AHL synthase RhlI (59). It is currently not
other organisms, the genes for AHL synthases known whether the same specificity occurs
are highly regulated. For example, in P. aerugi- with LasI or for other AHL synthases where
nosa the expression of the RhlI enzyme is under multiple ACP genes exist. However, analyses of
the positive control of the LasR transcriptional mutants of the VqsR gene (a novel quorum-
regulator, which requires 3-oxo-C12-HSL, the sensing transcriptional regulator) show the
product of LasI. Furthermore, LasR also regu- important consequences of regulation of the
lates LasI expression, which forms a positive fatty acid biosynthesis cycle and metabolic state
feedback loop that rapidly amplifies the amount on AHL production. Null mutants of the gene
of LasI to increase the amount of AHLs pro- encoding VqsR fail to produce and/or secrete
duced early in the quorum-sensing process AHLs, which appears to be due to the fact
(65). One consequence of the production of that their ACP genes are significantly down-
multiple distinct types of AHLs is that their cog- regulated (37, 38).
nate R proteins regulate distinct sets of genes The levels and the types of AHLs in the cel-
that are required at different times during lular environment also change because of dif-
biofilm formation and virulence factor produc- ferent conditions of pH and temperature. As
tion. Other purposes of this complex pattern of AHLs are not very stable molecules, they are
signal production remain unclear, because there susceptible to hydrolysis by both chemical
are still functions as well as AHL receptors that and enzymatic means (19). Under conditions
remain to be identified. of alkaline pH, AHLs are hydrolyzed rapidly
AHL synthesis is sensitive to cellular meta- (9, 84). As such, in Erwinia species the AHLs
bolic changes. Both of the substrates of AHL may serve as a type of pH sensor, which alters
synthesis, SAM and acyl-ACP, are metabolic the effectiveness of their quorum-sensing sig-
products that are central to many cellular naling in an alkaline environment.Interestingly,
processes. Significant amounts of acyl-ACP, this sensitivity to hydrolysis of the AHL signal
particularly those carrying longer acyl chains, may be exploited by plants that respond to
are destined for lipopolysaccharide and mem- infection by producing alkaline substances at
brane phospholipid synthesis, and most likely the site of infection (9, 52).There is also some
do not accumulate to levels where they evidence that temperature modulates the AHLs
would be used by AHL synthases (34). How- available in the cellular environment for Yersinia
ever, at times in the cell when an AHL synthase species (1, 41).
is very highly expressed,it is likely that the pools AHL degradation machinery plays an
of acyl-ACP could be altered as the best sub- unclear role in terminating the AHL signal.
strates for the AHL synthase are depleted. AHLs are susceptible to enzymatic degradation
This would lead to shifts in the types of AHLs by hydrolysis of the lactone ring and deacyla-
that the bacterium could produce. Indeed, in P. tion. Both acylases (36, 43, 69) and lactonases
aeruginosa the pool of acyl-ACPs can be shifted have been discovered (17) (reviewed in chapter
toward those with shorter acyl-chain lengths 24) in a variety of species (60).The importance
by manipulating keto-acyl-ACP reductase of these systems for recycling the AHL signal
(FabG) activity (34). The exact distribution of into products of greater utility to the cell has
acyl-chain lengths in pools of P. aeruginosa been demonstrated (35). Furthermore, the util-
acyl-ACP is not known, and it is not known ity of these enzymes as AHL-degrading quo-
how these may change during the quorum- rum-quenching therapeutic or agricultural
sensing process. agents is promising (69). However, the role of
The population of substrates can be modu- these enzymes in bacterial virulence is not
lated as a means to alter AHL production. P. completely clear (60).The intriguing potential
aeruginosa has at least three ACP genes, which for these enzymes to function in AHL regula-
tion in vivo and as quorum-quenching agents is Vibrio harveyi: sequence and function of genes regu-
an area of continued exploration. lating expression of luminescence. Mol. Microbiol.
9:773–786.
3. Beck von Bodman, S., and S. K. Farrand.
CONCLUSIONS AND PERSPECTIVES 1995. Capsular polysaccharide biosynthesis and
Although the phenotypes controlled by the pathogenicity in Erwinia stewartii require induction
AHL-dependent quorum-sensing system are by an N-acylhomoserine lactone autoinducer. J.
diverse and unique to each species, the key fea- Bacteriol. 177:5000–5008.
4. Beck von Bodman, S., G.T. Hayman, and S.
tures of AHL synthesis are relatively conserved K. Farrand. 1992.Opine catabolism and conjugal
among more than 90 gram-negative species. transfer of the nopaline Ti plasmid pTiC58 are
The structural and mechanistic analyses coordinately regulated by a single repressor. Proc.
described here have provided a basis for the Natl.Acad. Sci. USA 89:643–647.
understanding of the intrinsic specificity of 5. Beck von Bodman, S., D. R. Majerczak, and
D. L. Coplin. 1998.A negative regulator mediates
AHL synthases. In addition, they have revealed quorum-sensing control of exopolysaccharide
features of the enzymes that are important production in Pantoea stewartii subsp stewartii. Proc.
for correct AHL synthesis and have suggested Natl.Acad. Sci. USA 95:7687–7692.
new mechanisms by which AHL synthesis 6. Brader, G., S. Sjoblom, H. Hyytiainen, K.
may be modulated. Importantly, natural Sims-Huopaniemi, and E. T. Palva. 2005.
Altering substrate chain length specificity of an
and synthetic mechanisms that inhibit or mis- acylhomoserine lactone synthase in bacterial com-
regulate quorum sensing can provide broad- munication. J. Biol. Chem. 280: 10403–10409.
spectrum control of particular bacterial diseases 7. Brencic, A., and S. C.Winans. 2005. Detection
(5, 17, 43, 70). AHL-mediated signaling also of and response to signals involved in host-microbe
functions in mixed bacterial populations; in interactions by plant-associated bacteria. Microbiol.
Mol. Biol. Rev. 69:155–194.
bacterial interactions with eukaryotic hosts 8. Burton, E. O., H. W. Read, M. C. Pellitteri,
such as the bobtail squid, plants, and yeast; and W. J. Hickey. 2005. Identification of acyl-
and also in patients infected with P. aeruginosa homoserine lactone signal molecules produced by
(reviewed in reference 72). Therefore, how Nitrosomonas europaea strain Schmidt. Appl. Envi-
the metabolic and environmental perturbations ron. Microbiol. 71:4906–4909.
9. Byers, J. T., C. Lucas, G. P. Salmond, and M.
alter the “language” spoken by a particular Welch. 2002. Nonenzymatic turnover of an
bacterium has broad and important biological Erwinia carotovora quorum-sensing signaling mole-
consequences. cule. J. Bacteriol. 184:1163–1171.
10. Chernin, L. S., M. K. Winson, J. M.
ACKNOWLEDGMENTS Thompson, S. Haran, B. W. Bycroft, I. Chet,
P. Williams, and G. S. Stewart. 1998. Chiti-
We are grateful to current and past contributors to this nolytic activity in Chromobacterium violaceum:
work, including Susanne Beck von Bodman, John Cro- substrate analysis and regulation by quorum
nan, Ty Gould, Robert Murphy, Herbert Schweizer, sensing. J. Bacteriol. 180:4435–4441.
Dale Val, and Bill Watson, as well as Paul Kirwan,Tim 11. Churchill, M. E. A. 2006.A new GNAT in bac-
Minogue, Linda Farb, and other members of the terial signaling? Structure 14:1342–1344.
Churchill laboratory.We thank Susanne Beck von Bod- 12. Clements, A., and R. Marmorstein. 2003.
man for helpful comments on the manuscript. We Insights into structure and function of
appreciate the support from the American Heart Asso- GCN5/PCAF and yEsa 1 histone acetyltransferase
ciation, Cystic Fibrosis Foundation, and the National domains. Methods Enzymol. 371:545–564.
Institutes of Health. 13. Cullinane, M., C. Baysse, J. P. Morrissey,
and F. O’Gara. 2005. Identification of two
REFERENCES lysophosphatidic acid acyltransferase genes with
1. Atkinson, S., J. P. Throup, G. Stewart, and P. overlapping function in Pseudomonas fluorescens.
Williams. 1999. A hierarchical quorum-sensing Microbiology 151:3071–3080.
system in Yersinia pseudotuberculosis is involved in the 14. Davies, D. G., M. R. Parsek, J. P. Pearson,
regulation of motility and clumping. Mol. Microbiol. B. H. Iglewski, J. W. Costerton, and E. P.
33:1267–1277. Greenberg. 1998.The involvement of cell-to-cell
2. Bassler, B. L., M.Wright, R. E. Showalter, and signals in the development of bacterial biofilm.
M. R. Silverman. 1993. Intercellular signalling in Science 280:295–298.
15. De Kievit,T. R., and B. H. Iglewski. 2000.Bac- aeruginosa acyl-homoserinelactone synthase LasI.
terial quorum sensing in pathogenic relationships. Mol. Microbiol. 53:1135–1146.
Infect. Immun. 68:4839–4849. 29. Gould, T. A., W. T. Watson, S. H. Choi, H. P.
16. De Kievit,T. R.,Y. Kakai, J. K. Register, E. C. Schweizer, and M. E.A. Churchill. 2004.Crys-
Pesci, and B. H. Iglewski. 2002. Role of the tallization of Pseudomonas aeruginosa AHL synthase
Pseudomonas aeruginosa las and rhl quorum-sensing LasI using beta-turn crystal engineering. Acta
systems in rhlI regulation. FEMS Microbiol. Lett. Crystallogr. D60:518–520.
212:101–106. 30. Hanzelka, B. L., and E. P. Greenberg. 1996.
17. Dong, Y.-H., L.-H. Wang, J.-L. Xu, H.-B. Quorum sensing in Vibrio fischeri: evidence
Zhang, X.-F. Zhang, and L.-H. Zhang. 2001. that S-adenosylmethionine is the amino acid
Quenching quorum-sensing-dependent bacterial substrate for autoinducer synthesis. J. Bacteriol.
infection by an N-acyl homoserine lactonase. 178:5291–5294.
Nature 411:813–817. 31. Hanzelka, B. L., M. R. Parsek, D. L.Val, P.V.
18. Dunphy, G., C. Miyamoto, and E. Meighen. Dunlap, J. E. J. Cronan, and E. P. Greenberg.
1997.A homoserine lactone autoinducer regulates 1999.Acylhomoserine lactone synthase activity of
virulence of an insect-pathogenic bacterium, the Vibrio fischeri AinS protein. J. Bacteriol.
Xenorhabdus nematophilus (Enterobacteriaceae). J. Bac- 181:5766–5770.
teriol. 179:5288–5291. 32. Hanzelka, B. L., A. M. Stevens, M. R. Parsek,
19. Eberhard,A.,A. L. Burlingame, C. Eberhard, T. J. Crone, and E. P. Greenberg. 1997. Muta-
G. L. Kenyon, K. H. Nealson, and N. J. tional analysis of the Vibrio fischeri LuxI polypep-
Oppenheimer. 1981. Structural identification of tide: critical regions of an autoinducer synthase. J.
autoinducer of Photobacterium fischeri. Biochemistry Bacteriol. 179:4882–4887.
20:2444–2449. 33. Hickman, A. B., M. A. Namboodiri, D. C.
20. Eberl, L. 2006. Quorum sensing in the genus Klein, and F. Dyda. 1999. The structural basis
Burkholderia. Int. J. Med. Microbiol. 296:103–110. of ordered substrate binding by serotonin N-
21. Engebrecht, J., and M. Silverman. 1984. Iden- acetyltransferase: enzyme complex at 1.8 Å resolu-
tification of genes and gene products necessary for tion with a bisubstrate analog. Cell 97:361–369.
bacterial bioluminescence. Proc. Natl. Acad. Sci. 34. Hoang, T. T., S. A. Sullivan, J. K. Cusick, and
USA 81:4154–4158. H. P. Schweizer. 2002. Beta-ketoacyl acyl carrier
22. Freeman, J. A., B. N. Lilley, and B. L. Bassler. protein reductase (FabG) activity of the fatty acid
2000.A genetic analysis of the functions of LuxN: biosynthetic pathway is a determining factor of 3-
a two-component hybrid sensor kinase that regu- oxo-homoserine lactone acyl chain lengths. Micro-
lates quorum sensing in Vibrio harveyi. Mol. Micro- biology 148:3849–3856.
biol. 35:139–149. 35. Huang, J. J., J. I. Han, L. H. Zhang, and J. R.
23. Fuqua, C., and A. Eberhard. 1999. Signal gen- Leadbetter. 2003.Utilization of acyl-homoserine
eration in autoinduction systems: synthesis of acy- lactone quorum signals for growth by a soil
lated homoserine lactones by LuxI-type proteins, pseudomonad and Pseudomonas aeruginosa PAO1.
p. 211–230. In G. Dunny and S. C. Winans (ed.), Appl. Environ. Microbiol. 69:5941–5949.
Cell-Cell Communication in Bacteria. ASM Press, 36. Huang, J. J.,A. Petersen, M.Whiteley, and J. R.
Washington, DC. Leadbetter. 2006. Identification of QuiP, the
24. Fuqua, C., M. R. Parsek, and E. P. Greenberg. product of gene PA1032, as the second acyl-
2001. Regulation of gene expression by cell- homoserine lactone acylase of Pseudomonas aerugi-
to-cell communication: acyl-homoserine lactone nosa PAO1.Appl.Environ.Microbiol. 72:1190–1197.
quorum sensing. Annu. Rev. Genet. 35:439–468. 37. Juhas, M., L. Eberl, and B. Tummler. 2005.
25. Fuqua, W. C., S. C. Winans, and E. P. Quorum sensing: the power of cooperation in the
Greenberg. 1996. Census and consensus in world of Pseudomonas. Environ. Microbiol.
bacterial ecosystems: the LuxR-LuxI family of 7:459–471.
quorum-sensing transcriptional regulators. Annu. 38. Juhas, M., L. Wiehlmann, P. Salunkhe, J.
Rev. Microbiol. 50:727–751. Lauber, J. Buer, and B. Tummler. 2005.
26. Gilson, L.,A. Kuo, and P.V. Dunlap. 1995.AinS GeneChip expression analysis of the VqsR regulon
and a new family of autoinducer synthesis proof Pseudomonas aeruginosaTB.FEMS Microbiol.Lett.
teins. J. Bacteriol. 177:6946–6951. 242:287–295.
27. Gould,T. A., J. Herman, J. Krank, R. C. Mur- 39. Kaplan, H. B., and E. P. Greenberg. 1985. Dif-
phy, and M. E.A. Churchill. 2006. Specificity of fusion of autoinducer is involved in the regulation
acyl-homoserine lactone syntheses examined by of the Vibrio fischeri luminescence system. J. Bacte-
mass spectrometry. J. Bacteriol. 188:773–783. riol. 163:1210–1214.
28. Gould, T. A., H. P. Schweizer, and M. E. A. 40. Khan, S. R., D.V. Mavrodi, G. J. Jog, H. Suga,
Churchill. 2004. Structure of the Pseudomonas L. S. Thomashow, and S. K. Farrand. 2005.
Activation of the phz operon of Pseudomonas evidence for a repressor function. Mol. Microbiol.
fluorescens 2–79 requires the LuxR homolog PhzR, 44:1625–1635.
N-(3-OH-hexanoyl)-L-homoserine lactone pro- 51. Moré, M. I., L. D. Finger, J. L. Stryker, C.
duced by the LuxI homolog PhzI, and a cis-acting Fuqua, A. Eberhard, and S. C. Winans. 1996.
phz box. J. Bacteriol. 187:6517–6527. Enzymatic synthesis of a quorum-sensing autoin-
41. Kirwan, J. P.,T. A. Gould, H. P. Schweizer, S. ducer through use of defined substrates. Science
W. Bearden, R. C. Murphy, and M. E. 272:1655–1658.
Churchill. 2006. Quorum-sensing signal synthe- 52. Nachin, L., and F. Barras. 2000.External pH:an
sis by the Yersinia pestis acyl-homoserine lactone environmental signal that helps to rationalize pel
synthase YspI. J. Bacteriol. 188:784–788. gene duplication in Erwinia chrysanthemi. Mol.
42. Laue, B. E.,Y. Jiang, S. R. Chhabra, S. Jacob, Plant-Microbe Interact. 13:882–886.
G. S. Stewart, A. Hardman, J. A. Downie, 53. Nealson, K. H. 1977.Autoinduction of bacterial
F. O’Gara, and P. Williams. 2000.The biocon- luciferase: occurrence, mechanism and signifi-
trol strain Pseudomonas fluorescens F113 produces cance. Arch. Microbiol. 112:73–79.
the Rhizobium small bacteriocin, N-(3-hydroxy- 54. Parsek, M. R., A. L. Schaefer, and E. P.
7-cis-tetradecenoyl)homoserine lactone, via Greenberg. 1997. Analysis of random and site-
HdtS, a putative novel N-acylhomoserine lactone directed mutations in rhlI, a Pseudomonas aeruginosa
synthase. Microbiology 146:2469–2480. gene encoding an acyl homoserine lactone
43. Leadbetter, J. R., and E. P. Greenberg. 2000. synthase. Mol. Microbiol. 26:301–310.
Metabolism of acyl-homoserine lactone quorum- 55. Parsek, M. R., D. L.Val, B. L. Hanzelka, J. E. J.
sensing signals by Variovorax paradoxus. J. Bacteriol. Cronan, and E. P. Greenberg. 1999. Acyl
182:6921–6926. homoserine-lactone quorum-sensing signal gen-
44. Lerat, E., and N. A. Moran. 2004. The evolu- eration. Proc. Natl.Acad. Sci. USA 96:4360–4365.
tionary history of quorum-sensing systems in bac- 56. Pearson, J. P., L. Passador, B. H. Iglewski,
teria. Mol. Biol. Evol. 21:903–913. and E. P. Greenberg. 1995.A second N-acylho-
45. Lewenza, S., B. Conway, E. P. Greenberg, and moserine lactone signal produced by Pseudomonas
P. A. Sokol. 1999. Quorum sensing in Burkholde- aeruginosa. Proc Natl Acad Sci USA 92:1490–1494.
ria cepacia: identification of the LuxRI homologs 57. Pearson, J. P., E. C. Pesci, and B. H. Iglewski.
CepRI. J. Bacteriol. 181:748–756. 1997. Roles of Pseudomonas aeruginosa las and rhl
46. Lithgow, J. K., A. Wilkinson, A. Hardman, quorum-sensing systems in control of elastase and
B. Rodelas, F. Wisniewski-Dye, P. Williams, rhamnolipid biosynthesis genes. J. Bacteriol.
and J. A. Downie. 2000. The regulatory locus 179:5756–5767.
cinRI in Rhizobium leguminosarum controls a net- 58. Puskas,A., E. P. Greenberg, S. Kaplan, and A.
work of quorum-sensing loci. Mol. Microbiol. L. Schaefer. 1997. A quorum-sensing system in
37:81–97. the free-living photosynthetic bacterium
47. Marketon, M. M., M. R. Gronquist, A. Rhodobacter sphaeroides. J. Bacteriol. 179:7530–7537.
Eberhard, and J. E. Gonzalez. 2002. Character- 59. Raychaudhuri, A., A. Jerga, and P. A.
ization of the Sinorhizobium meliloti sinR/sinI locus Tipton. 2005.Chemical mechanism and substrate
and the production of novel N-acyl homoserine specificity of RhlI, an acylhomoserine lactone
lactones. J. Bacteriol. 184:5686–5695. synthase from Pseudomonas aeruginosa. Biochemistry
48. Milton, D. L.,V. J. Chalker, D. Kirke,A. Hard- 44:2974–2981.
man, M. Camara, and P. Williams. 2001.The 60. Roche, D. M., J. T. Byers, D. S. Smith, F. G.
LuxM homologue VanM from Vibrio anguillarum Glansdorp, D. R. Spring, and M.Welch. 2004.
directs the synthesis of N-(3-hydroxyhexanoyl) Communications blackout? Do N-acylhomoser-
homoserine lactone and N-hexanoylhomoserine ine-lactone-degrading enzymes have any rolein
lactone. J. Bacteriol. 183:3537–3547. quorum sensing? Microbiology 150:2023–2028.
49. Milton, D. L., A. Hardman, M. Camara, S. R. 61. Rojas, J. R., R. C.Trievel, J. Zhou,Y. Mo, S. L.
Chhabra, B. W. Bycroft, G. S. Stewart, and Berger, C. D. Allis, and R. Marmorstein.
P. Williams. 1997. Quorum sensing in Vibrio 1999. Structure of Tetrahymena GCN5 bound to
anguillarum: characterization of the vanI/vanR coenzyme A and a histone H3 peptide. Nature
locus and identification of the autoinducer N-(3- 401:93–98.
oxodecanoyl)-L-homoserine lactone. J. Bacteriol. 62. Schaefer, A. L., D. L.Val, B. L. Hanzelka, J. E.
179:3004–3012. Cronan, Jr., and E. P. Greenberg. 1996. Gener-
50. Minogue, T. D., M. Wehland-von Trebra, F. ation of cell-to-cell signals in quorum sensing: acyl
Bernhard, and S. B. von Bodman. 2002. homoserine lactone synthase activity of a purified
The autoregulatory role of EsaR, a quorum- Vibrio fischeri LuxI protein. Proc. Natl. Acad. Sci.
sensing regulator in Pantoea stewartii ssp. stewartii: USA 93:9505–9509.
63. Schuster, M., and E. P. Greenberg. 2006. A genicity of Burkholderia mallei. Infect. Immun.
network of networks: quorum-sensing gene regu- 72:6589–6596.
lation in Pseudomonas aeruginosa. Int. J. Med. Micro- 75. Val, D. L., and J. E. J. Cronan. 1998. In vivo evi-
biol. 296:73–81. Epub 14 Feb 2006. dence that S-adenosylmethionine and fatty acid
64. Schuster, M., C. P. Lostroh, T. Ogi, and E. P. synthesis intermediates are the substrates for the
Greenberg. 2003. Identification, timing, and sig- LuxI family of autoinducer synthases. J. Bacteriol.
nal specificity of Pseudomonas aeruginosa quorum- 180:2644–2651.
controlled genes: a transcriptome analysis. J. 76. Van Wagoner, R. M., and J. Clardy. 2006.FeeM,
Bacteriol. 185:2066–2079. an N-acyl amino acid synthase from an uncultured
65. Seed, C. P., L. Passador, and B. H. Iglewski. soil microbe: structure, mechanism, and acyl carrier
1995.Activation of the Pseudomonas aeruginosa lasI protein binding. Structure 14:1425–1435.
gene by LasR and the Pseudomonas autoinducer PAI: 77. von Bodman, S. B., J. K. Ball, M. A. Faini, C.
an autoinduction regulatory hierarchy. J. Bacteriol. M. Herrera, T. D. Minogue, M. L.
177:654–659. Urbanowski, and A. M. Stevens. 2003. The
66. Shadel, G. S., and T. O. Baldwin. 1991.The Vib- quorum sensing negative regulators EsaR and
rio fischeri LuxR protein is capable of bidirectional ExpR(Ecc), homologues within the LuxR family,
stimulation of transcription and both positive and retain the ability to function as activators of tran-
negative regulation of the luxR gene. J. Bacteriol. scription. J. Bacteriol. 185:7001–7007.
173:568–574. 78. Wagner,V. E., D. Bushnell, L. Passador, A. I.
67. Shaw, P. D., G. Ping, S. L. Daly, C. Cha, J. E. J. Brooks, and B. H. Iglewski. 2003. Microarray
Cronan, K. L. Rinehart, and S. K. Farrand. analysis of Pseudomonas aeruginosa quorum-sensing
1997. Detecting and characterizing N-acyl- regulons:effects of growth phase and environment.
homoserine lactone signal molecules by thin-layer J. Bacteriol. 185:2080–2095.
chromatography. Proc. Natl. Acad. Sci. USA 79. Waters, C. M., and B. L. Bassler. 2005. Quo-
94:6036–6041. rum sensing: cell-to-cell communication in bacte-
68. Shih, G. C., C. M. Kahler, J. S. Swartley, M. M. ria. Annu. Rev. Cell Dev. Biol. 21:319–346.
Rahman, J. Coleman, R.W. Carlson, and D. S. 80. Watson,W.T.,T. D. Minogue, D. L.Val, S. Beck
Stephens. 1999. Multiple lysophosphatidic acid von Bodman, and M. E. A. Churchill. 2002.
acyltransferases in Neisseria meningitidis. Mol. Micro- Structural basis and specificity of acyl-homoserine
biol. 32:942–952. lactone signal production in bacterial quorum
69. Sio, C. F., L. G. Otten, R. H. Cool, S. P. Diggle, sensing. Mol. Cell. 9:685–694.
P. G. Braun, R. Bos, M. Daykin, M. Camara, 81. Watson, W. T., F. V. Murphy, T. A. Gould, P.
P. Williams, and W. J. Quax. 2006. Quorum Jambeck, D. L.Val, J. E. Cronan, S. Beck von
quenching by an N-acyl-homoserine lactone acy- Bodman, and M. E. A. Churchill. 2001. Crys-
lase from Pseudomonas aeruginosa PAO1. Infect. tallization and rhenium MAD phasing of the acyl-
Immun. 74:1673–1682. homoserine lactone synthase EsaI. Acta Crystallogr.
70. Smith, K. M.,Y. Bu, and H. Suga. 2003. Induc- D57:1945–1949.
tion and Inhibition of Pseudomonas aeruginosa quo- 82. Whitehead, N. A., A. M. Barnard, H. Slater,
rum sensing by synthetic autoinducer analogs. N. J. Simpson, and G. P. Salmond. 2001. Quo-
Chem. Biol. 10:81–89. rum sensing in gram-negative bacteria. FEMS
71. Stevens, A. M., and E. P. Greenberg. 1999. Microbiol. Rev. 25:365–404.
Transcriptional activation by LuxR, p. 231–242. In 83. Yan,Y., S. Harper, D.W. Speicher, and R. Mar-
G.Dunny and S.C.Winans (ed.),Cell-Cell Commu- morstein. 2002. The catalytic mechanism of the
nication in Bacteria. ASM Press,Washington, DC. ESA1 histone acetyltransferase involves a self-acety-
72. Taga, M. E., and B. L. Bassler. 2003. Chemical lated intermediate. Nat. Struct. Biol. 9:862–869.
communication among bacteria. Proc. Natl. Acad. 84. Yates, E. A., B. Philipp, C. Buckley, S.
Sci. USA 100:14549–14554. Atkinson, S. R. Chhabra, R. E. Sockett, M.
73. Tanner, K. G., R. C. Trievel, M.-H. Huo, R. Goldner, Y. Dessaux, M. Camara, H. Smith,
M. Howard, S. L. Berger, C. D. Allis, R. and P. Williams. 2002. N-acylhomoserine lac-
Marmorstein, and J. M. Denu. 1999. Catalytic tones undergo lactonolysis in a pH-, temperature-,
mechanism and function of invariant glutamic acid and acyl chain length-dependent manner during
173 from the histone acetyltransferase GCN5 growth of Yersinia pseudotuberculosis and Pseudomonas
transcriptional coactivator. J. Biol. Chem. aeruginosa. Infect. Immun. 70:5635–5646.
274:18157–18160. 85. Zhang, L., P. J. Murphy, A. Kerr, and M. E.
74. Ulrich, R. L., D. Deshazer, H. B. Hines, and J. Tate. 1993.Agrobacterium conjugation and gene
A. Jeddeloh. 2004. Quorum sensing: a transcrip- regulation by N-acyl-L-homoserine lactones.
tional regulatory system involved in the patho- Nature 362:446–448.
CELL-CELL SIGNALING
WITHIN CROWN GALL TUMORS
Stephen C.Winans
18
It has been 100 years since Agrobacterium tumefa- gation systems for horizontal transfer of the
ciens was demonstrated to cause crown gall plasmid, although these systems are expressed
tumors at plant wound sites (57). General inter- only in the presence of specific opines (24, 36).
est in this organism increased with the observa- Conjugation therefore can only occur in the
tion, published in a series of papers almost 30 crown gall tumor environment, as these tumors
years ago, that it can directly transform plant are the only natural sources of these com-
cells by transferring oncogenic fragments of pounds.The type of opine required for conju-
DNA (17). This transformation causes the gation depends on the type of Ti plasmid (24,
infected plant cells to overproduce phytohor- 36), and octopine is the conjugal opine for so-
mones, causing cell proliferation, which results called octopine-type Ti plasmids, while agro-
in neoplasias called crown galls.The transform- cinopines A and B are the conjugal opines for
ing DNA (T-DNA) also encodes genes for the the so-called nopaline-type Ti plasmids.
production of novel compounds called opines, Since these discoveries, it has been shown
which are sources of nutrients for the coloniz- that the Ti plasmids also carry a quorum-
ing bacteria. By inducing tumors and directing sensing system related to the LuxR-LuxI sys-
these tumorous cells to produce specific nutri- tem of Vibrio fischeri (described in chapter 15).
ents that few other bacteria can use, A. tumefa- The basic components of the system include
ciens makes a novel niche for itself in its TraI, which synthesizes N-3-oxooctanoyl-L-
environment.Since these early discoveries,plant homoserine lactone (OOHL) (22, 29, 43),
transformation using A. tumefaciens has become and the OOHL receptor TraR (22, 51, 74).
a cornerstone for plant molecular genetics (23). Within or near crown galls, OOHL accumu-
Almost all of the genes required for tumori- lates to a level that permits the formation of
genesis are carried on large (approximately TraR-OOHL complexes, which then activate
200 kb) plasmids (34, 35, 61). These plasmids, transcription of target genes on the Ti plasmid,
called Ti plasmids,also encode complete conju- including those required for Ti plasmid conju-
gation (20) and for vegetative replication (49).
Stephen C. Winans Department of Microbiology, Cornell It was later shown that expression of traR
University, Ithaca, New York 14853. genes requires particular opines called conjugal
291
292 ■ WINANS
opines (1, 21), which fully explained why con- homologues are found in the genome sequence
jugation requires opines. Two additional proof C58.Virtually nothing is known about these
teins, TraM and TrlR, act as antiactivators of latter genes, and they can be considered
TraR (8, 48, 78). In this review, we first present “orphan” receptors, as there are no correspon-
background information relevant to quorum ding luxI-type genes found on the chromo-
sensing in A. tumefaciens and then focus on our some, and traI mutants appear not to produce
current knowledge of the molecular biology of detectable levels of any acyl-homoserine lac-
the TraR-TraI system. tone (AHL). Genome sequences of two addi-
tional strains, the pathogen A. vitis and the
A.TUMEFACIENS, THE Ti PLASMIDS, nonpathogenic biocontrol strain A. radiobacter,
AND CROWN GALL DISEASE have recently been completed (http://depts.
A. tumefaciens is a member of the Rhizobiaceae, washington.edu/agro/).
which is in the alpha subgroup of the pro- Ti plasmids from different isolates of A.tume-
teobacteria. All members of this family are soil faciens are traditionally classified by opine type.
bacteria and are traditionally divided into two However, a variety of different opine synthases
groups.The rhizobia, which include the genera and their corresponding catabolism genes are
Rhizobium, Sinorhizobium, Mesorhizobium, encoded by each isolate (77).The best-studied
Azorhizobium, and Bradyrhizobium, are all are the nopaline-type Ti plasmids (including
nitrogen-fixing symbionts of legumes and form pTi37 and pTiC58) and the octopine-type Ti
root nodules that are packed with diazotrophic plasmids (including pTiA6, pTiB6, pTiAch5,
bacteroids. The agrobacteria include the plant pTi15966, and pTiR10, which are virtually
pathogens A. tumefaciens,A. rubi,A. vitis, and A. identical).A composite DNA sequence of these
rhizogenes, as well as the nonpathogenic A. plasmids has been published (77). The genes
radiobacter (2). The most important differences required for virulence (the vir genes),for conju-
between members of the rhizobia and the gation (tra and trb genes), and for vegetative
agrobacteria lie in their plasmids rather than replication (rep) are highly conserved between
their chromosomes, and it was therefore pro- the octopine-type and nopaline-type Ti plas-
posed that the genus Agrobacterium be aban- mids, whereas genes required for opine synthe-
doned and that all species be placed in the genus sis and catabolism are dissimilar.
Rhizobium (73),although this proposal has so far The expression of genes required for
not gained widespread acceptance. LuxR- tumorigenesis (vir genes) is induced in response
LuxI-type systems have been described for a to signals released from plant wounds (pheno-
number of the rhizobial symbiosis plasmids (see lics, sugars, and a pH range of 5.0 to 5.5) (31).
chapter 14). These signals are detected byVirA, a transmem-
The complete genome sequence of A. tume- brane sensor kinase, which then phosphorylates
faciens strain C58 was published by two groups its cognate response regulator, VirG (58, 69).
(25, 71). The genome is unusual in that it has Phospho-VirG then activates expression of the
both a circular and a linear chromosome that vir gene promoters, and Vir proteins direct T-
together carry all of the essential and house- DNA processing from the Ti plasmid and its
keeping genes of the organism.The linear chro- transfer into plant host cells (23). In the host
mosome contains a plasmid-like replication cell cytoplasm, the single-stranded T-DNA is
system, suggesting that it may have evolved targeted to the nucleus and integrated into
from a plasmid (25, 71). C58 also carries two host genomic DNA. T-DNA transfer requires
dispensable circular plasmids, pAtC58 (also direct contact between the bacterium and plant
called the “cryptic” plasmid, 545 kb), and the Ti cell, and sequence analysis shows that the trans-
plasmid pTiC58 (214 kb). The traR and traI fer apparatus has evolved from a bacterial con-
genes of the quorum-sensing system are both jugation system (33, 37, 38, 53). However, the
located on pTiC58, while four additional luxR vir system and the tra and trb genes required
18. CELL-CELL SIGNALING WITHIN CROWN GALL TUMORS ■ 293
for Ti plasmid conjugation are not closely Control of traR expression by opines there-
related. T-DNA transfer and integration into fore has evolved independently in these two
plant nuclear DNA have been the subjects of types of Ti plasmids.The genes of the arc operon
intense study, and we refer the reader to recent and occ operon are not similar, except for traR.
reviews (17, 23). LysR and AccR are also dissimilar, as OccR is a
LysR-type transcriptional activator that binds
QUORUM SENSING to promoter DNA both in the presence and
IN A.TUMEFACIENS absence of the inducing signal (65). In contrast,
Regulation of TraR activity is complex and AccR is a somewhat distant relative of the Lac
occurs at the levels of transcription, protein repressor of Escherichia coli (1). Binding of agro-
folding, resistance to proteolysis, and the cinopine to AccR blocks its DNA binding and
formation of quaternary complexes with other relieves AccR-mediated repression (39).
TraR subunits or with two different anti- Control of traR expression by opines is a fea-
activators of TraR. In this section, we provide ture of all Ti plasmids that have been studied to
an overview of the many factors involved in date (Fig. 2). Regulation of traR expression on
regulation. We then discuss current advances pTiChry5 is also thought to be through an
in the biochemical and structural biology of AccR homologue, although in this case dere-
some of these individual proteins.The proper- pression occurs in response to agrocinopines
ties of TraI and its relatives are discussed in C and D (46). Agrocinopines C and D are
chapter 17. also known to induce conjugal transfer of
pTiBo542,although in this case the mechanism
Regulation of traR Gene Expression of regulation is not known (16).The nonpatho-
Two groups independently discovered that the genic strain A. radiobacter K84 carries a plasmid
traR gene is regulated by conjugal opines on called pATK84, whose conjugation is regulated
both the nopaline and octopine-type Ti plas- by two different traR genes (47). One copy,
mids. On nopaline-type Ti plasmids, traR is the traRnoc, is in the nox operon, which is induced
fourth gene in the five-gene arc operon, which by nopaline, whereas traRacc expression is
is divergently transcribed from the acc operon induced by agrocinopines A and B. It is striking
(1). The transcriptional regulator AccR is that the control of traR by opines could have
encoded by the first gene of the acc operon, the evolved independently so many times, espe-
other members of which are required for catab- cially given that we have no idea why there
olism of the opines agrocinopine A and B. In should be any selection whatever.
the presence of agrocinopine A or B, repression
of both the arc and acc operons by AccR is Posttranscriptional Control
relieved, resulting in transcription of all of the of TraR Activity
genes of these operons,including traR (1;52).In It is becoming more clear that many bacterial
octopine-type Ti plasmids, the conjugal opine and eukaryotic proteins are intrinsically
is octopine (24, 36). Octopine binds to its intra- unstructured in vivo and require either low-
cellular target, the transcriptional regulator molecular-weight ligands or other proteins to
OccR, resulting in activation of the occ operon assist in correct folding. These unstructured
(Fig. 1) (21).The traR gene is at the distal end of proteins are often nonfunctional and sometimes
this operon, while many of the genes upstream are targeted for proteolysis. Some of these are
of traR are required for uptake and catabolism involved in time-dependent processes, includ-
of octopine (21).When traR is expressed from a ing control of transcription (56). Incorporation
constitutive promoter, conjugation no longer of ligand into the folding process may also opti-
requires octopine.Therefore, regulation of traR mize the specificity of the protein-ligand inter-
expression by OccR fully explains the require- action. The term “intrinsically unstructured”
ment of octopine for Ti plasmid conjugation. was coined by Dyson and Wright to describe
294 ■ WINANS
FIGURE 1 A model of the quorum-sensing system in octopine-type Ti plasmids.TraR is expressed in response

to the tumor-released nutrient octopine, whereas the TraR antiactivator TrlR is expressed in response to
mannopine,which is also released from tumors.Apo-TraR is rapidly degraded by proteases but is rescued from pro-
teolysis by binding OOHL,the quorum-sensing signal that is produced by TraI.TraR-OOHL dimers activate tran-
scription of traM and the tra, trb, and rep operons. TraR-OOHL complexes can be inactivated through direct
interactions with TraM or TrlR. OOHL can be destroyed by the BlcC protein (formerly AttM).
proteins that require ligands or other interac- in vivo.When TraR is strongly overexpressed in
tions for correct folding and maturation (14). either E. coli or A. tumefaciens, it accumulates but
Accumulating evidence indicates that TraR forms only insoluble inclusion bodies. If the
is intrinsically unstructured and requires protein is only mildly expressed, it does not
OOHL for folding, stability, and accumulation accumulate in the absence of OOHL,but rather
FIGURE 2 A comparison of traR regulation via opines on different types of Ti plasmids. On octopine-type Ti
plasmids, traR is activated by OccR in response to octopine.The TraR antiactivator TrlR is expressed in response to
mannopine,probably via inactivation of the MocR repressor.On nopaline-type Ti plasmids,traR is expressed when
AccR repression is relieved by agrocinopines A and B. Regulation of traR on the chrysopine-type plasmid is simi-
lar, except the inducing opines are agrocinopines C and D.There are two copies of traR on pAtK84b, one thought
to be activated by NocR in response to nopaline, while the other is activated in response to agrocinopines A and B
via derepression of AccR.
is degraded by cytoplasmic proteases (80).This apoprotein was targeted for proteolysis

led to the hypothesis that TraR requires OOHL extremely rapidly and that OOHL added
for stabilization against proteolysis. Therefore, directly after the pulse of radiolabel did not res-
OOHL must trigger some conformational cue the protein from proteolysis. OOHL must
change in the protein upon binding. Pulse- therefore stabilize TraR during or immediately
chase experiments were used to show that the after translation and,therefore,is likely to be part
296 ■ WINANS
of the protein-folding process itself. It was also expression of trlR also results in a decrease in
demonstrated that purified apo-TraR is rapidly conjugation, demonstrating that the inhibitory
degraded by trypsin, whereas purified TraR- effect of mannopine occurs solely through TrlR
OOHL complexes were much more resistant to (8). Favored catabolites, including succinate,
degradation, and is cleaved preferentially at a glutamine, and tryptone, block trlR expression
linker sequence that bridges two domains (80). (8).This led to speculation that TrlR functions
The data described above suggested that if to attenuate the energetically expensive process
inducing concentrations of OOHL are present, of conjugation when nutrients are limiting.
TraR is stabilized in the cell and can activate
transcription of target genes (79, 80). However, BIOCHEMICAL AND STRUCTURAL
the activity of stable TraR-OOHL complexes STUDIES OF TraR
can still be directly inhibited by two antiactiva- The realization that TraR requires OOHL for
tors, TraM and TrlR. Of these, TraM is con- stability, folding, and solubility allowed overex-
served among all known Ti and Ri plasmids of pression and purification of stable TraR-
Agrobacterium spp. as well as several other plas- OOHL complexes from E. coli. Purified
mids found in rhizobia (33), while TrlR appears TraR-OOHL complexes were used to show
to be found only on octopine-type Ti plasmids that OOHL binds TraR monomers in a 1:1
(26). For both the nopaline and octopine-type mole ratio and that these complexes form
Ti plasmids, null mutations in traM result in a homodimers in solution (54, 80).TraR-OOHL
hyperconjugal phenotype, while traM overex- complexes also bind as dimers to 18-bp
pression has the opposite effect (18, 28). TraR sequences (called tra boxes) on the Ti plasmid
activity in TraM mutants still requires OOHL with relatively high affinity and specificity (79).
but occurs at lower OOHL concentrations The crystal structure of TraR-OOHL-DNA
than in wild-type cells. Thus, it is clear that complexes was published by two groups in the
TraM blocks TraR activity, and subsequent same year, and to date, they are the only crystal
studies showed that this occurs through direct structures of LuxR-type proteins available
protein-protein interactions (30, 59). In both Ti (63, 75), although the structure of the SdiA
plasmids,traM expression is activated by TraR in protein of E. coli was solved by nuclear mag-
the presence of OOHL, creating a negative netic resonance spectroscopy (72). The two
feedback loop (18,28).Transcriptional profiling TraR structures are nearly identical and con-
experiments showed that traM is also modestly firmed earlier predictions that the protein binds
induced by plant-released phenolic compounds to DNA as a dimer and that each monomer has
that induce the vir regulon (12). Perhaps this is a two domains, an N-terminal OOHL-binding
mechanism to avoid concurrent expression and domain and a C-terminal DNA-binding
activity of the vir and conjugation systems.Both domain (Color Plate 11A).The N-terminal and
include a type IV secretion system that is C-terminal domains are connected by a 12-
required for DNA transfer. It is possible that residue unstructured linker. The N-terminal
these two systems could interfere with each domain has an -- sandwich structure, with
other if expressed at the same time. one molecule of OOHL embedded between
The trlR gene lies near the distal end of the the -sheet and a layer of -helix (63, 75). In
mot operon of octopine-type Ti plasmids that fact, the ligand is completely engulfed in the
includes the genes required for mannopine core of the protein and makes no contact with
uptake and catabolism (Fig. 1 and 2) (48, 78). bulk solvent.Therefore, it is likely that OOHL
Expression of this operon is activated by is incorporated during the protein-folding
mannopine, possibly via the MocR protein process itself as previously predicted or, less
(Fig. 1) (48, 78). Mannopine attenuates conju- likely, that major structural rearrangements
gation in at least two such plasmids in a TrlR- must occur in TraR upon OOHL binding.
dependent manner (48, 78). Constitutive The contacts themselves between TraR and
OOHL are extensive and include hydrophobic The two TraR structures were crystallized
interactions and hydrogen bonds between the with the 18-base-pair consensus tra box, which
polar groups of the OOHL and hydrophilic has perfect dyad symmetry (63, 75).This DNA
residues in the binding pocket (Color Plate in the crystal is B-form DNA, but with a
11B) (63, 75). These contacts further support smooth 30 bend toward the protein.Although
the model that OOHL stabilizes the TraR pro- this bend is modest compared to that of other
tein in vivo by participating in the protein- transcription factors, it still results in greater
folding process itself. buried surface area between the protein and
The smaller C-terminal domain of each DNA than would be possible with unbent
monomer consists of four helices joined by DNA (32).While the protein-DNA interface is
three random coils (Color Plate 11A) (63, 75). extensive, the sequence-specific contacts are
The second and third helices of this domain surprisingly few and are mediated by three
constitute a canonical helix-turn-helix DNA- amino acid residues of each TraR subunit
binding motif (44).The sequence and structure (found on the second helix of the helix-turn-
of this domain are highly conserved and place helix) and only four nucleotide bases of each tra
the LuxR family within the larger NarL-FixJ box half site, located in the major groove (63,
superfamily of bacterial transcription regula- 75). Additional protein-DNA contacts were
tors (19). All of these proteins have similar also found between TraR and the DNA back-
DNA-binding domains but differ in their signal bone,but these ought not contribute to binding
transduction domains. The high-resolution specificity.An exhaustive analysis of the contri-
structures of two additional proteins in this bution of each base to binding affinity showed
superfamily have been reported. One is the E. that every base in the tra box contributes to
coli NarL protein, crystallized with its DNA- affinity, even those that do not make sequence-
binding site, and the other is the Bacillus subtilis specific contact with TraR. These bases must
GerE protein (13, 42). The DNA-binding contribute by a phenomenon known as “indi-
domains of TraR, NarL, and GerE proteins can rect readout,” facilitating the DNA topology
be superimposed. required for high-specificity binding (66).
There are two dimerization interfaces in There is significant interest in understanding
the TraR-OOHL dimer (Color Plate 11A). how LuxR-type proteins discriminate cognate
The most extensive of these is composed pri- AHLs from heterologous ones. When TraR is
marily of a long -helix in the N-terminal expressed at native levels in vivo, it detects
domain of each subunit that is parallel with the OOHL with extremely high specificity and
same helix of the opposite subunit (63, 75). detects no other naturally occurring AHL (76).
A number of residues buried at this interface The polar contacts in the binding pocket
appear to contribute to dimer formation, a include four hydrogen bonds between residues
finding confirmed by mutational analysis of TraR and the polar groups of OOHL (Color
(41). A less-extensive dimerization interface Plate 11B). Two mutational studies of TraR
occurs between the C-terminal helices of confirmed that these hydrogen bonds are criti-
the DNA-binding domain. The interdomain cal to OOHL binding and stabilization of the
linker causes extensive flexibility of the two protein (6, 41). One of these studies considered
N-terminal domains of the TraR dimer and further the role of the 3-oxo group (which is
the two C-terminal domains. This flexibility not common to all AHLs) in binding specificity
caused a pronounced asymmetry in each com- (6).This group makes a water-mediated hydro-
plex. The C-terminal domains of each dimer gen bond to a threonine residue in the binding
have a twofold axis of symmetry, and the N- pocket of TraR (75). Point mutations of the
terminal domains also have an axis of twofold threonine residue were constructed in an
symmetry, but these two axes lie at a 90 angle attempt to increase hydrophobicity, which was
to each other. predicted to exclude the water molecule from
298 ■ WINANS
the binding pocket and therefore increase the of two antiparallel alpha-helices. A significant
relative affinity for a 3-unsubstituted AHL (C8- hydrophobic interface is buried at the interface
AHL). The data largely supported this predic- between the two subunits. This extensive
tion, though absolute affinity was decreased for dimerization interface had also been confirmed
both AHLs. In the same study, point mutations in a study using deletion mutants of TraM (55).
were also constructed in an attempt to increase The two groups that crystallized TraM pub-
the binding affinity for short-chain AHLs by lished different models of how TraR-TraM
increasing the hydrophobic bulk in the binding interactions might disrupt TraR-DNA interac-
pocket.A number of these mutations did result tions. One model suggests that formation of an
in altered specificity; however, the stability of inactive heterodimer requires dissociation of
these mutants was also decreased.This result is TraR and TraM (9). This is consistent with
not surprising, as these residues are buried in mutational studies that showed that residues of
the hydrophobic core of the protein and are TraM that are important for initial binding to
likely to be important for protein folding and TraR are different from those required for TraR
stabilization in ways that cannot be compen- inactivation (59). Furthermore, when the two
sated with a short-chain AHL. subunits of TraM dimers are covalently cross-
linked to each other, the altered dimers are still
The Antiactivators TrlR and TraM able to bind TraR but are not able to disrupt
and the OOHL Lactonase BlcC DNA binding (9). A different model proposed
Above, we introduced TraM as an inhibitor of that two dimers of TraM bind to two dimers of
TraR activity (18, 28). Overproduction of TraR TraR, such that the TraM dimers are “clamped”
can fully overcome inhibition, suggesting that between the TraR dimers in such as way as to
TraM acts by making stoichiometric contacts block DNA binding (62). Structural studies of
with TraR. Direct interactions were confirmed TraR-TraM complexes may resolve these mod-
by yeast two-hybrid assays and by far Western els, although static structures may not provide a
immunoblots (37). The same study used dele- complete picture, as TraM binding to TraR and
tion and point mutations to show that the pro- inactivation of TraR are expected to occur in
tein-protein interactions occurred at the sequential steps rather than simultaneously.
C-terminal regions of both proteins (30).TraR- The other antiactivator of TraR activity,
TraM complexes form with high affinity (in the TrlR, appears to use a much less-complicated
nanomolar range) and are stable (59).The num- mechanism for TraR inhibition.The trlR gene
ber of TraR and TraM subunits per complex is thought to have originated from a gene
remains controversial.Two groups used gel fil- duplication event of traR itself (48).This gene is
tration chromatography to determine the mass virtually identical to traR except for a single
of TraR-TraM complexes.The results from one frameshift difference upstream of the C-
study (9) suggested that one TraR monomer terminal DNA-binding domain. Genes identi-
binds to one or two TraM monomers (H.Cho cal to trlR have been identified on a number of
and S. C. Winans, unpublished data), whereas octopine-type Ti plasmids that were isolated
the results of the other study (62) were consis- independently, indicating that the truncation is
tent with two TraR dimers binding to two not a laboratory artifact (78). Like TraR, TrlR
TraM dimers (78). Purified TraM can block requires OOHL to fold into a stable, soluble
binding of TraR to tra box DNA and can also protein. Although TrlR cannot activate tra
cause dissociation of preformed TraR-DNA genes, correction of the frameshift mutation by
complexes (40).These data led to the suggestion site-directed mutagenesis resulted in a fully
that TraM binds directly to the DNA-binding active protein (78).The similarity of TrlR to the
surface of TraR,making it inaccessible to DNA. TraR N-terminal domain led to the suggestion
Two published studies describe the crystal that TrlR inhibits TraR activity directly through
structure of TraM (9, 62).TraM was crystallized formation of inactive heterodimers (which
as a dimer, with each subunit consisting largely would have only one functional DNA-binding
domain rather than two). This prediction was merase (RNAP)-binding site, is reminiscent of
confirmed using purified proteins (8). TrlR class II promoters, first described for cyclic
could,in principle,also inhibit TraR by titrating AMP receptor protein (CRP) (3).
OOHL. The trb genes of the traI-trb operon encode
The BlcC protein (formerly AttM) is also the type IV secretion system required for con-
thought to inhibit TraR activity, although jugal DNA transfer (20).The OOHL synthase,
through an entirely different mechanism. BlcC traI, is the first gene of this operon.Activation of
encodes a lactonase that destroys the biological traI expression by TraR-OOHL complexes rep-
activity of OOHL by hydrolytic ring opening. resents a positive feedback loop, as has been
The gene encoding BlcC is part of the blcABC described for a number of other LuxR-LuxI-
operon (formerly attKLM), which encodes a type systems (68). The traI-trb operon is diver-
pathway for the catabolism of gamma-butyro- gently transcribed from the repABC operon
lactone, and is discussed in chapter 24. The (20, 49). The traI-trb operon and one repABC
ability of BlcC to inactivate OOHL and other promoter (repAP1) are activated by TraR bind-
AHLs is most likely to be incidental to the ing to tra box II. Both of these promoters are
activity for which this enzyme was selected. On also class II promoters. An additional TraR-
the other hand, induction of this gene by the dependent class II promoter (repAP3) has been
catabolic intermediates 3-hydroxybutyrate and identified for repABC,activated from tra box III.
succinic semialdehyde blocks TraR activity, Some promoters, designated class I pro-
which could have ecologically significant moters, are activated by proteins that bind to
effects during tumor colonization. sites centered approximately 65 nucleotides
upstream of the transcription start site (3).
TraR as an Activator of Transcription There are two class I TraR-dependent promot-
We recently used gene arrays to profile the TraR ers on the octopine-type Ti plasmid (3). The
transcriptome and found that all induced genes repAP2 promoter is activated when TraR binds
were Ti plasmid-encoded (11). These genes to tra box II, which is centered 66.5 nucleotides
include the tra and trb genes, which are involved upstream of the transcription start site. The
in conjugal transfer; the rep genes, which are traM promoter is activated when TraR binds
required for vegetative replication and plasmid to a site designated tra box IV, which lies 66.5
partitioning into daughter cells; and traM. The nucleotides upstream of the traM transcription
operon structure, promoters, and tra boxes of start site.
these genes have all been described.There are a Most tra boxes themselves are highly similar
total of seven known TraR-dependent promot- and differ from the canonical sequence by one
ers and four tra boxes for five operons (20, 49, or two nucleotides. The affinities of TraR-
67). The tra genes, required for DNA transfer OOHL complexes for tra boxes I, II, and III are
and replication, are arranged in two operons: very similar and in the nanomolar range (49,
traAFBH and traCDG-yci (20). These operons 79).The tra box IV is the least conserved of all
are divergently transcribed, and the origin of four tra boxes, perhaps explaining the observa-
conjugation lies directly upstream of traA. Both tion that the traM promoter is expressed more
operons are activated by TraR binding to a sin- weakly than any other (67).Activation of TraR-
gle tra box, called tra box I (whose sequence is dependent promoters requires only promoter
identical to the consensus tra box sequence and DNA, TraR-OOHL complexes, and 70-
consists of a perfect dyad symmetry).At both of RNAP (79), indicating that TraR makes direct
these promoters, the tra box is centered approx- contacts with RNAP. These contacts should
imately 45 nucleotides upstream from the tran- serve to recruit RNAP to the promoter or may
scription start site and overlaps the 35 facilitate a subsequent step in initiation. A
element of each promoter. This type of pro- potential RNAP contact site, or activating
moter structure, with the activator-binding site region, which includes at least six residues, was
overlapping the distal end of the RNA poly- identified by mutagenesis of the surface of the
300 ■ WINANS
TraR C-terminal domain (67). This activating

region overlaps with a patch on the surface of
LuxR that is also critical for activation (15, 67),
and its position is reminscent of the AR3 region
of CRP. Both the AR3 region of CRP and the
AR region of LuxR are thought to contact the
sigma subunit of RNAP. However, mutants of
the TraR activation region are defective at both
class II and class I promoters, suggesting that
they interact with the alpha subunit rather than
the sigma subunit.These data are hard to recon-
cile. On the one hand, the TraR AR region
should bind the alpha-CTD, as it is needed for
class I promoters. On the other hand, its posi-
tion on the surface of TraR suggests an interac-
tion with the sigma subunit.This question will
be answered empirically in future studies.
As second possible activating region was
identified by mutations of asparate 10 and
glycine 123, which lie closely together on the
NTD of this protein (49,67).The author’s labo-
ratory has begun a systematic analysis of this
part of the protein,and preliminary data suggest
that additional residues may contribute to this
potential activating region. By analogy to CRP,
such a region could interact with the N-
terminal domain of the alpha subunit of
RNAP. However, the flexibility of the TraR
dimer could allow other interactions as well.
Expression of Some TraR- FIGURE 3 (A) Induction of the tra regulon using
Activated Genes Is Also Influenced full-genome microarrays. Black circles indicate genes
by Other Regulators located on the circular or linear chromosome or on the
It has recently become clear that some of the cryptic Ti plasmid pAtC58, while gray circles indicate
operons known to be directly regulated by TraR genes located on the octopine-type Ti plasmid pTiR10.
The x axis shows the range of expression in the absence
are also subject to regulation by additional proof TraR, while the y axis shows the range of expression
teins, some of which are activators, while others when TraR is overproduced, which suffices to induce
are repressors. A good example of this is found the entire regulon (22).Note that virtually all Ti plasmid
in the repABC operon.As described above,TraR genes are expressed more strongly in the presence than
binds to two sites upstream of repA to regulate in the absence of TraR, as TraR activates the repABC
operon, increasing Ti plasmid copy number.(B) Induc-
three promoters of the rep operon in addition to tion of the vir regulon using Ti plasmid microarrays.
the divergent traI-trb promoter. Activation of Open symbols represent non-Ti plasmid genes, while
repABC by TraR causes the copy number of the filled symbols represent Ti plasmid genes. The x axis
Ti plasmid to increase about sixfold (49) and represents the range of gene expression in the absence
indirectly causes the expression of every Ti plas- of acetosyringone (AS), while the y axis represents the
range of expression in the presence of 100 M AS.Note
mid-encoded gene to increase a similar amount that all Ti plasmid genes are slightly induced by AS,as AS
(Fig. 3A and 4A).This operon is also subject to acts through VirG to activate the repABC operon,
at least three additional levels of control. First, increasing the Ti plasmid copy number.
FIGURE 4 Interactions between

TraR and other regulators. (A) TraR
and VirG both regulate the repABC
operon and therefore influence the
Ti plasmid copy number. TraR-
OOHL complexes bind to tra boxes
II and III (tbII and tbIII) to activate
promoters P1, P2, and P3, as well as
the divergent promoter PtraI.
P~VirG bind to a vir box (vb) to acti-
vate promoter P4. Of these, P4 is
active in the absence of either pro-
tein, causing a basal level of gene
expression sufficient for low-level Ti
plasmid replication. (B) RepA and
RepB form a complex that binds to
an operator directly downstream of
promoter P4 and also binds to a site
between repA and repB (not shown).
Binding represses all four promoters
of the repABC operon.RepC is neg-
atively regulated at the transcript
elongation or translational level by a
small antisense RNA encoded by
the repE gene. (C) The divergent
traCDG and traAFGB operons are
activated by TraR-OOHL com-
plexes bound to the tra box I (tbI).
Both promoters are also repressed by
the TraA protein bound to the origin
of conjugal transfer (oriT).TraA also
processes one DNA strand at oriT in
a step that is essential for conjuga-
tion. TraC and TraD assist TraA in
repression, in DNA processing, and
in conjugation but are not essential
for any of these events.
activated VirG binds to a fourth promoter (P4) posttranscriptionally by a small antisense RNA
to activate repABC, causing a similar increase in that is encoded by repE, located on the opposite
plasmid copy number in response to wound- strand between repB and repC (Fig. 4B) (7).
released chemical signals (Fig. 3B and 4A) (12). Clearly, TraR is only one player in the overall
The repABC operon is also subject to negative regulation of this operon.
autoregulation by the RepA and RepB pro- A second example of promoters at which
teins, which bind to a site downstream of pro- TraR shares the stage with other regulators is
moter P4, and also to a site between repA and found in the divergent traA and traC promoters.
repB (Fig. 4B). Binding represses expression of As described above, TraR binds to a single tra
all four promoters (5, 50). Finally, repC, which box that overlaps the 35 region of each pro-
encodes the replication initiator, is regulated moter.The origin of conjugal transfer (oriT) lies
302 ■ WINANS
directly downstream of the traA transcription have seen that two antiactivators block TraR
start site.This site ought, in principle, to provide activity, but we do not know much about how
a binding site for the conjugal relaxase, which they work, especially how TraM disrupts pre-
is encoded by the traA gene.We have recently formed TraR-DNA complexes.Why does this
demonstrated that TraA represses both PtraA system benefit from these antiactivators, given
and PtraC by binding to this site and is essen- that they are not conserved? We are just begin-
tial for single-stranded scissions within this ning to glimpse the interactions between TraR
site. Repression of both promoters, oriT pro- and RNAP and have yet to learn which sub-
cessing, and conjugation itself are facilitated units of RNAP are contacted by TraR and
by TraC and TraD, although the latter two which steps in initiation are facilitated. Studies
processes occur at reduced rates in traC or traD with CRP suggest that both the initial binding
mutants (Fig. 4C). and promoter melting are enhanced (45).
A third possible example of two proteins act- We also wonder why quorum sensing is used
ing on one gene is found in traM.As described to regulate conjugation genes. Are conjugal
above,this gene is directly activated by TraR and donors truly measuring a quorum of other
encodes a TraR antiactivator.Microarray studies donors? Are they coordinating their conjuga-
in two laboratories have shown that traM is also tion activities? If so, why? As plasmids are gen-
weakly induced by activated VirG (12; D. W. erally thought to be self-interested genetic
Wood and E.W. Nester, personal communica- elements, they ought, in principle, benefit from
tion).This suggests that induction of the vir reg- conjugation at all times, not only in the tumor
ulon blocks induction of the tra regulon and environment or in the presence of a quorum of
helps explain the existence of TraM, given that donors. Perhaps these plasmids regulate conju-
most LuxR-type proteins function in the gation to minimize the energetic costs to their
absence of antiactivators.As described above,we host bacteria.Either way,it makes little sense for
already know that both TraR and VirG directly donors to monitor the activities of other
activate repABC, so the idea that they might donors. Perhaps conjugation intrinsically is
both activate traM has a clear precedent. How- more efficient when two or more donors mate
ever, to date it has been difficult to confirm this with a single recipient.This would be unprece-
activation using direct measurements of traM dented but is testable. Perhaps the intended
mRNA.The reasons for this are not clear. recipient is itself a donor rather than a plasmid-
free cell, although our preliminary experiments
PERSPECTIVES AND indicate that Ti plasmids, like other conjugal
FUTURE STUDIES
plasmids, block the conjugal entry of sibling
This review describes a current view of a work
plasmids (H. Cho and S. C. Winans, unpub-
in progress.The combined efforts of a number
lished data).While we can speculate about these
of geneticists, biochemists, and structural biolo-
riddles, they may never be fully solved.
gists have propelled the Ti plasmid quorum-
sensing system to the forefront of this family of ACKNOWLEDGMENTS
regulators. However, much work remains to be I thank the members of my laboratory for helpful dis-
done.We are just beginning to glimpse the fold- cussions and critical review of the manuscript.
ing process of TraR and the role of OOHL in Research in my laboratory is supported by the National
folding.We would like to know whether other Institutes of Health (grant GM41892).
LuxR homologues require AHLs for folding,
and the available evidence indicates that AHLs REFERENCES
are required for most of these proteins but that a 1. Beck von Bodman, S., G.T. Hayman, and S. K.
Farrand. 1992. Opine catabolism and conjugal
subset is stable (and fully active) as apoproteins transfer of the nopaline Ti plasmid pTiC58 are
and are actually inhibited by AHLs (4, 64; C. coordinately regulated by a single repressor. Proc.
Tsai and S. C. Winans, unpublished data). We Natl.Acad. Sci. USA 89:643–647.
2. Binns, A. N., and P. Costantino. 1998. The 15. Egland, K. A., and E. P. Greenberg. 2001.
Agrobacterium oncogenes, p. 251–266. In H. P. Quorum sensing in Vibrio fischeri: analysis of the
Spaink,A. Kondorosi, and P. J. Hooykaas (ed.), The LuxR DNA binding region by alanine-scanning
Rhizobiaceae: Molecular Biology of Model Plant- mutagenesis. J. Bacteriol. 183:382–386.
Associated Bacteria. Kluwer Academic Publishers, 16. Ellis, J. G., A. Kerr, A. Petit, and J. Tempe.
Dordrecht,The Netherlands. 1982. Conjugal transfer of nopaline and agropine
3. Busby, S., and R. H. Ebright. 1999.Transcrip- Ti-plasmids—the role of agrocinopines. Mol. Gen.
tion activation by catabolite activator protein Genet. 186:269–273.
(CAP). J. Mol. Biol. 293:199–213. 17. Escobar, M. A., and A. M. Dandekar. 2003.
4. Carlier, A. L., and S. B. von Bodman. 2006. Agrobacterium tumefaciens as an agent of disease.
The rcsA promoter of Pantoea stewartii subsp. stew- Trends Plant. Sci. 8:380–386.
artii features a low-level constitutive promoter and 18. Fuqua, C., M. Burbea, and S. C.Winans. 1995.
an EsaR quorum-sensing-regulated promoter. J. Activity of the Agrobacterium Ti plasmid conjugal
Bacteriol. 188:4581–4584. transfer regulator TraR is inhibited by the product
5. Chai,Y., and S. C.Winans. 2005. RepB protein of the traM gene. J. Bacteriol. 177:1367–1373.
of an Agrobacterium tumefaciens Ti plasmid binds to 19. Fuqua, C., and E. P. Greenberg. 2002. Listen-
two adjacent sites between repA and repB for plas- ing in on bacteria: acyl-homoserine lactone sig-
mid partitioning and autorepression. Mol. Micro- nalling. Nat. Rev. Mol. Cell. Biol. 3:685–695.
biol. 58:1114–1129. 20. Fuqua, C., and S. C. Winans. 1996. Conserved
6. Chai,Y., and S. C. Winans. 2004. Site-directed cis-acting promoter elements are required for
mutagenesis of a LuxR-type quorum-sensing density-dependent transcription of Agrobacterium
transcription factor: alteration of autoinducer tumefaciens conjugal transfer genes. J. Bacteriol.
specificity. Mol. Microbiol. 51:765–776. 178:435–440.
7. Chai,Y., and S. C. Winans. 2005. A small anti- 21. Fuqua, C., and S. C.Winans. 1996. Localization
sense RNA downregulates expression of an essen- of OccR-activated and TraR-activated promoters
tial replicase protein of an Agrobacterium tumefaciens that express two ABC-type permeases and the
Ti plasmid. Mol. Microbiol. 56:1574–1585. traR gene of Ti plasmid pTiR10. Mol. Microbiol.
8. Chai,Y., J. Zhu, and S. C.Winans. 2001.TrlR, a 20:1199–1210.
defective TraR-like protein of Agrobacterium tume- 22. Fuqua, W. C., and S. C. Winans. 1994. A
faciens, blocks TraR function in vitro by forming LuxR-LuxI type regulatory system activates
inactive TrlR:TraR dimers. Mol. Microbiol. Agrobacterium Ti plasmid conjugal transfer in the
40:414–421. presence of a plant tumor metabolite. J. Bacteriol.
9. Chen, G., J. W. Malenkos, M. R. Cha, C. 176:2796–2806.
Fuqua, and L. Chen. 2004. Quorum-sensing 23. Gelvin, S. B. 2003. Agrobacterium-mediated plant
antiactivator TraM forms a dimer that dissociates to transformation: the biology behind the “gene-
inhibit TraR. Mol. Microbiol. 52:1641–1651. jockeying”tool.Microbiol.Mol.Biol.Rev. 67:16–37.
10. Reference deleted. 24. Genetello, C., N.Van Larebeke, M. Holsters,
11. Cho, H., and S. C. Winans. 2007. TraA, TraC, A. De Picker, M.Van Montagu, and J. Schell.
and TraD autorepress two divergent quorum- 1977. Ti plasmids of Agrobacterium as conjugative
regulated promoters near the transfer origin of plasmids. Nature 265:561–563.
the Ti plasmid of Agrobacterium tumefaciens. Mol. 25. Goodner, B., G. Hinkle, S. Gattung, N. Miller,
Microbiol. 63:1769–1782. M. Blanchard, B. Qurollo, B. S. Goldman,Y.
12. Cho, H., and S. C.Winans. 2005.VirA and VirG Cao, M. Askenazi, C. Halling, L. Mullin, K.
activate the Ti plasmid repABC operon, elevating Houmiel, J. Gordon, M. Vaudin, O.
plasmid copy number in response to wound- Iartchouk, A. Epp, F. Liu, C. Wollam, M.
released chemical signals. Proc. Natl.Acad. Sci. USA Allinger, D. Doughty, C. Scott, C. Lappas, B.
102:14843–14848. Markelz, C. Flanagan, C. Crowell, J. Gurson,
13. Ducros, V. M., R. J. Lewis, C. S. Verma, C. Lomo, C. Sear, G. Strub, C. Cielo, and S.
E. J. Dodson, G. Leonard, J. P. Turkenburg, Slater. 2001. Genome sequence of the plant
G. N. Murshudov, A. J. Wilkinson, and J. A. pathogen and biotechnology agent Agrobacterium
Brannigan. 2001. Crystal structure of GerE, tumefaciens C58. Science 294:2323–2328.
the ultimate transcriptional regulator of spore 26. He, X., W. Chang, D. L. Pierce, L. O. Seib, J.
formation in Bacillus subtilis. J. Mol. Biol. Wagner, and C. Fuqua. 2003. Quorum sensing
306:759–771. in Rhizobium sp.strain NGR234 regulates conjugal
14. Dyson, H. J., and P. E.Wright. 2002. Coupling transfer (tra) gene expression and influences
of folding and binding for unstructured proteins. growth rate. J. Bacteriol. 185:809–822.
Curr. Opin. Struct. Biol. 12:54–60. 27. Reference deleted.
304 ■ WINANS
28. Hwang, I., D. M. Cook, and S. K. Farrand. Correlating function with molecular structure of a
1995. A new regulatory element modulates quorum-sensing transcriptional activator. J. Biol.
homoserine lactone-mediated autoinduction Chem. 278:13173–13182.
of Ti plasmid conjugal transfer. J. Bacteriol. 42. Maris, A. E., M. R. Sawaya, M. Kaczor-
177:449–458. Grzeskowiak, M. R. Jarvis, S. M. Bearson, M.
29. Hwang, I., P. L. Li, L. Zhang, K. R. Piper, D. L. Kopka, I. Schroder, R. P. Gunsalus, and R.
M. Cook, M. E.Tate, and S. K. Farrand. 1994. E. Dickerson. 2002. Dimerization allows DNA
TraI, a LuxI homologue, is responsible for produc- target site recognition by the NarL response regu-
tion of conjugation factor, the Ti plasmid N-acyl- lator. Nat. Struct. Biol. 9:771–778.
homoserine lactone autoinducer. Proc. Natl. Acad. 43. More, M. I., L. D. Finger, J. L. Stryker, C.
Sci. USA 91:4639–4643. Fuqua, A. Eberhard, and S. C. Winans. 1996.
30. Hwang, I., A. J. Smyth, Z. Q. Luo, and S. K. Enzymatic synthesis of a quorum-sensing autoin-
Farrand. 1999. Modulating quorum sensing by ducer through use of defined substrates. Science
antiactivation:TraM interacts with TraR to inhibit 272:1655–1658.
activation of Ti plasmid conjugal transfer genes. 44. Nelson, H. C. 1995. Structure and function of
Mol. Microbiol. 34:282–294. DNA-binding proteins. Curr. Opin. Genet. Dev.
31. Johnson,T. M., and A. Das. 1998. Organization 5:180–189.
and regulation of expression of the Agrobacterium 45. Niu, W., Y. Kim, G. Tau, T. Heyduk, and
virulence genes, p. 265–279. In H. P. Spaink, A. R. H. Ebright. 1996.Transcription activation at
Kondorosi, and P. J. Hooykaas (ed.), The Rhizobi- class II CAP-dependent promoters: two interac-
aceae: Molecular Biology of Model Plant-Associated tions between CAP and RNA polymerase. Cell
Bacteria. Kluwer Academic Publishers, Dordrecht, 87:1123–1134.
The Netherlands. 46. Oger, P., and S. K. Farrand. 2001.Co-evolution
32. Jones, S., P. van Heyningen, H. M. Berman, of the agrocinopine opines and the agrocinopine-
and J. M. Thornton. 1999. Protein-DNA inter- mediated control of TraR, the quorum-sensing
actions: a structural analysis. J. Mol. Biol. activator of the Ti plasmid conjugation system.
287:877–896. Mol. Microbiol. 41:1173–1185.
33. Kado, C. I. 1994. Promiscuous DNA transfer sys- 47. Oger, P., and S. K. Farrand. 2002.Two opines
tem of Agrobacterium tumefaciens: role of the virB control conjugal transfer of an Agrobacterium plas-
operon in sex pilus assembly and synthesis. Mol. mid by regulating expression of separate copies
Microbiol. 12:17–22. of the quorum-sensing activator gene traR. J.
34. Kerr, A. 1971. Acquisition of virulence by non- Bacteriol. 184:1121–1131.
pathogenic isolates of Agrobacterium radiobacter. 48. Oger, P., K. S. Kim, R. L. Sackett, K. R. Piper,
Physiol. Plant Pathol. 1:241–246. and S. K. Farrand. 1998. Octopine-type Ti plas-
35. Kerr, A. 1969.Transfer of virulence between iso- mids code for a mannopine-inducible dominant-
lates of Agrobacterium. Nature 223:1175–1176. negative allele of traR, the quorum-sensing
36. Kerr, A., P. Manigault, and J. Tempe. 1977. activator that regulates Ti plasmid conjugal trans-
Transfer of virulence in vivo and in vitro in fer. Mol. Microbiol. 27:277–288.
Agrobacterium. Nature 265:560–561. 49. Pappas, K. M., and S. C. Winans. 2003. A
37. Lessl, M., D. Balzer, W. Pansegrau, and E. LuxR-type regulator from Agrobacterium tumefa-
Lanka. 1992. Sequence similarities between the ciens elevates Ti plasmid copy number by activating
RP4 Tra2 and the Ti VirB region strongly support transcription of plasmid replication genes. Mol.
the conjugation model for T-DNA transfer. J. Biol. Microbiol. 48:1059–1073.
Chem. 267:20471–20480. 50. Pappas, K. M., and S. C. Winans. 2003. The
38. Lessl, M., and E. Lanka. 1994. Common mech- RepA and RepB autorepressors and TraR play
anisms in bacterial conjugation and Ti-mediated opposing roles in the regulation of a Ti plasmid
T-DNA transfer to plant cells. Cell 77:321–324. repABC operon. Mol. Microbiol. 49:441–455.
39. Lewis, M. 2005. The lac repressor. C. R. Biol. 51. Piper, K. R., S. Beck von Bodman, and S. K.
328:521–548. Farrand. 1993. Conjugation factor of Agrobac-
40. Luo, Z. Q., Y. Qin, and S. K. Farrand. 2000. terium tumefaciens regulates Ti plasmid transfer by
The antiactivator TraM interferes with the autoinduction. Nature 362:448–450.
autoinducer-dependent binding of TraR to DNA 52. Piper, K. R., S. Beck Von Bodman, I. Hwang,
by interacting with the C-terminal region of and S. K. Farrand. 1999. Hierarchical gene
the quorum-sensing activator. J. Biol. Chem. regulatory systems arising from fortuitous gene
275:7713–7722. associations: controlling quorum sensing by the
41. Luo, Z. Q., A. J. Smyth, P. Gao,Y. Qin, and S. opine regulon in Agrobacterium. Mol. Microbiol.
K. Farrand. 2003. Mutational analysis of TraR. 32:1077–1089.
53. Pohlman, R. F., H. D. Genetti, and S. C. OccR causes a bend at a target promoter, which is
Winans. 1994. Common ancestry between IncN partially relaxed by a plant tumor metabolite. Cell
conjugal transfer genes and macromolecular 69:659–667.
export systems of plant and animal pathogens. Mol. 66. White, C. E., and S. C. Winans. 2007.
Microbiol. 14:655–668. The quorum-sensing transcription factor TraR
54. Qin, Y., Z. Q. Luo, A. J. Smyth, P. Gao, S. decodes its DNA binding site by direct contacts
Beck von Bodman, and S. K. Farrand. 2000. with DNA bases and by detection of DNA
Quorum-sensing signal binding results in dimer- flexibility. Mol. Microbiol. 64:245–256.
ization of TraR and its release from membranes 67. White, C. E., and S. C.Winans. 2005. Identifi-
into the cytoplasm. EMBO J. 19:5212–5221. cation of amino acid residues of the Agrobacterium
55. Qin,Y., A. J. Smyth, S. Su, and S. K. Farrand. tumefaciens quorum-sensing regulator TraR that are
2004. Dimerization properties of TraM, the antiac- critical for positive control of transcription. Mol.
tivator that modulates TraR-mediated quorum- Microbiol. 55:1473–1486.
dependent expression of the Ti plasmid tra genes. 68. Whitehead, N. A., A. M. Barnard, H. Slater,
Mol. Microbiol. 53:1471–1485. N. J. Simpson, and G. P. Salmond. 2001.
56. Shoemaker, B. A., J. J. Portman, and P. G. Quorum-sensing in gram-negative bacteria.
Wolynes. 2000. Speeding molecular recognition FEMS Microbiol. Rev. 25:365–404.
by using the folding funnel: the fly-casting mecha- 69. Winans, S. C., R. A. Kerstetter, and E. W.
nism. Proc. Natl.Acad. Sci. USA 97:8868–8873. Nester. 1988. Transcriptional regulation of the
57. Smith, E. F., and C. O. Townsend. 1907. virA and virG genes of Agrobacterium tumefaciens. J.
A plant-tumor of bacterial origin. Science Bacteriol. 170:4047–4054.
24: 671–673. 70. Reference deleted.
58. Stachel, S. E., and P. C. Zambryski. 1986. virA 71. Wood, D. W., J. C. Setubal, R. Kaul, D. E.
and virG control the plant-induced activation of Monks, J. P. Kitajima,V. K. Okura,Y. Zhou, L.
the T-DNA transfer process of A. tumefaciens. Cell Chen, G. E.Wood, N. F.Almeida, Jr., L.Woo,Y.
46:325–333. Chen, I.T. Paulsen, J. A. Eisen, P. D. Karp, D.
59. Swiderska, A., A. K. Berndtson, M. R. Cha, Bovee, Sr., P. Chapman, J. Clendenning, G.
L. Li, G. M. Beaudoin III, J. Zhu, and C. Deatherage, W. Gillet, C. Grant, T. Kutyavin,
Fuqua. 2001. Inhibition of the Agrobacterium R. Levy, M. J. Li, E. McClelland, A. Palmieri,
tumefaciens TraR quorum-sensing regulator. Inter- C. Raymond, G. Rouse, C. Saenphimmachak,
actions with the TraM anti-activator. J. Biol. Chem. Z. Wu, P. Romero, D. Gordon, S. Zhang, H.
276:49449–49458. Yoo,Y.Tao, P. Biddle, M. Jung,W. Krespan, M.
60. Reference deleted. Perry, B. Gordon-Kamm, L. Liao, S. Kim, C.
61. Van Larebeke, N., G. Engler, M. Holsters, Hendrick, Z.Y. Zhao, M. Dolan, F. Chumley,
S. Van den Elsacker, I. Zaenen, R. A. S. V. Tingey, J. F. Tomb, M. P. Gordon, M. V.
Schilperoort, and J. Schell. 1974. Large plasmid Olson, and E. W. Nester. 2001.The genome of
in Agrobacterium tumefaciens essential for crown the natural genetic engineer Agrobacterium tumefa-
gall-inducing ability. Nature 252:169–170. ciens C58. Science 294:2317–2323.
62. Vannini,A., C.Volpari, and S. Di Marco. 2004. 72. Yao, Y., M. A. Martinez-Yamout, T. J.
Crystal structure of the quorum-sensing protein Dickerson,A. P. Brogan, P. E.Wright, and H.
TraM and its interaction with the transcriptional J. Dyson. 2006. Structure of the Escherichia coli
regulator TraR. J. Biol. Chem. 279:24291–24296. quorum sensing protein SdiA: activation of the
63. Vannini, A., C. Volpari, C. Gargioli, E. folding switch by acyl homoserine lactones. J. Mol.
Muraglia, R. Cortese, R. De Francesco, P. Biol. 355:262–273.
Neddermann, and S. D. Marco. 2002.The crys- 73. Young, J. M., L. D. Kuykendall, E. Martinez-
tal structure of the quorum sensing protein TraR Romero, A. Kerr, and H. Sawada. 2003.
bound to its autoinducer and target DNA. EMBO Classification and nomenclature of Agrobacterium
J. 21:4393–4401. and Rhizobium. Int. J. Syst. Evol. Microbiol.
64. von Bodman, S. B., J. K. Ball, M. A. Faini, 53:1689–1695.
C. M. Herrera, T. D. Minogue, M. L. 74. Zhang, L., P. J. Murphy, A. Kerr, and M. E.
Urbanowski, and A. M. Stevens. 2003. The Tate. 1993. Agrobacterium conjugation and gene
quorum sensing negative regulators EsaR and regulation by N-acyl-L-homoserine lactones.
ExpR(Ecc), homologues within the LuxR family, Nature 362:446–448.
retain the ability to function as activators of tran- 75. Zhang, R. G., T. Pappas, J. L. Brace, P. C.
scription. J. Bacteriol. 185:7001–7007. Miller, T. Oulmassov, J. M. Molyneaux, J. C.
65. Wang, L., J. D. Helmann, and S. C. Winans. Anderson, J. K. Bashkin, S. C.Winans, and A.
1992. The A. tumefaciens transcriptional activator Joachimiak. 2002. Structure of a bacterial
306 ■ WINANS
quorum-sensing transcription factor complexed terium tumefaciens is inhibited by a truncated, domi-

with pheromone and DNA. Nature 417:971–974. nant defective TraR-like protein. Mol. Microbiol.
76. Zhu, J., J. W. Beaber, M. I. More, C. Fuqua, 27:289–297.
A. Eberhard, and S. C. Winans. 1998. Analogs 79. Zhu, J., and S. C. Winans. 1999. Autoin-
of the autoinducer 3-oxooctanoyl-homoserine ducer binding by the quorum-sensing regulator
lactone strongly inhibit activity of the TraR TraR increases affinity for target promoters
protein of Agrobacterium tumefaciens. J. Bacteriol. in vitro and decreases TraR turnover rates in
180:5398–5405. whole cells. Proc. Natl. Acad. Sci. USA
77. Zhu, J., P. M. Oger, B. Schrammeijer, P. J. 96:4832–4837.
Hooykaas, S. K. Farrand, and S. C. Winans. 80. Zhu, J., and S. C. Winans. 2001.The quorum-
2000.The bases of crown gall tumorigenesis. J. Bac- sensing transcriptional regulator TraR requires its
teriol. 182:3885–3895. cognate signaling ligand for protein folding, pro-
78. Zhu, J., and S. C. Winans. 1998. Activity of tease resistance, and dimerization. Proc. Natl.Acad.
the quorum-sensing regulator TraR of Agrobac- Sci. USA 98:1507–1512.
A NEW LOOK AT SECONDARY
METABOLITES
Michael G. Surette and Julian Davies
19
Microbes have evolved to exist in complex ordered assembly of communities such as
communities composed of other microbes and microbial biofilms that form on dental surfaces
a variety of eukaryotic organisms.Within these (46, 47, 54). In such a community, only a subset
communities there exist competitive, coopera- of organisms can adhere to dental enamel (the
tive, and neutral interactions. At the simplest primary colonizers), secondary colonizers can
level limitations in nutritional resources and recognize and bind specifically to surface mole-
real estate will drive competition for specific cules or extracellular polymers produced by the
nutrients and specialization within the com- primary colonizers, tertiary colonizers to sec-
munity. Coincidentally, exploitation of com- ondary colonizers, and so on. Through these
plex resources may often require coordinated physical interactions, spatially and temporally
behavior of an organism and its interaction ordered communities can be assembled. Super-
with multiple specialists. In syntrophic interac- imposed on these physical and nutritional
tions, metabolic pathways are integrated over interactions are a multitude of chemical inter-
different cell types.A simple scenario would be actions that can occur in microbial communi-
where a microbe’s ability to utilize a particular ties. The topic of this chapter is the chemical
nutrient source is limited by the buildup of interactions between cells and the diverse
toxic by-products; consumption of the latter by nature of those interactions.We emphasize the
one or more other strains would benefit the roles of secondary metabolites. (Secondary
community.These types of interactions would metabolites are low-molecular-weight organic
allow for the self-organization of complex molecules not essential for maintenance of cel-
communities without any specific communi- lular function or for normal growth of an
cation or signaling between organisms. More organism. These compounds have a variety of
specific physical interactions can also lead to biological activities and include antibiotics,
quorum-sensing molecules, surfactants, pig-
Michael G. Surette Department of Microbiology and Infec- ments, and siderophores.) Interactions through
tious Diseases, Department of Biochemistry and Molecular primary metabolites are not considered,
Biology, University of Calgary, Calgary, Alberta, Canada T2N although the contributions of these interactions
4N1. Julian Davies Department of Microbiology and
Immunology, University of British Columbia, Vancouver, to the stability and dynamics of microbial com-
British Columbia, Canada V6T 1Z3. munities cannot be understated. We focus on
307
308 ■ SURETTE AND DAVIES
two seemingly well-defined groups of second- plants, and animals and must have defined
ary metabolites, quorum-sensing signals and endogenous bioactivities has been of limited
antibiotics, to demonstrate that their described concern,but what functions are encompassed in
biological activities do not necessarily define this vast chemical collection in living systems?
their functional roles in microbial communi- Are the activities exploited in clinical and indus-
ties. Rather than categorizing secondary trial applications of these compounds equiva-
metabolites in distinct groups, it is apparent that lent to their natural functions?
they can have multiple functions and represent The production of bioactive secondary
a continuum of activities that contribute to metabolites by microbes (and plants) is a highly
cooperative, competitive, and manipulative sophisticated process. Some of the largest and
interactions within complex microbial com- most complex biosynthetic pathways known
munities.Within the microbial world these sec- concern their production; significant amounts
ondary metabolites comprise large numbers of of energy are required and the pathways are
small molecules that have been screened and tightly regulated. In most cases the biosynthetic
honed by natural selection to act on a large pathways are activated during normal growth
number of biological receptors. A thorough and involve at least a partial switch from pri-
understanding of the full biological spectrum of mary metabolic production since a number of
activities of these secondary metabolites is nec- primary metabolites are used as precursors for
essary to delineate the mechanisms that lead to the production of the bioactive secondary
establishment of complex microbial communi- metabolites (3). The complex multienzyme
ties that exhibit both long-term stability and pathways consist of a variety of different,
dynamics. Identifying the various roles and defined classes of enzymatic reactions, includ-
modes of action of secondary metabolites in ing those for core structure synthesis such as
communities provides new experimental polyketide synthases or nonribosomal peptide
avenues for the discovery of novel antimicro- synthases together with a host of modifying
bials and other strategies for therapeutic inter- enzymes (dehydrogenases, methylases, etc.)
vention in human diseases. (77). However, the products made by a polyke-
tide pathway in one organism may differ
SECONDARY METABOLITES considerably in structure and function from the
The study of the biology of living organisms products of a polyketide pathway in another
and associated biochemical processes has, to organism. Indeed, most actinomycetes strains
date, focused primarily on the structures and possess multiple pathways of the same enzy-
functions of DNA,RNA,proteins,lipids,carbo- matic class but generate distinctly different mol-
hydrates, and their macromolecular complexes. ecules employing similar types of biosynthetic
This large body of work has provided the basis reactions. In addition, combinatorial products
of the existing understanding of the molecular of two different pathways in the same organism
genetic basis of life. However, these studies have may generate even greater molecular diversity.
essentially ignored the presence of an enormous The above description should make it very
number of low-molecular-weight (3,000) clear that the chemical space of biological sec-
bioactive molecules that are produced by living ondary metabolites is vast and largely unex-
organisms as part of secondary metabolism (12, plored. This fact has been emphasized when
75). Small molecules have long been known for comparing the drug potential of natural prod-
their properties as therapeutic agents having uct libraries compared to synthetic chemical
seen use for all manner of human diseases. Since libraries (91). Based on the (low) estimate that
the 1950s they have been best known for their streptomycetes produce a quarter of a million
much-applied activities as antibiotics, antivirals, bioactive small molecules, the actinomycetes
antifungals, and anticancer, immunosuppressive, would increase this number 10-fold and the
and cardiovascular agents. The fact that these actinobacteria another 10-fold, providing a
molecules are produced naturally by microbes, total number of secondary metabolites in excess
19. A NEW LOOK AT SECONDARY METABOLITES ■ 309
of 25 million. This rough calculation ignores range invariably shows stimulatory action at
the contributions of the proteobacteria, other subinhibitory concentrations; this illustrates the
prokaryotes,and the fungi.The largest synthetic principle of hormesis (Fig. 1) (101)—the phe-
chemical libraries suffer badly by compari- nomenon where a particular compound
son, especially since all microbial products exhibits distinct activities at different concentra-
are biologically active, having been subjected tion. It is reasonable to assume that the activities
to screening, modification, and rescreening of secondary metabolites at low concentrations
through the natural processes of evolution. represent their true natural function.
The biological space of secondary metabo- Signaling processes are the fabric of biology.
lites is also large in the sense that they possess a Within and between all cells,“signal and linked
large variety of biochemical functions; it should responses” are essential to the response of cells
be noted that the majority of these assignments to their environment and to control processes
are made on the basis of laboratory tests of differentiation and development. In the
such as enzyme inhibition and as such may or majority of cases, the signal is a small molecule
may not indicate their true or only biological with the capacity for stereochemical interaction
functions in nature. Many of these compounds with a receptor or ligand leading to a response
exhibit inhibitory activity at elevated concen- that is primarily transcriptional. Our thesis is
trations, but whether such activity is important that the conclusions of the majority of the stud-
in a natural context is not known. For example, ies of the biology of secondary metabolites are
examination of the activity of putative anti- specious; they are interpretations of artificial
biotic molecules over a wide concentration (laboratory) experimentation. This applies to
FIGURE 1 Hormesis and small molecules. The nature and extent of transcription
responses to bioactive small molecules are concentration dependent.The cellular response
at low concentration will be observed where there are no significant growth effects and
can often be observed as changes in patterns of gene expression.The pattern of expressed
genes will change when growth becomes inhibited. Reprinted from reference 101.
© Elsevier (2006).
studies of secondary metabolite production, extent and nature of microbial diversity in the
activity, and target specificity. For example, it is biosphere are major questions in biological sci-
widely championed that secondary metabolites ence, and the chemical and biochemical diver-
are made by microbes only when cell growth sity of natural products is an intrinsic aspect of
approaches stationary phase and cell multiplica- these questions.
tion is slowing (3). Under these conditions key
nutrients may be lacking in the culture ANTIBIOTICS
medium; this may equate to the situation in a One of the major classes of characterized natu-
soil environment. However, if secondary rally occurring small molecules is the antibi-
metabolite production by actinomycetes is otics. There has been much discussion of
assayed with very sensitive methods for the appropriate descriptors for secondary metabo-
detection of bioactivity, biosynthesis is often lites and their formation, usually based on their
found to commence during early stages of growth-related production in the laboratory.
growth and continue throughout the growth The definition of “antibiotic” has been modi-
phase. Is production dependent on medium fied since the original description of a com-
composition and its depletion or on specific pound produced by a microbe that inhibits the
growth-phase triggers? This will be difficult growth of other microbes (93). Many natural
to resolve until the parameters controlling products have this activity; it is important to
microbial growth in natural environments can realize that antibiotic is a definition of an activ-
be reproduced in the laboratory. ity, not the definition of a compound. A large
The preeminence of secondary metabolite number of synthetic chemicals with antimicro-
small-molecule therapeutics is illustrated by the bial activity have been produced, and they are
fact that the current world market for antibi- not considered to be antibiotics. As we have
otics has grown to more than $30 billion per noted previously,the antibiotic activity of a sec-
year (19). In 2000, the production of antibiotics ondary metabolite is concentration dependent,
in the United States alone totaled 50 million and a secondary metabolite may (and usually
pounds (100,000 metric tons). Given that does) have other bioactivities when screened in
China, Russia, and India are the largest produc- the laboratory. There are notable examples of
ers of therapeutics by fermentation,it is difficult compounds that were discovered to have
to estimate the quantities made worldwide. antimicrobial activity (originally classified as
Since the introduction of penicillin (fungal antibiotics) but which have had highly success-
product) and streptomycin (bacterial) in the ful and profitable applications in other forms of
late 1940s, the amounts of bioactive secondary therapeutic intervention; these include many
metabolite molecules produced from microbial important therapeutics, such as cyclosporine
sources must be in the range of many millions (immunosuppressive), daunomycin (anti-
of metric tons. One wonders how this com- cancer), and the herbicide bialaphos.This is not
pares with the natural production of low- at all surprising; organic compounds may react
molecular-weight bioactive molecules in the differently with all manner of receptors to gen-
biosphere over the same period. erate different responses in living organisms or
Studies of the native roles of natural products systems. In considering chemical interactions
have been of minor consideration compared to between cells, it suffices for one cell to produce
their application in medicine and as commer- a secondary metabolite that interacts with a
cial entities. Nonetheless, they have constantly component of another to trigger a response;the
provided intellectual challenges to chemists and latter is often metabolic but could take a num-
biologists, and even with the advances in tech- ber of different forms.As we shall see, most sec-
nology currently available,the determination of ondary metabolites at defined concentrations
their complex structures and their total synthe- modulate transcription processes in target cells
sis often poses difficult problems to solve. The or organisms.At other concentrations they have
different responses that may have physiological nature of chemical interactions should be con-
or therapeutic consequences, such as antibiotic sidered more broadly, as not all interactions will
activity. Our thesis is that the latter may be rare be cooperative. Accordingly, the perspective of
in microbial interactions in nature. signaling and communication, as used more
generally in ecological or evolutionary terms,
CELL-CELL COMMUNICATION provides a better conceptual framework to
AND QUORUM SENSING define the nature of these interactions (40). It is
Another group of secondary metabolites that useful to consider the sender (the producer of
has gained increasing attention in the past two the secondary metabolite) and the receiver (the
decades includes a group of molecules involved target of the secondary metabolite) separately
in coordinated gene regulation among popula- in terms of consequences of the chemical inter-
tions of cells. Many bacteria regulate gene action. In this context a signal refers to a chem-
expression in response to accumulation of sec- ical that alters the behavior and/or gene
ondary metabolites, and this behavior has been expression of other organisms that evolved
collectively referred to as quorum sensing or because of that effect and is effective because
cell-cell communication (5,29,33,40,98).In its the receiver’s response has also evolved specifi-
simplest form, this process results from the pro- cally for that purpose (40). In general, but not
duction and accumulation of signaling mole- always, quorum sensing occurs where the
cules (also referred to as autoinducers) in the sender and receiver are the same strain or
surrounding environment. At some threshold species. This gives rise to the problem of
concentration, the signaling molecules bind to cheaters,cells within the population that do not
receptors on or in the bacterial cell, initiating a participate in the coordinated behavior but
signal transduction pathway leading to changes nonetheless reap the benefits. The topic of
in gene expression. Quorum sensing is gener- cheaters is beyond the scope of this review;
ally thought to act as a mechanism for the coor- however, several recent reviews have discussed
dinated regulation of behavior at the level of this topic in depth (83, 89, 94).
populations of cells.In general,quorum-sensing Interspecies signaling is a much rarer phe-
signals have the following properties: their pro- nomenon. More often, interspecies chemical
duction occurs during specific growth condi- interactions can be described as chemical
tions, they accumulate extracellularly, and they manipulation or sensing of cues (40). The
are recognized by a specific receptor triggering phenomenon of chemical manipulation occurs
a cellular response at some threshold concentra- when the secondary metabolite alters the
tion.The cellular response involves physiologi- behavior and gene expression of other organ-
cal changes distinct from processes required isms where it has a negative effect on the fitness
simply to metabolize or detoxify the signal and of the receiver. A secondary metabolite may
most often involve some behavior that requires be acting as a cue when it alters the behavior
or is advantageous when coordinated between and gene expression of a receiver but where
groups of cells. it did not evolve in the sender specifically for
Cell-cell communication implies a deliber- that effect.
ate evolved process of sending and receiving Precisely defining the nature of chemical
information between cells.The process of com- interactions as signaling, manipulation, or
munication in ecological and evolutionary cues is not necessarily a trivial exercise.
terms has a well-defined and specific defini- Benefits and costs may be context dependent.
tion. In the context of an individual strain or However, understanding the ecology and
species of bacteria, this can be viewed as a evolutionary roles of these secondary metabo-
method for cooperative behavior. However, in lites in their natural context is necessary for a
the context of polymicrobial communities comprehensive understanding of microbial
composed of different genera and species, the communities.
Currently, there are five well-defined classes genes.The LuxI protein carries out synthesis of
of molecules that serve as the paradigms for the signaling molecule. Over 40 members of
chemical signaling in bacteria:N-acyl homoser- the luxI/luxR family have been identified in
ine lactones (AHLs), quinolones, oligopeptides, gram-negative bacteria and are involved in a
AI-2 furanones, and the -butyrolactones (Fig. wide range of responses (29, 33, 60). A second
2).The classical example of AHL quorum sens- family of AHL biosynthetic enzymes is known.
ing is the autoinduction of luminescence in Vib- The best-studied examples of this family are the
rio fischeri (31, 57, 72, 92). In this system an AHL LuxLM proteins from Vibrio harveyi (2, 60).
interacts directly with a regulatory protein, Many gram-positive bacteria communicate
LuxR, to regulate expression of luminescence in a cell-density-dependent manner using
FIGURE 2 Representative structures of small molecule signaling compounds.Two exam-

ples of N-acyl homoserine lactones from P. aeruginosa: (A) N-3-oxododecanoyl-homoserine
lactone (70) and (B) N -butanoyl-homoserine lactone (71). (C) The P. aeruginosa quinolone
signal 2-heptyl-3-hydroxy-4-quinolone (PQS) (58).Two examples of -butyrolactones: (D)
factor 1 from Streptomyces viridochromogenes (81) and (E) IM-2 from Streptomyces lavendulae (81).
(F) AI-2 is formed from 4,5-dihydroxy-2,3-pentanedione (DPD), the product generated by
LuxS (4, 73), which spontaneously cyclizes into a family of furanones. (G) The furanosyl
borate diester complex of (2S,4S)-2-methyl-2,3,3,4-tetrahydroxytetrahydrofuran is the form
of AI-2 bound to LuxP receptor of V. harveyi (9). (H) The (2R,4S)-2-methyl-2,3,3,4-tetrahy-
droxytetrahydrofuran isomer interacts with the LsrB receptor in S. enterica serovar
Typhimurium (61).
oligopeptides (44, 45, 51). These systems specific cellular responses to -butyrolactones
involve synthesis of a precursor polypeptide that have been characterized to date are limited
that is cleaved during or after export, and in to control of antibiotic production and differen-
some cases,the mature signaling peptide may be tiation. The signaling characteristics of the -
further modified. Receptors for the peptides butyrolactones have been well established.The
are typically found in the inner membrane and -butyrolactones act by binding cytoplasmic
belong to the histidine kinase family of two- receptors and inhibit DNA binding (64, 66, 81).
component signal transduction systems. In In general, these receptors are transcriptional
Bacillus subtilis and Streptococcus pneumoniae, repressors and the cellular response is activation
oligopeptide signaling is involved in regulation of target genes (66). Although they are struc-
of competence (10, 11, 30, 32). In Enterococcus turally similar to AHLs, there is no known over-
faecalis and related species, oligopeptide signal- lap in synthesis or cross-talk in response (81).
ing regulates aggregation and conjugational Although to date they have been well character-
transfer between strains (6, 7, 25, 35). In Entero- ized in Streptomyces, recent reports indicate that
coccus, both stimulatory and inhibitory peptides they may be more widespread, and it can be
are made, and the active signaling occurs assumed that they are important pheromones in
between related but not identical cells. In the actinomycete world (81).
Staphylococcus aureus, a modified peptide A fourth class is referred to as LuxS/autoin-
cyclized into a thiolactone ring is involved in ducer-2 (AI-2) and is produced by a variety
cell-cell signaling and many virulence factors of gram-negative and gram-positive bacteria
are under control of this global regulatory sys- (18, 78, 79, 90, 98).The signal produced by all
tem (51,55,65,67).Different groups of S.aureus strains is thought to be an identical product
produce different signaling peptides, and each (4,5-dihydroxy-2,3-petanedione) that is chem-
group is distinguished by the observation that ically unstable and cyclizes into a family of
the peptides produced by isolates within the furanones in chemical equilibrium (Fig. 2) (4,
group are recognized as signaling molecules but 73). The receptors for AI-2 in V. harveyi and
the peptides from other groups are competitive Salmonella enterica serovar Typhimurium bind
inhibitors (37, 55). Similar competition has to different stereoisomers (the S and R forms
been demonstrated between S. aureus and of 2-methyl-2,3,3,4-tetrahydroxytetrahydro-
Staphylococcus epidermidis strains (67). furan, respectively (9, 61). The signal is gener-
The -butyrolactones produced by the ated from an alternative pathway in the
Streptomyces represent the first cell-cell signaling degradation of the key metabolic compound
systems identified. In 1967 Khokhlov and S-adenosylhomocysteine.As such,it may repre-
coworkers reported the isolation of a -butyro- sent a primary metabolite,although the enzyme
lactone, A factor, from a streptomycete (41). LuxS appears dispensable for normal growth
This compound was found to restore antibiotic in all cases where it has been examined. In
production and aerial mycelium formation to a some bacteria, AI-2 clearly acts as a classically
mutant deficient in these properties. Subse- defined quorum-sensing signal; however, in
quent studies have shown that a number of such most bacteria it has been harder to define a role
compounds are produced by actinomycetes; for AI-2.This has led to considerable debate as
they have a wide range of pleiotropic effects and to whether LuxS/AI-2 represents a quorum-
are active at extremely low concentrations sensing system.This debate is beyond the scope
(nanomolar). Interestingly, these compounds of this chapter, and the reader is referred else-
have antibiotic activity at higher concentrations where for more thorough discussion of this
(micromolar). The most extensive work on A subject (18, 24, 59, 78, 90, 95, 98). However,
factor has been done by Horinouchi and col- the difficulty in defining a specific role for AI-2
leagues, and many structurally related com- fits in the larger premise of our thesis presented
pounds have been identified (36, 66). The in this chapter.
Quorum-sensing systems are not generic, The lantibiotics are short peptides contain-
widespread mechanisms to regulate gene ing thioethers generated by the processing of a
expression but highly selected systems that precursor protein (8, 56, 63, 85).They are pro-
evolved in particular organisms to function in duced by a number of gram-positive organisms
particular environments and circumstances. and act as antimicrobials against other gram-
They developed to function in a natural con- positive bacteria.These small peptides are pro-
text that is far removed from the typical labora- duced and released into the environment of the
tory conditions in which they are normally producing cell. Such properties are reminiscent
examined. These systems are often regulated of auto-inducing peptides, and, indeed, nisin,
not only by themselves (autoinduction) but also subtilin, and mercascidin have all been shown
by environmental factors.The generalization of to regulate their own synthesis (43, 48, 74, 103).
quorum sensing as a density-dependent process It is easy to see how antibiotic production could
ignores the reality that most bacteria do not benefit as a coordinated population behavior
exist in well-stirred reactors and the signaling that controls production through a quorum-
will largely be a local event between small sensing mechanism.
groups of cells. Of course, exceptions do exist The dual role of quorum-sensing signal and
such as in the light organ of bobtail squid where antibiotic is not exclusive to these peptide
the the symbiont V. fischeri uses AHL signaling antibiotics. A bactericidal activity produced by
to regulate light production (31, 57, 72, 92). a strain of Rhizobium leguminosarum that inhib-
This was one of the first examples of cell-cell ited the growth of several related strains (known
signaling, and the paradigm established with as the “small bacteriocin”) was purified and
this organism seems to be hard to break. It is demonstrated to be a typical AHL [N-(3-
also hard to dismiss the importance of other hydroxy-7-cis-tetradecenoyl)-L-homoserine
functions that these chemicals may play in lactone] (76). More recently, it has been shown
mixed microbial communities. that the 3-oxododecanyl homoserine lactone
produced by the lasIR system of Pseudomonas
aeruginosa can undergo spontaneous rearrange-
ALTERNATIVE ROLES:
SIGNALING MOLECULES AS ments to form a molecule belonging to a class
ANTIBIOTICS/ANTIBIOTICS AS of antibiotics known as the tetrameric acids
SIGNALING MOLECULES (39). Indeed, it was demonstrated that these
As discussed above, many bioactive molecules compounds have potent antibiotic activity
may exhibit different activities at different con- against gram-positive bacteria (39). Moreover,
centrations. This notion of hormesis may be these compounds can also chelate metals in
of particular importance in the context of sec- what has been proposed to be a trimeric com-
ondary metabolites and distinguishing their plex.These rearrangements do not require any
roles in nature as compared to their roles in the unusual conditions and occur spontaneously in
laboratory or clinical applications (101). Most water.The role of 3-oxododecanyl homoserine
quorum-sensing signals act to induce specific lactone as a quorum-sensing signal in P. aerugi-
transcriptional changes in their target cells at nosa is well established. The possible functions
very low concentrations but exhibit different of the antibacterial activity and the metal-
activities against other cells (usually at higher chelating activity of the tetrameric acid deriva-
concentrations), most notably exhibiting tives are not known, but they could act in the
antibiotic-like activities. Likewise, many antibi- competitive environments of mixed microbial
otics modulate gene expression in target cells at communities.The metal-chelating activity may
concentrations below those required to inhibit indeed have additional roles in competition for
cell growth. Indeed, the distinction between metal ions in such an environment.
quorum-sensing signals and antibiotics is some- In P. aeruginosa, it has been demonstrated that
what blurred. in addition to the two AHL-signaling pathways,
an additional quinolone-signaling system As might be expected, there is no general rule;

is present (20, 58).This system acts to regulate in some cases, antibiotics inhibit expression of
the expression of a number of target genes, virulence-associated genes, whereas in other
including those involved in the production cases, they may activate virulence-associated
of pyocanin, which has potent inhibitory genes (15).
effects on target cells. Moreover, it is clear More systematic analysis of transcriptional
that this organism produces not one but a num- responses to subinhibitory concentrations of
ber of quinolones (as many as 50) that antibiotics has been carried out in recent years
have additional activities, including potent (27, 84, 100, 101). At concentrations below
antibiotic activities (21). those that have a marked effect on growth,
As a corollary to the above, compounds antibiotics induce specific transcriptional
typically considered as antibiotics may in fact responses that are independent of stress
have other activities at other concentrations. responses. These often include virulence-
Significantly,at concentrations below those that associated genes but also induce pathways
have an effect on growth or viability, antibiotics involved in basic metabolism.The responses are
can act to modulate gene expression in target dependent on the compounds’ interactions
cells independent of stress responses. The with their known cellular targets.The patterns
response of bacteria to antibiotics has been of transcriptional response are specific to chem-
actively investigated. Pioneering studies of ical classes of antibiotics (Fig. 3).The specificity
bacterial responses to inhibitory levels of of the response is highlighted by the observa-
antibiotics revealed distinct patterns of stress tion that macrolide-lincosamide-streptogramin
responses characteristic for specific classes of antibiotics, which act on the same cellular
antibiotics (1,86,87).One area that has received target, can be readily distinguished from their
considerable attention is the effect of antibiotics transcriptional responses at subinhibitory con-
on their clinical targets, bacterial pathogens. centrations (84).
FIGURE 3 Transcriptional response to subinhibitory antibiotics. S. enterica serovar Typhimurium containing a

promoter-luxCDABE fusion for an amino acid biosynthesis operon was plated on LB agar with antibiotics added to
sterile filter disks as indicated in the first panel.The middle panel is a photograph of the plate after 20 h showing zones
of inhibition.The third panel is a photograph taken in the dark showing strong induction of luciferase at subin-
hibitory concentrations for some but not all antibiotics.The dashed lines in this panel indicate the zone of inhibition
for each antibiotic.The abbreviations for the antibiotics are Erm, erythromycin; Cam, chloramphenicol;Tet, tetracy-
cline; Spc, spectinomycin; Nal, nalidixic acid; Str, streptomycin; Gat, gatifloxicin;Tmp, trimethoprim ( J. Davies and
M. G. Surette, unpublished data).
We would argue that such activities single protein or RNA component may provide
observed at low concentrations have evolved the ligand-binding site.There is great structural
for other biological purposes. In natural envi- variation in ligands and receptors, and their
ronments, antibiotics are likely to be present at specificity may be rooted in different stereo-
low concentrations and may act on target cells chemical or physical interactions. In microbes,
not to inhibit growth but to manipulate their the signals are primarily chemical. It is impor-
behavior. Chemical manipulation through tant to realize that where chemical interactions
secondary metabolites such as antibiotics may are between species (in the cases of cues and
play an important role in polymicrobial com- manipulation), the receptors may have other,
munity dynamics.An example where chemical independent, biological functions in the
manipulation of one organism on another target cells. The specificity of ligand-receptor
may act through modulating primary metabo- interaction is dependent on ligand concentra-
lism is that described for Streptococcus gordonii tion, for example, AHLs, -butyrolactones,
and Veillonella atypica, two species found in the and signaling peptides interact with their recep-
dental plaque biofilms (26). V. atypica requires tors at nanomolar concentrations. Secondary
lactic acid to grow and relies on S. gordonii metabolites, such as erythromycin binding to a
to ferment sugars and release lactic acid. Inter- receptor site in the ribosome and rifampicin
estingly, it has been shown that V. atypica pro- binding to RNA polymerase, occur at similar
duces a soluble chemical that induces amylase low concentrations. Do all secondary metabo-
expression in S. gordonii, thereby increasing the lites function in cell biology as ligands with spe-
degradation of complex carbohydrates and cific receptors? Given that they are the products
lactic-acid production (26). of natural selection, one would predict that they
An interesting exception to the observation have indeed evolved to interact with specific
that subinhibitory antibiotics produce tran- targets. If this is the case, then all secondary
scriptional responses independent of stress metabolite chemical space is binding space,
concerns the fluoroquinolone antimicrobials. which implies an equally large, complementary
We have observed that even at subinhibitory space for receptors. If inhibitory activities are to
concentrations, these compounds produce pri- be used to define the receptor sites for second-
marily stress responses associated with double- ary metabolites with antibiotic activity, then
stranded DNA breaks ( J. McClure, M. G. almost any macromolecular structure in a
Surette, and J. Davies, unpublished results).This microbial cell (and perhaps any type of cell)
is likely due to the ability of the cells to repair may act as a receptor for different chemical
damage at low-level DNA damage induced by classes of secondary metabolite (13, 14). In real-
the antibiotics and the activation of the stress ity, although the macromolecule may be com-
response in a fraction of the population at any plex, it is often possible to identify a single,
one time. This class of antimicrobials is not specific component that is the receptor through
derived from natural compounds, and the which the secondary metabolite signal operates.
corollary to the argument presented above is It is remarkable that in a complex molecule such
that, unlike naturally derived antibiotics, they as a ribosome, defined sequences of rRNA can
have not undergone any evolutionary selection be identified as primary receptor sites, usually in
for other activities. concert with one or more of the more than 50
ribosomal proteins. The binding of secondary
RECEPTORS metabolites to ribosomes has been analyzed in
Specific ligand-receptor interactions are the exquisite detail using modern structural meth-
basis of life. The secondary metabolites are ods (80, 82, 102), yet molecular details of the
of low molecular weight, and the receptors consequences of the small-molecule interac-
may be small (proteins or RNA) or very large tions remain to be elucidated.The fact that dis-
(macromolecules), although in the latter case a crete rRNA sites are the receptors can be argued
to support the evolutionary coincidence of the antibiotic. Since this time, many inactivating
receptor-ligand complexes. mechanisms have been identified, and it was
Another RNA receptor is mRNA; the found in 1974 that secondary metabolite-
process of riboswitching is an effective mecha- producing streptomycete strains have the capac-
nism for regulating gene expression at the tran- ity to inactivate their cognate antibiotics (16).
scription level (62).The ligands so far identified This association is general and led to the sugges-
are internal, being products of the associated tion that plasmid-encoded antibiotic resistance
biosynthetic pathway. It seems highly likely that mechanisms to secondary metabolites originate
structurally related secondary metabolites pro- in secondary-metabolite-producing microbes
duced externally would participate in such reac- in the environment.When the genes are “picked
tions although none have been reported so far up” by R-elements and transferred into a for-
(and perhaps not looked for?). Many antibiotic- eign cytoplasm, their overexpression establishes
like secondary metabolites have been shown to a resistance phenotype. A wide distribution of
regulate secondary metabolite production and streptomycin and gentamicin resistance in the
microbial morphology; in most cases the chem- environment was subsequently shown (34, 88).
ical nature of the signals has not been deter- The ubiquity of environmental secondary
mined. As mentioned previously, certain of the metabolite resistance has received strong sup-
autoinducers (homoserine lactones) have been port from the recent studies of D’Costa et al.,
shown to exert cell growth-inhibitory activity who isolated and defined a large family of
at higher concentrations than those required for naturally occurring secondary-metabolite-
their activity as quorum-sensing agents. This modifying functions described as the “resis-
further emphasizes the fine (concentration) dis- tome” (17, 38, 96). However, despite the fact
tinction between cell signaling activity and that this gene pool may be the source of all clin-
antibiotic activity. ically significant transferable antibiotic resist-
ance determinants, there is no evidence that
RESISTANCE these genes are, naturally, resistance genes.This
Resistance or recalcitrance to secondary point is worth emphasizing. In all bacteria (and
metabolite inhibitors has been identified since fungi) that produce secondary metabolite, there
the introduction of antimicrobial therapy. The are genes encoding mechanisms for efflux,
introduction of any therapeutic agent was modification, or inactivation that are often
followed shortly by the appearance of recalci- included in the biosynthetic gene cluster. Efflux
trant pathogens (49, 50, 69).The beginning of or pump systems are required to release com-
penicillin use in 1942 was accompanied by pounds from the cell, perhaps to avoid the
treatment-associated resistance (1945) as was buildup of toxic concentrations in the cyto-
the case for mycobacterial infections treated plasm.The modification and inactivation mech-
with streptomycin; this pattern continues to anisms take a number of forms, from enzymatic
plague antibiotic usage to this day. modification of the endogenous target, enzy-
In 1955 a new and totally unexpected series matic inactivation of toxic compounds, the pro-
of epidemic multidrug-resistant Shigellae was duction of secondary metabolite sequestering
identified in Japan (42); the antimicrobials (inactivating) proteins, and other mechanisms.
included streptomycin, penicillin, tetracycline, The current thinking is that these are all self-
chloramphenicol, and sulfonamide. Genetic protection mechanisms to prevent the produc-
studies of these strains indicated that the mul- ing strain from suicide during production.
tidrug resistance (mdr) was extrachromosomal However, the evidence for this is anecdotal and
and encoded by specific resistance plasmids. is based on little hard evidence.The main evi-
The streptomycin and chloramphenicol resist- dence in support of the secondary metabolite/
ance determinants were later shown to catalyze antibiotic/resistance relationship is the finding
chemical inactivation of the corresponding that heterologous expression of the presumed
resistance genes on multicopy plasmids pro- nals are highly specific, and S. aureus strains can
duces high-level resistance phenotypes in be classified according to their oligopeptide sig-
pathogens. The resistance genes of producers nals. Groups are defined as strains that produce
and pathogens are evolutionarily related, but signaling peptides that can cross-activate. How-
interestingly, the transcription signals on R- ever, the signaling peptides within a particular
plasmids are distinctly different and only some, group act as potent inhibitors of signaling in
such as those found associated with resistance other groups (37, 55).This is an example where
integrons, appear to have evolved for very effi- the signals have evolved a role not only in sig-
cient expression of resistance.Also, it should be naling but also in intraspecific competition.
noted that many orthologs of antibiotic resist-
ance genes are found in bacteria (other than CONCLUSION
producing organisms), and in the majority of In this chapter we addressed the complex world
cases tested, they do not confer significant levels of secondary metabolites and highlight that
of resistance to the host; nevertheless, they do there is much to be learned. It is clear that
constitute part of the resistome. It seems that chemical interactions between cells play an
natural resistance is more likely involved in the important role in microbial communities
regulation of secondary metabolite synthesis, through both primary and secondary metabo-
synchronization of cellular networks, or other lites.We have focused on two established types
modulations of secondary metabolite signals. of secondary metabolites (antibiotics and quo-
Of course, their roles in protection cannot be rum-sensing molecules) and emphasized that
eliminated. even within these well-characterized groups
In addition to the widespread occurrence of the paradigm of their function understates their
antibiotic resistance mechanisms, it has become biological activities; the various observed activ-
apparent in recent years that there are many ities are dependent on concentration.
naturally occurring systems that interfere with Associated with the enormous microbial
cell-cell signaling pathways. To date, this has diversity, the chemical space of secondary
been explored mostly with respect to AHL sig- metabolites produced in the microbial world is
naling. Two general strategies have emerged: vast and largely unexplored. Our ability to
signal degradation and chemical antagonists search this rich chemical resource is often con-
(22, 23, 28).A number of studies have demon- strained by the availability of the assays being
strated that many organisms produce enzymes used.The observations that many of these com-
capable of degrading these signals. Both AHLs pounds can have a variety of different biochem-
and AHL-acylases have been described (22, 23, ical activities should be exploited. For example,
28). These activities have been isolated from the potent transcriptional responses induced by
bacteria as well as eukaryotes (68,99).In the lat- subinhibitory antibiotics would seem to offer
ter case, it is postulated that the ability of a host an efficient approach to screen for new bioac-
to degrade quorum-sensing signals of an invad- tive compounds, as compared to conventional
ing pathogen may play a role in host defense. antibiotic screens that require elevated concen-
Similarly, a number of antagonists of quorum- trations. Moreover, it is likely that many more
sensing signals have been found to be produced natural products will have modulatory activity
by eukaryotic organisms,where their role is also than antibiotic activity.These natural bioactive
postulated to be involved in host defense. An molecules might serve as new scaffolds for the
example is the halogenated furanones produced development of novel semisynthetic antibiotics
by marine algae that have been shown to be or have other valuable therapeutic activity by
inhibitors of AHL signaling (28, 52, 53, 97). modulating gene expression and cellular physi-
Inhibition between strains using closely ology in target organisms rather than as growth
related signals also occurs and has been best inhibitors.The exquisite characteristics of small
characterized in S. aureus.The oligopeptide sig- molecules must be applied in a more subtle and
intelligent manner if we are to benefit from 12. Davies, J. 1994. Microbial molecular diversity:
their therapeutic potential. The language of past and present; Thom Award Lecture. J. Ind.
microbial communities is based on small- 13. Davies, J. 2006. Are antibiotics naturally antibi-
molecule interactions. Unfortunately, we do otics? J. Ind. Microbiol. Biotechnol. 33:496–499.
not yet have the dictionary/cipher to interpret 14. Davies, J. 2007. Small molecules: the lexicon of
this; we know that they are broadcasting but biodiversity. J. Biotechnol. 129:3–5.
cannot tune in to the right wavelength. 15. Davies, J., G. B. Spiegelman, and G. Yim.
2006.The world of subinhibitory antibiotic con-
centrations. Curr. Opin. Microbiol. 9:445–453.
REFERENCES 16. Davies, J. E., and R. E. Benveniste. 1974.
1. Bandow, J. E., H. Brotz, L. I. Leichert, Enzymes that inactivate antibiotics in transit to
H. Labischinski, and M. Hecker. 2003. Pro- their targets. Ann. N.Y.Acad. Sci. 235:130–136.
teomic approach to understanding antibiotic 17. D’Costa, V. M., K. M. McGrann, D. W.
action. Antimicrob.Agents Chemother. 47:948–955. Hughes, and G. D.Wright. 2006. Sampling the
2. Bassler, B. L., M.Wright, R. E. Showalter, and antibiotic resistome. Science 311:374–377.
M. R. Silverman. 1993. Intercellular signalling in 18. De Keersmaecker, S. C., K. Sonck, and
Vibrio harveyi: sequence and function of genes reg- J. Vanderleyden. 2006. Let LuxS speak up in
ulating expression of luminescence. Mol. Microbiol. AI-2 signaling. Trends Microbiol. 14:114–119.
9:773–786. 19. Demain, A. L. 2000. Small bugs, big business: the
3. Bibb, M. J. 2005. Regulation of secondary economic power of the microbe. Biotechnol. Adv.
metabolism in streptomycetes. Curr. Opin. Micro- 18:499–514.
biol. 8:208–215. 20. Deziel, E., F. Lepine, S. Milot, J. He, M. N.
4. Burgess, N. A., D. F. Kirke, P. Williams, Mindrinos, R. G. Tompkins, and L. G.
K. Winzer, K. R. Hardie, N. L. Meyers, Rahme. 2004. Analysis of Pseudomonas aeruginosa
J. Aduse-Opoku, M. A. Curtis, and 4-hydroxy-2-alkylquinolines (HAQs) reveals a
M. Camara. 2002. LuxS-dependent quorum role for 4-hydroxy-2-heptylquinoline in cell-to-
sensing in Porphyromonas gingivalis modulates pro- cell communication. Proc. Natl. Acad. Sci. USA
tease and haemagglutinin activities but is not 101:1339–1344.
essential for virulence. Microbiology. 148:763–772. 21. Diggle, S. P., P. Cornelis, P.Williams, and M.
5. Camilli, A., and B. L. Bassler. 2006. Bacterial Camara. 2006. 4-quinolone signalling in
small-molecule signaling pathways. Science Pseudomonas aeruginosa: old molecules, new per-
311:1113–1116. spectives. Int. J. Med. Microbiol. 296:83–91.
6. Chandler, J. R., and G. M. Dunny. 2004. 22. Dong, Y. H., L. H. Wang, and L. H. Zhang.
Enterococcal peptide sex pheromones: synthesis 2007. Quorum-quenching microbial infections:
and control of biological activity. Peptides. mechanisms and implications. Philos.Trans. R Soc.
25:1377–1388. Lond. B Biol. Sci. 362:1201–1211.
7. Chandler, J. R., H. Hirt, and G. M. Dunny. 23. Dong,Y. H., and L. H. Zhang. 2005. Quorum
2005.A paracrine peptide sex pheromone also acts sensing and quorum-quenching enzymes. J. Micro-
as an autocrine signal to induce plasmid transfer biol. 43:101–109.
and virulence factor expression in vivo. Proc. Natl. 24. Duan, K., and M. Surette. 2006. LuxS in cel-
Acad. Sci. USA 102:15617–15622. lular metabolism and cell-to-cell signaling,
8. Chatterjee, C., M. Paul, L. Xie, and W. A. van p. 117–150. In D. Demuth and R. Lamont (ed.),
der Donk. 2005. Biosynthesis and mode of action Bacterial Cell-to-Cell Communication: Role in Viru-
of lantibiotics. Chem. Rev. 105:633–684. lence and Pathogenesis. CambridgeUniversity Press,
9. Chen, X., S. Schauder, N. Potier, A. Van New York, NY.
Dorsselaer, I. Pelczer, B. L. Bassler, and F. M. 25. Dunny, G. M., M. H. Antiporta, and H. Hirt.
Hughson. 2002. Structural identification of a 2001. Peptide pheromone-induced transfer of
bacterial quorum-sensing signal containing boron. plasmid pCF10 in Enterococcus faecalis: probing the
Nature 415:545–549. genetic and molecular basis for specificity of the
10. Claverys, J. P., M. Prudhomme, and B. pheromone response. Peptides 22:1529–1539.
Martin. 2006. Induction of competence regulons 26. Egland, P. G., R. J. Palmer, Jr., and P. E.
as a general response to stress in gram-positive bac- Kolenbrander. 2004. Interspecies communica-
teria. Annu. Rev. Microbiol. 60:451–475. tion in Streptococcus gordonii-Veillonella atypica
11. Cvitkovitch, D. G. 2001. Genetic competence biofilms: signaling in flow conditions requires jux-
and transformation in oral streptococci. Crit. Rev. taposition. Proc. Natl. Acad. Sci. USA 101:
Oral. Biol. Med. 12:217–243. 16917–16922.
27. Goh, E. B., G. Yim, W. Tsui, J. McClure, 41. Khokhlov, A. S., I. I. Tovarova, L. N.
M. G. Surette, and J. Davies. 2002. Transcrip- Borisova, S. A. Pliner, L. N. Shevchenko,
tional modulation of bacterial gene expression by E. Kornitskaia, N. S. Ivkina, and I. A.
subinhibitory concentrations of antibiotics. Proc. Rapoport. 1967. [The A-factor, responsible for
Natl.Acad. Sci. USA 99:17025–17030. streptomycin biosynthesis by mutant strains
28. Gonzalez, J. E., and N. D. Keshavan. 2006. of Actinomyces streptomycini]. Dokl. Akad. Nauk.
Messing with bacterial quorum sensing. Microbiol. SSSR, 177:232–235.
Mol. Biol. Rev. 70:859–875. 42. Kitamoto, O., N. Kusia, N. Fukaya, and A.
29. Greenberg, E. P. 2003. Bacterial communica- Kawashima. 1956. Drug sensitivity of the Shigella
tion: tiny teamwork. Nature 424:134. strains isolated in 1955. J. Jpn. Assoc. Infect. Dis.
30. Hamoen, L.W., G.Venema, and O. P. Kuipers. 30:403–404.
2003. Controlling competence in Bacillus subtilis: 43. Kleerebezem, M., R. Bongers, G. Rutten, W.
shared use of regulators. Microbiology 149:9–17. M. de Vos, and O. P. Kuipers. 2004. Autoregula-
31. Hastings, J. W., and E. P. Greenberg. 1999. tion of subtilin biosynthesis in Bacillus subtilis: the
Quorum sensing: the explanation of a curious role of the spa-box in subtilin-responsive promot-
phenomenon reveals a common characteristic of ers. Peptides 25:1415–1424.
bacteria. J. Bacteriol. 181:2667–2668. 44. Kleerebezem, M., O. P. Kuipers, W. M. de
32. Havarstein, L. S. 1998. Identification of a com- Vos, M. E. Stiles, and L. E. Quadri. 2001. A
petence regulon in Streptococcus pneumoniae by two-component signal-transduction cascade in
genomic analysis. Trends Microbiol. 6:297–299; dis- Carnobacterium piscicola LV17B: two signaling
cussion, 299–300. peptides and one sensor-transmitter. Peptides
33. Henke, J. M., and B. L. Bassler. 2004. Bacterial 22:1597–1601.
social engagements. Trends Cell. Biol. 14:648–656. 45. Kleerebezem, M., L. E. Quadri, O. P. Kuipers,
34. Heuer, H., E. Krögerrecklenfort, E. M. H. and W. M. de Vos. 1997. Quorum sensing by
Wellington, S. Egan, J. D. van Elsas, L. S. van peptide pheromones and two-component signal-
Overbeek, J. M. Collard, G. Guillaume, A. D. transduction systems in gram-positive bacteria.
Karagouni, T. L. Nikolakopoulou, and K. Mol. Microbiol. 24:895–904.
Smalla. 2002. Gentamicin resistance genes in 46. Kolenbrander, P. E., R. N. Andersen,
environmental bacteria: prevalence and transfer. D. S. Blehert, P. G. Egland, J. S. Foster, and
FEMS Microbiol. Ecol. 42:289–302. R. J. Palmer, Jr. 2002. Communication
35. Hirt, H., D. A. Manias, E. M. Bryan, J. R. among oral bacteria. Microbiol. Mol. Biol. Rev.
Klein, J. K. Marklund, J. H. Staddon, M. L. 66:486–505.
Paustian, V. Kapur, and G. M. Dunny. 2005. 47. Kolenbrander, P. E., P. G. Egland, P. I. Diaz,
Characterization of the pheromone response of and R. J. Palmer, Jr. 2005. Genome-genome
the Enterococcus faecalis conjugative plasmid pCF10: interactions:bacterial communities in initial dental
complete sequence and comparative analysis of the plaque. Trends Microbiol. 13:11–15.
transcriptional and phenotypic responses of 48. Kuipers, O. P., M. M. Beerthuyzen, P. G.
pCF10-containing cells to pheromone induction. de Ruyter, E. J. Luesink, and W. M. de Vos.
J. Bacteriol. 187:1044–1054. 1995.Autoregulation of nisin biosynthesis in Lacto-
36. Horinouchi, S. 2002. A microbial hormone, coccus lactis by signal transduction. J. Biol. Chem.
A-factor, as a master switch for morphological dif- 270:27299–27304.
ferentiation and secondary metabolism in Strepto- 49. Levy, S. B. 1998. The challenge of antibiotic
myces griseus. Front. Biosci. 7:d2045–d2057. resistance. Sci.Am. 278:46–53.
37. Ji, G., R. Beavis, and R. P. Novick. 1997. Bac- 50. Levy, S. B. 2001.Antimicrobial resistance poten-
terial interference caused by autoinducing peptide tial. Lancet 358:1100–1101.
variants. Science 276:2027–2030. 51. Lyon, G. J., and R. P. Novick. 2004. Peptide
38. Josephson, J. 2006. The microbial “resistome.” signaling in Staphylococcus aureus and other gram-
Environ. Sci.Technol. 40:6531–6534. positive bacteria. Peptides 25:1389–1403.
39. Kaufmann, G. F., R. Sartorio, S. H. Lee, C. J. 52. Manefield, M., R. de Nys, N. Kumar, R.
Rogers, M. M. Meijler, J. A. Moss, B. Read, M. Givskov, P. Steinberg, and S.
Clapham, A. P. Brogan, T. J. Dickerson, and Kjelleberg. 1999. Evidence that halogenated
K. D. Janda. 2005. Revisiting quorum sensing: furanones from Delisea pulchra inhibit acylated
discovery of additional chemical and biological homoserine lactone (AHL)-mediated gene
functions for 3-oxo-N-acylhomoserine lactones. expression by displacing the AHL signal from its
Proc. Natl.Acad. Sci. USA 102:309–314. receptor protein. Microbiology 145:283–291.
40. Keller, L., and M. G. Surette. 2006. Communi- 53. Manefield, M., T. B. Rasmussen, M.
cation in bacteria: an ecological and evolutionary Henzter, J. B. Andersen, P. Steinberg, S.
perspective. Nat. Rev. Microbiol. 4:249–258. Kjelleberg, and M. Givskov. 2002. Halogenated
furanones inhibit quorum sensing through the accessory gene regulator agr system. Peptides
accelerated LuxR turnover. Microbiology 148: 22:1603–1608.
1119–1127. 68. Ozer, E. A., A. Pezzulo, D. M. Shih, C. Chun,
54. Marsh, P. D. 2004. Dental plaque as a microbial C. Furlong, A. J. Lusis, E. P. Greenberg, and
biofilm. Caries Res. 38:204–211. J. Zabner. 2005. Human and murine paraoxonase
55. Mayville, P., G. Ji, R. Beavis, H. Yang, M. 1 are host modulators of Pseudomonas aeruginosa
Goger, R. P. Novick, and T. W. Muir. 1999. quorum-sensing.FEMS Microbiol.Lett. 253:29–37.
Structure-activity analysis of synthetic autoinduc- 69. Palumbi, S.R. 2001.Humans as the world’s great-
ing thiolactone peptides from Staphylococcus aureus est evolutionary force. Science 293:1786–1790.
responsible for virulence. Proc. Natl.Acad. Sci. USA 70. Pearson, J. P., K. M. Gray, L. Passador, K. D.
96:1218–1223. Tucker,A. Eberhard, B. H. Iglewski, and E. P.
56. McAuliffe, O., R. P. Ross, and C. Hill. 2001. Greenberg. 1994. Structure of the autoinducer
Lantibiotics: structure, biosynthesis and mode of required for expression of Pseudomonas aeruginosa
action. FEMS Microbiol. Rev. 25:285–308. virulence genes. Proc. Natl. Acad. Sci. USA
57. McFall-Ngai, M. J. 2000. Negotiations between 91:197–201.
animals and bacteria: the ‘diplomacy’ of the squid- 71. Pearson, J. P., L. Passador, B. H. Iglewski,
vibrio symbiosis. Comp. Biochem. Physiol. A Mol. and E. P. Greenberg. 1995. A second N-
Integr. Physiol. 126:471–480. acylhomoserine lactone signal produced by
58. McKnight, S. L., B. H. Iglewski, and E. C. Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA
Pesci. 2000. The Pseudomonas quinolone signal 92:1490–1494.
regulates rhl quorum sensing in Pseudomonas aerug- 72. Ruby, E. G. 1999. The Euprymna scolopes-Vibrio
inosa. J. Bacteriol. 182:2702–2708. fischeri symbiosis: a biomedical model for the study
59. McNab, R., and R. J. Lamont. 2003. Microbial of bacterial colonization of animal tissue. J. Mol.
dinner-party conversations: the role of LuxS in Microbiol. Biotechnol. 1:13–21.
interspecies communication. J. Med. Microbiol. 73. Schauder, S., K. Shokat, M. G. Surette, and B.
52:541–545. L. Bassler. 2001. The LuxS family of bacterial
60. Miller, M. B., and B. L. Bassler. 2001. Quorum autoinducers:biosynthesis of a novel quorum-sens-
sensing in bacteria. Annu. Rev. Microbiol. ing signal molecule. Mol. Microbiol. 41:463–476.
55:165–199. 74. Schmitz, S., A. Hoffmann, C. Szekat, B.
61. Miller, S.T., K. B. Xavier, S. R. Campagna, M. Rudd, and G. Bierbaum. 2006.The lantibiotic
E.Taga, M. F. Semmelhack, B. L. Bassler, and mersacidin is an autoinducing peptide. Appl. Envi-
F. M. Hughson. 2004. Salmonella typhimurium ron. Microbiol. 72:7270–7277.
recognizes a chemically distinct form of the bacte- 75. Schreiber, S.L. 2005. Small molecules: the miss-
rial quorum-sensing signal AI-2. Mol. Cell. ing link in the central dogma. Nat. Chem. Biol.
15:677–687. 1:64–66.
62. Muller, S. 2003. Another face of RNA: 76. Schripsema, J., K. E. de Rudder, T. B. van
metabolite-induced “riboswitching” for regula- Vliet, P. P. Lankhorst, E. de Vroom, J. W.
tion of gene expression. Chembiochem. 4:817–819. Kijne, and A. A. van Brussel. 1996. Bacteriocin
63. Nagao, J., S. M. Asaduzzaman, Y. Aso, K. small of Rhizobium leguminosarum belongs to the
Okuda, J. Nakayama, and K. Sonomoto. class of N-acyl-L-homoserine lactone molecules,
2006. Lantibiotics: insight and foresight for new known as autoinducers and as quorum sensing co-
paradigm. J. Biosci. Bioeng. 102:139–149. transcription factors. J. Bacteriol. 178:366–371.
64. Namwat, W., Y. Kamioka, H. Kinoshita, Y. 77. Strohl,W. R. 1997. The Biotechnology of Antibiotics.
Yamada, and T. Nihira. 2002. Characterization Marcel Decker, New York, NY.
of virginiamycin S biosynthetic genes from Strepto- 78. Sun, J., R. Daniel, I.Wagner-Dobler, and A. P.
myces virginiae. Gene 286:283–290. Zeng. 2004.Is autoinducer-2 a universal signal for
65. Novick, R. P. 2003. Autoinduction and signal interspecies communication: a comparative
transduction in the regulation of staphylococcal genomic and phylogenetic analysis of the synthesis
virulence. Mol. Microbiol. 48:1429–1449. and signal transduction pathways. BMC Evol. Biol.
66. Ohnishi, Y., H. Yamazaki, J. Y. Kato, A. 4:36.
Tomono, and S. Horinouchi. 2005. AdpA, 79. Surette, M. G., M. B. Miller, and B. L. Bassler.
a central transcriptional regulator in the A-factor 1999.Quorum sensing in Escherichia coli,Salmonella
regulatory cascade that leads to morphological typhimurium, and Vibrio harveyi: a new family of
development and secondary metabolism in genes responsible for autoinducer production.
Streptomyces griseus. Biosci. Biotechnol. Biochem. Proc. Natl.Acad. Sci. USA 96:1639–1644.
69:431–439. 80. Sutcliffe, J.A. 2005. Improving on nature: antibi-
67. Otto, M. 2001. Staphylococcus aureus and Staphylo- otics that target the ribosome.Curr.Opin.Microbiol.
coccus epidermidis peptide pheromones produced by 8:534–542.
81. Takano, E. 2006.Gamma-butyrolactones:Strepto- 93. Waksman, S. 1961. The role of antibiotics

myces signalling molecules regulating antibiotic in nature, p. 271–272. In Perspectives in Biology
production and differentiation. Curr. Opin. Micro- and Nature. The Johns Hopkins University
biol. 9:287–294. Press, Baltimore, MD.
82. Tenson, T., and A. Mankin. 2006. Antibiotics 94. West, S. A., A. S. Griffin, A. Gardner, and
and the ribosome. Mol. Microbiol. 59:1664–1677. S. P. Diggle. 2006. Social evolution theory
83. Travisano, M., and G. J.Velicer. 2004.Strategies for microorganisms. Nat. Rev. Microbiol. 4:
of microbial cheater control. Trends Microbiol. 597–607.
12:72–78. 95. Winzer, K., K. R. Hardie, and P. Williams.
84. Tsui, W. H., G. Yim, H. H. Wang, J. E. 2003. LuxS and autoinducer-2: their contribu-
McClure, M. G. Surette, and J. Davies. 2004. tion to quorum sensing and metabolism in bacte-
Dual effects of MLS antibiotics: transcriptional ria. Adv.Appl. Microbiol. 53:291–396.
modulation and interactions on the ribosome. 96. Wright, G. D. 2007. The antibiotic resistome:
Chem. Biol. 11:1307–1316. the nexus of chemical and genetic diversity. Nat.
85. Twomey, D., R. P. Ross, M. Ryan, B. Meaney, Rev. Microbiol. 5:175–186.
and C. Hill. 2002. Lantibiotics produced by lactic 97. Wu, H., Z. Song, M. Hentzer, J. B.Andersen,
acid bacteria: structure, function and applications. S. Molin, M. Givskov, and N. Hoiby. 2004.
Antonie Van Leeuwenhoek 82:165–185. Synthetic furanones inhibit quorum-sensing and
86. VanBogelen, R.A., and F. C. Neidhardt. 1990. enhance bacterial clearance in Pseudomonas aerug-
Ribosomes as sensors of heat and cold shock in inosa lung infection in mice. J. Antimicrob.
Escherichia coli. Proc. Natl. Acad. Sci. USA Chemother. 53:1054–1061.
87:5589–5593. 98. Xavier, K. B. and B. L. Bassler. 2003. LuxS
87. VanBogelen, R. A., E. E. Schiller, J. D. quorum sensing: more than just a numbers game.
Thomas, and F. C. Neidhardt. 1999. Diagnosis Curr. Opin. Microbiol. 6:191–197.
of cellular states of microbial organisms using pro- 99. Yang, F., L. H.Wang, J.Wang,Y. H. Dong, J.
teomics. Electrophoresis 20:2149–2159. Y. Hu, and L. H. Zhang. 2005. Quorum
88. van Overbeek, L. S., E. M. H. Wellington, quenching enzyme activity is widely conserved
S. Egan, K. Smalla, H. Heuer, J. M. Collard, in the sera of mammalian species. FEBS Lett.
G. Guillaume, A. D. Karagouni, T. L. 579:3713–3717.
Nikolakopoulou, and J. D. van Elsas. 2002. 100. Yim, G., F. de la Cruz, G. B. Spiegelman,
Prevalence of streptomycin-resistance genes in and J. Davies. 2006. Transcription modulation
bacterial populations in European habitats. FEMS of Salmonella enterica serovar Typhimurium pro-
Microbiol. Ecol. 42:277–288. moters by sub-MIC levels of rifampin. J. Bacteriol.
89. Velicer, G. J. 2003. Social strife in the microbial 188:7988–7991.
world. Trends Microbiol. 11:330–337. 101. Yim, G., H. H.Wang, and J. Davies. 2006.The
90. Vendeville, A., K. Winzer, K. Heurlier, C. M. truth about antibiotics. Int. J. Med. Microbiol.
Tang, and K. R. Hardie. 2005. Making ‘sense’ of 296:163–170.
metabolism: autoinducer-2, LuxS and pathogenic 102. Yonath, A. 2005. Antibiotics targeting ribo-
bacteria. Nat. Rev. Microbiol. 3:383–396. somes: resistance, selectivity, synergism and cellu-
91. Verdine, G. L. 1996. The combinatorial chem- lar regulation. Annu. Rev. Biochem. 74:649–679.
istry of nature. Nature 384:11–13. 103. Zhou, X. X.,W. F. Li, G. X. Ma, and Y. J. Pan.
92. Visick, K. L. 2005. Layers of signaling in 2006. The nisin-controlled gene expression sys-
a bacterium-host association. J. Bacteriol. 187: tem: construction, application and improve-
3603–3606. ments. Biotechnol.Adv. 24:285–295.
SIGNAL INTEGRATION IN THE VIBRIO
HARVEYI AND VIBRIO CHOLERAE
QUORUM-SENSING CIRCUITS
Brian Hammer and Bonnie L. Bassler
20
The marine animal pathogen Vibrio harveyi and THE V. HARVEYI
the human pathogen Vibrio cholerae are aquatic QUORUM-SENSING CIRCUIT
bacteria that engage in a process of cell-cell The marine pathogen V. harveyi associates with
communication called quorum sensing (QS). many types of invertebrates and fish in the ocean
Both microbes orchestrate a QS response by (1).These bacterial-host interactions often cause
the production, secretion, and response to mul- serious disease in the eukaryote, but V. harveyi
tiple autoinducer (AI) molecules. Responses to does not cause disease in humans. V. harveyi QS-
AI inputs are modulated by a variety of addi- regulated genes are controlled by the produc-
tional internal and external factors that overlay tion, detection, and response to three AIs:
the basic QS signaling scaffolds. Factors includ- HAI-1,AI-2, and CAI-1. HAI-1 is a 3-OH-C4-
ing the prevailing chemistry of the niche, the homoserine lactone and is synthesized by
composition of the biota present, the presence LuxM. HAI-1 facilitates intraspecies communi-
or absence of additional chemical signaling cation among V. harveyi and apparently also the
molecules, and also the metabolic state of the closely related species Vibrio parahaemolyticus (4,
organism all influence the Vibrio QS response. 7). The structure of CAI-1 has not yet been
Surprisingly, the molecular mechanisms under- determined. CAI-1 is produced by the CqsA
lying the integration of sensory information synthase,and is proposed to enable communica-
differ in V. harveyi and V. cholerae, although the tion broadly among members of the genus Vib-
architectures of the two organisms’ QS systems rio (19, 31). AI-2 is a family of interconverting
appear quite similar. Differences in the net- molecules that require LuxS for their synthesis,
works likely enable each microbe to uniquely and AI-2 is proposed to be a rather universal sig-
tailor its QS response for adaptation to its par- nal that fosters interspecies communication in
ticular niche. the bacterial world (3, 8, 32, 48).
The V. harveyi QS circuit is shown in
Brian Hammer Department of Molecular Biology, Princeton Fig. 1. The AIs are detected by cognate two-
University, Princeton, New Jersey 08544-1014. Bonnie L. component sensor kinase proteins called CqsS
Bassler Howard Hughes Medical Institute, Chevy Chase, (CAI-1), LuxN (HAI-1), and LuxQ (AI-2).
MD 20815, and Department of Molecular Biology, Prince-
ton University, Princeton, New Jersey 08544-1014. While HAI-1 and CAI-1 apparently bind
323
324 ■ HAMMER AND BASSLER
FIGURE 1 Model of the V. harveyi quorum-sensing system. Three parallel sensory systems
converge to regulate quorum-sensing gene expression by controlling the activity of LuxO.The
three autoinducers are CAI-1 (circles), HAI-1 (pentagons), and AI-2 (triangles). LuxO~P, along
with 54, activates transcription of the genes encoding the Qrr sRNAs that indirectly regulate
LuxR protein levels by destabilizing the luxR mRNA.This process is mediated by the sRNA
chaperone Hfq. Alterations in the transcription of the multiple sRNAs, in turn, produce an
increasing gradient of LuxR protein as the cells transition from low to high cell density. Question
marks denote additional regulators proposed to control qrr expression.OM,outer membrane;IM,
inner membrane.
directly to their cognate sensors, AI-2 binds to 14). LuxO~P activates the transcription of
LuxP, a periplasmic binding protein that associ- genes encoding small regulatory RNAs (Qrr1-
ates with the sensor kinase protein LuxQ 5, for quorum-regulatory RNAs), in a 54-
(5, 36). At low cell density (in the absence of dependent manner (25, 26). The Qrr sRNAs
appreciable AIs), information in the form of participate with the RNA chaperone Hfq in
phosphate flows from each sensor kinase to the the posttranscriptional destabilization of the
phosphotransfer protein, LuxU, which phos- mRNA encoding the master regulator of QS,
phorylates the response regulator, LuxO (13, LuxR (44, 47). Presumably, negative regulation
20. SIGNAL INTEGRATION IN QUORUM-SENSING CIRCUITS ■ 325
occurs through the mutual destruction of the As described above, each AI is detected by a
Qrr sRNAs and the luxR mRNA (25). cognate membrane-bound sensor kinase pro-
At high cell density (when there is an abun- tein: CAI-1 by CqsS, and AI-2 by LuxPQ (19,
dance of AIs), the binding of the signaling mol- 31).Again, in a manner paralleling V. harveyi, at
ecules to their cognate sensors switches the low cell density, the sensor kinases phosphory-
sensors from acting as kinases to acting as phos- late LuxO via the phosphotransfer protein
phatases (13, 14). Phosphate is removed from LuxU. In V. cholerae, LuxO~P acts with 54 in
LuxO,via LuxU,and inactive LuxO is unable to the transcriptional activation of genes encoding
activate expression of the qrr genes.The mRNA four small regulatory RNAs (Qrrs 1 to 4).The
of luxR is translated, and LuxR protein is pro- four sRNAs, along with Hfq, negatively regu-
duced. late the stability of the mRNA of HapR (the
The LuxR-dependent activation of the homologue of LuxR in V. harveyi) (Fig. 2) (25).
luciferase operon (lux) of V. harveyi accounts for Phosphate flow is reversed at high cell density,
the hallmark expression of bioluminescence in leading to the production of HapR.
this organism only at high cell density (29, 35, In addition to the sensor kinases CqsS and
39). Additional target genes are also under the LuxPQ, V. cholerae also possesses a sensor kinase
control of QS in V. harveyi, such as those and response regulator pair, VarS/VarA (24),
required for the type III virulence secretion sys- that are homologous to two-component sen-
tem, siderophore production, metalloprotease sory systems found in other bacteria such as
production, and biofilm formation (18, 26, 33, Escherichia coli (BarA/UvrY), Pseudomonas
47).Like luciferase,all of the V.harveyi QS target aeruginosa (GacS/GacA), and Legionella pneu-
genes identified to date require LuxR for their mophila (LetS/LetA) (17, 34). In V. cholerae, at
response to AIs (47). Some target genes are low cell density,VarS is inactive and is unable
directly controlled by LuxR, while others are to phosphorylate VarA. Inactive VarA cannot
controlled indirectly. The steps linking LuxR activate transcription of the genes encoding the
to expression of the indirectly controlled target Hfq-independent sRNAs called CsrB, CsrC,
genes are currently unknown. and CsrD. This allows the posttranscriptional
regulatory protein CsrA to further activate
THE V. CHOLERAE QUORUM- LuxO. The mechanism for this activation is
SENSING CIRCUIT unclear, but CsrA-dependent activation of
The human pathogen V. cholerae lives in marine LuxO, in turn, further activates the expression
environments, where it often is found attached of the qrr genes, and as mentioned, the Qrr
to the surface of phytoplankton and zooplank- sRNAs destabilize the mRNA of hapR at low
ton (9).These associations are thought to aid in cell density (25).
the survival of this pathogen between epi- At high cell density, an unidentified stimulus
demics. When ingested by humans who con- activates VarS kinase activity (46, 54). Active
sume contaminated seafood or impure water, V. VarS phosphorylates VarA, which subsequently
cholerae colonizes the intestine with the aid of activates the transcription of genes encoding
the toxin-coregulated pilus (TCP), and it the CsrBCD sRNAs (24).The multiple stem-
secretes cholera enterotoxin (CT), which is loops of these sRNAs display AGGA motifs
responsible for the often fatal diarrhea associ- that resemble ribosome-binding sites and are
ated with the disease cholera (10). proposed to bind and sequester CsrA (27). In V.
Attachment and CT production are two cholerae, sequestration of CsrA promotes inacti-
major processes used by V. cholerae for associa- vation of LuxO, permitting the translation of
tion with human and aquatic host cells, and HapR (24). It is not yet known whether V. har-
both of these processes are coordinated by QS veyi encodes a homologous VarS/VarA system.
(16, 52, 53). Analogous to V. harveyi,V. cholerae Recent evidence indicates that growth
utilizes CAI-1 and AI-2 as AI signals, but unlike phase also influences V. cholerae QS in a manner
V. harveyi,V. cholerae does not use HAI-1 (Fig. 2). independent of the AIs. In addition to the
FIGURE 2 Model of the V. cholerae quorum-sensing system. Multiple sensory systems converge to regulate quorum-sensing gene expression by
controlling the activity and/or levels of LuxO.The two autoinducers are CAI-1 (circles) and AI-2 (triangles). (A) Low cell density:The two quorum-
sensing circuits (CAI- 1/CqsS and AI-2/LuxPQ) and the VarS/VarA-CsrA/BCD global regulatory system function in conjunction with Fis to
increase the amount and/or activity of LuxO-phosphate. LuxO~P, along with 54, activates transcription of the genes encoding the Qrr sRNAs.The
Qrr sRNAs indirectly regulate HapR protein levels by destabilizing the hapR mRNA.This process is mediated by the RNA chaperone Hfq. (B)
High cell density: Phosphate is drained from LuxO. Fis and VarS/A-CsrA/BCD are inactive. Under these conditions, hapR mRNA is stabilized and
HapR protein is produced.The lightning bolt represents the putative signal detected by VarS. Dotted lines denote hypothetical interactions. OM,
outer membrane; IM, inner membrane.
CqsS/CAI-1, LuxPQ/AI-2, and VarS/VarA ample, the PrtV protease is activated by HapR
inputs, the small nucleoid protein, Fis, acts in and protects V. cholerae from grazing by bacteri-
the QS cascade (2, 20). Fis is proposed to bend ovorous predators (45). HapR also activates
the DNA of promoter regions and aid in ini- expression of the vc1917 gene, which is
tiation of transcription by binding to the - required for DNA uptake (i.e., natural compe-
subunit of RNA polymerase (12). In V. cholerae, tence). However, this process only occurs when
expression of fis is maximal at low cell density V. cholerae is attached to the chitinous exoskele-
and decreases at high cell density. At low cell ton of zooplankton. Apparently, chitin serves
density, when Fis is produced, it binds to the as an additional signal that triggers expression
qrr promoter regions, enhancing their tran- of the regulator TfoX, which simultaneously
scription. This further ensures that HapR is induces genes for chitin degradation and for
not produced when AI levels are low (23). In competence (30). The stationary-phase sigma
this manner, signals that alert bacteria to factor RpoS also plays a role in expression of
changes in growth phase also impinge on the HapR in response to stress,linking QS to nutri-
Qrr sRNAs, which control HapR levels and, ent limitation (50).Thus, the model for natural
ultimately, the entire V. cholerae QS regulon.The competence control in V. cholerae requires the
role of Fis in the V. harveyi QS circuit has not presence of three signals: the AIs, nutrient stress,
been investigated. and chitin. Together, all of these findings
All QS information in V. cholerae is funneled demonstrate that QS, via HapR, coordinates
into the regulation of HapR. Once expressed, behaviors in V. cholerae that aid in its success in
HapR binds promoter DNA elements and both aquatic systems and in the human host.
either activates or represses gene expression
(21, 22). Many of the QS target genes identified CHEMISTRY AND MICROBIOTA
are required for association with eukaryotes AFFECT THE VIBRIO QS RESPONSES
(53). For example, expression of the genes Many factors in the environment have the
encoding both CT and TCP is repressed by potential to alter bacterial responses to AIs. In
HapR at high cell density. In this case, HapR the cases germane to Vibrio species, the chemi-
binds directly to and represses the promoter of cal composition of the environment has been
aphA, the product of which is required for shown to alter the balance and blend of AI-2
expression of the ctx and tcp genes (22). HapR molecules present. AI-2 is derived from S-
also indirectly represses expression of adenosylmethionine in three enzymatic steps
exopolysaccharide biosynthesis genes (vpsA–K (37). First, S-adenosylmethionine serves as a
and vpsL–Q) that are required for attachment to methyl donor for many biochemical processes,
surfaces, allowing the bacteria to form biofilms and these methyltransferase-dependent reac-
(16). Vps-dependent biofilms are proposed to tions yield S-adenosylhomocysteine. Second,
be important for efficient colonization of the S-adenosylhomocysteine is metabolized to
human host (52). The current model predicts adenine and S-ribosylhomocysteine by the
that following colonization of the host at low enzyme Pfs, and third, S-ribosylhomocysteine
cell density, QS allows V. cholerae to establish is the substrate for the LuxS enzyme. LuxS
biofilms and express CT and TCP. However, cleaves S-ribosylhomocysteine to generate
once high cell density is achieved, the pathogen homocysteine and the unstable 4,5-dihydroxy-
stops expressing CT and TCP and terminates 2,3-pentanedione (DPD) molecule (Fig. 3), the
biofilm production, and these events promote precursor of all AI-2 molecules (32, 37).
the release of V. cholerae back into the environ- DPD interconverts into different moieties in
ment (16, 52). solution (32, 38). In aquatic habitats, the con-
In the marine environment, HapR-depend- centration of boron averages 0.4 mM (6).
ent QS-target genes also play a role in the Under this condition, DPD readily cyclizes and
association of V. cholerae with host cells. For ex- reacts with borate to form (2S, 4S)-2-methyl-
FIGURE 3 Chemistry of AI-2 signaling molecules. Model showing the proposed pathways for the formation of
AI-2 signaling molecules recognized by V. harveyi (upper branch) and enteric bacteria (lower branch). Both
signals are derived from the common precursor DPD, the product of the LuxS reaction. S-THMF-borate binds to
the V. harveyi receptor LuxP, whereas R-THMF binds to the enteric receptor LsrB.Adapted from reference 32 with
permission from Elsevier.
2,3,3,4-tetrahydroxytetrahydrofuran borate AI-2 levels, and other classes of AIs are clearly
(S-THMF borate) (Fig. 3, upper pathway). manipulated (51), but here we restrict the dis-
Consistent with this, the crystal structure of the cussion to AI-2 and how that pertains to Vibrio
AI-2 receptor LuxP from V. harveyi identified QS. Appropriate and distinct responses to
S-THMF borate as the signaling ligand (8). potentially different communities are possible
Interestingly, in vitro studies show that under because of signal integration in the Vibrio cir-
conditions with limited boron, the S-THMF cuits. For example, activation of the master reg-
moiety cannot be found, and V. harveyi does not ulator LuxR (HapR) occurs in response to
respond to AI-2 (32, 38).These results suggest AI-2 produced by LuxS from other bacteria
that the QS response of V. harveyi,V. cholerae, and grown in coculture with vibrios (40, 48). And
likely all other AI-2-responsive bacteria is half of all sequenced bacterial genomes possess
highly affected by the chemical environment. luxS, suggesting that many bacterial species
Consistent with this notion, enteric bacteria, contribute to the total level of AI-2 in any given
which live in environments limited for boron, environment.Due to the rapid equilibrium that
use R-THMF, an unborated rearranged DPD exists between rearranging AI-2 moieties, irre-
moiety, as their AI-2 signal (Fig. 3, lower path- spective of which AI-2 derivative a particular
way) (32, 49). bacterium detects, it can nonetheless monitor
The architectures of the V. harveyi and V. the total concentration of AI-2 supplied by all
cholerae QS circuits are particularly well suited the producers in a niche.
to allow the vibrios to monitor the species AI-2-producing members of a bacterial
composition of the nascent bacterial consortia consortium aid in promoting the activation of
because they are sensitive to multiple signals the master regulator (LuxR/HapR) of the V.
derived from multiple sources. Thus, a second harveyi or V. cholerae QS systems and, in turn, the
changing environmental parameter that influ- downstream target genes. Importantly, how-
ences the Vibrio AI response is the other species ever, depletion of AI-2 pools also occurs in
or genera that vibrios encounter in their partic- mixed species consortia. Enteric bacteria, such
ular niches (43). In mixed species consortia, as E.coli,which V.cholerae must encounter in the
other microbes also have the potential to alter human intestine during infection, are LuxS.
E. coli not only secretes AI-2, but it also removes Host cells to which vibrios attach may also
AI-2 from the surroundings.Enterics possess an manipulate the QS response of the bacteria
Lsr (LuxS-regulated) import system responsible with which they associate. For example, the
for binding and importing AI-2 from the extra- red seaweed, Delisea pulchra, produces a variety
cellular environment (41, 42, 49).The Lsr trans- of halogenated furanones that are structurally
porter is composed of LsrB, a periplasmic similar to acyl-homoserine lactone AI mole-
binding protein that binds R-THMF (Fig. 3) cules. Synthetic furanones and those extracted
(32, 42). Channel proteins LsrC and LsrD from D. pulchra protect several eukaryotic
mediate the delivery of the ligand across the hosts, such as the black tiger prawn and brine
membrane. LsrA is an ATPase that supplies the shrimp, from infection by pathogenic vibrios
energy required for transport. Rapid Lsr- (11, 15, 28). While it is thought that the fura-
dependent transport of R-THMF into the cell nones act through the canonical QS circuits,
occurs at high cell densities (41, 49).This event the exact mechanism by which the furanones
further activates expression of the lsr genes. act is not defined. They appear to antagonize
Therefore, enteric bacteria can both contribute signals in vibrios that possess distinct QS
to and deplete AI-2 from their surroundings. regulatory cascades and use structurally dissim-
The consequences of the alternations in lev- ilar AI signals.
els of AI-2 that occur in mixed species consor-
tia depend on the growth phase of the bacteria DIFFERENCES IN THE
present. In model V. cholerae-E. coli mixtures, at REGULATION OF THE Qrr sRNAs
early times during growth, the contribution of IN V. HARVEYI AND V. CHOLERAE
AI-2 by E. coli activates QS, i.e., HapR produc- The basic architectures of the V. harveyi and V.
tion in V. cholerae (48). However, at later times cholerae QS circuits are similar; however, the dif-
during growth, the Lsr transporter is induced, ferent environmental and signal inputs detected
and it removes AI-2 from the extracellular by each bacterium (described above) produce
surroundings. This latter process significantly differences in their respective QS outputs.
decreases HapR levels and terminates QS Equally important are distinct regulatory differ-
behaviors in V. cholerae.Thus, early on, V. cholerae ences embedded in each of the QS networks.
counts both its own cell number and the One particular regulatory difference that has
number of E. coli present; however, at later a dramatic effect on the QS responses of
times, V. cholerae miscounts and fails to carry out the two vibrios is the distinct roles the Qrr
a QS response appropriate for the cell density sRNAs play in controlling the master regula-
present (48).This AI-2 interaction is two way, tors LuxR/HapR.
because the AI-2 produced by V. cholerae in The function of the Qrr sRNAs was first
these mixtures also alters expression of the AI- analyzed in V. cholerae (25). As mentioned, the
2-dependent Lsr transporter in E. coli. Such Qrr sRNAs act with Hfq to control the stability
communication is likely common in niches of the hapR mRNA (Fig. 2). Importantly, the V.
where multiple species exist in close proximity. cholerae Qrr sRNAs act redundantly. Only a V.
In the human gut, these interactions could have cholerae mutant with all four qrr genes deleted
significant consequences for the maintenance displays the QS null phenotype. Mutants of V.
of the normal microflora. As more bacterial cholerae with deletions of any three qrr genes
species are identified and sequenced from the maintain the wild-type QS phenotype. On the
environment and from the human intestine,it is basis of these findings, a model was proposed in
likely that other microbes will be uncovered which a balance between the Qrr sRNAs and
that contribute to and/or interfere with AI-2- the hapR mRNA is crucial for an ultrasensitive
mediated QS. Understanding these bacterial (switch-like) response to AIs. Specifically, this
chemical interactions could play an important model predicts that the mutual destruction of
role in the prevention of bacterial diseases. the Qrr sRNAs and the hapR mRNA target
provides a mechanism for the bacteria to luxR expression that could fine-tune the V.
respond rapidly to changes in the relative levels harveyi QS response to different environmen-
of the sRNAs and their target mRNA.A slight tal niches (44).
increase of the rate of sRNA synthesis tips the Why do the Qrr sRNAs function so differ-
balance in favor of the sRNAs and leads to low ently in two bacterial species that appear to
levels of hapR mRNA. In contrast, a slight share a highly conserved QS circuit? It is possi-
increase in hapR mRNA synthesis leads to ble that V. harveyi and V. cholerae have evolved
rapid accumulation of hapR mRNA at the unique molecular strategies to tailor their QS
expense of the sRNAs. This model, therefore, responses for survival in their distinct habitats.
predicts an “all-or-none” response to AIs in V. V. harveyi is a shrimp and fish pathogen, whereas
cholerae (25). V. cholerae is a human pathogen.Thus, although
Recent studies in V. harveyi show that it pos- both organisms use similar QS components
sesses five qrr genes, like its closest Vibrio rela- to regulate gene expression, each species has
tives. Examination of their functions reveals modified this basic scaffold,and the target genes
that, in stark contrast to V. cholerae, in V. harveyi controlled by it, to precisely meet its specific
the Qrr sRNAs act additively to control luxR needs.
mRNA levels (44). Each V. harveyi Qrr sRNA
appears to repress luxR mRNA to a different REFERENCES
degree, and the cumulative concentration of all 1. Austin, B., and X. H. Zhang. 2006. Vibrio har-
the Qrr sRNAs equals that of the luxR veyi: a significant pathogen of marine vertebrates
and invertebrates. Lett. Appl. Microbiol. 43:
mRNA.The result of this mechanism is not an 119–124.
“all-or-none” response, as in V. cholerae, but 2. Ball, C. A., R. Osuna, K. C. Ferguson, and R.
rather a more graded response to AIs in V. har- C. Johnson. 1992. Dramatic changes in Fis levels
veyi (44). These results are consistent with a upon nutrient upshift in Escherichia coli. J. Bacteriol.
recent study showing that many QS target 174:8043–8056.
3. Bassler, B. L., E. P. Greenberg, and A. M.
genes in V. harveyi display a graded response to Stevens. 1997.Cross-species induction of lumines-
AI accumulation (48). cence in the quorum-sensing bacterium Vibrio
In both V. harveyi and V. cholerae, the qrr harveyi. J. Bacteriol. 179:4043–4045.
sRNA genes appear to be controlled by func- 4. Bassler, B. L., M.Wright, R. E. Showalter, and
tions in addition to QS components.For exam- M. R. Silverman. 1993. Intercellular signalling
in Vibrio harveyi: sequence and function of genes
ple, in V. harveyi, deletion of qrr5 does not alter regulating expression of luminescence. Mol. Micro-
the luxR mRNA or LuxR protein levels. biol. 9:773–786.
However, qrr5, when expressed in E. coli or 5. Bassler, B. L., M. Wright, and M. R.
overexpressed in V. harveyi, does control luxR Silverman. 1994. Multiple signalling systems con-
mRNA expression. These results suggest trolling expression of luminescence in
Vibrio harveyi: sequence and function of genes
that a V. harveyi-specific repressor, which is encoding a second sensory pathway. Mol. Microbiol.
not present in E. coli, impinges on qrr5 expres- 13:273–286.
sion in vivo (44). In addition, the strength of qrr 6. Bowen, H. J. M. 1966. Trace Elements in Biochem-
expression in V. harveyi (qrr4 2 3 1 5) istry. Academic, London, United Kingdom.
is not preserved in E. coli (qrr3 4 1 5 7. Cao, J. G., and E.A. Meighen. 1989. Purification
and structural identification of an autoinducer for
2), also suggesting that unidentified regula- the luminescence system of Vibrio harveyi.
tory factors function at the distinct qrr pro- J. Biol. Chem. 264:21670–21676.
moter regions in V. harveyi. Preliminary 8. Chen, X., S. Schauder, N. Potier, A. Van
bioinformatics analysis bolsters this prediction Dorsselaer, I. Pelczer, B. L. Bassler, and
by identifying putative binding motifs in some F. M. Hughson. 2002. Structural identification
of a bacterial quorum-sensing signal containing
qrr promoter regions that are absent from other boron. Nature 415:545–549.
qrr promoter regions. These findings suggest a 9. Colwell, R. R. 1996. Global climate and
possible mechanism for integrating additional infectious disease: the cholera paradigm. Science
(i.e., metabolic) signals into the control of 274:2025–2031.
10. Cotter, P. A., and V. J. DiRita. 2000. Bacterial 24. Lenz, D. H., M. B. Miller, J. Zhu, R. V.
virulence gene regulation: an evolutionary per- Kulkarni, and B. L. Bassler. 2005. CsrA and
spective. Annu. Rev. Microbiol. 54:519–565. three redundant small RNAs regulate quorum
11. Defoirdt, T., R. Crab, T. K. Wood, P. sensing in Vibrio cholerae. Mol. Microbiol. 58:
Sorgeloos, W.Verstraete, and P. Bossier. 2006. 1186–1202.
Quorum sensing-disrupting brominated fura- 25. Lenz, D. H., K. C. Mok, B. N. Lilley, R. V.
nones protect the gnotobiotic brine shrimp Kulkarni, N. S. Wingreen, and B. L. Bassler.
Artemia franciscana from pathogenic Vibrio harveyi, 2004.The small RNA chaperone Hfq and multiple
Vibrio campbellii,and Vibrio parahaemolyticus isolates. small RNAs control quorum sensing in Vibrio har-
Appl. Environ. Microbiol. 72:6419–6423. veyi and Vibrio cholerae. Cell 118:69–82.
12. Dorman, C. J., and P. Deighan. 2003. Regula- 26. Lilley, B. N., and B. L. Bassler. 2000. Regula-
tion of gene expression by histonelike proteins in tion of quorum sensing in Vibrio harveyi by LuxO
bacteria. Curr. Opin. Genet. Dev. 13:179–184. and sigma-54. Mol. Microbiol. 36:940–954.
13. Freeman, J. A., and B. L. Bassler. 1999. A 27. Liu, M.Y., G. Gui, B.Wei, J. F. Preston, 3rd, L.
genetic analysis of the function of LuxO, a two- Oakford, U. Yuksel, D. P. Giedroc, and T.
component response regulator involved in quo- Romeo. 1997.The RNA molecule CsrB binds to
rum sensing in Vibrio harveyi. Mol. Microbiol. 31: the global regulatory protein CsrA and antago-
665–677. nizes its activity in Escherichia coli. J. Biol. Chem.
14. Freeman, J. A., and B. L. Bassler. 1999. 272:17502–17510.
Sequence and function of LuxU:a two-component 28. Manefield, M.,T. B. Rasmussen, M. Henzter,
phosphorelay protein that regulates quorum J. B.Andersen, P. Steinberg, S. Kjelleberg, and
sensing in Vibrio harveyi. Mol. Microbiol. 31: M. Givskov. 2002. Halogenated furanones inhibit
665–677. quorum sensing through accelerated LuxR
15. Givskov, M., R. de Nys, M. Manefield, L. turnover. Microbiology 148:1119–1127.
Gram, R. Maximilien, L. Eberl, S. Molin, P. D. 29. Martin, M., R. Showalter, and M. Silverman.
Steinberg, and S. Kjelleberg. 1996. Eukaryotic 1989. Identification of a locus controlling expres-
interference with homoserine lactone-mediated sion of luminescence genes in Vibrio harveyi. J. Bac-
prokaryotic signalling. J. Bacteriol. 178:6618–6622. teriol. 171:2406–2414.
16. Hammer, B. K., and B. L. Bassler. 2003. Quo- 30. Meibom, K. L., M. Blokesch, N.A. Dolganov,
rum sensing controls biofilm formation in Vibrio C. Y. Wu, and G. K. Schoolnik. 2005. Chitin
cholerae. Mol. Microbiol. 50:101–104. induces natural competence in Vibrio cholerae. Sci-
17. Heeb, S., and D. Haas. 2001.Regulatory roles of ence 310:1824–1827.
the GacS/GacA two-component system in plant- 31. Miller, M. B., K. Skorupski, D. H. Lenz, R. K.
associated and other gram-negative bacteria. Mol. Taylor, and B. L. Bassler. 2002. Parallel quorum
Plant-Microbe Interact. 14:1351–1363. sensing systems converge to regulate virulence in
18. Henke, J. M., and B. L. Bassler. 2004. Quorum Vibrio cholerae. Cell 110:303–314.
sensing regulates type III secretion in Vibrio harveyi 32. Miller, S.T., K. B. Xavier, S. R. Campagna, M.
and Vibrio parahaemolyticus. J. Bacteriol. 186: E.Taga, M. F. Semmelhack, B. L. Bassler, and
3794–3805. F. M. Hughson. 2004. Salmonella typhimurium rec-
19. Henke, J. M., and B. L. Bassler. 2004. Three ognizes a chemically distinct form of the bacterial
parallel quorum-sensing systems regulate gene quorum-sensing signal AI-2.Mol.Cell 15:677–687.
expression in Vibrio harveyi. J. Bacteriol. 186: 33. Mok, K. C., N. S.Wingreen, and B. L. Bassler.
6902–6914. 2003. Vibrio harveyi quorum sensing: a coincidence
20. Ishihama, A. 1999. Modulation of the nucleoid, detector for two autoinducers controls gene
the transcription apparatus, and the translation expression. EMBO J. 22:870–881.
machinery in bacteria for stationary phase survival. 34. Molofsky,A. B., and M. S. Swanson. 2004.Dif-
Genes Cells 4:135–143. ferentiate to thrive: lessons from the Legionella
21. Jobling, M. G., and R. K. Holmes. 1997. Char- pneumophila life cycle. Mol. Microbiol. 53:29–40.
acterization of hapR, a positive regulator of the 35. Nealson, K. H., and J.W. Hastings. 1979. Bac-
Vibrio cholerae HA/protease gene hap, and its iden- terial bioluminescence: its control and ecological
tification as a functional homologue of the Vibrio significance. Microbiol. Rev. 43:496–518.
harveyi luxR gene. Mol. Microbiol. 26:1023–1034. 36. Neiditch, M. B., M. J. Federle,A. J. Pompeani,
22. Kovacikova, G., and K. Skorupski. 2002. Reg- R. C. Kelly, D. L. Swem, P. D. Jeffrey, B. L.
ulation of virulence gene expression in Vibrio Bassler, and F. M. Hughson. 2006. Ligand-
cholerae by quorum sensing: HapR functions at the induced asymmetry in histidine sensor kinase
aphA promoter. Mol. Microbiol. 46:1135–1147. complex regulates quorum sensing. Cell 126:
23. Lenz, D. H., and B. L. Bassler. 2007.The small 1095–1108.
nucleoid protein Fis is involved in Vibrio cholerae 37. Schauder, S., K. Shokat, M. G. Surette, and B.
quorum sensing. Mol. Microbiol. 63:859–871. L. Bassler. 2001. The LuxS family of bacterial
autoinducers: biosynthesis of a novel quorum- 46. Valverde, C., S. Heeb, C. Keel, and D.
sensing signal molecule. Mol. Microbiol. 41: Haas. 2003. RsmY, a small regulatory RNA, is
463–476. required in concert with RsmZ for GacA-
38. Semmelhack, M. F., S. R. Campagna, C. dependent expression of biocontrol traits in
Hwa, M. J. Federle, and B. L. Bassler. 2004. Pseudomonas fluorescens CHA0. Mol. Microbiol.
Boron binding with the quorum sensing signal 50:1361–1379.
AI-2 and analogues. Org. Lett. 6:2635–2637. 47. Waters, C. M., and B. L. Bassler. 2006. The
39. Showalter, R. E., M. O. Martin, and M. R. Vibrio harveyi quorum-sensing system uses
Silverman. 1990. Cloning and nucleotide shared regulatory components to discriminate
sequence of luxR, a regulatory gene controlling between multiple autoinducers. Genes Dev.
bioluminescence in Vibrio harveyi. J. Bacteriol. 20:2754–2767.
172:2946–2954. 48. Xavier, K. B., and B. L. Bassler. 2005. Interfer-
40. Surette, M. G., M. B. Miller, and B. L. Bassler. ence with AI-2-mediated bacterial cell-cell com-
1999.Quorum sensing in Escherichia coli,Salmonella munication. Nature 437:750–753.
typhimurium, and Vibrio harveyi: a new family of 49. Xavier, K. B., and B. L. Bassler. 2005. Regula-
genes responsible for autoinducer production. tion of uptake and processing of the quorum-
Proc. Natl.Acad. Sci. USA 96:1639–1644. sensing autoinducer AI-2 in Escherichia coli.
41. Taga, M. E., S. T. Miller, and B. L. Bassler. J. Bacteriol. 187:238–248.
2003. Lsr-mediated transport and processing of 50. Yildiz, F. H., X. S. Liu,A. Heydorn, and G. K.
AI-2 in Salmonella typhimurium. Mol. Microbiol. 50: Schoolnik. 2004. Molecular analysis of rugosity
1411–1427. in a Vibrio cholerae O1 El Tor phase variant. Mol.
42. Taga, M. E., J. L. Semmelhack, and B. L. Microbiol. 53:497–515.
Bassler. 2001.The LuxS-dependent autoinducer 51. Zhang, L. H., and Y. H. Dong. 2004. Quorum
AI-2 controls the expression of an ABC trans- sensing and signal interference: diverse implica-
porter that functions in AI-2 uptake in Salmonella tions. Mol. Microbiol. 53:1563–1571.
typhimurium. Mol. Microbiol. 42:777–793. 52. Zhu, J., and J. J. Mekalanos. 2003. Quorum
43. Thompson, F. L.,T. Iida, and J. Swings. 2004. sensing-dependent biofilms enhance colonization
Biodiversity of vibrios. Microbiol. Mol. Biol. Rev. in Vibrio cholerae. Dev. Cell 5:647–656.
68:403–431. 53. Zhu, J., M. B. Miller, R. E. Vance, M.
44. Tu, K. C., and B. L. Bassler. 2007. Multiple Dziejman, B. L. Bassler, and J. J. Mekalanos.
small RNAs act additively to integrate sensory 2002. Quorum-sensing regulators control viru-
information and control quorum sensing in Vibrio lence gene expression in Vibrio cholerae. Proc. Natl.
harveyi. Genes Dev. 21:221–233. Acad. Sci. USA 99:3129–3134.
45. Vaitkevicius, K., B. Lindmark, G. Ou, 54. Zuber, S., F. Carruthers, C. Keel, A. Mattart,
T. Song, C. Toma, M. Iwanaga, J. Zhu, A. C. Blumer, G. Pessi, C. Gigot- Bonnefoy, U.
Andersson, M. L. Hammarstrom, S. Tuck, Schnider-Keel, S. Heeb, C. Reimmann, and
and S. N. Wai. 2006. A Vibrio cholerae protease D. Haas. 2003. GacS sensor domains pertinent to
needed for killing of Caenorhabditis elegans has a the regulation of exoproduct formation and to the
role in protection from natural predator grazing. biocontrol potential of Pseudomonas fluorescens
Proc. Natl.Acad. Sci. USA 103:9280–9285. CHA0. Mol. Plant-Microbe Interact. 16:634–644.
SIGNAL TRAFFICKING WITH BACTERIAL
OUTER MEMBRANE VESICLES
Lauren Mashburn-Warren and Marvin Whiteley
21
QUORUM SENSING AND PQS factors critical for competitiveness are con-
The gram-negative bacterium Pseudomonas trolled by population density. It is hypothesized
aeruginosa is a ubiquitous opportunistic patho- that these factors are critical for colonization
gen that causes infection in immunocom- and persistence in the CF lung where P. aerugi-
promised individuals, including those with the nosa must compete with a wide array of bacter-
heritable disease cystic fibrosis (CF). Patients ial species.Recent evidence that QS-controlled
with CF manifest a host defense defect local- genes are significantly up-regulated in the CF
ized to the conducting airways of the lung that lung supports this hypothesis (38).
results in chronic colonization by several bacte- P.aeruginosa QS involves at least three distinct
rial species. Approximately 80 to 90% of indi- signaling systems (Fig. 1).These systems involve
viduals with CF are infected with P. aeruginosa, biosynthesis of a signaling molecule from com-
and these infections are virtually impossible to mon metabolic intermediates within the cell,
eradicate with conventional therapeutics. Simi- followed by sensing of the signaling com-
lar to many bacteria, P. aeruginosa uses chemical pounds by transcriptional regulatory proteins.
signals (autoinducers) to monitor its population Binding of the signaling molecule to the tran-
density and coordinate group behavior, a scriptional activator elicits changes in gene
process referred to as quorum sensing (QS) transcription through binding of the protein/
(39). QS has been proposed to be important for signal complex to DNA operators.Two of the P.
colonization of the CF lung, and virulence aeruginosa QS systems utilize acyl-homoserine
studies indicate that inactivation of QS in P. lactone signaling molecules and are referred to
aeruginosa significantly reduces virulence in as the las and rhl QS systems (48).The las system
mammalian, plant, and insect models (21, 40, involves production of the signaling molecule
51). QS has also been implicated in P. aeruginosa N-(3-oxododecanoyl)-L-homoserine lactone
competition with other bacteria as numerous (3-O-C12-HSL) by the autoinducer synthase
LasI and sensing of 3-O-C12-HSL by the tran-
scriptional regulator LasR. The rhl system is
Lauren Mashburn-Warren and Marvin Whiteley Section of
Molecular Genetics and Microbiology, The University of similar to the las QS system and involves syn-
Texas at Austin,Austin,Texas 78712. thesis of the signal N-butyryl-L-homoserine
333
334 ■ MASHBURN-WARREN AND WHITELEY
FIGURE 1 A simplified model for P. aeruginosa QS. Each system utilizes a distinct chemical signal that is sensed by
a corresponding transcriptional regulator. The signal/transcriptional regulator complex binds DNA and elicits
changes in gene expression.Approximately 5% of all P. aeruginosa genes are regulated by QS (47, 50).
lactone (C4-HSL) by RhlI and sensing by are likely present under natural circumstances.
RhlR. The third signaling system utilizes PQS biosynthesis proceeds through a head-
a chemically distinct signaling molecule, 2- to-head condensation of anthranilic acid and
heptyl-3-hydroxy-4-quinolone, referred to as -keto-decanoic acid to form the immediate
the pseudomonas quinolone signal (PQS) (42), precursor of PQS, 2-heptyl-4-quinolone
which interacts with the transcriptional regula- (HHQ).Although the precise enzymatic func-
tor PqsR (MvfR) (49).These three systems do tions are not known, PqsABCD are required
not act independently of one another (Fig. 1) for HHQ biosynthesis. HHQ is then hydroxy-
but instead constitute an integrated signaling lated, presumably by the putative hydroxylase
network that controls transcription of approxi- PqsH, to form PQS. It should be noted that
mately 5% of all P. aeruginosa genes (46, 50). HHQ also interacts with PqsR and serves as a
PQS is a member of a large collection of signaling molecule to activate transcription of
quinolone molecules (at least 56 molecules in PqsR-regulated genes (54).
total) produced by P. aeruginosa (30). Many of Since QS signals are synthesized within the
these molecules exist in nature as enantiomeric bacterial cell, the QS paradigm dictates that
mixtures of the quinolone and quinoline these molecules be exuded to the extracellular
forms. For simplicity, PQS will be referred to as milieu for trafficking between cells. For some
a quinolone throughout this chapter, although QS molecules (such as C4-HSL), diffusion
the reader should keep in mind that both forms mediates entry and exit from the cell (41),while
21. SIGNAL TRAFFICKING WITH OUTER MEMBRANE VESICLES ■ 335
more hydrophobic signals (3-O-C12-HSL) uti- for this enrichment is not known. Although
lize active efflux (41). Recent evidence from cytoplasmic proteins have not been identified
our laboratory indicates that the hydrophobic in MVs, it is clear that the process of MV for-
signal PQS is trafficked between bacterial cells mation leads to packaging of periplasmic com-
within membrane vesicles (MVs) liberated ponents (1, 2).
from the bacterial outer membrane (36). The To be utilized as trafficking vehicles, MVs
focus of this chapter is a discussion of MVs, must (i) have the ability to deliver their cargo
including their potential use as trafficking vehi- to other cells and (ii) possess physiologically
cles for a variety of cargo, including cell- relevant cargo, necessitating transfer between
cell signals. cells. This section summarizes current knowl-
edge of the delivery mechanism of MVs and
MEMBRANE VESICLES AS outlines the wide array of molecules within
TRAFFICKING VEHICLES MVs that are trafficked to prokaryotic and
MVs are produced by many gram-negative eukaryotic cells.
bacteria (28), including Escherichia coli and P.
aeruginosa.These bilayered MVs are spherical in Cargo and the Mechanism of Delivery
shape and range in size from 50 to 250 nm in Very little is currently known about how MVs
diameter depending on the bacterial species deliver their cargo; however, it is clear that MVs
from which they are derived (Fig. 2) (37). Elec- produced by a variety of gram-negative bacte-
tron microscopy reveals that MVs are produced ria are capable of delivering their contents to
by “ballooning” and “pinching off ” of the bac- gram-negative and gram-positive bacteria as
terial outer membrane into the surrounding well as eukaryotic cells.The mechanism of how
medium (2, 3, 7, 10, 13, 23, 24, 31, 43). As MVs fuse to the outer surfaces of these distinct
expected, the composition of MVs is similar to cell types is unknown. In the case of gram-
the outer membrane and consists of a bilayered positive bacteria, it has been proposed that the
membrane with an outer leaflet composed of anionic LPS of MVs is cross-bridged to the
lipopolysaccharide (LPS) and an inner leaflet anionic gram-positive surface by the divalent
of phospholipids.MVs also contain a wide array cations Mg2 and Ca2, which are normally
of outer membrane proteins, and there is some present on the surfaces of these bacteria.These
evidence that specific outer membrane proteins cation bridges then permeate the outer portion
may be enriched within MVs as compared to of the anionic LPS, thereby forcing the LPS
the outer membrane; however, the mechanism into a low-curvature structure that ruptures the
MV, causing release of its contents (22). For
gram-negative bacteria, it has been proposed
that the initial interaction between the LPS of
the bacterial cell and the MV LPS would also
be mediated by divalent cation salt bridges;
however, this would be followed by more com-
plex lipid interactions resulting in fusion of the
MV to the bacterial outer membrane (22).The
situation is likely different for MV fusion to
eukaryotic cells. Recent evidence with MVs
from E. coli indicates that a specific protein
within MVs, the heat-labile enterotoxin, serves
as a specific receptor to mediate fusion and
FIGURE 2 Transmission electron micrograph of delivery of MV cargo to eukaryotic cells (26).
negatively stained P. aeruginosa MVs. Diameter of a Overall, the understanding of the molecular
single MV is shown for scale. mechanism of how MVs fuse to these distinct
cell surfaces is in its infancy, and other mecha- and gram-negative bacteria results in degrada-
nisms likely exist; however, it is clear that MVs tion of the peptidoglycan layer surrounding the
are competent to serve as trafficking and deliv- bacterium and cell lysis (22). This MV lytic
ery vehicles for a variety of components. activity was not specific to P. aeruginosa as MVs
Included below is a list of components that have from several gram-negative bacterial species
been shown to use MVs for this purpose. including Citrobacter, Enterobacter, Escherichia,
Klebsiella, Morganella, Proteus, Pseudomonas, Sal-
PQS monella, and Shigella were able to lyse metaboli-
Recent evidence in our laboratory indicates cally inactive gram-positive and gram-negative
that PQS is associated with MVs and that MVs bacteria. As expected, this lysis correlated with
are competent to traffic this molecule between the peptidoglycan chemotype of the target
bacterial cells (36).As discussed previously,PQS cells, with the A1 chemotype being the most
is a highly hydrophobic molecule, and this susceptible to lysis (31).
property likely impairs trafficking of this signal P. aeruginosa MVs also contain an array of
between cells. For planktonically growing P. small molecules including numerous quin-
aeruginosa, approximately 86% of PQS is local- olones. P. aeruginosa produces at least 56 quin-
ized with MVs while approximately 9% is asso- olones that show structural similarity to PQS.
ciated with the bacterial cells and less than 5% is These molecules include HHQ, 4-hydroxy-2-
“free” in the supernatant.This is in contrast to nonyl-quinoline (HNQ), and 4-hydroxy-2-
the more hydrophilic P. aeruginosa signals C4- heptyl-quinoline N-oxide (HQNO) (11) (Fig.
HSL and 3-O-C12-HSL, which are primarily 3). These three molecules have been shown
present (~99%) in the culture supernatant.This to have antimicrobial activity against gram-
observation demonstrates that PQS, but not positive bacteria (34); thus it is not surprising
HSLs,is associated within P.aeruginosa MVs,and that MVs have antimicrobial activity against
these MVs are competent to deliver this signal actively growing Staphylococcus spp. (Fig. 4)
to other bacterial cells within the population (36). Most of the antimicrobial activity was
(36). It is not known if PQS is the only QS sig- extractable with an organic solvent, indicating
naling molecule associated with MVs, but it is that under the experimental condition utilized,
possible (and we believe likely) that hydropho- the MV lytic activity was primarily due to the
bic signaling molecules produced by other bac- antimicrobial quinolones present and not the
terial species will also use MVs as trafficking murein hydrolase packaged within MVs (36).
vehicles for QS signals. The ecological importance of packaging
antimicrobial factors within MVs is not known,
ANTIMICROBIAL FACTORS but several possibilities exist.The most obvious
MVs isolated from P. aeruginosa have significant benefit may be competition with other bacteria
antimicrobial activity,particularly against gram- during growth in complex multispecies envi-
positive bacteria (22, 31, 32).This antimicrobial ronments. For example, the ability to lyse other
activity is multifaceted, including both small- bacteria will likely reduce competition for
molecule and protein components. One pro- nutrients, providing the lytic bacterium with
teinaceous lytic component within P. aeruginosa more nutrients for growth. Another benefit
MVs is a murein hydrolase normally involved in would be utilization of lysed bacteria as nutri-
cell division and peptidoglycan turnover. This ent sources. Recent evidence from our
enzyme is found in the cell envelope and laboratory indicates that P. aeruginosa lyses
periplasm of P. aeruginosa and is presumably Staphylococcus aureus in vivo and uses it as an iron
packaged into MVs along with other periplas- source (35).The ability to utilize lysed bacteria
mic proteins during “pinching off ” of the MV as an iron source likely provides a significant
(32). Delivery of this autolysin to gram-positive benefit in vivo as iron is the limiting nutrient
during infection. One situation where this is

likely important is the CF lung, where P. aerugi-
nosa chronically resides with several other bac-
terial species, including Staphylococcus spp.
PROTECTIVE FACTORS
MVs have also been shown to promote survival
of bacterial populations against specific stresses
including antibiotics. P. aeruginosa strains iso-
lated from the CF lung often exhibit high levels
of antibiotic drug resistance including resist-
ance to commonly utilized -lactam antibi-
otics. This resistance to -lactams is mediated
by increased production of the chromosomally
encoded -lactamase protein that enzymati-
cally inactivates -lactam antibiotics (15, 33).
Much of the -lactamase activity from these
bacterial strains is not associated with the bacte-
rial cells but is instead present in the extracellu-
lar milieu within MVs. -lactamase-loaded
MVs are competent to inactivate exogenously
added -lactam antibiotics, which should
effectively lower the antibiotic concentration
in the CF lung (8). Since MVs are capable of
fusing to surrounding bacteria, there is also
potential for transfer of antibiotic resistance
proteins from a -lactam-resistant bacterium
to a -lactam-sensitive bacterium.Thus, pack-
aging the -lactamase enzyme into MVs may
provide multiple mechanisms for protection
from -lactam treatment, both of which
obviate the need for transfer of the -lactamase
gene to sensitive bacteria. It should be
FIGURE 3 Structures of relevant quinolones pro- noted that packaging -lactamase into MVs
duced by P. aeruginosa. may also protect -lactamase-sensitive non-
FIGURE 4 P. aeruginosa MVs have

antimicrobial activity against gram-positive
bacteria. P. aeruginosa cells (left) and P. aerug-
inosa MVs (right) were spotted onto a
confluent lawn of Staphylococcus epidermidis.
S. epidermidis killing is indicated by zones
of lysis.
pseudomonad bacteria within the population recipient cells (12), thereby providing neigh-
from antibiotic treatment. boring bacteria with the genetic potential to
Along with antibiotics, MVs also protect survive antimicrobial treatment. In E. coli
bacterial cells from the host immune system. 0157:H7 clinical isolates, MVs harbor a plasmid
Capnocytophaga ochracea and Prevotella loescheii containing genes necessary for replication,
are two oral bacterial pathogens often found mobilization, and partitioning. However, this
associated with periodontal diseases.These bac- plasmid does not contain the ability to synthe-
teria are normally killed in human serum by the size the pilus needed for conjugation. Packag-
antibody-activated classical complement cas- ing this plasmid within MVs allows efficient
cade; however, addition of MVs from the oral DNA transfer to recipient cells and eliminates
bacterium Porphyromonas gingivalis to C. ochracea the need for the close proximity of donor and
and P. loescheii inhibited the killing of these two recipient cells normally needed for conjuga-
microorganisms by serum (16).The mechanism tion.This method constitutes a fourth mecha-
for this protection is likely mediated by the nism for DNA transfer that is different from
MVs serving as “decoys”for complement.Since transfection, transformation, and conjugation
MVs are composed of the outer membrane of (55). Thus, DNA transfer via MVs likely pro-
their corresponding cells, MVs likely sequester vides some bacteria with a novel, clinically
complement components, thus reducing the relevant method for transfer of genes critical
number of complement molecules available to for survival in the host.
interact with intact cells. These two examples Since cytoplasmic proteins have not been
provide evidence that in addition to serving as identified in MVs, it is important to understand
trafficking vehicles for specific cargo, MVs may how DNA is packaged into MVs.Although the
also provide protective functions for the bacter- mechanism has not been elucidated, two mod-
ial population through packaging of antibiotic els have been proposed (44). The first model
resistance proteins or through their use as bait involves movement of DNA from the cyto-
to lure away lytic components. plasm into the periplasmic space where it is
entrapped during “pinching off ” of the MVs;
DNA however, it is not known how DNA moves
The composition of MVs has been the subject from the cytoplasm to the periplasm.The sec-
of intense study,and these studies reveal that the ond model involves extracellular DNA enter-
protein components of MVs are derived from ing the periplasm and becoming encapsulated
the outer membrane and periplasm. Although into MVs.The origin of this extracellular DNA
these data suggest that cytoplasmic components and the mechanism by which this DNA enters
are not packaged into MVs, several studies have the periplasm are also unknown.Recent exper-
demonstrated that double-stranded DNA is iments by Renelli et al. support the notion that
present in MVs isolated from the gram- both mechanisms are important for incorpora-
negative bacteria P. aeruginosa, E. coli 0157:H7, tion of DNA into MVs (44).
and Neisseria gonorrhoeae (12,27,44,55).P.aerug-
inosa MVs have been shown to contain double- TOXINS
stranded plasmid DNA, and the MVs serve to The gram-negative oral bacterium Aggregatibac-
protect the plasmid DNA from DNases and ter actinomycetemcomitans is associated with the
exonucleases (44). Although P. aeruginosa MVs early onset of periodontal diseases and endo-
were unable to transform plasmid DNA into carditis. To cause disease, A. actinomycetemcomi-
recipient cells under the conditions tested,MVs tans produces a variety of virulence factors
from other bacteria are capable of delivering including a potent leukotoxin, LtxA. LtxA dif-
plasmids between bacterial cells (12, 27, 44, 55). fers from other secreted toxins in that it lacks a
MVs isolated from N. gonorrhoeae were able to signal peptide sequence required for secretion
transform -lactamase-encoding plasmids into via the type II secretion system. This toxin
requires the help of two other proteins, LtxB nism exists to enrich a receptor into MVs to
and LtxD, for translocation across the cytoplas- facilitate MV binding to target cells.
mic membrane and into the periplasm (29). Another toxin produced by E. coli, ClyA
Once in the periplasm, LtxA partitions into the is also abundantly found in MVs. This pore-
outer membrane and is packaged into MVs. forming cytolysin must exist as oligomers to be
LtxA is trafficked to eukaryotic cells by MVs biologically active. In the periplasm, ClyA is
and is six times more active in MVs. LtxA is also present as inactive monomers; however, within
enriched in MVs compared to other A. actino- MVs, ClyA is present as active oligomers. In
mycetemcomitans outer membrane proteins, sug- fact,ClyA is eight-fold more active when deliv-
gesting a specific sorting mechanism for LtxA ered to eukaryotic cells via MVs than purified
packaging within MVs (25). monomeric protein (52). Collectively these
In the stomach pathogen Helicobacter pylori, it examples signify the diversity of toxins traf-
was proposed that bacterial autolysis was the ficked by MVs.The use of MVs to traffic toxins
primary means of secreting and distributing has a number of potential benefits for the bac-
virulence proteins to host cell targets. Recent terium, including concentrated delivery to
evidence indicates that the cytotoxin VacA is their targets cells; protection of toxins from
present in MVs released by H. pylori both in extracellular degradation; and, in the case of LT,
vitro and in vivo. In vitro, MVs containing VacA targeting of MVs to eukaryotic cell receptors.
were shown to bind gastric epithelial cells at the
plasma membrane. MVs were also present in Mechanisms of MV Formation
epithelial cell endosomes and vacuoles induced Although a wide variety of gram-negative bac-
by VacA. Human gastric biopsies showed the teria produce MVs, the molecular mechanism
presence of VacA containing MVs surrounding underlying their formation is unknown.Several
H. pylori. MVs were also attached to the mucosa models have been proposed, but to understand
and were present in cytoplasmic vacuoles of the these models we must first understand the
gastric epithelium (13). structure of the gram-negative outer mem-
Enterotoxigenic E. coli (ETEC) produces brane from which they are derived. Gram-
MVs enriched with periplasmic components, negative bacteria contain an inner membrane
outer membrane proteins, and a heat-labile composed of phospholipids and an outer mem-
enterotoxin (LT) (14, 19, 26). LT is very similar brane composed of an inner leaflet of phospho-
to Vibrio cholera toxin (CT). Although both LT lipids and an outer leaflet of LPS.The individual
and CT cross the cytoplasmic membrane into LPS molecules are highly anionic, and charge-
the periplasm, ETEC does not have the secre- charge repulsion between individual LPS mol-
tory apparatus to secrete LT; therefore, soluble ecules is negated by the presence of divalent
LT is present in the periplasm (19, 26). Recent cations (Mg2 and Ca2) in the outer mem-
evidence indicates that LT is located inside and brane.These divalent cations form salt bridges
on the surface of ETEC MVs (19, 26). LT is between individual LPS molecules, thereby sta-
tightly associated with MVs and exhibits signif- bilizing the outer membrane. Between the
icant biological activity against eukaryotic tar- inner and outer membranes is a thin layer of
get cells when associated with MVs. Similar to peptidoglycan, which provides structural
LtxA from A. actinomycetemcomitans, LT is integrity for the cell.Since the outer membrane
enriched in MVs compared to other periplas- lies outside the peptidoglycan, it is intrinsically
mic proteins. unstable. To stabilize the outer membrane,
LT also serves as a receptor for eukaryotic specific outer membrane proteins (lipopro-
cells, therefore targeting ETEC vesicles to these teins, peptidoglycan-associated proteins) bind
cells, where they bind and are endocytosed (19, to the underlying peptidoglycan, effectively
26).These data provide an interesting and novel tethering the outer membrane to the rigid
paradigm in which a protein-sorting mecha- peptidoglycan layer. This basic knowledge of
outer membrane regions not tethered to pepti-

doglycan will then “bulge”and be released from
the outer membrane as MVs.This model is sup-
ported by the observations that significantly less
lipoprotein is present in E. coli MVs and that
proteins known to interact with lipoprotein are
largely excluded from MVs (18,53).This model
is not predicated on the complete absence of
lipoprotein from the MVs since approximately
66% of the lipoprotein in E. coli is not cova-
lently bound to peptidoglycan (5, 18, 20); thus,
it is plausible that lipoprotein present within
E. coli MVs is primarily this unbound form.
MODEL 2
MVs are formed when products arising from
normal peptidoglycan turnover are not effi-
ciently internalized by the bacterial cell.These
turnover products build up in the periplasmic
space (56) and exert a turgor pressure on the
outer membrane. This causes the outer mem-
brane to “bulge” and be released as MVs. The
evidence supporting this model is that muramic
FIGURE 5 Proposed mechanisms of MV forma-
tion. (Model 1) In areas lacking peptidoglycan-outer acid, a component of peptidoglycan, is present
membrane protein linkages, MVs form when the outer within MVs (56). Subsequently, Hayashi et al.
membrane grows faster than the underlying peptido- reported that mutations in a P. gingivalis pepti-
glycan layer.This causes the outer membrane to bulge doglycan hydrolase (autolysin) enhance MV
and eventually “pinch off ” to form MVs. (Model 2) formation in P. gingivalis (17). These authors
During normal peptidoglycan turnover, soluble low-
molecular-weight peptidoglycan fragments are not hypothesize that loss of this autolysin may
internalized efficiently by the cell, resulting in accumu- prevent complete degradation of cell wall com-
lation of these fragments in the periplasm. Accumula- ponents, leading to accumulation of peptido-
tion of these small fragments exerts turgor pressure on glycan intermediates in the periplasm. It should
the outer membrane, causing it to swell and form MVs. be noted that this model will also require
(Model 3) Ionic interactions between PQS and Mg2
in the P. aeruginosa outer membrane enhances anionic peptidoglycan buildup in regions of the outer
repulsion between LPS molecules, resulting in mem- membrane containing few peptidoglycan-
brane blebbing. LPS, lipopolysaccharide; OM, outer outer membrane linkages.
membrane; PG, peptidoglycan; IM, inner membrane.
Figure reprinted from reference 36 with permission MODEL 3
from Blackwell Publishing.
Recent studies in our laboratory indicate that
the outer membrane is critical to understand- PQS is not only associated with MVs but is also
ing the three mechanistic models proposed for required for MV formation.These conclusions
MV formation (Fig. 5). are based on observations that P. aeruginosa
mutants unable to synthesize PQS produce sig-
MODEL 1 nificantly fewer MVs; however, addition of
MVs are formed when the outer membrane exogenous PQS at a physiologically relevant
expands faster than the underlying peptidogly- concentration to these mutants elicited MV
can layer (53).This asymmetric growth prevents formation (36).The mechanism by which PQS
lipoprotein tethering of the outer membrane to elicits MVs is not known; however, we recently
the peptidoglycan in these areas (18, 53). The proposed a model involving the interactions
between PQS and LPS (37). P. aeruginosa LPS, as dominant mode of growth of most bacteria in
with LPS from many gram-negative cells, is the natural environment, it is important to
highly anionic, and this anionic nature is neu- understand the roles of MVs in biofilms.
tralized by Mg2 and Ca2 salt bridges. On the Recent studies using thin sectioning and
basis of the structure of PQS and the observa- transmission electron microscopy revealed that
tion that this molecule binds trivalent iron (6), MVs were consistently present in the biofilm
we proposed that PQS binds Mg2 and Ca2 in EPS matrix.This was observed in vitro using a
the outer membrane, thus sequestering these variety of laboratory conditions as well as in
cations and preventing their interaction with environmental samples (45). Compared to
LPS.This disruption of the outer membrane salt planktonic cells, biofilm bacteria produced
bridges would result in localized outer mem- significantly more MVs, and biofilm MVs were
brane instability and blebbing. As with the smaller in size and possessed greater proteolytic
other proposed models, this is likely to occur in activity than planktonic MVs (45). These data
regions of the outer membrane that are not indicate that MVs are also produced by bacteria
tethered to the underlying peptidoglycan by during biofilm growth in the laboratory and in
lipoproteins. the natural environment, providing strong evi-
At the present time, there is little experi- dence that these trafficking vehicles are impor-
mental evidence for each of these models. It tant in natural settings.
should be pointed out that these three models
are not mutually exclusive, and a plausible CONCLUSIONS
model invoking aspects of each of these models MVs have been studied for over 50 years, when
could be proposed. The basic mechanism of the first transmission electron micrograph of
outer membrane blebbing dictates that MV Veillonella showed the presence of endotoxic
formation will not occur in regions of the outer vesicles (4). Since their discovery, MVs have
membrane containing strong outer membrane- been recognized as trafficking vehicles for a
peptidoglycan linkages; thus, one component number of cargo including toxins, protective
that is likely consistent for any MV formation factors, cell signals, DNA, and antimicrobials.
model is the lack of peptidoglycan-associated Packaging these contents within MVs likely
outer membrane proteins in the blebbing protects them from degradation until they
region. reach their destination. Future studies under-
standing the roles of MVs in natural bacterial
MEMBRANE VESICLES ARE MAJOR populations as well as defining, at the molecular
COMPONENTS IN BIOFILMS level, the mechanism of MV formation are keys
Historically, the study of bacteria has involved to understanding the importance of this novel
the evaluation of planktonic (free-swimming) trafficking mechanism as a mediator of bacterial
bacteria grown in batch culture in the labora- group behavior.
tory. However, in recent years, it has become
apparent that most bacteria in nature do not REFERENCES
exist in the planktonic state but are instead asso- 1. Bauman, S. J., and M. J. Kuehn. 2006. Purifica-
ciated with surfaces in bacterial biofilms. tion of outer membrane vesicles from Pseudomonas
aeruginosa and their activation of an IL-8 response.
Biofilms are sessile communities of microor-
Microb. Infect. 8:2400–2408.
ganisms enclosed in a self-produced polymeric 2. Beveridge, T. J. 1999. Structures of gram-
matrix (EPS) and attached to an inert or living negative cell walls and their derived membrane vesi-
surface (9).The EPS matrix encases the bacter- cles. J. Bacteriol. 181:4725–4733.
ial population and protects the bacteria from 3. Beveridge, T. J., S. A. Makin, J. L.
Kadurugamuwa, and Z. Li. 1997. Interactions
numerous stresses including antibiotics and
between biofilms and the environment. FEMS
host immune factors. Also within this matrix Microbiol. Rev. 20:291–303.
are myriad nutrients including proteins, DNA, 4. Bladen, H. A., and S. E. Mergenhagen. 1964.
lipids, and LPS (9). Since biofilms are the pre- Ultrastructure of Veillonella and morphological cor-
relation of an outer membrane with particles associ- 16. Grenier, D., and M. Belanger. 1991. Protective
ated with endotoxic activity. J. Bacteriol. 88: effect of Porphyromonas gingivalis outer membrane
1482–1492. vesicles against bactericidal activity of human
5. Braun, V. 1975. Covalent lipoprotein from the serum. Infect. Immun. 59:3004–3008.
outer membrane of Escherichia coli. Biochim. Bio- 17. Hayashi, J., N. Hamada, and H. K.
phys.Acta 415:335–377. Kuramitsu. 2002.The autolysin of Porphyromonas
6. Bredenbruch, F., R. Geffers, M. Nimtz, gingivalis is involved in outer membrane vesicle
J. Buer, and S. Haussler. 2006.The Pseudomonas release. FEMS Microbiol. Lett. 216:217–222.
aeruginosa quinolone signal (PQS) has an 18. Hoekstra, D., J. W. van der Laan, L. de
iron-chelating activity. Environ. Microbiol. Leij, and B. Witholt. 1976. Release of outer
8:1318–1329. membrane fragments from normally growing
7. Chatterjee, S. N., and J. Das. 1967. Electron Escherichia coli.Biochim.Biophys.Acta 455: 889–899.
microscopic observations on the excretion of cell- 19. Horstman, A. L., and M. J. Kuehn. 2000.
wall material by Vibrio cholerae. J. Gen. Microbiol. 49: Enterotoxigenic Escherichia coli secretes active heat-
1–11. labile enterotoxin via outer membrane vesicles.
8. Ciofu, O.,T. J. Beveridge, J. Kadurugamuwa, J. Biol. Chem. 275:12489–12496.
J. Walther-Rasmussen, and N. Hoiby. 2000. 20. Inouye, M. 1975. Biosynthesis and assembly of
Chromosomal beta-lactamase is packaged into the outer membrane proteins of Escherichia coli, p.
membrane vesicles and secreted from Pseudomonas 351–391.In A.Tzagoloff (ed.),Membrane Biogenesis.
aeruginosa. J.Antimicrob. Chemother. 45:9–13. Plenum Publishing Corporation, New York, NY.
9. Davey, M. E., and A. O’Toole G. 2000. Micro- 21. Jander, G., L. G. Rahme, and F. M. Ausubel.
bial biofilms: from ecology to molecular genetics. 2000. Positive correlation between virulence of
Microbiol. Mol. Biol. Rev. 64:847–867. Pseudomonas aeruginosa mutants in mice and
10. Deich, R. A., and L. C. Hoyer. 1982. Genera- insects. J. Bacteriol. 182:3843–3845.
tion and release of DNA-binding vesicles by 22. Kadurugamuwa, J. L., and T. J. Beveridge.
Haemophilus influenzae during induction and loss 1996. Bacteriolytic effect of membrane vesicles
of competence. J. Bacteriol. 152:855–864. from Pseudomonas aeruginosa on other bacteria
11. Deziel, E., F. Lepine, S. Milot, J. He, M. N. including pathogens:conceptually new antibiotics.
Mindrinos, R. G. Tompkins, and L. G. J. Bacteriol. 178:2767–2774.
Rahme. 2004. Analysis of Pseudomonas aeruginosa 23. Kadurugamuwa, J. L., and T. J. Beveridge.
4-hydroxy-2-alkylquinolines (HAQs) reveals a 1995. Virulence factors are released from
role for 4-hydroxy-2-heptylquinoline in cell-to- Pseudomonas aeruginosa in association with mem-
cell communication. Proc. Natl. Acad. Sci. USA brane vesicles during normal growth and exposure
101:1339–1344. to gentamicin: a novel mechanism of enzyme
12. Dorward, D.W., C. F. Garon, and R. C. Judd. secretion. J. Bacteriol. 177:3998–4008.
1989. Export and intercellular transfer of DNA via 24. Kahn, M. E., G. Maul, and S. H. Goodgal.
membrane blebs of Neisseria gonorrhoeae. J. Bacteriol. 1982.Possible mechanism for donor DNA binding
171:2499–2505. and transport in Haemophilus. Proc. Natl. Acad. Sci.
13. Fiocca, R., V. Necchi, P. Sommi, V. Ricci, J. USA 79:6370–6374.
Telford, T. L. Cover, and E. Solcia. 1999. 25. Kato, S.,Y. Kowashi, and D. R. Demuth. 2002.
Release of Helicobacter pylori vacuolating cytotoxin Outer membrane-like vesicles secreted by Acti-
by both a specific secretion pathway and budding nobacillus actinomycetemcomitans are enriched in
of outer membrane vesicles. Uptake of released leukotoxin. Microb. Pathog. 32:1–13.
toxin and vesicles by gastric epithelium. J. Pathol. 26. Kesty, N. C., K. M. Mason, M. Reedy,
188:220–226. S. E. Miller, and M. J. Kuehn. 2004. Entero-
14. Gankema, H., J.Wensink, P. A. Guinee,W. H. toxigenic Escherichia coli vesicles target toxin
Jansen, and B.Witholt. 1980. Some characteris- delivery into mammalian cells. EMBO J.
tics of the outer membrane material released by 23:4538–4549.
growing enterotoxigenic Escherichia coli. Infect. 27. Kolling, G. L., and K. R. Matthews. 1999.
Immun. 29:704–713. Export of virulence genes and Shiga toxin by
15. Giwercman, B., P. A. Lambert,V. T. Rosdahl, membrane vesicles of Escherichia coli O157:H7.
G. H. Shand, and N. Hoiby. 1990. Rapid Appl. Environ. Microbiol. 65:1843–1848.
emergence of resistance in Pseudomonas aerugi- 28. Kuehn, M. J., and N. C. Kesty. 2005. Bacterial
nosa in cystic fibrosis patients due to in-vivo outer membrane vesicles and the host-pathogen
selection of stable partially derepressed beta- interaction. Genes Dev. 19:2645–2655.
lactamase producing strains.J.Antimicrob.Chemother. 29. Lally, E. T., E. E. Golub, I. R. Kieba, N. S.
26:247–259. Taichman, J. Rosenbloom, J. C. Rosen-
bloom, C. W. Gibson, and D. R. Demuth. involved in transport of Pseudomonas aeruginosa

1989. Analysis of the Actinobacillus actinomycetem- cell-to-cell signals. J. Bacteriol. 181:1203–1210.
comitans leukotoxin gene. Delineation of unique 42. Pesci, E. C., J. B. Milbank, J. P. Pearson, S.
features and comparison to homologous toxins. J. McKnight, A. S. Kende, E. P. Greenberg, and
Biol. Chem. 264:15451–15456. B. H. Iglewski. 1999. Quinolone signaling in the
30. Lepine, F., S. Milot, E. Deziel, J. He, and cell-to-cell communication system of Pseudomonas
L. G. Rahme. 2004. Electrospray/mass aeruginosa. Proc. Natl. Acad. Sci. USA
spectrometric identification and analysis of 4- 96:11229–11234.
hydroxy-2-alkylquinolines (HAQs) produced by 43. Pettit, R. K., and R. C. Judd. 1992.The interac-
Pseudomonas aeruginosa. J. Am. Soc. Mass. Spectrom. tion of naturally elaborated blebs from serum-
15:862–869. susceptible and serum-resistant strains of Neisseria
31. Li, Z., A. J. Clarke, and T. J. Beveridge. 1998. gonorrhoeae with normal human serum. Mol.
Gram-negative bacteria produce membrane vesi- Microbiol. 6:729–734.
cles which are capable of killing other bacteria. J. 44. Renelli, M., V. Matias, R. Y. Lo, and T. J.
Bacteriol. 180:5478–5483. Beveridge. 2004. DNA-containing membrane
32. Li, Z., A. J. Clarke, and T. J. Beveridge. 1996. vesicles of Pseudomonas aeruginosa PAO1 and their
A major autolysin of Pseudomonas aeruginosa: sub- genetic transformation potential. Microbiology
cellular distribution, potential role in cell growth 150:2161–2169.
and division and secretion in surface membrane 45. Schooling, S. R., and T. J. Beveridge. 2006.
vesicles. J. Bacteriol. 178:2479–2488. Membrane vesicles: an overlooked component
33. Livermore, D. M. 1987. Clinical significance of of the matrices of biofilms. J. Bacteriol. 188:
beta-lactamase induction and stable derepression 5945–5957.
in gram-negative rods. Eur. J. Clin. Microbiol. 46. Schuster, M., C. P. Lostroh, T. Ogi, and E. P.
6:439–445. Greenberg. 2003. Identification, timing, and sig-
34. Machan, Z. A., G. W. Taylor, T. L. Pitt, P. J. nal specificity of Pseudomonas aeruginosa quorum-
Cole, and R.Wilson. 1992. 2-Heptyl-4-hydrox- controlled genes: a transcriptome analysis. J.
yquinoline N-oxide, an antistaphylococcal agent Bacteriol. 185:2066–2079.
produced by Pseudomonas aeruginosa. J. Antimicrob. 47. Schuster, M., C. P. Lostroh, T. Ogi, and E. P.
Chemother. 30:615–623. Greenberg. 2003. Identification, timing, and sig-
35. Mashburn, L. M., A. M. Jett, D. R. Akins, and nal specificity of Pseudomonas aeruginosa quorum-
M. Whiteley. 2005. Staphylococcus aureus serves as controlled genes: a transcriptome analysis. J.
an iron source for Pseudomonas aeruginosa during in Bacteriol. 185:2066–2079.
vivo coculture. J. Bacteriol. 187:554–566. 48. Smith, R. S., and B. H. Iglewski. 2003. P. aerug-
36. Mashburn, L. M., and M. Whiteley. 2005. inosa quorum-sensing systems and virulence. Curr.
Membrane vesicles traffic signals and facilitate Opin. Microbiol. 6:56–60.
group activities in a prokaryote. Nature 437: 49. Wade, D. S., M.W. Calfee, E. R. Rocha, E. A.
422–425. Ling, E. Engstrom, J. P. Coleman, and E. C.
37. Mashburn-Warren, L. M., and M. Whiteley. Pesci. 2005. Regulation of Pseudomonas quin-
2006. Special delivery: vesicle trafficking in olone signal synthesis in Pseudomonas aeruginosa. J.
prokaryotes. Mol. Microbiol. 61:839–846. Bacteriol. 187:4372–4380.
38. Palmer, K. L., L. M. Mashburn, P. K. Singh, 50. Wagner,V. E., D. Bushnell, L. Passador, A. I.
and M. Whiteley. 2005. Cystic fibrosis sputum Brooks, and B. H. Iglewski. 2003. Microarray
supports growth and cues key aspects of analysis of Pseudomonas aeruginosa quorum-sensing
Pseudomonas aeruginosa physiology. J. Bacteriol. 187: regulons:effects of growth phase and environment.
5267–5277. J. Bacteriol. 185:2080–2095.
39. Parsek, M. R., and E. P. Greenberg. 2000. 51. Wagner,V. E., L. L. Li,V. M. Isabella, and B. H.
Acyl-homoserine lactone quorum sensing in Iglewski. 2007.Analysis of the hierarchy of quo-
gram-negative bacteria: a signaling mechanism rum-sensing regulation in Pseudomonas aeruginosa.
involved in associations with higher organisms. Anal. Bioanal. Chem. 387:469–479.
Proc. Natl.Acad. Sci. USA 97:8789–8793. 52. Wai, S. N., B. Lindmark, T. Soderblom, A.
40. Pearson, J. P., M. Feldman, B. H. Iglewski, Takade, M. Westermark, J. Oscarsson, J. Jass,
and A. Prince. 2000. Pseudomonas aeruginosa cell- A. Richter-Dahlfors, Y. Mizunoe, and B. E.
to-cell signaling is required for virulence in a Uhlin. 2003.Vesicle-mediated export and assem-
model of acute pulmonary infection. Infect. Immun. bly of pore-forming oligomers of the enterobacte-
68:4331–4334. rial ClyA cytotoxin. Cell 115:25–35.
41. Pearson, J. P., C. Van Delden, and B. H. 53. Wensink, J., and B. Witholt. 1981. Outer-
Iglewski. 1999. Active efflux and diffusion are membrane vesicles released by normally growing
Escherichia coli contain very little lipoprotein.Eur. J. 55. Yaron, S., G. L. Kolling, L. Simon, and K. R.
Biochem. 116:331–335. Matthews. 2000. Vesicle-mediated transfer of
54. Xiao, G., E. Deziel, J. He, F. Lepine, B. virulence genes from Escherichia coli O157:H7 to
Lesic, M. H. Castonguay, S. Milot, A. other enteric bacteria. Appl. Environ. Microbiol.
P. Tampakaki, S. E. Stachel, and L. G. 66:4414–4420.
Rahme. 2006. MvfR, a key Pseudomonas aerug- 56. Zhou, L., R. Srisatjaluk, D. E. Justus, and R. J.
inosa pathogenicity LTTR-class regulatory Doyle. 1998. On the origin of membrane vesicles
protein, has dual ligands. Mol. Microbiol. in gram-negative bacteria. FEMS Microbiol. Lett.
62:1689–1699. 163:223–228.
COOPERATIVE REGULATION OF
COMPETENCE DEVELOPMENT IN
STREPTOCOCCUS PNEUMONIAE:
CELL-TO-CELL SIGNALING VIA A
PEPTIDE PHEROMONE AND AN
ALTERNATIVE SIGMA FACTOR
Marco R. Oggioni and Donald A. Morrison
22
It is generally known that even a setting perfectly lions of separate cells depends not only on their
suited for the growth of Pneumococci is not neces- shared chemical environment but also on com-
sarily conducive to their transformation. So it is that
one must incorporate a small quantity of albumin
munication via a secreted pheromone peptide
into the culture. On the other hand, a favorable cul- signal. Recognized as a protease-sensitive moi-
ture of Pneumococcus does not always contain ety in several streptococcal species capable of
“competent” bacteria capable of producing, when genetic transformation (71, 100), it appeared
acted on by foreign DNA, transformed colonies. to elicit two responses: elaboration of more sig-
Competence appears only in the course of the phase
of exponential replication and disappears before the
nal and differentiation into the competent
end of growth.The significance of this property is state.The first property was early recognized as
far from being elucidated. key to the explosive spread of competence
– René Thomas, 1955 (97) throughout a culture; in a positive feedback
In laboratory cultures of Streptococcus pneumo- loop it could create a rapid amplification of a
niae (pneumococcus), the capacity for genetic small initiating dose of the signal (98).The sec-
transformation, one of the most widely known ond was understood to provide competent
traits of this common commensal human cells with apparatus for DNA transport and a
pathogen, is often, perhaps typically, completely highly efficient gene replacement mechanism
absent. However, this capacity can suddenly be of genetic recombination.
displayed by virtually all cells of a logarithmi- The previous volume on bacterial cell-cell
cally growing culture and then,nearly as rapidly, signaling (24) identified this pheromone as a
disappear. The coordination that brings about small peptide produced by pneumococcus and
this nearly simultaneous behavior among mil- capable of inducing development of the com-
petent state by acting through a receptor and
Marco R.Oggioni Laboratorio di Microbiologia Molecolare e response regulator to stimulate the production
Biotecnologia, Dipartimento di Biologia Molecolare, of numerous proteins required for DNA trans-
Universita’ di Siena 53100 Siena, Italy. Donald A. Morrison port and recombination. The signal transduc-
Laboratory for Molecular Biology, Department of Biological
Sciences, University of Illinois at Chicago, Chicago, Illinois tion pathway has now been further defined; a
60607. more comprehensive list of genes up-regulated
345
346 ■ OGGIONI AND MORRISON
in response to the competence-stimulating swapping experiments and sequence compar-

peptide (CSP) has been established, and a isons among comD alleles; comparison of comD
broader picture of the biology of competent allele sequences shows that within ComD, the
cells has begun to emerge. In this chapter, we CSP specificity is conferred by a few residues
outline this regulatory network, as well as a within its amino-terminal transmembrane
number of recent discoveries indicating that the domains (37, 42). In the case of S. pneumoniae,
coordination of gene expression mediated just two alleles of comC are known (79, 83).
among neighboring pneumococcal cells by Consistent with a central role of ComD in the
CSP makes important contributions to the CSP response,six mutations of ComD (Y312C,
interaction of this pathogen with its human T290A, E214K, Q228L, D299N, and T233I)
host, and thus has significance beyond the have been identified that increase expression
acquisition of new genetic information that from the comC promoter and stimulate compe-
originally drew attention to the system. tence (46, 55).
Outside of competence regulation, the most
closely related homologues of ComD and
THE PHEROMONE SIGNAL
TRANSDUCTION PATHWAY: ComE are associated with genes for production
SUCCESSIVE TRANSCRIPTIONAL of peptide bacteriocins, where they participate
REGULATORS in the cooperative production of these toxins,
The competence pheromone CSP is synthe- mediated by peptide cell-cell signals (86, 92). In
sized ribosomally as a propeptide that depends several of these cases, the ComE paralogue has
on a specific ATP-binding cassette transporter been found to bind at direct repeat sequences
for export and for removal of an N-terminal upstream of promoters of peptide-responsive
leader (36).The first specific hint of the pathway genes, where it acts to stimulate transcription.
that connects the binding of CSP to changes in In pneumococcus, recombinant ComE protein
gene expression (Fig. 1) came from the discov- binds to similar direct repeats (CAxTT–
ery that the structural gene (comC) for the 16–CAxTT) at the three CSP-responsive oper-
pheromone propeptide was closely linked to ons containing comA, comC, and SP1717 (106)
two genes, now named comD and comE, that and is thought to stimulate transcription at
encode members of the histidine protein kinase these sites and at similar direct repeats found at
receptor and response regulator families,respec- the promoter sequences of several other CSP-
tively (73). Complete loss of the pheromone responsive genes (Fig. 2).
response in comD and comE mutants shows that Among the products of genes induced on
these genes are critical for CSP sensing and that stimulation by CSP and linked to CAxTT
there is no alternative to this pathway for CSP repeats is an alternative sigma factor, ComX or
signal transduction in pneumococcus.Naturally sigma X, apparently the only alternative sigma
transformable strains are common among the in this species (47, 52). Interestingly, this sigma
mitis and anginosus groups of streptococcal factor is encoded by two identical genes, which
species, and they carry the comCDE compe- are located upstream of two separate 16s rRNA
tence regulatory locus (38).Each such species in operons. Because of the unusual duplication, a
the mitis group exhibits from 2 to 10 or more role for this sigma in competence regulation
allelic forms of the mature comC product, while was discovered not by genetic approaches, but
three species of the anginosus group share a sin- only by reverse genetics: the protein was found
gle pheromone allele among them.In each case, as a minor component of RNA polymerase
a single strain typically responds only to the sin- uniquely present in competent cultures. How-
gle allelic form of ComC that it elaborates (38, ever, once the duplicate genes were discovered,
79).This receptor specificity was mapped to the it was shown, using double mutants, that the
comD gene, and specifically to the transmem- protein is absolutely required for development
brane domains of ComD by means of domain of competence and that either copy is sufficient
22. PEPTIDE PHEROMONE AND SIGMA FACTOR REGULATE COMPETENCE ■ 347
FIGURE 1 Network of regulation leading to the X state. Regulatory interactions among the
early genes, shown at left, coordinate the response to pheromone, stress, and other unknown sig-
nals. ComABCDE establish an autocatalytic positive feedback circuit, while ComX and ComW
establish a link to transcription of downstream late genes. Principal functions provided by respre-
sentative late gene operons are illustrated at the right.
for a full response to CSP. Homologues of genes whose induction is required for compe-
ComX are widely distributed among strepto- tence are under the control of ComX (52).
cocci and close relatives.Their biological signif- Nevertheless, another sigma X-independent
icance outside control of competence has not but ComE-activated gene, comW, does con-
been determined, although artificially activated tribute to competence development (53) (Fig.
comX in Streptococcus pyogenes does provoke 1). Deletion of comW affects competence dra-
transcription of many homologues of the pneu- matically, reducing transformation to less than
mococcal competence genes (107). 0.01% of wild type. In the comW mutant, two
Sigma X is largely reduced in or absent from effects of the loss can be distinguished (93).
noncompetent cells; its amount increases at First, the level of ComX protein achieved in
least 100-fold in response to CSP, and it assem- response to CSP is reduced approximately ten-
bles with core polymerase to recognize a set of fold. Second, even when the level of ComX is
noncanonical promoter sites marked by the increased above normal by inactivation of the
sequence TACGAATA, known as the combox ClpP protease, comW is needed for normal lev-
or cinbox consensus (54). The full promoter els of transformation (93). Thus, the level of
structure required for recognition by sigma X is ComX may be subject to one or more post-
not yet fully defined, since there are at least 12 transcriptional mechanisms of regulation, as is
copies of this consensus sequence in the typical of alternative sigma factors. The rapid
genome that are not active as promoters (10, disappearance of this protein soon after compe-
77). Artificial expression of ComX in a strain tence development indicates that it is an unsta-
lacking ComE produces competent cells, ble protein. Indeed, a specific ATP-dependent
showing that no other gene is essential for this protease, ClpEP, has been implicated in this
transcriptional switch, and establishing that all instability (53). It has been noted that in a clpP
up-regulated in competent cells,consistent with

earlier data showing that development of com-
petence in response to pheromone is blocked by
inhibitors of RNA or protein synthesis (99,
101). Since the earliest evidence of regulation at
the transcriptional level (3, 72, 73), multiple
studies built a picture of the complete regulated
set of genes by using promoter probe plasmids,
plasmid clone hybridization libraries, and DNA
microarrays to monitor gene expression on a
genomic scale (5, 20, 74, 77, 78, 85).Together,
these studies establish that in cultures suddenly
exposed to a high level of CSP, more than 123
genes are transiently induced. Two subsets
among them can be related to the two regula-
tory proteins ComE and ComX (Color Plate
12).Induction of the genes of a small set is inde-
pendent of the activity of ComX but depends
on ComE. A larger set of genes also depends
on ComX for induction in response to CSP.
As the former messages reach a maximal level
a few minutes before that of the latter, these
two sets have been named early and late, respec-
tively. cis regulatory sites at which ComE and
ComX act have been identified are marked
by short consensus sequences (Fig. 2). A few
genes in each class have no apparent correspon-
ding promoter consensus sites; they may reflect
distinct binding specificities, or they may
FIGURE 2 cis-acting regulatory sites at early and late depend on different intermediate regulatory
CSP-induced loci. (A) Bases matching the direct repeat proteins. (If the latter is correct, however, it is
ComE consensus upstream of early genes are bold. perplexing that their temporal expression pat-
Apparent canonical 10 promoter sites are underlined.
(B) DNA sequences located upstream of late gene clus-
terns are exactly the same as those of other early
ters. Base pairs to the start codon of the first open read- or late genes.) CSP-responsive genes in a third
ing frame are shown. Matches to combox consensus group are induced with a further lag of several
octamer are in bold; mismatches, in lowercase. minutes and have been designated delayed.
The link of the delayed genes to CSP has not
been identified, but since several of them
mutant, ComX is stabilized and persists after
encode stress chaperones, it may be related
transient competence induction, but compe-
to the rapid global switch in protein synthesis as
tence still shuts off quickly, possibly indicating
cells become competent, or to the fundamental
yet another level of control (51).
role of competence as a stress response (see
below).
THE COMPETENCE PHEROMONE
REGULONS INCLUDE 39 GENE Although about 24 early and about 81 late
CLUSTERS, MANY RELATED TO genes are formally represented by elevated
GENETIC EXCHANGE mRNA levels after CSP treatment of naive
Protein pulse-labeling studies revealed directly cells, several features of gene organization
(61) that synthesis of many proteins is strongly suggest that some of these represent cases of
transcriptional read-through and may not be two genes (comM and cibC) that are also induced
biologically significant. Particularly suspect are at competence and that provide competent
genes read in the antisense direction. Genes for cells with immunity from lytic attack by other
which transcription increases to a lesser extent competent cells. Some indications that the
and that are found in operons with putative important natural sources of DNA are both
canonical promoters separating them from close relatives and closely related species may be
more highly induced upstream genes may rep- gleaned from observing which kinds of donor
resent additional cases of read-through into DNA molecules are most successful in transfor-
operons that are independently regulated. mation.Long homologous donor molecules are
Color Plate 12 illustrates the organization of optimal,but there is little discrimination against
CSP-induced genes and the location of such frequent mismatches or against large inserts or
apparent cases of read-through. deletions (41, 81). It may also be significant that
Many of the CSP-induced genes contribute long donor insertions, which would be
to various aspects of transformation.About half expected to become susceptible to restriction
of the genes highlighted in Color Plate 12 have by the DpnII restriction nuclease after the
known or suspected roles in DNA exchange (8, first round of replication, are protected from
44). First, a dozen genes in four operons are potential restriction by competence-specific
required for assembly of the DNA transport expression of a cognate ssDNA-specific DNA
machine, which operates in conjunction with a methyltransferase (45).
constitutive membrane-associated nuclease, The class of early genes is dominated by
EndA, to effect a “divide-and-conquer” strat- genes that act in regulation of competence
egy for uptake of portions of donor DNA development. These include those for
molecules in a threading process that starts at pheromone production (ComC, ComAB) and
new 3′-OH ends made by the competent cell pheromone-sensing elements (ComDE),as well
(7, 59). Second, five induced genes in five oper- as two that act to transmit the signal to late
ons are required for efficient replacement of genes (comW and comX). An unidentified late
recipient by donor DNA strands. Deletion of gene also appears to have a regulatory role, par-
coiA, dalA (dprA), radA, recA, or ssbB reduces ticipating in shutoff of early gene expression,
recombinant yields by 70 to 99.9% (7, 8, 9a, 22, as comX mutants fail to shut off early gene
61a). The high frequency of genetic recombi- expression (47).
nation between donor and recipient DNA Nine operons (two early, seven late) appear
in competent cells is specific for single stranded to be regulated as strictly as those already dis-
DNA (ssDNA) (or for recently imported cussed with known functions in competence
and processed ssDNA), as chromosomal and are apparently controlled through the same
DNA introduced by electroporation into com- classes of regulatory cis-acting sites, yet have no
petent cells is not incorporated to a significant known part in DNA uptake or in regulation.
degree (48). We suggest that they participate in new aspects
This efficient gene uptake system is coordi- of the biology of pneumococcus not yet
nated with a highly orchestrated mechanism by understood as cooperative.While some of these
which pneumococcal cells sample their neigh- may be still-undiscovered aspects of DNA
bors’ genomes (32, 39, 43, 90, 91). Recent exchange, others may be important keys to the
reports have revealed four genes induced at interaction of pneumococcus with its host, as
competence, but not needed for DNA incor- suggested by the discussion below.
poration, that participate in lysis of noncompe-
tent cells.Are these lytic mechanisms important PHEROMONE
in pneumococcal life history or could they be REGULATORY REGIMES
artifacts of laboratory culture? A strong sugges- The circumstances in which this regulatory sys-
tion of the former case is given by the finding of tem acts in nature are not known. However, a
succession of responses to varying levels of CSP CiaH and the response regulator CiaR affects
can be distinguished in vitro. Examination of competence regulation by two or more routes
these can begin to reveal parts of the regulatory (27, 30, 57, 108).The system was named cia for
repertoire that may be deployed in more natural competence induction and antibiotic resist-
contexts. Even in the simplified arena of the ance, when a point mutation in ciaH was found
laboratory culture, one can already distinguish to reduce endogenous competence induction
at least eight distinct regulatory regimes, and increase resistance to cefotaxime. Specifi-
including those exhibited during naïve growth, cally,the ciaH* mutant,which is thought to have
response to low levels of pheromone, response increased levels of phosphorylation and activity
to high levels of pheromone (three phases), of CiaR, is deficient in competence but could
shutoff of competence, and steady-state growth still respond to high doses of CSP. Conversely,
in high CSP,as well as a loss of competence dur- null mutants of ciaH or ciaR led to increased
ing stationary phase of growth. Not one of expression of many genes of the competence
these regimes is fully characterized, either as to regulons (109). This suppressive influence of
regulated expression levels or as to regulatory CiaR on competence depends on HtrA, the
interactions maintaining the regime; however, product of a gene regulated by ciaR, as it is
in several regimes, some important interactions relieved by an active site mutation in htrA.
have already been identified. However, the stimulation of competence in
In growing noncompetent pneumococcal ciaR mutants does not appear to occur through
cultures, the com genes are expressed at low and a reduction of HtrA activity, as htrA mutants are
possibly variable levels, but neither their not similarly stimulated to competence induc-
absolute expression levels nor their regulatory tion (89).
inputs are known.There are, however, numer- The response to low levels of accumulated
ous indications that the “basal” rate of CSP syn- CSP is thought to be restricted to accelerated
thesis—absent cell-cell amplification—may production of CSP (98). In the low-CSP
itself be regulated. For example, increased response regime, CSP production is accelerated
expression of ComC via transcriptional read- following activation of ComE by ComD and
through (55), mutation of the ATP-dependent consequent transcriptional activation of both
protease ClpP (11, 87), or loss of oligopeptide the pheromone structural gene, comC, and
transporter activity (2) can each affect the initi- comAB, encoding the CSP transporter and mat-
ation of development of competence. Espe- uration peptidase. The manner in which
cially interesting is the recent discovery that responses to low and high CSP levels are distin-
sublethal levels of several antibiotics (e.g., lev- guished is not known. However, it would
ofloxacin, norfloxacin, kanamycin, mitomycin appear a priori that maximal cooperation may
C, and streptomycin), but not others (e.g., be ensured by such a distinction,so that “effort”
ampicillin, erythromycin, tetracycline, novo- is first devoted to establishing the highest level
biocin, rifampin, or vancomycin), also stimulate of signal possible, then turned toward the coop-
competence development in exponentially erative endeavor itself with an increased likeli-
growing cultures (80). Additional genes in hood that even the least responsive cells
which mutations alter the expression of the participate.The set of genes activated in such a
comC operon are as varied as luxS, stkP, vicK, restricted initial response to low amounts of
purA,guaA,guaB,cls,and pbp1b (14,25,88,103). CSP has not been defined experimentally.It has
In each of these cases,the stimulation may result been suggested that comX is not expressed in
from changes in basal expression, but in no case this regime,due to low affinity of ComE for the
has the effect yet been traced to a specific point variant direct repeat at the comX promoter (14),
of the signal transmission and production chain. but neither the relative expression level of
The two-component signal transduction ComX nor the relative affinity of ComE for its
system comprising the protein histidine kinase promoter has been determined directly.
The temporal pattern of gene expression the shutoff (31). Cells then enter a refractory
during development of competence coordi- state in which CSP can no longer elicit compe-
nated by endogenous pheromone production tence.This refractory period may represent one
has not been determined. However, it is known or more distinct expression states.
that sudden exposure of naïve cells to high lev- During succeeding generations of further
els of CSP leads in rapid succession to a series of exponential growth, competence for DNA
three transient expression regimes during a uptake remains low. During this postcompe-
period of less than 40 min (20, 78). In the first, tence refractory period, the culture medium
transcripts of all early genes rise sharply, leading remains rich in CSP, as freshly added naïve cells
to rapid accumulation of thousands of mole- rapidly become competent in it (18).Yet most
cules of sigma X. The early gene product of the postcompetent cells themselves remain
ComW is important for the accumulation of incapable of DNA uptake, whether in the
ComX, acting posttranscriptionally, apparently same medium or after transfer to fresh CSP-
as an antiprotease, but it stimulates ComX supplemented medium.After the initial refrac-
activity as well. tory period, cells do become somewhat
In the second, transcription of late genes responsive to CSP again, with a steady state of
driven by sigma X leads to rapid synthesis of low competence being established. Direct
many proteins that are absent from noncompe- measurement shows that during the postcom-
tent cells and to increases over the constitutive petence period of exponential growth, only 4%
levels of others.This regime culminates in what of cells are competent at any one moment (60).
are commonly designated as competent cells. The state of the remaining 96% is unknown but
Since much of their phenotype is determined must differ from that of naïve cells, as they do
by the ComX-dependent late genes, it is useful not recapitulate the global coordinated compe-
to term it the X-state, as proposed (15), to tence induction mounted by naïve cells on
emphasize that DNA uptake is just one of a contact with high levels of CSP.In this new reg-
repertoire of activities associated with such cells. ulatory equilibrium in excess CSP, a select sub-
In what may be loosely considered a third set of CSP-dependent gene products may be
phase of the high-dose response, transcription important, even while competence is low and
of early and late genes ceases and their mRNA restricted to a small subpopulation.The proper-
decays rapidly; unstable competence-specific ties of such “postcompetent” cultures are
proteins such as ComX and CoiA decay (22, poorly defined or understood. However, cells
54); and expression of the delayed genes cultured continuously in CSP may differ from
increases two- to threefold.Neither the mecha- naïve cells in more respects than their being
nism of induction of delayed genes nor the partially refractory to CSP.There is, for exam-
mechanism of transcription shutdown in this ple, a long period of continuing DNA release
third phase is known. However, either ComX after early competence induction in dilute cul-
itself or a ComX-dependent gene product tures and a special sensitivity to loss of CiaR or
probably participates in the rapid reversal of the DprA after competence (6, 20, 62), which sug-
high-dose CSP response, since expression of gests that the metabolism of such cells is differ-
early genes continues at a high rate in comX ent from that of cells never exposed to CSP.As
mutants (47). The shutdown is postponed for only DNA uptake and cell lysis have been sys-
many minutes in a ComER120S mutant, sug- tematically determined in postcompetent cul-
gesting one part of the shutoff mechanism may tures, better understanding of this phase of
act directly on ComE (31, 55). Furthermore, growth will require determining whether it
the observation that overexpression of ComD entails a new pattern of expression of the
and ComE can suppress response to CSP sug- competence-related genes. It will also be valu-
gests that the strong induction of these genes at able to learn whether such postcompetent cells
competence may itself directly contribute to “remember” their ancestors’ competence or
continuously respond to CSP, that is, whether genetic diversity, as well as other functions for
CSP continues to play a role in communication reducing or repairing damage (15).
after the acute but transient initial response to The idea that quorum sensing is one func-
elevated CSP. tion of the competence pheromone is broadly
If there are two alternative steady-state consistent with the development of compe-
modes of exponential growth, one with CSP tence in dilute cultures in the early exponential
and another without CSP, then the phase com- phase of growth, conditions rarely associated
monly described as transient competence may with stress, and with an experimental pattern in
properly be viewed as a transition process lead- which cultures exhibit a delay in competence
ing from one mode to the other. The latter inversely related to inoculum size. Early discus-
mode, not well recognized yet, may correspond sions of pneumococcal competence for genetic
to the more general view of a state controlled transformation emphasized that competence
by quorum sensing, i.e., one that is off at low was regulated in a manner that was sensitive to
cell densities but continuously on at high cell cell density, as well as to other environmental
densities, and it may be more similar to the parameters (e.g., see references 26 and 97).The
behavior of cells in natural biofilms than the idea of a dependence on population density was
transient state of competence. reinforced by discovery of a protease-sensitive
In stationary phase, cells become unrespon- signal elaborated by competent cells (71, 100).
sive to CSP, even in cultures grown in compe- In two rare cases, competence appeared at an
tence-nonpermissive conditions, where cells approximately constant density independent of
remain receptive to CSP throughout the expo- inoculum dilution (3, 98). However, the more
nential growth phase (35). The mechanism of generally observed pattern is a delay much
this loss of pheromone receptivity is not shorter than required to establish a constant cell
known, but since it also occurs in comA density. René Thomas, for example, reported a
mutants, it is clearly not another postcompe- delay of only 3 generations for a 100-fold dilu-
tence phenomenon but reflects a distinct regu- tion and only 5 generations for a 1,000-fold
latory input. Moreover, it is not even known dilution (97), and Alex Tomasz (100) reported
whether the failure of CSP to enable transfor- that a 100-fold dilution delayed competence by
mation at stationary phase reflects some specific somewhat less than one full generation.Thomas
defect in DNA processing or a more general explicitly pointed out that the effect did not
failure of pheromone signal transduction result in competence appearing at a constant
cell density; indeed, the more dilute cultures
IS A MICROBIAL CELL-CELL SIGNAL became competent at lower densities. Thus,
IPSO FACTO FOR QUORUM SENSING? competence was never strongly linked to a spe-
Despite the wealth of data on the mechanisms cific fixed “critical” cell density in the early
of competence regulation, the conditions lead- studies, and although the regulation of compe-
ing to the development of competence remain tence in pneumococcus has been widely
obscure.One commonly held view is that com- described as a case of quorum sensing (3, 13, 14,
petence regulation is a clear case of quorum 16, 37, 60, 73), the mechanism underlying the
sensing, such that actively growing cultures inverse dependence on cell density fails to bring
develop competence when their members about competence development at a constant
accumulate to a specific population density or cell density even in parallel cultures growing in
produce a specific concentration of pheromone a single medium. Rather than quorum sensing,
(40,84,95,102,105).A recently proposed alter- then, the inverse dependence on inoculum size
native view is that development of competence should instead be understood to reflect the sim-
is triggered as a general stress response that ple fact that at lower cell densities cell-cell sig-
includes recombinational mechanisms for naling is less efficient, and its consequences
repair of damaged DNA and for enhancing appear less rapidly. In that case, other aspects of
the culture history and manipulation must pro- heat-inactivated smooth cells of serotype 1 and
vide the fundamental triggering impulse. living rough serotype 2 cells (29).These experi-
It was recently proposed that competence is a ments used the mouse not only for selection of
general stress response (15, 80).This idea has the revertants but, most importantly, to enable
advantage that it may rationalize under one development of competence and selection of
umbrella the wide array of signals that can pro- transformants. Neufeld and Levinthal and
voke competence, as discussed above.Thus, sig- Dawson, also in 1928, published similar obser-
nals as varied as nutrient imbalance,cell wall and vations using in vivo transformation experi-
membrane integrity, nucleotide pools, DNA ments as well (21, 64). It is thus 80 years since
damage, and translational stalling could all point the first report on pneumococcal cell-cell
to a present or impending stress.While it is plau- signaling within a living animal. After about
sible to view DNA uptake and targeted gene 25 years of work exclusively focused on in
replacement (especially when the donor is an vitro description of transformation, Austrian
identical sibling) as provisions for repair of dam- described transformation of pneumococci in a
aged DNA, the nature of the “stress” experi- variety of mammals,including primates (4),and
enced by exponentially growing cells in cultures showed that more than one bacterial cell in the
at less than 1% of saturation densities, for exam- infecting population is transformed in vivo
ple, remains a puzzle. Even this may be accom- by following two different markers. Hall
modated by the proposal that an alkaline stress is and Gale in 1959 (33) and Ottolenghi-
created under the common protocols for com- Nightingale and MacLeod in 1963 (68)
petence induction (15), as alkaline stress has not demonstrated that DNA-mediated transforma-
been well defined for pneumococcus and it is tion can occur spontaneously in genetically
possible that for this species a shift from a late mixed populations of pneumococci growing in
exponential culture at pH 6.8 to a pH of 7.2 a living host after subcutaneous infection.A few
imposes a significant stress (12).Identification of years later,Conant and Sawyer and Ottolenghi-
the peptide pheromone, the corresponding sig- Nightingale in two papers (17, 69) described
nal transduction pathway, and several down- the phenomenon of in vivo transformation in
stream regulons has revealed many elements more detail with the use of drug resistance
of the competence regulatory mechanism. markers. They demonstrated (i) that transfor-
Broader understanding of the functions pro- mation occurs within the host, both by inocu-
vided by this signal coordinated system will lating donor and recipient pneumococci 6 h
arise from identification of the points at which apart and by inoculating donor and recipient
this pathway is affected by additional internal or pneumococci in different sites (intraperitoneal
external signals, as well as from identification of and subcutaneous), and (ii) that transformation
the conditions under which some or all of these can occur in distinct anatomical sites of infec-
genes are expressed in vivo. Equally important tion, which include the lung, peritoneum, and
will be a clearer characterization of the nature of
subcutaneously.These authors further indicated
the relief that the X-state affords to stressed cells.
that subacute long-lasting infections increase
the opportunity for transformation and that
RELATION TO INVASIVE DISEASES
interspecies transformation also occurs. In
Competence during Infection 1972, the last paper culminating this series of
Viewing transformation as a reporter system for studies on transformation in vivo describes a
cell-cell signaling, the historical papers on type 9 pneumococcus colonizing a human vol-
transformation during infection gain a new unteer that is transformed to streptomycin
aspect.The work of Griffith in 1928 introduced resistance after spray administration of the
a completely new concept into science, named rough streptomycin-resistant nonvirulent strain
transformation of type, as he isolated live R36 (70). In summary, these papers show that
serotype 1 cells from mice he infected with in a variety of hosts, including humans, and in
various infectious models (colonization, pneu- yet been identified. In static 24-h biofilms the
monia, abscess, peritonitis, etc.), pneumococci attachment of cells to the wells’ surface was
are transformable, indicating that in vivo the dependent on addition of CSP to the growth
cell-cell communication system of competence medium and was completely abolished in
is generally switched on,or at least that there are mutants lacking the CSP receptor ComD (66).
frequent opportunities for its activation. Consistent with the CSP-ComD interdepen-
dance of the biofilm phenotype, comA, comE,
Expression and Roles and comX were found to be up-regulated in
in Biofilms In Vitro attached cells when compared to cells in liquid
Work on the competence regulatory system has culture (Table 1) (66). This CSP dependence
focused exclusively on its function during the was not observed in early biofilms formed dur-
exponential phase of growth of planktonic ing the exponential phase of growth (63).These
pneumococcal cells. Recently the CSP-ComD results indicate that there are different phases in
signal-sensor system has also been linked to the pneumococcal biofilm formation, of which the
biofilm mode of growth. Pneumococcal earlier seem to be competence independent,
biofilms have been described in vitro when while the later are competence dependent.The
grown on filters (9, 58, 104), in continuous cul- competence system seems thus to be involved
ture flow cells (1, 23), in static microtiter mod- not in biofilm formation, but rather in biofilm
els (63, 66), and in vivo by microscopic maturation or maintenance.
examination of otitis media exudates in chil- The novelty of this new competence-
dren (34). In one standard model of a static dependent regulation rests in the fact that it does
biofilm in microtiter plates, biofilm formation not occur during a phase of rapidly changing
after overnight growth was dependent on addi- cell density and that it does not seem to be lim-
tion of CSP to the growth medium (66).With ited in time, as in competence for genetic trans-
the exception of this static biofilm, for none of formation. The situation in a mature biofilm
the biofilm models has a regulatory mechanism appears to be more similar to a steady-state
TABLE 1 Competence gene expression in vitro in liquid culture and biofilm and during infection in vivo by S.
pneumoniaeTIGR4 (65, 66)
Median fold change (SD) in gene expressiona
Gene 5 min after 10 min after Static 24-h Lungs of Brains of Blood of
CSP addition CSP addition biofilmc mice with mice with septic micec
in liquidb in liquidb pneumoniac meningitisc
Early
SP0042 comA 3.1 (0.7) 1.3 (0.1) 7.1 (0.7) 7.1 (1.0) 10.0(1.5) 1.3 (0.4)
SP2235 comE ND ND ND 4.1(1.5) 6.0(2.1) 1.3(0.3)
SP0014 comX1 3.9 (1.4) 1.4 (0.1) 5.1 (0.8) 7.3 (4.2) 12.0(8.4) 1.6 (0.3)
Late
SP1266 dprA 5.4 (1.3) 2.9 (0.6) ND 7.7 (6.9) 12.2(5.9) 0.9 (0.7)
SP1937 lytA 3.4 (0.03) 10.5 (1.8) ND 1.3 (0.2) 2.0(1.1) 2.3 (0.4)
Delayed
SP0515 hrcA 1.6 (02) 1.4 (0.01) 0.1 (0.03) 0.1 (0.1) 0.1(0.1) 1.2 (0.1)
SP0798 ciaR 1.1 (0.2) 1.0 (0.2) 3.6 (0.3) 1.3 (0.4) 1.7(0.8) 1.5 (0.3)
a
Median values of fold change of expression are calculated for three biological replicas. ND, not done.
b
Changes with respect to time zero.The values reported for liquid medium match those obtained by microarray (20, 78).
c
Gene expression during infection and in biofilm was compared to mid-exponential phase gene expression.This experimental series
therefore has reference conditions similar to those in the experiments in liquid (left two columns). Gene expression in biofilm and during
infection of lung and meninges matches the data in liquid culture for the early competence genes, but it differs with respect to the few late
and delayed genes analyzed.
regulation of the cell density appropriate for assayed on the well bottom, a reduction of total
stabilization of biofilm over time, whereas in cell numbers was observed at just those doses of
genetic transformation, rapidly changing cell CSP most suitable to maintain or stabilize pneu-
density in the exponential phase appears to be mococcal biofilms (66). The number of sessile
responsible for reaching the threshold of CSP and planktonic cells in these overnight cultures
concentration necessary for the transient activa- was not constant, as the reduction of planktonic
tion of X-state genes, which, due to a still cells was by hundreds of millions of cells, while
unknown mechanism,is limited to few minutes. the biofilm in this static model reached at most
Similarly, in biofilm the attachment of cells to only tens of millions of cells (Fig. 3). Still, these
solid surfaces and microcolony formation could data indicate that there is an effect of CSP, not
be the mechanism responsible to reach locally a only on the cells within the biofilm, but also on
suitable CSP concentration for the induction of the cells in suspension. How these data on late
a positive feedback responsible for cell density CSP-dependent phenomena in stationary cul-
maintenance. In such a model, CSP concentra- tures relate to competence development or lysis
tion above the appropriate concentration could of noncompetent cells by competent cells
be inhibitory or have a negative feedback effect. remains to be elucidated (32, 39).
Experimentally, this was indirectly confirmed The importance of cell-cell signaling via the
when only defined concentrations were found competence system in biofilm formation is
to permit biofilm development (Fig.3) (66).The common in the related oral streptococci. In
narrow range of a biofilm-active CSP concen- species including Streptococcus gordonii, Strepto-
tration is in accordance with the model on how coccus sanguis, Streptococcus oralis, and Streptococcus
cell-cell signaling should work in the absence of intermedius, the competence regulatory systems
cell density changes. are organized in a similar fashion as in S. pneu-
When assaying the numbers of planktonic moniae (56). When searching for the genetic
cells in the microwells in which biofilm was basis of biofilm formation in S. gordonii, the first
FIGURE 3 Comparative counts of S. pneumoniae sessile and planktonic cells of different strains incubated for 18 h
with CSP. Counts of sessile cells are reported in gray, and counts of planktonic cells of the same wells are reported in
black. Graphs A and B report the same values with different scales. (A) The CFUs are reported on a log scale that
highlights differences in sessile cells. (B) The CFUs are reported on a linear scale, evidencing variations affecting
planktonic growth. Strains carrying the comC1 allele were incubated with CSP1; D39 (diamond), G54 (circle),
ATCC 6302 (open square),ATCC 6303 (open triangle) and ATCC 6305 (open circle). Strains carrying the comC2
allele were incubated with CSP2;TIGR4 (square),A66 (triangle), and ATCC 6307 (inverted triangle).
mutant isolated was, as in the case of pneumococci in biofilms in humans was demonstrated
coccus, mutated in CSP receptor ComD (50). directly in middle ear infection in humans (34).
The link between the competence system and In acute models of infections in mice,it was also
biofilm formation was described most exten- shown that biofilm bacteria are more virulent
sively for Streptococcus mutans (19, 49, 76, 82). in tissue infection (pneumonia and meningitis)
Albeit carrying the same names (comAB and than bacteria from liquid culture (66).An indi-
comCDE), the genes of the S. mutans compe- rect evidence for the role of biofilm in pneu-
tence system may be evolutionarily more monia and meningitis was also given by
closely related to the second pneumococcal demonstrating that CSP rendered these infec-
peptide-sensing system Blp, which, at least in S. tions more severe,whereas a ComD mutant was
pneumoniae, is not linked to biofilm (56). The attenuated in these disease models (66). An
consistent pattern in these species was that opposite behavior of pneumococci was found
biofilm is induced by the CSP peptide and that in blood during bacteremic sepsis in which (i)
competence mutants in comC (CSP peptide planktonic bacteria were more virulent than
gene), comD (CSP receptor histidine kinase), or biofilm cells, (ii) CSP decreased the virulence
comE (response regulator) are biofilm defective of bacteria, and (iii) ComD mutants were more
(19, 49, 50, 75, 76, 82, 94). When assaying for virulent than wild-type cells (65, 66).
biofilm-specific gene expression patterns, over- Quantitative analysis of pneumococcal gene
laps are evident showing up-regulation of com- expression in pneumonia and meningitis in
petence genes in biofim both in S. pneumoniae mice revealed coordinated up-regulation of all
and in S. gordoni (28, 66). three competence genes assayed (comA, comE,
and comX),while a low level of expression of the
Expression of Competence same genes was found in sepsis (66). In the
Genes during Infection murine experiments, invariant high level of
Horizontal gene transfer has long been recog- competence gene expression in tissues was
nized as the main mechanism by which pneu- found from 6 h after challenge (first sample
mococci evolve, and competence for genetic assayed) until 48 h.Microarray data indicate sig-
transformation has been postulated to be the nificant down-regulation of the competence
principal mode for gene exchange used by this operon in a rabbit model of meningitis with
species.Thus, genetic diversity is the first reason high inoculum and early sacrifice, but no chal-
why competence genes must be expressed at lenge data with selected mutants have been
some stages during the life cycle of pneumo- reported to complement the gene expression
cocci in the host.The recognition of pneumo- data in this model so far (67).The comparison
coccal biofilms in humans and the importance of pneumococcal gene expression in lung and
of CSP-ComD interaction in biofilm indicate a on meninges of mice shows an identical pattern
second possible reason for competence gene to gene expression in pneumococcal biofilms
expression in vivo. The presence of pneumo- (Table 1). This overlap in the gene expression
TABLE 2 Competence-related phenotypes in vitro and during infection

Competence Model situation Effect of CSP Effect of ComD mutation
Off
In vivo Sepsis Decreases virulence Increases virulence
In vitro Liquid medium Inhibits growth and induces None
transformation
On
In vivo Meningitis and pneumonia Increases virulence Decreases virulence
In vitro Mature biofilm Increases virulence Inhibits biofilm
profile,the phenotypic modulation of virulence most important? When is the X state impor-
in mice using CSP or the comD mutant, and the tant? When is a post-X state important? If it is
increased infectivity of biofilm cells all indicate important, as Sun Tzu famously remarked of
either directly or indirectly that competence military preparations, to know your enemy,
genes are expressed at high levels during these then surely it is especially important to know
types of infections.Whether competence gene the strategies that your enemy has for coopera-
expression in this context is also related to trans- tion in large numbers.
formation, as in biofilm, or whether transfor-
mation happens in moments of transition of ACKNOWLEDGMENTS
competence gene expression, possibly linked to Parts of recent work discussed that was carried out in
changes of ecological niche and/or disease Chicago were supported by the National Science
stage, is still unknown. Some of the principal Foundation (MCB-011-0311, MCB-054-3187).Work
carried out in Siena was supported in part by the Euro-
pneumococcal phenotypes related to compe- pean Commission grants LSHM-CT-2005-512099
tence gene expression in biofilm and during and LSHB-CT-2005-512061 and grant from MIUR
infection are summarized in Table 2. Although FIRB_RBAU01X9TB.
expression in bacteremic sepsis was found to
be very low for competence genes, no data REFERENCES
are yet available for colonization or even the 1. Allegrucci, M., F. Z. Hu, K. Shen, J. Hayes,
intracellular life of pneumococci. Further work G. D. Ehrlich, J. C. Post, and K. Sauer. 2006.
elucidating the regulatory circuit connected Phenotypic characterization of Streptococcus pneumo-
niae biofilm development. J. Bacteriol. 188:
to competence genes in this situation still has 2325–2335.
to be done, as does the identification of the 2. Alloing, G., C. Granadel, D. A. Morrison, and
stimuli guiding competence gene expression J. P. Claverys. 1996. Competence pheromone,
in vivo. oligopeptide permease, and induction of compe-
tence in Streptococcus pneumoniae. Mol. Microbiol.
21:471–478.
OPEN QUESTIONS 3. Alloing, G., B. Martin, C. Granadel, and J. P.
While recent progress has revealed many Claverys. 1998. Development of competence in
aspects of how pneumococcus can act coopera- Streptococcus pneumonaie: pheromone autoinduction
tively to regulate gene expression, and points to and control of quorum sensing by the oligopeptide
several strategic goals served by the coopera- permease. Mol. Microbiol. 29:75–83.
4. Austrian, R. 1952. Observations on the transfor-
tion, much remains to be explained. What mation of pneumococcus in vivo. Bull. Johns
metabolic signals trigger competence or modu- Hopkins Hosp. 91:189–196.
late it? Does the cell-cell signaling system serve 5. Bartilson, M., A. Marra, J. Christine, J.
for cell counting or for a cooperative response S. Asundi, W. P. Schneider, and A. E.
to other signals? How is sudden activation of Hromockyj. 2001. Differential fluorescence
induction reveals Streptococcus pneumoniae loci regu-
the pheromone response coordinated? How is lated by competence stimulatory peptide. Mol.
donor DNA protected, processed, and targeted Microbiol. 39:126–35.
for gene replacement? What are the roles of 6. Bergé, M. 2002. Compétence pour la transforma-
CSP-induced genes not required for regulation tion gènétique chez la bactérie à Gram positif
or genetic exchange? How is shutoff of the Streptococcus pneumoniae: étude du gène tardif dprA.
Ph.D. thesis, Université Paul Sabatier de Toulouse
CSP response accomplished and modulated? Is (Sciences).
there a specific post-X state? What bacterial 7. Bergé, M., I. Mortier-Barrière, B. Martin, and
species are targets of the lytic action of compe- J. P. Claverys. 2003.Transformation of Streptococcus
tent cells? Fundamental understanding of the pneumoniae relies on DprA- and RecA-dependent
biology of this pheromone will also depend on protection of incoming DNA single strands. Mol.
finding approaches to more general questions 8. Bergé, M., M. Moscoso, M. Prudhomme,
about the biology of interaction with the host: B. Martin, and J. P. Claverys. 2002. Uptake of
How is CSP important in vivo? What DNA is transforming DNA in gram-positive bacteria: a
view from Streptococcus pneumoniae. Mol. Microbiol. Martin, J.Wells, and J. P. Claverys. 2004. Inter-
45:411–421. connection of competence, stress and CiaR regu-
9. Budhani, R. K., and J. K. Struthers. 1997.The lons in Streptococcus pneumoniae: competence
use of sorbarod biofilms to study the antimicrobial triggers stationary phase autolysis of ciaR mutant
susceptbility of a strain of Streptococcus pneumoniae. cells. Mol. Microbiol. 51:1071–1086.
J.Antimicrob. Chemother. 40:601–602. 21. Dawson, M. H. 1928.The interconvertibility of
9a. Burghout, P., H. J. Bootsma, T. G. Klooster- R and S forms of pneumococcus. J. Exp. Med.
man, J. J. E. Bijlsma, C. E. de Jongh, O. P. 47:577–591.
Kuipers, and P. W. Hermans. 2007. Search for 22. Desai, B. V., and D. A. Morrison. 2006. An
genes essential for pneumococcal transformation: unstable competence-induced protein, CoiA,
the RADA DNA repair protein plays a role in promotes processing of donor DNA after uptake
genomic recombination of donor DNA. J. Bacte- during genetic transformation in Streptococcus pneu-
riol. 189:6540–6550. moniae. J. Bacteriol. 188:5177–5186.
10. Campbell, E.A., S.Y. Choi, and H. R. Masure. 23. Donlan, R. M., J. A. Piede, C. D. Heyes, L.
1998. A competence regulon in Streptococcus pneu- Sanii, R. Murga, P. Edmonds, I. El-Sayed, and
moniae revealed by genomic analysis.Mol.Microbiol. M.A. El-Sayed. 2004. Model system for growing
27:929–939. and quantifying Streptococcus pneumoniae biofilms
11. Chastanet,A., M. Prudhomme, J. P. Claverys, in situ and in real time. Appl. Environ. Microbiol.
and T. Msadek. 2001. Regulation of Streptococcus 70:4980–4988.
pneumoniae clp genes and their role in competence 24. Dunny, G. M. and S. C. Winans (ed.). 1999.
development and stress survival. J. Bacteriol. Cell-Cell Signalling in Bacteria. ASM Press, Wash-
183:7295–7307. ington, DC.
12. Chen, J. D., and D. A. Morrison. 1987. Modu- 25. Echenique, J., A. Kadioglu, S. Romao, P. W.
lation of competence for genetic transformation in Andrew, and M. C.Trombe. 2004. Protein ser-
Streptococcus pneumoniae. J. Gen. Microbiol. ine/threonine kinase StkP positively controls viru-
133:1959–1967. lence and competence in Streptococcus pneumoniae.
13. Claverys, J. P., B. Grossiord, and G. Alloing. Infect. Immun. 72:2434–2437.
2000. Is the Ami-AliA/B oligopeptide per- 26. Ephrussi-Taylor, H. 1955.Current status of bac-
mease of Streptococcus pneumoniae involved in sens- terial transformations. Adv.Virus Res. 3: 275–307.
ing environmental conditions? Res. Microbiol. 27. Giammarinaro, P., M. Sicard, and A. M. Gasc.
151:457–463. 1999. Genetic and physiological studies of the
14. Claverys, J. P., and L. S. Havarstein. 2002. CiaH-CiaR two-component signal-transducing
Extracellular-peptide control of competence for system involved in cefotaxime resistance and com-
genetic transformation in Streptococcus pneumoniae. petence of Streptococcus pneumoniae. Microbiology
Front. Biosci. 7:1798–1814. 145:1859–1869.
15. Claverys, J. P., M. Prudhomme, and B. 28. Gilmore, K. S., P. Srinivas, D. R. Akins, K. L.
Martin. 2006. Induction of competence regulons Hatter, and M. S. Gilmore. 2003. Growth,
as a general response to stress in gram-positive development, and gene expression in a persistent
bacteria. Annu. Rev. Microbiol. 60:451–475. Streptococcus gordonii biofilm. Infect. Immun.
16. Claverys, J. P., M. Prudhomme, I. Mortier- 71:4759–4766.
Barriere, and B. Martin. 2000. Adaptation to 29. Griffith,F. 1928.The significance of pneumococ-
the environment: Streptococcus pneumoniae, a para- cal types. J. Hyg. 27:113–159.
digm for recombination-mediated genetic plastic- 30. Guenzi, E., A. M. Gasc, M. A. Sicard, and R.
ity? Mol. Microbiol. 35:251–259. Hakenbeck. 1994. A two-component signal-
17. Conant, J.E., and W. D. Sawyer. 1967.Transfor- transducing system is involved in competence and
mation during mixed pneumococcal infection of penicillin susceptibility in laboratory mutants of
mice. J. Bacteriol. 6:1869–1875. Streptococcus pneumoniae. Mol. Microbiol.
18. Coomaraswamy, G. 1996. Control of compe- 12:505–515.
tence in Streptococcus pneumoniae by synthetic pep- 31. Guiral, S.,V. Henard, C. Granadel, B. Martin,
tide pheromones. Ph.D. thesis, University of and J. P. Claverys. 2006. Inhibition of compe-
Illinois at Chicago, Chicago. tence development in Streptococcus pneumoniae by
19. Cvitkovitch, D. G.,Y. H. Li, and R. P. Ellen. increased basal-level expression of the ComDE
2003. Quorum sensing and biofilm formation in two-component regulatory system. Microbiology
streptococcal infections. J. Clin. Invest. 112: 152:323–331.
1626–1632. 32. Guiral, S., T. J. Mitchell, B. Martin, and J.-P.
20. Dagkessamanskaia, A., M. Moscoso, V. Claverys. 2005. Competence-programmed
Henard, S. Guiral, K. Overweg, M. Reuter, B. predation of noncompetent cells in the human
pathogen Streptococcus pneumoniae: genetic require- niae is essential for competence-induced cell lysis.
ments. Proc. Natl.Acad. Sci. USA 102:8710–8715. J. Bacteriol. 187:4338–4345.
33. Hall, R. H., and G. O. Gale. 1959. New obser- 44. Lacks, S.A. 2004.Transformation,p.89–115.In E.
vations on pneumococcal transformation in vivo. I. Tuomanen (ed.), The Pneumococcus. ASM Press,
Proc. Soc. Exp. Biol. Med. 101:487–491. Washington, DC.
34. Hall-Stoodley, L., F. Z. Hu, A. Gieseke, L. 45. Lacks, S.A., S.Ayalew,A. G. de la Campa, and
Nistico, D. Nguyen, J. Hayes, M. Forbes, D. P. B. Greenberg. 2000. Regulation of competence
Greenberg, B. Dice, A. Burrows, P. A. for genetic transformation in Streptococcus pneumo-
Wackym, P. Stoodley, J. C. Post, G. D. Ehrlich, niae: expression of dpnA, a late competence
and J. E. Kerschner. 2006. Direct detection of gene encoding a DNA methyltransferase of the
bacterial biofilms on the middle-ear mucosa of DpnII restriction system. Mol. Microbiol. 35:
children with chronic otitis media. JAMA 1089–1098.
296:202–211. 46. Lacks, S.A., and B. Greenberg. 2001.Constitu-
35. Håvarstein, L. S., G. Coomaraswamy, and D. tive competence for genetic transformation in
A. Morrison. 1995. An unmodified heptade- Streptococcus pneumoniae caused by mutation of a
capeptide pheromone induces competence for transmembrane histidine kinase. Mol. Microbiol.
genetic transformation in Streptococcus pneumoniae. 42:1035–1045.
Proc. Natl.Acad. Sci. USA 92:11140–11144. 47. Lee, M. S., and D. A. Morrison. 1999. Identifi-
36. Håvarstein, L. S., D. B. Diep, and I. F. Nes. cation of a new regulator in Streptococcus pneumoniae
1995. A family of bacteriocin ABC transporters linking quorum sensing to competence for genetic
carry out proteolytic processing of their substrates transformation. J. Bacteriol. 181:5004–5016.
concomitant with export. Mol. Microbiol. 16: 48. Lefrançois, J., M. M. Samrakandi, and A. M.
229–240. Sicard. 1998. Electrotransformation and natural
37. Håvarstein, L. S., P. Gaustad, I. F. Nes, and D. transformation of Streptococcus pneumoniae: require-
A. Morrison. 1996. Identification of the strepto- ment of DNA processing for recombination.
coccal competence-pheromone receptor. Mol. Microbiology 144:3061–3068.
Microbiol. 21:863–869. 49. Li,Y. H., N.Tang, M. B.Aspiras, P. C. Lau, J. H.
38. Håvarstein, L. S., R. Hakenbeck, and P. Lee, R. P. Ellen, and D. G. Cvitkovitch. 2002.A
Gaustad. 1997. Natural competence in the genus quorum-sensing signaling system essential for
Streptococcus: evidence that streptococci can change genetic competence in Streptococcus mutans is
pherotype by interspecies recombinational involved in biofilm formation. J. Bacteriol.
exchanges. J. Bacteriol. 179:6589–6594. 184:2699–2708.
39. Håvarstein, L. S., B. Martin, O. Johnsborg, 50. Loo, C.Y., D. Corliss, and N. Ganeshkumar.
C. Granadel, and J.-P. Claverys. 2006. New 2000. Streptococcus gordonii biofilm formation: iden-
insights into the pneumococcal fratricide:relation- tification of genes that code for biofilm pheno-
ship to clumping and identification of a novel types. J. Bacteriol. 182: 1374–1382.
immunity factor. Mol. Microbiol. 59:1297–1307. 51. Luo, P. 2003. Genetic transformation in Streptococ-
40. Håvarstein, L. S., and D. A. Morrison. 1998. cus pneumoniae:regulation by ComX,an alternative
Quorum-sensing and peptide pheromones, p. sigma factor. Ph.D. thesis, University of Illinois at
9–26. In G. M. Dunny and S. C. Winans (ed.), Chicago, Chicago.
Cell-Cell Signaling in Bacteria.ASM Press,Washing- 52. Luo, P., H. Li, and D. A. Morrison. 2003.
ton, DC. ComX is a unique link between multiple quorum
41. Humbert, O., M. Prudhomme, R. Haken- sensing outputs and competence in Streptococcus
beck, C. G. Dowson, and J. P. Claverys. 1995. pneumoniae. Mol. Microbiol. 50:623–633.
Homeologous recombination and mismatch 53. Luo, P., H. Li, and D.A. Morrison. 2004. Iden-
repair during transformation in Streptococcus tification of ComW as a new component in the
pneumoniae: saturation of the Hex mismatch regulation of genetic transformation in Streptococ-
repair system. Proc. Natl. Acad. Sci. USA 92: cus pneumoniae. Mol. Microbiol. 54:172–183.
9052–9056. 54. Luo, P., and D. A. Morrison. 2003. Transient
42. Iannelli, F., M. R. Oggioni, and G. Pozzi. association of an alternative sigma factor, ComX,
2005. Sensor domain of histidine kinase ComD with RNA polymerase during the period of com-
confers competence pherotype specificity in Strep- petence for genetic transformation in Streptococcus
toccoccus pneumoniae. FEMS Microbiol. Lett. pneumoniae. J. Bacteriol. 185:349–358.
252:321–326. 55. Martin, B., M. Prudhomme, G. Alloing, C.
43. Kausmally, L., O. Johnsborg, M. Lunde, E. Granadel, and J. P. Claverys. 2000. Cross-
Knutsen, and L. S. Håvarstein. 2005. Choline- regulation of competence pheromone production
binding protein D (CbpD) in Streptococcus pneumo- and export in the early control of transformation
in Streptococcus pneumoniae. Mol. Microbiol. 38: 67. Orihuela, C. J., J. N. Radin, J. E. Sublett, G.
867–878. Gao, D. Kaushal, and E. I. Tuomanen. 2004.
56. Martin, B.,Y. Quentin, G. Fichant, and J.-P. Microarray analysis of pneumococcal gene expres-
Claverys. 2006. Independent evolution of com- sion during invasive disease. Infect. Immun.
petence regulatory cascades in streptococci? 72:5582–5596.
Trends Microbiol. 14:339–345. 68. Ottolenghi, E. 1963. Genetic transformation
57. Mascher, T., D. Zahner, M. Merai, N. among living pneumococci in the mouse. Genetics
Balmelle, A. B. de Saizieu, and R. 50:417–419.
Hakenbeck. 2003. The Streptococcus pneumoniae 69. Ottolenghi-Nightingale, E. 1969. Sponta-
cia regulon: CiaR target sites and transcription neously occurring bacterial transformation in
profile analysis. J. Bacteriol. 185:60–70. mice. J. Bacteriol. 100:445–452.
58. McEllistrem, M. C., J. Ransford, and S. 70. Ottolenghi-Nightingale, E. 1972. Competence
Khan. 2007. Characterisation of in vitro biofilm- of pneumococcal isolates and bacterial transforma-
associated pneumococcal phase variants of a tion in man. Infect. Immun. 6:785–792.
clinically-relevant serotype 3 clone. J. Clin. Micro- 71. Pakula, R., and W. Walczak. 1963. On the
biol. 45:97–101. nature of competence of transformable strepto-
59. Méjean,V., and J. P. Claverys. 1993.DNA pro- cocci. J. Gen. Microbiol. 31:125–133.
cessing during entry in transformation of Strepto- 72. Pearce, B. J., A. M. Naughton, E. A.
coccus pneumoniae. J. Biol. Chem. 268:5594–5599. Campbell, and H. R. Masure. 1995. The rec
60. Morrison, D. A. 1997. Streptococcal compe- locus, a competence-induced operon in Streptococ-
tence for genetic transformation: regulation by cus pneumoniae. J. Bacteriol. 177:86–93.
peptide pheromones.Microb.Drug Resist. 3:27–37. 73. Pestova, E. V., L. S. Håvarstein, and D. A.
61. Morrison, D.A., and M. F. Baker. 1979.Com- Morrison. 1996. Regulation of competence for
petence for genetic transformation in pneumo- genetic transformation in Streptococcus pneumoniae
coccus depends on synthesis of a small set of by an auto-induced peptide pheromone and a
proteins. Nature 282:215–217. two-component regulatory system. Mol. Microbiol.
61a. Mortier-Barrière, I., M.Velten, P. Dupaigne, 21:853–862.
N. Mirouze, O. Piétrement, S. McGovern, G. 74. Pestova, E.V., and D.A. Morrison. 1998. Isola-
Fichant, B. Martin, P. Noirot, E. Le Cam, P. tion and characterization of three Streptococcus
Polard, and J. P. Claverys. 2007. A key presy- pneumoniae transformation-specific loci by use of a
naptic role in transformation for a widespread lacZ reporter insertion vector. J. Bacteriol.
bacterial protein: DprA conveys incoming 180:2701–2710.
ssDNA to RecA. Cell 130:824–836. 75. Petersen, F. C., D. Pecharki, and A.A. Scheie.
62. Moscoso, M., and J. P. Claverys. 2004.Release 2004.Biofilm mode of growth of Streptococcus inter-
of DNA into the medium by competent Strepto- medius favored by a competence-stimulating sig-
coccus pneumoniae: kinetics, mechanism and stabil- naling peptide. J. Bacteriol. 186:6327–6331.
ity of the liberated DNA. Mol. Microbiol. 76. Petersen, F. C., L.Tao, and A.A. Scheie. 2005.
54:783–794. DNA binding-uptake system: a link between
63. Moscoso, M., E. Garcia, and R. Lopez. 2006. cell-to-cell communication and biofilm forma-
Biofilm formation by Streptococcus pneumoniae: tion. J. Bacteriol. 187:4392–4400.
role of choline, extracellular DNA, and capsular 77. Peterson, S., R.T. Cline, H.Tettelin,V. Sharov,
polysaccharide in microbial accretion. J. Bacteriol. and D. A. Morrison. 2000. Gene expression
188:7785–7795. analysis of the Streptococcus pneumoniae competence
64. Neufeld, F., and W. Levinthal. 1928. Beitraege regulons by use of DNA microarrays. J. Bacteriol.
zur variabilitaet der pneumokokken. Zeitschrift 182:6192–6202.
fuer Immunitaetsforschung 55:324–340. 78. Peterson, S. N., C. K. Sung, R. Cline, B. V.
65. Oggioni, M. R., F. Iannelli, S. Ricci, D. Chi- Desai, E. C. Snesrud, P. Luo, J.Walling, H. Li,
avolini, R. Parigi, C.Trappetti, J. P. Claverys, M. Mintz, G. Tsegaye, P. C. Burr, Y. Do, S.
and G. Pozzi. 2004. Antibacterial activity of a Ahn, J. Gilbert, R. D. Fleischmann, and D. A.
competence-stimulating peptide in experimental Morrison. 2004. Identification of competence
sepsis caused by Streptococcus pneumoniae. Antimi- pheromone responsive genes in Streptococcus pneu-
crob.Agents Chemother. 48:4725–4732. moniae by use of DNA microarrays. Mol. Microbiol.
66. Oggioni, M. R., C. Trappetti, A. Kadioglu, 51:1051–1070.
M. Cassone, F. Iannelli, S. Ricci, P. W. 79. Pozzi, G., L. Masala, F. Iannelli, R.
Andrew, and G. Pozzi. 2006. Switch from Manganelli, L. S. Håvarstein, L. Piccoli, D.
planktonic to sessile life: a major event in pne Simon, and D.A. Morrison. 1996. Competence
mococcal pathogenesis. Mol. Microbiol. 61:1196– for genetic transformation in encapsulated
1210. strains of Streptococcus pneumoniae: two allelic
variants of the peptide pheromone. J. Bacteriol. 178: 91. Steinmoen, H., A. Teigen, and L. S.
6087–6090. Håvarstein. 2003.Competence-induced cells of
80. Prudhomme, M., L. Attaiech, G. Sanchez, B. Streptococcus pneumoniae lyse competence-defi-
Martin, and J. P. Claverys. 2006. Antibiotic cient cells of the same strain during cocultivation.
stress induces genetic transformability in the J. Bacteriol. 185:7176–7183.
human pathogen Streptococcus pneumoniae. Science 92. Sturme, M. H., M. Kleerebezem, J.
313:89–92. Nakayama, A. D. Akkermans, E. E.Vaugha,
81. Prudhomme, M., V. Libante, and J. P. and W. M. de Vos. 2002. Cell to cell communi-
Claverys. 2002. Homologous recombination at cation by autoinducing peptides in gram-positive
the border: insertion-deletions and the trapping of bacteria. Antonie Van Leeuwenhoek 81:233–243.
foreign DNA in Streptococcus pneumoniae.Proc.Natl. 93. Sung, C. K., and D. A. Morrison. 2005.Two
Acad. Sci. USA 99:2100–2105. distinct functions of ComW in stabilization
82. Qi, F., J. Kreth, C. M. Levesque, O. Kay, R.W. and activation of the alternative sigma factor
Mair, W. Shi, D. G. Cvitkovitch, and S. D. ComX in Streptococcus pneumoniae. J. Bacteriol.
Goodman. 2006. Peptide pheromone induced 187:3052–3061.
cell death of Streptococcus mutans. FEMS Microbiol. 94. Suntharalingam, P., and D. G. Cvitkovitch.
Lett. 251:321–326. 2005. Quorum sensing in streptococcal biofilm
83. Ramirez, M., D.A. Morrison, and A.Tomasz. formation. Trends Microbiol. 13:3–6.
1997. Ubiquitous distribution of the competence 95. Surette, M. G., M. B. Miller, and B. L.
related genes comA and comC among isolates of Bassler. 1999.Quorum sensing in Escherichia coli,
Streptococcus pneumoniae. Microb. Drug Resist. Salmonella typhimurium, and Vibrio harveyi: a new
3:39–52. family of genes responsible for autoinducer pro-
84. Redfield, R. J. 2002. Is quorum sensing a side duction. Proc. Natl.Acad. Sci. USA 96:1639–1644.
effect of diffusion sensing? Trends Microbiol. 96. Tettelin, H., K. E. Nelson, I.T. Paulsen, J. A.
10:365–370. Eisen,T. D. Read, S. Peterson, J. Heidelberg,
85. Rimini, R., B. Jansson, G. Feger,T. C. Roberts, R.T. DeBoy, D. H. Haft, R. J. Dodson, A. S.
M. de Francesco, A. Gozzi, F. Faggioni, E. Durkin, M. Gwinn, J. F. Kolonay, W. C.
Domenici, D. M.Wallace, N. Frandsen, and A. Nelson, J. D. Peterson, L. A. Umayam, O.
Polissi. 2000. Global analysis of transcription White, S. L. Salzberg, M. R. Lewis, D.
kinetics during competence development in Strep- Radune, E. Holtzapple, H. Khouri, A. M.
tococcus pneumoniae using high density DNA arrays. Wolf, T. R. Utterback, C. L. Hansen, L. A.
Mol. Microbiol. 36: 1279–1292. McDonald, T. V. Feldblyum, S. Angiuoli, T.
86. Risoen, P. A., M. B. Brurberg, V. G. Eijsink, Dickinson, E. K. Hickey, I. E. Holt, B. J.
and I. F. Nes. 2000. Functional analysis of pro- Loftus, F.Yang, H. O. Smith, J. C.Venter, B.
moters involved in quorum sensing-based regula- A. Dougherty, D. A. Morrison, S. K.
tion of bacteriocin production in Lactobacillus. Mol. Hollingshead, and C. M. Fraser. 2001. Com-
Microbiol. 37:619–628. plete genome sequence of a virulent isolate of
87. Robertson, G. T., W. L. Ng, J. Foley, Streptococcus pneumoniae. Science 293:498–506.
R. Gilmour, and M. E. Winkler. 2002. 97. Thomas, R. 1955. Recherches sur la cinétique
Global transcriptional analysis of clpP mutations des transformations bactériennes. Biochim. Bio-
of type 2 Streptococcus pneumoniae and their phys.Acta 18:467–481.
effects on physiology and virulence. J. Bacteriol. 98. Tomasz, A. 1966. Model for the mechanism
184:3508–3520. controlling the expression of competent state in
88. Romao, S., G. Memmi, M. R. Oggioni, Pneumococcus cultures. J. Bacteriol. 91:1050–1061.
and M. C. Trombe. 2006. LuxS impacts on 99. Tomasz, A. 1970. Cellular metabolism in
LytA-dependent autolysis and on competence genetic transformation of pneumococci: require-
in Streptococcus pneumoniae. Microbiology 152: ment for protein synthesis during induction of
333–341. competence. J. Bacteriol. 101:860–871.
89. Sebert, M. E., K. P. Patel, M. Plotnick, and J. 100. Tomasz,A., and R. D. Hotchkiss. 1964. Reg-
N. Weiser. 2005. Pneumococcal HtrA protease ulation of the transformability of pneumococcal
mediates inhibition of competence by the CiaRH cultures by macromolecular cell products. Proc.
two-component signaling system. J. Bacteriol. Natl.Acad. Sci. USA 51:480–487.
187:3969–3979. 101. Tomasz, A., and J. L. Mosser. 1966. On the
90. Steinmoen, H., E. Knutsen, and L. S. nature of the pneumococcal activator substance.
Håvarstein. 2002. Induction of natural compe- Proc. Natl.Acad. Sci. USA 55:58–66.
tence in Streptococcus pneumoniae triggers lysis and 102. Visick, K. L., and C. Fuqua. 2005. Decoding
DNA release from a subfraction of the cell popula- microbial chatter: cell-cell communication in
tion. Proc. Natl.Acad. Sci. USA 99:7681–7686. bacteria. J. Bacteriol. 187:5507–5519.
103. Wagner, C., A. Saizieu Ad, H. J. Schonfeld, Streptococcus pneumoniae. Mol. Microbiol.
M. Kamber, R. Lange, C. J. Thompson, 33:817–827.
and M. G. Page. 2002. Genetic analysis and 107. Woodbury, R. L., X.Wang, and C. P. Moran,
functional characterization of the Streptococcus Jr. 2006. Sigma X induces competence gene
pneumoniae vic operon. Infect. Immun. 70: expression in Streptococcus pyogenes. Res. Microbiol.
6121–6128. 157:851–856.
104. Waite, R. D., J. K. Struthers, and C. G. 108. Zahner, D., T. Grebe, E. Guenzi, J. Krauss,
Dowson. 2001. Spontaneous sequence duplica- M. van der Linden, K. Terhune, J. B. Stock,
tion within an open reading frame of the and R. Hakenbeck. 1996. Resistance determi-
pneumococcal type 3 capsule locus causes high- nants for beta-lactam antibiotics in laboratory
frequency phase variation. Mol. Microbiol. mutants of Streptococcus pneumoniae that are
42:1223–1232. involved in genetic competence. Microb. Drug
105. Waters, C. M., and B. L. Bassler. 2005. Quo- Resist. 2:187–191.
rum sensing: cell-to-cell communication in bac- 109. Zahner, D., K. Kaminski, M. van der
teria. Ann. Rev. Cell Devel. Biol. 21:319–346. Linden, T. Mascher, M. Meral, and R.
106. Ween, O., P. Gaustad, and L. S. Håvarstein. Hakenbeck. 2002. The ciaR/ciaH regulatory
1999. Identification of DNA binding sites for network of Streptococcus pneumoniae. J. Mol.
ComE, a key regulator of natural competence in Microbiol. Biotechnol. 4:211–216.
THE A FACTOR REGULATORY CASCADE
THAT TRIGGERS SECONDARY
METABOLISM AND MORPHOLOGICAL
DIFFERENTIATION IN STREPTOMYCES
Sueharu Horinouchi
23
The genus Streptomyces comprises gram- mits the eventual separation of adjacent spores.
positive, soil-dwelling, filamentous bacteria. It Spore chains usually consist of many tens of
shows complex morphological differentiation spores.The aerial spores formed in this way are
resembling that of filamentous fungi, which resistant to heat treatment and lysozyme diges-
makes this genus one of the model prokaryotes tion. Streptomyces strains have hence been called
to study multicellular differentiation. On agar “boundary organisms” between prokaryotes
medium, one or more substrate hyphae formed and eukaryotes.
from a germinating spore branch frequently Another characteristic of the genus Strepto-
and grow rapidly by cell-wall extension at the myces is the ability to produce a wide variety of
hyphal tips.Aerial hyphae subsequently emerge secondary metabolites, including antibiotics
by reuse of material assimilated into the sub- and biologically active substances. Secondary
strate mycelium, such as DNA, proteins, lipids, metabolism is sometimes termed “physiologi-
and storage compounds. Because many cells in cal” differentiation because it occurs during the
substrate hyphae lyse and die, this process is idiophase after the main period of rapid vegeta-
sometimes called cannibalism or apoptosis. tive growth and assimilative metabolism. The
When apical growth of aerial hyphae stops, in ability to produce a wide variety not only of
contrast to substrate mycelium, septa are secondary metabolites but also of industrially
formed at regular intervals along the hyphae to important enzymes makes the genus Strepto-
form many unigenomic compartments within myces an excellent supplier of industrially and
a sheath composed of elongated hollow or clinically important substances.
grooved elements, finer fibrillar elements, and Morphological development and secondary
amorphous material. The sporulation septa metabolism are simultaneously under the con-
consist of two membrane layers separated by a trol of various nutritional environments,such as
double layer of cell-wall material, which per- carbon, nitrogen, phosphorous nutrients, and
trace elements. In addition to the nutritional
control, chemical signaling molecules having
Sueharu Horinouchi Department of Biotechnology, Graduate
School of Agriculture and Life Sciences, The University of a -butyrolactone ring also control both mor-
Tokyo, Bunkyo-ku,Tokyo 113-8657, Japan. phological and physiological differentiation
363
364 ■ HORINOUCHI
of Streptomyces spp. The pioneer work structure of CprB, a homolog of the A factor
of a Russian microbiologist, A. S. Khokhlov receptor protein, and (iii) a key transcriptional
(28), in the 1960s on an autoregulatory factor factor AdpA in the A factor regulatory cascade.
(A factor; 2-isocapryloyl-3R-hydroxymethyl- We can now explain Khokhlov’s discovery 40
-butyrolactone;) (see Fig. 1 for chemical years ago of the mysterious link between A fac-
structure), which simultaneously induced tor and streptomycin biosynthesis at the molec-
sporulation and streptomycin biosynthesis ular level.
in a mutant of Streptomyces griseus, revealed
an exact link between secondary metabolism PROPERTIES OF A FACTOR AND ITS
HOMOLOGS IN ACTINOMYCETES
and morphological differentiation. His inter-
esting, but mysterious, observation had been Biological Activities of A Factor
neglected by Western scientists for more The strain we have studied is a historical one, S.
than 25 years until Hara and Beppu (9, 10) griseus IFO13350 ( ISP5236), which won the
confirmed the link via A factor using chemi- Nobel Prize for Selman A. Waksman. The
cally synthesized, optically active (3R)-A genetic study of A factor biosynthesis by Hara
factor. Since then, the data from several labora- and Beppu (10) with this strain showed that the
tories have shown that A factor and its deriv- chemically synthesized, optically active (3R)-
atives in various Streptomyces spp. are actually form of A factor (38) at a concentration as low
autoregulators or microbial hormones that as 109 M restores all the phenotypic defects in
switch on morphological differentiation or streptomycin production and sporulation of an
secondary metabolism, or both (reviewed in A factor-deficient mutant strain HH1. Subse-
references 15–18, 60, 73). In addition to the quent studies showed that A factor also controls
hormonal control by A factor, protein the production of almost all the secondary
serine/threonine kinases also control second- metabolites, including a yellow pigment named
ary metabolism and morphogenesis. A repre- grixazone (41, 59) and a melanin-like pigment
sentative of more than 30 serine/threonine named hexahydroxyperylenequinone (6, 7), in
kinases in a given Streptomyces sp. is AfsK, which addition to streptomycin, that are produced by
activates the DNA-binding activity of a tran- S. griseus. A factor-deficient mutants, like
scriptional factor AfsR by phosphorylating mutant HH1, were readily obtained by treat-
its threonine residue (14, 63, 68).A calmodulin- ment of the wild-type strain by UV irradiation
like protein (80) and cAMP, biosynthesized or incubation at 32C (9).The extreme instabil-
by a eukaryotic-type adenylcyclase (21, 58), ity of the A factor productivity was later
are also involved in morphological differentia- explained in terms of the location of afsA, a
tion. Hence, Streptomyces members are bound- gene encoding a key A factor biosynthesis
ary organisms from the viewpoint not only of enzyme AfsA (1, 23), which is located in the
their morphogenesis but also of gene regula- vicinity of one end of the linear chromosome
tory systems. (34).The ends of the Streptomyces chromosomes
This chapter deals with the study of the biol- are deleted at high frequency due to homolo-
ogy and chemistry of -butyrolactone-type gous recombination between a long repeat
autoregulators that switch on secondary sequence on both ends of the linear chromo-
metabolism and morphological differentiation some. The genome sequence of S. griseus in a
in Streptomyces. Because A factor has been the total of 8,545,929 bp (Y. Ohnishi et al., unpub-
representative of the -butyrolactones since its lished data) shows that afsA is located 272 kb
discovery, I summarize the study of the A factor apart from one end of the chromosome. The
regulatory cascade that leads to aerial mycelium location of afsA near one end of the linear chro-
formation and secondary metabolite formation mosome of S. griseus is the molecular basis of
in S. griseus, with emphases on (i) the whole the simultaneous loss of streptomycin produc-
biosynthesis pathway of A factor, (ii) the crystal tion and sporulation that has long been
23. A FACTOR REGULATORY CASCADE IN STREPTOMYCES ■ 365
empirically observed by those who are engaged with piecemeal sporulation of individual
in streptomycin fermentation. hyphae induced by environmental stimuli such
A factor and its homologs produced at a por- as nutritional limitation.
tion of a hypha can move freely within the indi-
vidual hypha and spread into neighboring A Factor Homologs in Actinomycetes
hyphae. Due to filamentous growth, Strepto- A factor homologs having a -butyrolactone
myces might have developed diffusible - structure have been found in various Strepto-
butyrolactone regulatory systems that facilitate myces species, such as S. bikiniensis, S. coelicolor
communication between the cells at a distance A3(2), S. cyaneofuscatus, S. lavendulae, S. virginiae,
within an individual hypha and between differ- and S. viridochromogenes (reviewed in references
ent hyphae.The filamentous mycelia of Strepto- 15–17, 60). Virginiae butanolides (VBs) con-
myces are close enough to communicate with trolling virginiamycin production in S. virginiae
one another. The signaling system between (74), IM-2 controlling pigment production
physically separate individual cells in the same in S. lavendulae (11), and SCB1 controlling
mycelium can be termed hormonal regulation, actinorhodin and undecylprodigiosin produc-
rather than quorum-sensing regulation found tion in S. coelicolor A3(2) (62) are examples.
in gram-negative single-cell bacteria growing Their chemical structures are shown in Fig. 1,
in liquid culture (3, 8). However,A factor is also and the chemistry is discussed below in relation
important in cell-cell communication between to the key biosynthesis enzyme AfsA for -
neighboring mycelia, similarly to the quorum- butyrolactones.These strains contain homologs
sensing system. Since a given Streptomyces strain of afsA, encoding the key A factor biosynthesis
contains its own -butyrolactone and its recep- enzyme,and arpA, encoding the A factor recep-
tor with strict ligand specificity, this system also tor, which implies that the mechanism of regu-
facilitates discrimination of signals originating lation by these -butyrolactones is the same as
from neighboring living things, thus allowing that for the A factor regulatory system includ-
the cell to recognize whether the neighbor is a ing afsA and arpA in S. griseus (see below).
member of the same species.This system is also As described above, afsA is located near one
advantageous to survival in the ecosystem; A end of the linear chromosome in S. griseus, and
factor produced by a cell is accepted by several arpA is located in the middle of the chromo-
hyphae and causes rapid sporulation of a whole some. Some afsA homologs are located near
population, which is advantageous compared arpA homologs, but others are not.The former
FIGURE 1 -Butyrolactones in Streptomyces.The differences in chemical structure among the -butyrolactones

are the length and branching of the acyl chain and the reduction state,either a keto or a hydroxyl group,at position 6.
366 ■ HORINOUCHI
includes, for example, the afsA/arpA homolo- (40). Once a Streptomyces strain acquired a -
gous system for production of virginiamycin butyrolactone as a chemical signaling molecule
by S. virginiae (30, 45), a pigment by S. venezue- during the evolution, the preexisting ArpA
lae (71), jadomycin B by S. venezuelae (79), ancestor employed it as a ligand to modulate its
and methylenomycin by S. coelicolor A3(2) own DNA-binding activity. Because the com-
(accession number AJ276673), and the latter bination of afsA and arpA in a given Streptomyces
includes the system for the production of strain is greatly different in topology of the
pristinamycin by S. pristinaespiralis (5) and phylogenetic tree (40), Streptomyces has changed
tylosin by S. fradiae (56). Apparently, the the combination of afsA and arpA when it
afsA/arpA system in the former case is specific acquired a new pair of afsA-arpA, during which
for the adjacent gene cluster for production of Streptomyces has selected the best-fit pair by trial
a certain secondary metabolite, and the system and error. This is a vivid contrast to the luxI-
in the latter case exerts pleiotropic effects on luxR systems that mediate quorum sensing
both secondary metabolism and morphological via N-acyl-homoserine lactones (3, 8). The
differentiation. inducer-receptor elements in a quorum-sensing
The A factor and receptor system in S.griseus system in various gram-negative bacteria have
acts as an all-or-nothing switch (i.e., a crucial evolved concomitantly, as revealed by phyloge-
switch) for both morphological and physiolog- netic analysis (33).
ical differentiation. CprA and CprB, both of
which are A factor receptor homologs, act as OVERALL PICTURE OF THE A
tuners for these processes in S. coelicolor A3(2) FACTOR REGULATORY CASCADE
(48). On the other hand, theVB system in S. vir- Figure 2 illustrates an overall picture of the A
giniae also controls the timing of antibiotic pro- factor regulatory cascade we so far have
duction but not morphological development. revealed (14, 42, 44). In S. griseus, A factor is
These observations imply that Streptomyces gradually accumulated in a growth-dependent
has evolved the -butyrolactone regulatory manner by the activity of AfsA that catalyzes
system to control different steps in the regula- -ketoacyl transfer between 8-methyl-
tory hierarchy for healthy growth, as an all- 3-oxononanoyl-acyl carrier protein (ACP)
or-nothing switch for some phenotypes and and dihydroxyacetone phosphate. When the
as just a tuner for other phenotypes.This may concentration of A factor reaches a critical
be why a Streptomyces strain contains redundant level at or near the middle of exponential
-butyrolactone regulatory systems. growth, it binds the A factor receptor protein
afsA encoding an A factor biosynthesis (ArpA), which has bound and repressed the
enzyme has been found only from Streptomyces promoter of a gene encoding a key transcrip-
(40), which is consistent with the idea that the tional factor (AdpA), and dissociates ArpA from
-butyrolactone regulatory cascades are con- the promoter, leading to transcription and
fined to Streptomyces. On the other hand, the translation of adpA.AdpA then activates a vari-
receptor protein ArpA and its homologs, espe- ety of genes of various functions that are
cially proteins having high similarity in DNA- required for secondary metabolism and mor-
binding domains, are distributed rather widely phological development. One of the targets of
among various bacteria.In addition,some Strep- AdpA is strR, encoding the pathway-specific
tomyces strains contain multiple afsA-arpA pairs. transcriptional factor for the streptomycin
Phylogenetic analysis of the -butyrolactone biosynthesis genes. Thus, the A factor signal is
synthases and receptors suggests that the ances- transferred to the streptomycin biosynthesis
tral ArpA protein had existed as a DNA-binding genes, starting with AfsA, then to ArpA to
protein, not as a -butyrolactone receptor, AdpA to StrR, and finally to all the strepto-
before the appearance of a -butyrolactone mycin biosynthetic genes, causing the onset of
receptor in the course of the bacterial evolution streptomycin biosynthesis (see below). Of the
FIGURE 2 The A factor regulatory cascade. The A factor signal, starting with the A factor biosynthesis gene afsA,
is transferred to the receptor ArpA, to a transcriptional activator AdpA, and finally to a variety of genes required for
morphological development and secondary metabolite formation. See the text for details of the target genes of
AdpA.Through this cascade, morphological and physiological differentiation occurs at a specific time of growth,
when the intracellular concentration reaches a critical level at or near the middle of the exponential growth.
major proteins in the A factor regulatory cas- tomycin production and sporulation. AfsA
cade, I will describe the functions of AfsA, appeared to be a key enzyme for A factor
ArpA, and AdpA. biosynthesis from primary metabolites present
commonly in bacterial cells because (i) afsA
mutants lost A factor productivity (10), (ii)
AfsA, a Key Enzyme for Biosynthesis
introduction of afsA into A factor nonpro-
of -Butyrolactones
ducing Streptomyces strains caused overproduc-
BIOSYNTHESIS OF A FACTOR tion of A factor with a gene dosage effect
By using the A factor-deficient mutant HH1 (19, 20), and (iii) introduction of afsA into
as the host, we cloned afsA that restored strep- Escherichia coli caused the host to produce a
368 ■ HORINOUCHI
substance having A factor activity (1). Con- [8]), which is then reduced by a reductase dif-
cerning the biosynthesis pathway, Sakuda ferent from BprA, resulting in A factor (Fig. 3
et al. (53) proposed a route on the basis of [1]). Because the introduction of afsA alone
feeding experiments in VB-producing Strep- into E. coli causes the host to produce a sub-
tomyces antibioticus. According to their pro- stance having A factor activity, the reductase
posal, the skeleton is formed by coupling and phosphatase are not specific to the A factor
between a 3-carbon (C3) unit derived from biosynthesis but commonly present in bacteria.
glycerol and a C10 -keto acid derivative. In addition, the afsA-bprA operon structure
Structure modeling of AfsA by S. Nakamura suggests that the former pathway is major for A
(personal communication) showed that, like factor biosynthesis.
-hydroxyacyl-ACP dehydratase, AfsA has a E. coli carrying afsA produces two new sub-
tunnel that can accept an acyl chain of acyl- stances that are absent in the broth of E. coli
ACP.The presence of such a tunnel in AfsA sup- without afsA with their m/z 241 and 213 and
ported the idea that AfsA is involved in the A the same MS/MS fragmentation pattern as A
factor biosynthesis at a step of condensation of a factor (1, 23). Because the culture broth shows
C3 compound and a C10 fatty acid derivative A factor activity and because E. coli does not
containing a -ketoacyl chain. produce branched fatty acids, we assume that
Since the fatty acids of Streptomyces con- these substances are A factor analogs with a C10
sist primarily of branched-chain fatty acids straight side chain (m/z 241) and a C8 straight
that are synthesized from isobutyryl- and side chain (m/z 213). AfsA is thus the key
methylbutyryl-coenzyme A (CoA) (72), we enzyme for the biosynthesis of -butyrolac-
supposed that the C10 branched side chain tones. Interestingly, a database search predicts
of A factor might be synthesized by conden- that afsA and its homologs are distributed only
sation of isobutyryl-CoA and three acetate in actinomycetes.
units. In fact, AfsA catalyzed acyl transfer
between 8-methyl-3-oxononanoyl-N-acetyl- STRUCTURAL VARIETY OF
cysteamine (NAC), a mimic of the correspon- -BUTYROLACTONES
ding -ketoacyl-ACP (Fig. 3 [3]) and There are two structural differences in the -
the hydroxyl group of dihydroxyacetone butyrolactone signal molecules in Streptomyces
phosphate (DHAP) (Fig. 3 [2]) (23). (Fig. 1); one is the length and branching of the
The fatty acid ester (Fig. 3 [4]) produced fatty acid side chain and the other is the reduc-
is nonenzymatically converted to a buteno- tion state of the 6-oxo group.The difference in
lide phosphate (Fig. 3 [5]) by intramole- side chain can be ascribed to the variety of the
cular aldol condensation. The butenolide -ketoacyl-ACPs used by AfsA and its
phosphate is then reduced by the bprA product homologs. SCB1 in S. coelicolor A3(2), having
that is encoded just downstream of afsA. The the same C10 branched side chain as A factor,
stereoselectivity, the R-form, at position 3 is must be formed by ScbA (AfsA ortholog) from
determined by this reduction step. The phos- the same -ketoacyl-ACP as for A factor
phate group on the resultant butanolide (Fig. 3 biosynthesis.The difference at position 6, either
[6]) is finally removed by a phosphatase, result- a keto or a hydroxyl group, can be ascribed to
ing in formation of A factor (Fig. 3 [1]).We also the existence and stereoselectivity of the reduc-
confirmed that the 8-methyl-3-oxononanoyl- tase that reduces this position. Because of the
DHAP ester produced (Fig. 3 [4]) by the action absence of such reductases in S. griseus, position
of AfsA is converted to A factor in an alternative 6 of A factor remains as a keto group. On the
way (23). The phosphate group on the ester other hand,VBs in S. virginiae and SCB1 in S.
(Fig.3 [4]) is first removed by a phosphatase and coelicolor A3(2) have a hydroxyl group at posi-
the dephosphorylated ester (Fig. 3 [7]) is con- tion 6. BarS1 in S. virginiae is an NADPH-
verted nonenzymatically to a butenolide (Fig. 3 dependent reductase that reduces the 6-oxo
FIGURE 3 The whole A factor biosynthesis pathway.The major pathway, highlighted by shadowing, and an alternative pathway are shown. In the major pathway,
AfsA catalyzes the condensation of DHAP [2] and a -keto acid derivative [3] to yield an 8-methyl-3-oxononanoyl-DHAP ester [4].The fatty acid ester is nonen-
zymatically converted to a butenolide phosphate [5] by intramolecular aldol condensation.The butenolide phosphate is then reduced by BprA to yield a butanolide
[6].The last dephosphorylation step yields A factor [1]. In the alternative pathway, the 8-methyl-3-oxononanoyl-DHAP ester [4] is first dephosphorylated to yield a
dephosphorylated ester [7], which is then nonenzymatically condensed, resulting in a butenolide [8].The CC double bond of the butenolide [8] is reduced to yield
A factor [1].
370 ■ HORINOUCHI
group of the penultimate intermediate in VB -butyrolactones as a positive factor in an

biosynthesis (55). autorepressive manner.
REGULATION OF A FACTOR
ArpA, the A Factor Receptor Protein
BIOSYNTHESIS
A factor is accumulated in a growth-dependent PROPERTIES OF ArpA
manner and reaches a maximum, 25 to 30 Miyake et al. (35, 36) detected A factor-binding
ng/ml (~100 nM) at or near the middle of activity in the cytoplasmic fraction of S. griseus
the exponential growth (2, 23).The transcrip- by using [3H]A factor and determined its
tion of afsA, starting at two points, is almost repressor function as to streptomycin produc-
constant throughout growth (23). Then how tion and aerial mycelium formation by genetic
is A factor biosynthesis in such an extremely studies. An A factor receptor gene, named arpA,
small amount controlled during growth? It is was then cloned,and the DNA-binding activity
most likely that the time course of A factor of ArpA and a consensus ArpA-binding
production at this extremely low concentration sequence were determined (46, 47). The con-
is due to the availability of the substrates for sensus ArpA-binding sequence was a 22-bp
AfsA. 8-Methyl-3-oxononanoyl-ACP, which palindromic site with the sequence 5′-
is synthesized by condensation of three acetate GG(T/C)CGGT(A/T)(T/C)G(T/G)-3′ as
units with the starter substrate isobutyryl-CoA, one-half of the palindrome (47). ArpA binds
is an intermediate in the primary fatty acid this site in the absence of A factor, and the
biosynthesis and is leaked from the pathway. exogenous addition of A factor to the ArpA-
Therefore, the intracellular pool of the -keto DNA complex induces immediate release of
acid derivatives must be extremely small. ArpA from the DNA.These observations led to
Both glycolysis and fatty acid synthesis are the idea that ArpA acts as a repressor-type regu-
active during exponential growth, and this lator for secondary metabolism and morpho-
is reflected on the time course of A factor logical differentiation by preventing the
production. Biosynthesis of -butyrolactones expression of a certain key gene(s) during the
reflective of the physiological conditions early growth phase. A factor, produced in a
of Streptomyces is analogous to that of N-acyl- growth-dependent manner, releases ArpA from
homoserine lactones in gram-negative the DNA, thus switching on the expression of
bacteria; N-acyl-homoserine lactones are the key genes,leading to the simultaneous onset
biosynthesized from S-adenosylmethionine of secondary metabolism and morphogenesis at
derived from the amino acid biosynthesis path- a certain timing during growth. Kato et al. (24)
way and the diverse intermediates in the fatty later identified adpA as the sole target of ArpA.
acid biosynthesis (37). Site-directed mutagenesis of the helix-turn-
A vivid contrast in the regulation of - helix DNA-binding motif of ArpA yielded a
butyrolactone production between the A mutant ArpA protein (Val41Ala) that lacks
factor regulatory cascade and the other - DNA-binding ability but still retains A factor-
butyrolactone regulatory systems is that the binding ability (57). Conversely, mutant
receptors,BarA forVB (30),FarA for IM-2 (31), Pro115Ser lacks A factor-binding ability but
and ScbR for SCBs (61), are required for pro- retains DNA-binding ability (49). Mutant
duction of the respective -butyrolactones, Trp119Ala also lacks A factor-binding ability
whereas ArpA for A factor does not affect A but retains DNA-binding ability,indicating that
factor production (25). In S. griseus, A factor Trp-119 is essential for A factor-binding (57).
production is controlled directly or indirectly These findings show that ArpA contains two
by adpA in a two-step regulatory feedback independently functional domains, a DNA-
loop (25). In the case of BarA, FarA, and ScbR, binding domain and an A factor-binding
these receptors control the production of the domain.
CRYSTALLOGRAPHY OF CprB, HOW A FACTOR DISSOCIATES

AN ArpA HOMOLOG ArpA FROM DNA
Our repeated attempts to crystallize ArpA A database search for structural comparison
failed, because it readily aggregated. We then revealed that the overall structure of CprB is
tried to crystallize CprB, an ArpA homolog in similar to those of the TetR family proteins
S. coelicolor A3(2) (48). CprB, consisting of 215 TetR and QacR. One of the crystal structures
amino acids, shows about 30% identity in of CprB, form Ib, is closely related to that of
amino acid sequence to ArpA. Although the QacR in complex with its target DNA (54).We
ligand of CprB is still unknown, it recognizes can hence predict how -butyrolactones disso-
and binds the same nucleotide sequence as ciate their cognate receptors from DNA upon
ArpA (57). In addition, CprB serves as a nega- binding the ligands, on the basis of the mecha-
tive regulator for both morphological differen- nism of the conformational changes of TetR
tiation and secondary metabolism in S. coelicolor upon tetracycline binding (50, 54). Ligand
A3(2), as ArpA does in S. griseus. The crystal binding induces the relocation of a long helix
structures of three different forms, Ia, Ib, and II, 4 that links the ligand-binding pocket with
were determined at 2.4 Å resolution (39), and the DNA-binding domain. As a result of the
they turned out to be a dimer with an “Ω” relocation of the DNA-binding domain,ArpA
shape (Color Plate 13). The two subunits are dissociates from the DNA (Color Plate 13).
bound via six hydrogen bonds and three water-
mediated hydrogen bonds. In addition, a disul-
AdpA, a Key Transcriptional Activator
fide bond via Cys-159 links the subunits.This
in the A Factor Regulatory Cascade
disulfide bridge is specific to CprB because typ-
ical -butyrolactone receptors contain no Cys PROPERTIES OF AdpA
residue at this position. The DNA-binding Vujaklija et al. (69, 70) studied the transcrip-
domain is composed of three N-terminal tional organization of part of the streptomycin
helices, 1, 2, and 3. Of the three, 2 and 3 biosynthesis gene cluster and found that one
form a typical helix-turn-helix motif. Consis- mRNA species covering a regulatory gene
tent with the observation that -butyrolactone (strR) and the streptomycin-6-phosphotrans-
receptors recognize the same nucleotide ferase (aphD) gene is dependent on A factor.
sequence, the residues on helix 3 are com- Ohnishi et al. (43) identified an A factor-
pletely conserved among the receptors. responsive protein (AdpA) able to bind the
A large cavity is present in the regulatory upstream activation sequence, about 270 bp
domain, which we assume is a ligand-binding upstream of the transcriptional start point of
pocket 5 Å in diameter and 20 Å long.Trp-127, strR.AdpA encoding a 405-amino-acid protein
corresponding to Trp-119 of ArpA, which has with a helix-turn-helix DNA-binding motif at
been found to be essential for A factor binding the central portion shows sequence similarity
by site-directed mutagenesis (57), participates to transcriptional regulators belonging to the
in forming the pocket. This pocket is com- AraC/XylS family.The 35 and 10 regions
pletely embedded in the molecule, and a of adpA contained a 22-bp palindrome,
flexible loop covers the entrance to it,serving as c a g g c A G G A AC G G AC C * G C G C G -
a lid for the pocket.The docking study suggests GTACGCt (the underlines indicate the 35
that a -butyrolactone molecule binds to the and 10 promoter elements; * indicates a dyad
pocket in an extended manner and Trp-127 axis), which showed similarity to the consen-
causes a hydrophobic interaction with the alkyl sus sequence of the ArpA-binding site,
chain of the -butyrolactone molecule. The (A/C)C(A/G)(T/A)ACCC(A/G)CC*GG(T/
hydrophobic interaction between Trp-127 and C)CGGT(A/T)(T/C)G(T/G). As expected,
the alkyl chain of the ligand stabilizes the lig- ArpA bound the promoter region of adpA
and binding. in the absence of A factor but did not in the
372 ■ HORINOUCHI
presence of A factor. In addition, exogenous tion and serving as a fine sensor of the intracel-
addition of A factor to the ArpA-DNA com- lular AdpA concentration.
plex immediately dissociated ArpA from the
DNA.Thus, the promoter of adpA turned out AdpA REGULON
to be a target of ArpA. Consistent with this, S1 Yamazaki et al. (75) collected more than 60
nuclease mapping showed that adpA is tran- DNA fragments that were specifically bound
scribed only in the presence of A factor and strR by AdpA by gel mobility shift assay in com-
is transcribed only in the presence of intact bination of immunoprecipitation and PCR.
adpA. Furthermore, adpA disruptants produce The presence of many genes, all of which are
no streptomycin and overexpression of adpA simultaneously activated by AdpA at a specific
causes the wild-type S. griseus strain to produce point in the growth phase, means that the signal
streptomycin at an earlier growth stage in a from A factor is greatly amplified at this regula-
larger amount (43). tory step via AdpA as an amplifier (Fig. 1).The
targets of AdpA required for morphological
adpA AS A SINGLE TARGET OF ArpA differentiation are adsA,encoding an extracyto-
Because ArpA acts as a repressor for aerial plasmic function (ECF) factor (75); amfR,
mycelium formation and secondary metabo- encoding a transcriptional factor that activates
lism, an arpA-disruptant forms aerial hyphae the amf operon (66, 77); extracellular proteases
and spores earlier than the wild-type strain and including a metalloendopeptidase (26), two
overproduces streptomycin and other second- trypsin-type proteases (22), and three chy-
ary metabolites. On the other hand, mutant motrypsin-type proteases (64); a Streptomyces
KM2, expressing a mutant ArpA (Trp119Ala), subtilisin inhibitor (SSI) gene (13); and ssgA
neither produces secondary metabolites nor essential for spore septum formation (76).The
forms aerial hyphae, since this A factor-insensi- A factor-dependent proteases in conjunction
tive mutant ArpA always binds to and represses with the SSI are supposed to control aerial
the adpA promoter.Trp-119 of ArpA is essential mycelium formation (13, 29), since Streptomyces
for A factor binding,and replacement of this Trp forms hyphae by reuse of material assimilated in
residue with Ala abolishes its A factor-binding substrate hyphae by apoptosis or cannibalism.
ability, resulting in the formation of a mutant The amf operon is for production of AmfS, a
ArpA that binds the target DNA irrespective of hydrophobin that is essential for the erection of
the presence of A factor (57).When adpA under aerial hyphae into the air (32, 67). For second-
the control of a foreign,constitutively expressed ary metabolism,AdpA indirectly activates griR,
promoter is introduced into mutant KM2, all a pathway-specific transcriptional activator for
the phenotypes that we can observe are restored grixazone production (12), a gene encoding a
(24).We can hence conclude that the only sig- transcriptional factor probably for biosynthesis
nificant target of ArpA is adpA. of a polyketide compound (78), in addition to
strR for streptomycin production (see below).
AUTOREPRESSION OF adpA Alignment of more than 10 AdpA-binding
The intracellular concentration of AdpA must sequences, which were determined by DNase I
be important for ordered gene expression for footprinting, did not predict an apparent con-
healthy growth. Consistent with this idea, sensus sequence for AdpA binding.Yamazaki et
AdpA represses its own transcription by form- al. (78) performed interference assays on several
ing a DNA loop via two molecules of AdpA AdpA-binding sites for determination of the
dimer that bind the operator sites in the adpA nucleotides directly associating with amino
promoter (25). Thus, AdpA self-controls the acids of AdpA and deduced a consensus AdpA-
intracellular concentration at an appropriate binding sequence, 5′-TGGCSNGWWY-3′ (S:
level by cooperative binding to the two opera- G or C;W:A or T;Y:T or C; N: any nucleotide).
tor sites, allowing effective regulation to result The AdpA-binding sites so far identified all
from small alterations in the AdpA concentra- contain a sequence similar to this consensus
sequence. Among the consensus sequence, the by the action of AfsA in a growth-dependent
fourth C is essential for AdpA binding, and manner binds ArpA that has bound and
replacement of this C residue with other repressed the promoter of adpA in an early
nucleotides results in abolishment of AdpA growth stage.When the concentration of A fac-
binding. In addition to the consensus sequence tor reaches a critical level, it binds the DNA-
of 10 nucleotides, eight additional nucleotides bound ArpA and dissociates ArpA from the
3′ to this sequence also contribute to the affin- DNA, thus causing transcription of adpA.Two
ity of AdpA,although there is no sequence con- AdpA dimers then bind the upstream activation
servation in this region. sequences of strR, approximately at nucleotide
For activation of target genes, a dimer of positions 270 and 50 with respect to the
AdpA binds various positions, for example, transcriptional start point of strR, and activate
more than 200 bp upstream and 25 bp down- its transcription (65). AdpA assists RNA poly-
stream of their transcriptional start points (78). merase in forming an open complex at an
In addition,AdpA binds a single site for activa- appropriate position for transcriptional initia-
tion of some genes and two or three sites for tion of strR, as determined by potassium per-
others. For transcriptional activation, some manganate footprinting. Simultaneous binding
genes require simultaneous binding of a dimer of AdpA to the two sites is necessary for full
of AdpA to multiple sites (78). Despite the induction of strR transcription, the reason of
differences in binding position and number which is still unknown. The pathway-specific
of binding sites, AdpA recruits RNA poly- transcriptional activator StrR induces tran-
merase to the specific promoter region of scription of all the streptomycin biosynthetic
the target genes and facilitates isomerization of genes by binding multiple sites in the gene clus-
the RNA polymerase-DNA complex into ter (52), thus leading to biosynthesis of strepto-
an open complex for transcriptional initiation mycin from glucose. The major streptomycin
(65, 78), as demonstrated for other transcrip- resistance determinant, aphD, located just
tional activators. downstream of strR, encoding streptomycin-6-
Our preliminary DNA microarray analy- phosphotransferase, is also transcribed by read-
sis on the basis of the S. griseus genome through from the A factor-dependent strR
sequence has shown that more than 600 genes promoter (70).The cotranscription of strR and
are activated by AdpA and about 200 genes are aphD accounts for the prompt induction of
relatively down-regulated in the adpA back- streptomycin resistance by A factor and
ground. It is amazing that such a large number achieves a rapid increase in self-resistance just
of genes are switched on or affected by AdpA before induction of streptomycin biosynthesis.
(actually A factor itself ). Especially, all gene
clusters for certain secondary metabolites in a CONCLUDING REMARKS
total of about 30, with an exception of only In addition to -butyrolactones, possible
one, are activated by AdpA (unpublished data). autoregulatory factors in Streptomyces species
Further study will reveal a more comprehensive that are involved in secondary metabolism and
picture of the A factor regulator cascade. morphological differentiation have been
reported, although most of them are not pre-
HOW A FACTOR TRIGGERS cisely defined as an autoregulator due to lack
STREPTOMYCIN PRODUCTION of genetic studies (17, 18). The following
strR, encoding the pathway-specific transcrip- two compounds can be termed a regulator.
tional activator for all the streptomycin biosyn- B-factor (3′-butylphosphoryl AMP) at a con-
thesis genes, is a member of the AdpA regulon. centration of a few nanomoles controls
The signal relay from A factor to the strepto- rifamycin biosynthesis in Nocardia mediterranei
mycin biosynthetic genes is from afsA to arpA (27). PI factor [pimaricin inducer; 2,3-
to adpA to strR to the streptomycin production diamino-2,3-bis(hydroxymethyl)-1,4-butane-
genes. As already described, A factor produced diol] at a few hundred nanomoler induces the
374 ■ HORINOUCHI
production of a macrolide,pimaricin,in Strepto- 2. Ando, N., K. Ueda, and S. Horinouchi. 1997.

myces natalensis (51). It is conceivable that dur- A Streptomyces griseus gene (sgaA) suppresses the
growth disturbance caused by high osmolality and
ing evolution Streptomyces spp. have employed a high concentration of A factor during early
various chemical substances as autoregulators growth. Microbiology 143:2715–2723.
to control their own secondary metabolism and 3. Bassler, B. L., and R. Losick. 2006. Bacterially
morphogenesis. The autoregulators must have speaking. Cell 125:237–246.
evolved concomitantly with their specific 4. Chater, K. F., and S. Horinouchi. 2003. Sig-
nalling early developmental events in two highly
receptors. Future studies may reveal chemical diverged Streptomyces species. Mol. Microbiol.
substances as novel autoregulators, together 48:9–15.
with their specific receptors. 5. Folcher, M., H. Gaillard, L. T. Nguyen,
Elucidation of the regulation by -butyro- K. T. Nguyen, P. Lacroix, N. Bamas-Jacques,
lactones is important not only for biology of M. Rinkel, and C. J. Thompson. 2001.
Pleiotropic functions of a Streptomyces pristinaespi-
the filamentous bacterial genus Streptomyces but ralis autoregulator receptor in development,antibi-
also for practical use of this genus for produc- otic biosynthesis, and expression of a superoxide
tion of useful secondary metabolites. The dismutase. J. Biol. Chem. 276:44297–44306.
sequencing of the S. griseus genome has been 6. Funa, N., M. Funabashi, Y. Ohnishi, and
completed in this laboratory (Y. Ohnishi et al., S. Horinouchi. 2005. Biosynthesis of hexahy-
droxyperylenequinone melanin via oxidative aryl
unpublished data). Comprehensive study by coupling by cytochrome P-450 in Streptomyces
employing DNA microarray techniques, pro- griseus. J. Bacteriol. 187:8149–8155.
teomics, and metabolomics will enable us to 7. Funa, N., Y. Ohnishi, I. Fujii, M. Shibuya,
depict an overall, precise picture of the A factor Y. Ebizuka, and S. Horinouchi. 1999. A new
regulatory cascade. Comparative genomics by pathway for polyketide synthesis in microorgan-
isms. Nature 400:897–899.
use of the genome sequences of S. coelicolor 8. Fuqua, W. C., S. C. Winans, and E. P.
A3(2) and S. avermitilis helps us to understand Greenberg. 1994. Quorum sensing in
the biology and chemistry of S. griseus and to bacteria: the LuxR-LuxI family of cell density-
mine and polish the treasure trove in this bacte- responsive transcriptional regulators. J. Bacteriol.
rial genus. Even between S. griseus and S. coeli- 176:269–275.
9. Hara, O., and T. Beppu. 1982. Mutants blocked
color A3(2), signaling pathways for early in streptomycin production in Streptomyces griseus—
developmental events and genes for secondary the role of A factor. J.Antibiot. 35: 349–358.
metabolite formation are different (4). Some 10. Hara, O., S. Horinouchi, T. Uozumi, and
genes for secondary metabolite formation are T. Beppu. 1983. Genetic analysis of A factor syn-
species-specific, but many genes are different thesis in Streptomyces coelicolor A3(2) and Strepto-
myces griseus. J. Gen. Microbiol. 129:2939–2944.
from strain to strain. Mining and polishing 11. Hashimoto, K., T. Nihira, S. Sakuda, and Y.
useful genes for practical purposes are impor- Yamada. 1992.IM-2,a butyrolactone autoregula-
tant, and at the same time, awakening of “sleep- tor, induces production of several nucleoside
ing” genes by various approaches is also antibiotics in Streptomyces sp. FRI-5. J. Ferment. Bio-
important. The genome projects for actino- eng. 73:449–455.
12. Higashi, T., Y. Iwasaki, Y. Ohnishi, and
mycetes, such as Actinoplanes and Kitasatospora, S. Horinouchi. 2007. A factor and phosphate-
showing more complex morphology than depletion signals are transmitted to the grixazone
Streptomyces, which are going on in Japan will biosynthesis genes via the pathway-specific tran-
further contribute to our understanding of the scriptional activator GriR. J. Bacteriol. 189:
boundary microorganisms. 3515–3524.
13. Hirano, S., J. Kato, Y. Ohnishi, and S.
Horinouchi. 2006. Control of the Streptomyces
REFERENCES subtilisin inhibitor gene by AdpA in the A factor
1. Ando, N., N. Matsumori, S. Sakuda, regulatory cascade in Streptomyces griseus. J. Bacte-
T. Beppu, and S. Horinouchi. 1997. Involve- riol. 188:6207–6216.
ment of AfsA in A factor biosynthesis as a key 14. Horinouchi, S. 2003.AfsR as an integrator of sig-
enzyme. J.Antibiot. 50:847–852. nals that are sensed by multiple serine/threonine
kinases in Streptomyces coelicolor A3(2). J. Ind. Micro- 27. Kawaguchi, T., M. Azuma, S. Horinouchi,
biol. Biotechnol. 30:462–467. and T. Beppu. 1988. Effect of B-factor and its
15. Horinouchi, S. 2007. Mining and polishing analogues on rifamycin biosynthesis in a Nocardia
of the treasure trove in the bacterial genus sp. J.Antibiot. 41:360–365.
Streptomyces. Biosci. Biotechnol. Biochem. 71: 28. Khokhlov,A. S., I. I.Tovarova, L. N. Borisova,
283–299. S. A. Pliner, L. A. Schevchenko, E. Y.
16. Horinouchi, S., and T. Beppu. 1990.Autoregu- Kornitskaya, N. S. Ivkina, and I.A. Rapoport.
latory factors of secondary metabolism and mor- 1967. A factor responsible for the biosynthesis of
phogenesis in actinomycetes. Crit. Rev. Biotechnol. streptomycin by a mutant strain of Actinomyces
10:191–204. streptomycini. Dokl. Akad. Nauk. SSSR 177:
17. Horinouchi, S., and T. Beppu. 1992.Autoregu- 232–235.
latory factors and communication in actino- 29. Kim, D.-W., K. Chater, K.-J. Lee, and A.
mycetes. Annu. Rev. Microbiol. 46:377–398. Hesketh. 2005. Changes in the extracellular
18. Horinouchi, S., and T. Beppu. 1994. A fac- proteome caused by the absence of the bldA gene
tor as a microbial hormone that controls cellular product, a developmentally significant tRNA,
differentiation and secondary metabolism in reveal a new target for the pleiotropic regulator
Streptomyces griseus. Mol. Microbiol. 12:859–864. AdpA in Streptomyces coelicolor. J. Bacteriol.
19. Horinouchi, S., Y. Kumada, and T. Beppu. 187:2957–2966.
1984. Unstable genetic determinant of A factor 30. Kinoshita, H., H. Ipposi, S. Okamoto,
biosynthesis in streptomycin-producing organ- H. Nakano, T. Nihira, and Y. Yamada. 1997.
isms: cloning and characterization. J. Bacteriol. Butyrolactone autoregulator receptor protein
158:481–487. (BarA) as a transcriptional regulator in Streptomyces
20. Horinouchi, S., H. Suzuki, M. Nishiyama, virginiae. J. Bacteriol. 179:6986–6993.
and T. Beppu. 1989. Nucleotide sequence and 31. Kitani, S., H. Kinoshita, T. Nihira, and
transcriptional analysis of the Streptomyces griseus Y. Yamada. 1999. In vitro analysis of the
gene (afsA) responsible for A factor biosynthesis. J. butyrolactone autoregulator protein (FarA) of
Bacteriol. 171:1206–1210. Streptomyces lavendulae FRI-5 reveals that FarA
21. Kang, D.-K., X.-M. Li, K. Ochi, and S. acts as a DNA-binding transcriptional regulator
Horinouchi. 1999. Possible involvement of that controls its own synthesis. J. Bacteriol.
cAMP in aerial mycelium formation and second- 181:5081–5084.
ary metabolism in Streptomyces griseus. Microbiology 32. Kodani, S., M. E. Hudson, M. C. Durrant,
145: 1161–1172. M. J. Buttner, J. R. Nodwell, and J. M.Willey.
22. Kato, J., W.-J. Chi, Y. Ohnishi, S.-K. Hong, 2004. The SapB morphogen is a lantibiotic-like
and S. Horinouchi. 2005. Transcriptional con- peptide derived from the product of the develop-
trol by A factor of two trypsin genes in Streptomyces mental gene ramS in Streptomyces coelicolor. Proc.
griseus. J. Bacteriol. 187:286–295. Natl.Acad. Sci. USA 101:11448–11453.
23. Kato, J., N. Funa, H.Watanabe,Y. Ohnishi, and 33. Lerat, E., and N. A. Moran. 2004. The evolu-
S. Horinouchi. 2007. Biosynthesis of g-butyro- tionary history of quorum-sensing systems in bac-
lactone autoregulators that switch on secondary teria. Mol. Biol. Evol. 21:903–913.
metabolism and morphological development in 34. Lezhava, A., D. Kameoka, H. Sugino, K.
Streptomyces. Proc. Natl. Acad. Sci. USA 104: Goshi, H. Shinkawa, O. Nimi, S. Horinouchi,
2378–2383. T. Beppu, and H. Kinashi. 1997. Chromosomal
24. Kato, J., I. Miyahisa, M. Mashiko,Y. Ohnishi, deletions in Streptomyces griseus that remove the
and S. Horinouchi. 2004.A single target is suffi- afsA locus. Mol. Gen. Genet. 253: 478–483.
cient to account for the biological effects of the A 35. Miyake, K., S. Horinouchi, M. Yoshida,
factor receptor protein of Streptomyces griseus.J.Bac- N. Chiba, K. Mori, N. Nogawa, N. Morikawa,
teriol. 186:2206–2211. and T. Beppu. 1989. Detection and properties of
25. Kato, J.,Y. Ohnishi, and S. Horinouchi. 2005. A factor-binding protein from Streptomyces griseus.
Autorepression of AdpA of the AraC/XylS family, J. Bacteriol. 171:4298–4302.
a key transcriptional activator in the A factor regu- 36. Miyake, K.,T. Kuzuyama, S. Horinouchi, and
latory cascade in Streptomyces griseus. J. Mol. Biol. T. Beppu. 1990.The A factor-binding protein of
350:12–26. Streptomyces griseus negatively controls strepto-
26. Kato, J., A. Suzuki, H.Yamazaki,Y. Ohnishi, mycin production and sporulation. J. Bacteriol.
and S. Horinouchi. 2002. Control by A factor of 172:3003–3008.
a metalloendopeptidase gene involved in aerial 37. Moré, M. I., L. D. Finger, J. L. Stryker,
mycelium formation in Streptomyces griseus. J. Bacte- C. Fuqua, A. Eberhard, and S. C. Winans.
riol. 184:6016–6025. 1996. Enzymatic synthesis of a quorum-sensing
376 ■ HORINOUCHI
autoinducer through use of defined substrates. Sci- factor receptor protein of Streptomyces griseus abol-
ence 272:1655–1658. ishes DNA-binding ability but not ligand-binding
38. Mori, K. 1983. Revision of the absolute ability. J. Bacteriol. 179:2748–2752.
configuration of A factor, the inducer of strepto- 50. Orth, P., D. Schnappinger, W. Hillen, W.
mycin biosynthesis, basing on the reconfirmed Saenger, and W. Hinrichs. 2000. Structural basis
(R)-configuration of ( )-paraconic acid. Tetrahe- of gene regulation by the tetracycline inducible Tet
dron 39:3107–3109. repressor-operator system. Nat. Struct. Biol. 7:
39. Natsume, R., Y. Ohnishi, T. Senda, and S. 215–219.
Horinouchi. 2004. Crystal structure of a -buty- 51. Recio, E., A. Colinas, A. Rumbero, J. F.
rolactone autoregulator receptor protein in Strepto- Aparicio, and J. F. Martin. 2004. PI factor, a
myces coelicolor A3(2). J. Mol. Biol. 336:409–419. novel type quorum-sensing inducer elicits
40. Nishida, H., Y. Ohnishi, T. Beppu, and S. pimaricin production in Streptomyces natalensis. J.
Horinouchi. 2007. Evolution of -butyrolactone Biol. Chem. 279:41586–41593.
synthases and receptors in Streptomyces. Environ. 52. Retzlaff, L., and J. Distler. 1995.The regulator
Microbiol. 9:1986–1994. of streptomycin gene expression, StrR, of Strepto-
41. Ohnishi, Y., Y. Furusho, T. Higashi, H.-K. myces griseus is a DNA binding activator protein
Chun, K. Furihata, S. Sakuda, and S. Hori- with multiple recognition sites. Mol. Microbiol.
nouchi. 2004. Structures of grixazone A and B,A 18:151–162.
factor-dependent yellow pigments produced 53. Sakuda, S., S. Tanaka, K. Mizuno, O.
under phosphate depletion by Streptomyces griseus. Sukcharoen, T. Nihira, and Y. Yamada. 1993.
J.Antibiot. 57:218–223. Biosynthetic studies on virginiae butanolide A, a
42. Ohnishi, Y., and S. Horinouchi. 2004. The butyrolactone autoregulator from Streptomyces. Part
A factor regulatory cascade that leads to morpho- 2. Preparation of possible biosynthetic intermedi-
logical development and secondary metabolism in ates and conversion experiments in a cell-free sys-
Streptomyces. Biofilms 1:319–328. tem. J. Chem. Soc., Perkin Trans. 1 1993:2309–2315.
43. Ohnishi, Y., S. Kameyama, H. Onaka, and 54. Schumacher, M. A., M. C. Miller, S. Grkovic,
S. Horinouchi. 1999.The A factor regulatory cas- M. H. Brown, R. A. Skurray, and R. G. Bren-
cade leading to streptomycin biosynthesis in Strep- nan. 2001. Structural mechanisms of QacR
tomyces griseus: identification of a target gene of the induction and multidrug recognition. Science
A factor receptor. Mol. Microbiol. 34:102–111. 294:2158–2163.
44. Ohnishi,Y., H.Yamazaki, J. Kato,A.Tomono, 55. Shikura, N., J.Yamamura, and T. Nihira. 2002.
and S. Horinouchi. 2005. AdpA, a central tran- barS1, a gene for biosynthesis of a -butyrolactone
scriptional regulator in the A factor regulatory cas- autoregulator, a microbial signaling molecule elic-
cade that leads to morphological development and iting antibiotic production in Streptomyces species.
secondary metabolism in Streptomyces griseus.Biosci. J. Bacteriol. 184:5151–5157.
Biotechnol. Biochem. 69:431–439. 56. Stratigopoulos, G., A. R. Gandecha, and
45. Okamoto, S., K. Nakamura, T. Nihira, and E. Cundliffe. 2002.Regulation of tylosin produc-
Y. Yamada. 1995. Virginiae butanolide binding tion and morphological differentiation in Strepto-
protein from Streptomyces virginiae: evidence that myces fradiae by TylR, a deduced -butyrolactone
VbrA is not the virginiae butanolide binding pro- receptor. Mol. Microbiol. 45:735–744.
tein and reidentification of the true binding pro- 57. Sugiyama, M., H. Onaka, T. Nakagawa, and
tein. J. Biol. Chem. 270:12319–12326. S. Horinouchi. 1998. Site-directed mutagenesis
46. Onaka, H., N. Ando, T. Nihira,Y.Yamada, T. of the A factor receptor protein:Val-41 important
Beppu, and S. Horinouchi. 1995. Cloning and for DNA-binding and Trp-119 important for
characterization of the A factor receptor gene from ligand-binding. Gene 222:133–144.
Streptomyces griseus. J. Bacteriol. 177:6083–6092. 58. Süsstrunk, U., J. Pidoux, S. Taubert, A. Ull-
47. Onaka, H., and S. Horinouchi. 1997. DNA- mann, and C. J. Thompson. 1998. Pleiotropic
binding activity of the A factor receptor protein effects of cAMP on germination, antibiotic
and its recognition DNA sequences.Mol.Microbiol. biosynthesis and morphological development in
24:991–1000. Streptomyces coelicolor. Mol. Microbiol. 30:33–46.
48. Onaka, H.,T. Nakagawa, and S. Horinouchi. 59. Suzuki, H.,Y. Ohnishi,Y. Furusho, S. Sakuda,
1998. Involvement of two A factor recep- and S. Horinouchi. 2006. Novel benzene ring
tor homologues in Streptomyces coelicolor A3(2) in biosynthesis from C3 and C4 primary metabolites
the regulation of secondary metabolism and mor- by two enzymes. J. Biol. Chem. 281:36944–36951.
phogenesis. Mol. Microbiol. 28:743–753. 60. Takano, E. 2006. g-Butyrolactones: Streptomyces
49. Onaka, H., M. Sugiyama, and S. Horinouchi. signaling molecules regulating antibiotic produc-
1997. A mutation at proline-115 in the A tion and differentiation.Curr.Opin.Microbiol.9:1–8.
61. Takano, E., R. Chakaraburtty, T. Nihira, mycin biosynthetic gene cluster of Streptomyces
Y.Yamada, and M. J. Bibb. 2001.A complex role griseus. Mol. Gen. Genet. 229:119–128.
for the g-butyrolactone SCB1 in regulating antibi- 71. Waki, M., T. Nihira, and Y. Yamada. 1997.
otic production in Streptomyces coelicolor A3(2). Cloning and characterization of the gene (farA)
Mol. Microbiol. 41:1015–1028. encoding the receptor for an extracellular regula-
62. Takano, E., T. Nihira, Y. Hara, J. J. Jones, tory factor (IM-2) from Streptomyces sp. strain FRI-
C. J. L. Gershater, Y. Yamada, and M. Bibb. 5. J. Bacteriol. 179:5131–5137.
2000. Purification and structure determination of 72. Wallace, K. K., B. Zhao, H.A. McArthur, and
SCB1,a g-butyrolactone that elicits antibiotic pro- K.A. Reynolds. 1995. In vivo analysis of straight-
duction in Streptomyces coelicolor A3(2). J. Biol. chain and branched-chain fatty acid biosynthesis
Chem. 275:11010–11016. in three actinomycetes. FEMS Microbiol. Lett.
63. Tanaka,A.,Y.Takano,Y. Ohnishi, and S. Hori- 131:227–234.
nouchi. 2007. AfsR recruits RNA polymerase to 73. Yamada,Y., and T. Nihira. 1998. Microbial hor-
the afsS promoter: a model for transcriptional acti- mones and microbial chemical ecology, p. 377–
vation by SARPs. J. Mol. Biol. 369:322–333. 413. In D. H. R. Barton and K. Nakanishi (ed.),
64. Tomono, A., Y. Tsai, Y. Ohnishi, and S. Comprehensive Natural Products Chemistry, vol. 8.
Horinouchi. 2005.Three chymotrypsin genes are Elsevier Science, Heidelberg, Germany.
members of the AdpA regulon in the A factor reg- 74. Yamada, Y., K. Sugamura, K. Kondo,
ulatory cascade in Streptomyces griseus. J. Bacteriol. M. Yanagimoto, and H. Okada. 1987. The
187:6341–6353. structure of inducing factors for virginiamycin
65. Tomono,A.,Y.Tsai, H.Yamazaki,Y. Ohnishi, production in Streptomyces virginiae. J. Antibiot.
and S. Horinouchi. 2005. Transcriptional con- 40:496–504.
trol by A factor of strR, the pathway-specific tran- 75. Yamazaki, H.,Y. Ohnishi, and S. Horinouchi.
scriptional activator for streptomycin biosynthesis 2000. An A factor-dependent extracytoplasmic
in Streptomyces griseus. J. Bacteriol. 187:5595–5604. function sigma factor (AdsA) that is essential for
66. Ueda, K., K. Miyake, S. Horinouchi, and morphological development in Streptomyces griseus.
T. Beppu. 1993.A gene cluster involved in aerial J. Bacteriol. 182:4596–4605.
mycelium formation in Streptomyces griseus encodes 76. Yamazaki, H.,Y. Ohnishi, and S. Horinouchi.
proteins similar to the response regulators of two- 2003.Transcriptional switch on of ssgA by A factor,
component regulatory systems and membrane which is essential for spore septum formation in
translocators. J. Bacteriol. 175:2006–2016. Streptomyces griseus. J. Bacteriol. 185:1273–1283.
67. Ueda, K., K. Oinuma, G. Ikeda, K. Hosono, 77. Yamazaki, H., Y. Takano, Y. Ohnishi, and
Y. Ohnishi, S. Horinouchi, and T. Beppu. S. Horinouchi. 2003. amfR, an essential gene for
2002. AmfS, an extracellular peptidic morphogen aerial mycelium formation, is a member of the
in Streptomyces griseus. J. Bacteriol. 184:1488–1492. AdpA regulon in the A factor regulatory cascade in
68. Umeyama, T., P.-C. Lee, and S. Horinouchi. Streptomyces griseus. Mol. Microbiol. 50:1173–1187.
2002. Protein serine/threonine kinases in signal 78. Yamazaki, H., A. Tomono, Y. Ohnishi, and
transduction for secondary metabolism and mor- S. Horinouchi. 2004. DNA-binding specificity
phogenesis in Streptomyces. Appl. Microbiol. Biotech- of AdpA, a transcriptional activator in the A factor
nol. 59:419–425. regulatory cascade in Streptomyces griseus. Mol.
69. Vujaklija, D., S. Horinouchi, and T. Beppu. Microbiol. 53:555–572.
1993. Detection of an A factor-responsive pro- 79. Yang, K., L. Han, and L. C. Vining. 1995.
tein that binds to the upstream activation sequence Regulation of jadomycin B production in Strepto-
of strR, a regulatory gene for streptomycin myces venezuelae ISP5230: involvement of a recep-
biosynthesis in Streptomyces griseus. J. Bacteriol. tor gene, jadR2. J. Bacteriol. 177:6111–6117.
175:2652–2661. 80. Yonekawa,T.,Y. Ohnishi, and S. Horinouchi.
70. Vujaklija, D., K. Ueda, S.-K. Hong,T. Beppu, 2005. A calmodulin-like protein in the bacterial
and S. Horinouchi. 1991. Identification of genus Streptomyces. FEMS Microbiol. Lett. 244:
an A factor-dependent promoter in the strepto- 315–321.
QUORUM QUENCHING: IMPACT
AND MECHANISMS
Lian-Hui Wang,Yi-Hu Dong, and Lian-Hui Zhang
24
Antibiotics,which act either by killing bacterial We must now view the single-celled bac-
pathogens or by inhibiting their growth, are terial pathogens as concerted communities,
commonly used to treat microbial infections. capable of synchronizing gene expression
Pathogens, however, can adapt to the selective among local members and coordinating their
pressure from antibiotics via genetic alteration, biological activities by sensing the population
leading to development of antibiotic resistance. density. This cell-cell communication mecha-
With the increase in bacterial resistance to mul- nism is known as quorum sensing.Various ver-
tiple antibiotics, it is becoming progressively sions of quorum-sensing mechanisms have
more difficult to treat infections with tradi- been adopted by different bacterial pathogens
tional antibiotics. Hence, there is a growing to modulate infection-associated activities,
urgency to search for novel targets and new including the production of virulence factors
ways to control and eradicate bacterial infec- and the formation of biofilms.These quorum-
tions.The discovery of pathogenic bacteria that sensing bacteria release, detect, and respond to
use quorum sensing to coordinate production the accumulation of small signal molecules, also
of virulence factors and to establish infection referred to as autoinducers, and hence coordi-
may offer a promising alternative for the devel- nate the expression of target genes in a cell
opment of novel therapeutics. Moreover, quo- density-dependent manner.There are two gen-
rum sensing is not only important for medically erally conserved quorum-sensing mechanisms:
important bacterial pathogens but also essential one is based on a signal ligand and its cognate
for community genetic regulation in animal receptor-like transcription factor and used pri-
pathogens, plant pathogens, and environmental marily by gram-negative bacteria; the other is
microorganisms. Exploration and exploitation based on a signal ligand and its corresponding
of bacterial quorum sensing have hence two-component sensor/response regulator,
become a common interest of various sectors, used primarily by gram-positive bacteria.
including medicine, agriculture, and environ- The first mechanism, represented by the
mental and biotechnology industries. AHL/LuxR system of Vibrio fischeri, has four
core elements, including acyl-homoserine lac-
Lian-Hui Wang,Yi-Hu Dong, and Lian-Hui Zhang Institute of
Molecular and Cell Biology, 61 Biopolis Drive, Singapore tone (AHL) signal, the signal synthase LuxI, the
138673. signal receptor LuxR, and the “lux box”—the
379
380 ■ WANG ET AL.
LuxR-binding motif on the promoters of benefit in interacting with their host organisms.
the target genes (Fig. 1).At low cell population Given that the virulence factor production by
density, each bacterial cell produces and releases each individual cell of many bacterial pathogens
a basal level of AHL signals. Upon reaching a is dependent on quorum-sensing regulation,
high level of density as bacterial cells proliferate, disruption of bacterial quorum sensing is
the accumulated AHL signals form functional expected to provide two obvious advantages to
LuxR-AHL complexes, which in turn boost host organisms:(i) reduced accumulation of vir-
AHL levels by autoinducing luxI expression ulence factors at the infection site, and (ii) the
and activate the expression of quorum-sensing- collective power of pathogens being disman-
regulated genes (70). The two-component- tled.The past decade has witnessed not only the
system-based quorum-sensing mechanism, rapid progress in our understanding of bacterial
represented by the AIP/ArgCA of Staphylococ- quorum-sensing systems but also the amazing
cus aureus, consists of AIP autoinducing peptide diversity of quorum-sensing interference
signal, produced by AgrD, and the sensor kinase mechanisms, collectively known as quorum
ArgC with its cognate response regulator, quenching. Unlike the conventional antibiotics
ArgA. The general mechanism is that the AIP used to kill bacteria or inhibit their growth,
signals are detected by its sensor ArgC. Upon quorum quenching acts by disrupting pathogen
AIP binding to ArgC,ArgC transfers the signal quorum-sensing systems, curtailing the bacter-
input via phosphorelay to ArgA, which ial potential for infection, and allowing the dis-
enhances the transcription of the arg operon armed pathogen to be cleared by the host’s own
for autoregulation and activates the target defense mechanisms. As a consequence, it is
genes through downstream regulatory ele- probable that the selective pressure for the
ments (45). Bacteria may use only one of these emergence of resistance will be lessened. In this
signaling systems but could also employ a com- chapter, we focus on the impact and molecular
bination of these systems for regulation of vari- mechanisms of quorum quenching, with
ous biological functions. For example, two emphasis on quorum-quenching enzymes and
AHL/LuxR systems are found in a series circuit other signal disruption mechanisms.
for the regulation of a number of virulence
gene expressions in Pseudomonas aeruginosa, QUORUM QUENCHING—IMPACT
whereas in Vibrio harveyi the AHL/LuxR and AND MODE OF ACTION
two-component systems are used in parallel for There are five key processes in quorum-sensing
modulating similar functions (70). circuit, i.e., (i) signal generation, (ii) signal trans-
Quorum sensing enables bacterial pathogens portation,(iii)signal accumulation,(iv) signal re-
to use their collective power to gain maximal cognition, and (v) signal autoinduction (Fig. 1).
FIGURE 1 Schematic representation of the

AHL-dependent quorum-sensing system in
gram-negative bacteria. The critical processes
in quorum-sensing communication that could
be targeted by quorum-quenching approaches
are indicated. I, the LuxI-type AHL synthase;
R, the LuxR-type transcription factor; trian-
gle,AHL signals.
24. QUORUM QUENCHING: IMPACT AND MECHANISMS ■ 381
The rapid progress in the search for potential signal is able to diffuse passively across bacterial
quorum-quenching mechanisms has signifi- membranes,whereas the long-chain AHL signal
cantly enriched our understanding of the design 3-O-C12-HSL efflux appears to rely on active
and development of high-potency interference transportation mechanisms. Several nonspecific
molecules but also of nature’s ability to evolve multidrug efflux pumps, including MexAB-
various forms of antagonizing mechanisms. It OprM (52), MexEF-OprN (29), and Mex-
appears that nature may have ready counteract- GHI-OpmD (1), are implicated in efflux of
ing strategies to most, if not all, of these key 3-O-C12-HSL signals. For example, mutation
processes of quorum sensing. of the MexGHI-OpmD efflux pump drastically
reduces the production of AHL signals and vir-
Blocking Signal Generation ulence factors (1). More recently, null mutation
Given the key role of signals in bacterial quo- of the efflux pump BpeAB-OprB in another
rum sensing, blocking signal production human bacterial pathogen, Burkholderia pseudo-
appears to be one of the most straightforward mallei, results in significant reduction of AHL
quorum-quenching strategies. A few potent signal accumulation in culture medium (6).
inhibitors that target various enzymes in the While the efficacy on inhibition of quorum-
signal biosynthesis pathway have been identi- sensing communication remains to be deter-
fied. Synthesis of AHL-type quorum-sensing mined, several groups of chemicals, including
signals requires substrates S-adenosyl methio- pyridopyrimidine and quinazolinone deriva-
nine,a particular fatty acid along with its carrier tives, were found to inhibit ABC-type efflux
protein, and a LuxI-type AHL synthase pumps (43, 72).
(41, 57). It was reported that triclosan can
inhibit enoyl-acyl carrier protein reductase, Preventing Signal Recognition
whose product is the essential intermediate Receptor antagonism with signal analogues is a
in AHL biosynthesis (20). An analogue of classic pharmacological approach. Competitive
S-adenosyl methionine has been proven to binding of a nonfunctional signal analogue to
be a potent inhibitor of RhlI, which is one the receptor may either block its signal-binding
of the two AHL synthases encoded by P. aerugi- site, preventing signal recognition, or cause
nosa. L-S-adenosylhomocysteine was found to accelerated receptor degradation due to drastic
lower the activity of RhlI by 97% (50). Methyl- changes in protein folding. The halogenated
anthranilate, an analogue of the Pseudomonas furanones, which are a structural mimic of AHL
quinolone signal (PQS) precursor anthranilate, signals, block quorum sensing by reducing the
inhibits the activity of anthranilate synthase half-life of LuxR receptor protein (36). Simi-
and hence interferes with PQS signal biosyn- larly, patulin, identified from Penicillium species,
thesis (3). Very recently, DADMe-Immucillins was found accelerating LuxR turnover,interfer-
was reported to block autoinducer-2 synthe- ing with quorum-sensing communication (54).
sis in Streptococcus pneumoniae by inhibiting In gram-positive bacteria, the truncated AIP-II
the activity of 5-methylthioadenosine/S- and AIP analogues block quorum-sensing sig-
adenosylhomocysteine hydrolase, which is the naling by interfering with AIP binding to the
only reaction known to generate ribosyl- receptor, which is a two-component-type
homocysteine, the LuxS substrate, and the pre- membrane-bound sensor kinase (35).
cursor of autoinducer-2 (60). In addition to the structural mimicry of sig-
nals, other nonrelated chemical inhibitors,
Disturbing Signal Exchange which can induce conformational changes in
Active transportation of quorum-sensing sig- receptor proteins,may also be potentially potent
nals has been proven important for some bacte- receptor inhibitors. Triphenyl compounds
rial pathogens. In the case of P. aeruginosa inhibit quorum-sensing signaling in P.aeruginosa
quorum-sensing system, the short-chain AHL by interacting with the long-chain AHL
382 ■ WANG ET AL.
receptor LasR at the AHL-binding site (42). Inactivating Quorum Sensing Signals
Several phenolic inhibitors, including RWJ- Quorum-sensing signal is the core of any bacte-
49815 and closantel, cause protein aggregation rial quorum-sensing system. In addition, some
of the receptor kinase KinA of Bacillus subtilis, quorum-sensing signals such as the 3-O-C12-
possibly by changing the structural conforma- HSL could even act directly as virulence factors
tions of the receptor protein (63). Given the through modulation of host defense systems
generally conserved structural features of sensor (62, 64). Four groups of enzymes, which
kinases, these phenolic chemicals may prove to degrade or modify the AHL-type quorum-
be broad-spectrum inhibitors. It was reported sensing signals, have been identified in the past
recently that closantel inhibits the autokinase few years (Fig. 2), including AHL-lactonase (10,
activities of three two-component sensor 12), AHL-acylase (30, 32, 55), paraoxonases
kinases from Ehrlichia chaffeensis and blocks the (PONs) (16, 46, 71), and AHL-oxoreductase
infection of host cells by the pathogen (7). (68). Significantly, these AHL-inactivating
enzymes, also known as quorum-quenching
Signal Trapping enzymes (13), are widely distributed in both
Two types of signal-trapping strategies have prokaryotic and eukaryotic kingdoms, suggest-
been tested recently. The first one is based ing their roles in microbe-microbe interactions
on signal-specific antibodies. The 3-O- and in host defense systems. These quorum-
C12-HSL signal produced by P. aeruginosa will at quenching enzymes have been explored and
low concentrations stimulate production evaluated as novel antimicrobial agents against
of immunoglobin IgG1 and IgE (65). This different pathogens. Expression of the AHL-
immunomodulating property can be improved lactonase encoded by the aiiA gene from Bacillus
by hapten design to generate AHL-specific sp. in transgenic potato and tobacco plants con-
monoclonal antibodies, which can then be fers strong resistance to bacterial pathogen
employed for trapping quorum-sensing signals. Erwinia carotovora, which requires AHL quo-
Treatment of the bacterial pathogen with rum-sensing signals for activating the expression
the 3-O-C12-HSL-specific antibody decreases of virulence genes (11). Similarly, natural or
pyocyanin, a cytotoxin, by up to 50% (26). In recombinant AHL-lactonase-producing bacter-
another similar study, the effect of active immu- ial strains, including Bacillus thuringiensis,
nization with carrier-conjugated 3-O-C12- Arthrobacter sp., and Pseudomonas fluorescens, pro-
HSL on the lethality of acute P. aeruginosa lung tect potato from E. carotovora infection when
infection in mice was investigated (38). The coinoculated with the pathogen (14, 39, 49).
result showed that all nonimmunized mice died In addition to the enzymes that degrade
by day 2 post bacterial challenge, whereas 36% AHLs, nature also seems to have mechanisms
of immunized mice survived to day 4, suggest- for other types of quorum-sensing signals.The
ing that specific antibody against 3-O-C12-HSL human NADPH oxidase was reported to gen-
provides a survival benefit in mice infected with erate reactive oxygen and nitrogen intermedi-
P. aeruginosa.The second strategy is to use poly- ates, including HOCl and ONOO-, which
mer cyclodextrins to mimic a receptor in order inactivate the AIP signals of Staphylococcus aureus,
to trap quorum-sensing signals (25).The results and thus providing protection from bacterial
show that addition of 10 mM 2-hydroxypropyl- infections in a mouse air pouch skin model
-cyclodextrin partially reduces the AHL- (56). The finding is consistent with previous
dependent production of pigment prodigiosin studies that humans and mice lacking NADPH
by Serratia marcescens. However, the potential oxidase are more susceptible to staphylococcal
difficulty of this signal-trapping strategy is that a infections (17, 24, 53). Structural analysis of the
high number of antibodies or polymers may be inactivated AIP signals showed that oxidation of
required to cope with the constantly produced the C-terminal methionine is primarily respon-
bacterial quorum-sensing signals. sible for loss of signaling activity (56) (Fig. 2).
FIGURE 2 Schematic representation of enzymatic reactions of various quorum-sensing signal degradation and
modification enzymes.(A) AHL signal inactivation,where R represents either 3-oxo substituent or absence of sub-
stitution. (B) AIP signal inactivation by the reactive oxygen or nitrogen intermediates generated by NADPH oxi-
dase complex.
MOLECULAR MECHANISMS OF sequence features, as well as the information

QUORUM-QUENCHING ENZYMES about their protein crystal structures, have sig-
Since the first reports of two AHL-inactivation nificantly facilitated our understanding of the
enzymes in 2000 (12, 30), numerous quorum- catalytic mechanisms and structural features
quenching enzymes have been identified that affect enzyme specificity and activity.
(Table 1). Some of them, although functionally
similar, share little sequence similarity; for AHL-Lactonase
example, both AHL-lactonases and PONs Since the first reported quorum-quenching
hydrolyze the lactone ring of AHL signals enzyme, encoded by the aiiA gene from a Bacil-
(13). Even among the members of the AHL- lus isolate 240B1 (hereafter referred to as
lactonase family, sequence homology could be AiiA240B1), was characterized to be an AHL-
very low. Nevertheless, as with many other lactonase that hydrolyzes the ester bond of the
enzymes and proteins, quorum-quenching homoserine lactone ring of AHL molecules
enzymes contain certain invariable signature (11, 12), its homologues have now been identi-
residues, which often serve as useful clues fied in a range of bacterial species, including
for probing the enzymatic mechanisms. Such at least four gram-positive Bacillus species and
384 ■ WANG ET AL.
TABLE 1 List of AHL quorum-quenching enzymes and antibody

Class Gene or chemicala Species Reference(s)
AHL-lactonase aiiA Bacillus sp. 240B1 12
aiiA homologues Bacillus thuringiensis 10, 31
aiiA homologues Bacillus cereus 10, 55
aiiA homologues Bacillus mycoides 10
aiiA homologues Bacillus anthracis 67
attM, aiiB Agrobacterium tumefaciens 5, 75
ahlD Arthrobacter sp. IBN110 49
ahlK Klebsiella pneumoniae 49
NA Rhodococcus spp. 47
Paraoxonases PON1, 2, 3 Mammalian species 8, 46, 71
AHL acylase NA Variovorax paradoxusVAI-C 30
aiiD Ralstonia strain XJ12B 21, 32
pvdQ Pseudomonas strain PAI-A 22, 61
P. aeruginosa PAO1
ahlM Streptomyces sp. 48
quiP P. aeruginosa PAO1 23
AHL-oxoreductase NA Rhodococcus erythropolis 68
mAb RS2-1G9 Monoclonal antibody raised by an AHL 26
structural mimic 38
Vaccine 3-O-C12-HSL conjugated to a protein
carrier
a
NA, information is not available.
gram-negative Agrobacterium tumefaciens, Kleb- crystal structure of AHL-lactonase from B.

siella pneumoniae, and Arthrobacter sp. (Table 1). thuringiensis subsp. kurstaki (hereafter referred to
Based on homology, these AHL-lactonases can as AiiABTK) by two independent research
be grouped into two clusters with overall groups has further enriched our understanding
homology at about 30% between clusters, of the enzyme architecture and catalytic mech-
which may suggest different origins of evolu- anism (27, 33). The crystal structure analysis
tion.The first one is the AiiA cluster with mem- shows that the enzyme contains two zinc ions
bers from Bacillus species sharing about 90% in the active site (27, 33), which agrees with the
sequence homology at the amino acid level. biochemical analysis that AHL-lactonase is a
The second one is the AttM cluster consisting metalloprotein (66). The structural analysis
of the enzymes from gram-negative bacterial shows that the two zinc ions are coordinated to
species sharing only about 25% peptide homol- several residues, including His104, His106,Asp108,
ogy (14).AiiA240B1 is one of the most character- His109, His169, and His235, in addition to an oxy-
ized enzymes in the first group. It has a gen ion of a bridging carboxylate from Asp191
250-residue-long peptide with a “His104-X- and a bridging water/hydroxide ion.These data
His106-X-Asp108-His109” zinc-binding motif from AiiABTK are highly consistent with the
that is conserved in several metallohydrolases previous mutagenesis study on AiiA240B1 (10).
including glyoxalase II, arylsulfatase, and - The only inconsistency is that substitution of
lactamase, and was proposed to be a member of His104 with serine was initially found nonessen-
the metallohydrolase superfamily (12). Site- tial for the AiiA240B1 activity (12) but critical for
directed mutagenesis based on sequence- AiiABTK, based on structural analysis and by
alignment of the AHL-lactonase homologues substitution with alanine (27). We have con-
showed the conserved motif “His106-X-Asp108- firmed recently that replacement of His104 with
His109-59X-His169-21X-Asp191” is essential for alanine in AiiA240B1 almost completely abol-
enzyme activity (10, 12). Elucidation of the ishes the enzyme activity (unpublished data).
By sequence alignment, we found that all the homologies of about 93% and 90% with
residues implicated in metal coordination are AiiA240B1 and AiiABTK, respectively. After sub-
conserved in the reported AHL-lactonases. stitution of Tyr194 with alanine, the enzyme
In spite of limited sequence similarity,AHL- activity was significantly reduced compared
lactonase may share the same catalytic mecha- with wild-type recombinant AiiAY2.This result
nism of binuclear metal-binding glyoxalase II suggests that the conserved residue Tyr194 is
and RNase Z. Crystal structure analysis of critical for catalytic activity of AHL-lactonase,
AHL-lactonase has revealed an / sand- in agreement with the postulated catalytic
wich fold in overall structure with a canonical mechanism.
dinuclear Zn2 center in their active sites, AHL-lactonase is by far the most specific
located in a loop-rich region on top of the AHL-degradation enzyme among known
/ fold (27, 33). These structural features quorum-quenching enzymes. It hydrolyzes
are remarkably similar to those of the metallo- both short- and long-chain AHL signals but
-lactamase superfamily, including glyoxalase II shows no or little residue activity to other
(4) and RNase Z proteins (9). More recently, chemicals including non-acyl lactones and aro-
the dinuclear metal site of AHL lactonase from matic carboxylic acid esters (69). However,
B. thuringiensis was proved to be essential both there are certain differences in the substrate
for structural stabilization and for catalysis (40). specificity between two well-characterized
Furthermore, isotopic labeling studies using AHL-lactonase homologues. Unlike AiiA240B1,
18
O and 2H demonstrate that the AHL lactonase which had no acyl chain length preference and
involves an addition-elimination pathway for exhibited a broad catalytic spectrum, AiiABTK
ring opening, in which a solvent-derived oxy- revealed an acyl chain length preference and
gen is incorporated into the product carboxy- maximum activity with C10-AHL. Further
late,identifying the alcohol as the leaving group determination of the crystal structure of AHL-
(40). A catalytic mechanism of AHL-lactonase lactonase/AHL complex would be essential to
has been proposed on the basis of its crystal elucidate its intriguing substrate specificity.
structures in the presence and absence of L- The AHL-lactonases in AttM cluster from
homoserine lactone and relevant experimental gram-negative bacteria,including A.tumefaciens,
evidence (27, 40). In this model, the enzyme Arthrobacter sp., Klebsiella pneumoniae, and
reacts with AHL signal through direct interac- Rhodococcus sp. (47, 49, 75), are less characterized
tion between the two zinc ions and the lactone compared with those in AiiA cluster.The overall
ring and carbonyl oxygens of AHL; the interac- peptide identities among these AHL-lactonases
tion results in enhanced polarization of the car- are in range of 21 to 26%, and less than 35% in
bonyl bond, making it more susceptible to comparison with AiiA240B1.Although the iden-
the nucleophilic attack by a nucleophilic tities were relatively low, the above-described
water/hydroxide, which bridges the two Zn2 “His-X-Asp-His~His~Asp” motif is well con-
ions; the nucleophilic attack on the carbonyl served in all these AHL-lactonases, suggesting
carbon of the substrate leads to formation of a that the AHL-lactonases from gram-negative
negatively charged intermediate, which may be bacteria could share the same catalytic mecha-
stabilized primarily by the interactions with nism as the AHL-lactonases in AiiA cluster.
Zn1 ion, and then an alcohol leaving group is
eliminated. In this process the Tyr194 of the PON Enzymes
enzyme may act as a general acid for protona- The serum PON family enzymes have been
tion of the leaving group. identified in mammals, other vertebrates, and
The catalytic role of Tyr194 was verified invertebrates. PONs, including PON1, PON2,
recently by site-directed mutagenesis of an and PON3, exhibit a wide range of physio-
AHL-lactonase from marine Bacillus cereus logically important hydrolytic activities,includ-
strain Y2 (34). This enzyme showed high ing drug metabolism and organophosphate
386 ■ WANG ET AL.
detoxification (15, 44). Interestingly, the PONs degradation of the C–O bond of the ester
of human and other mammals have been also intermediate.The negative charge of the inter-
reported to have lactonase-like enzyme activity, mediate generated in step (ii) is probably stabi-
capable of hydrolyzing the homoserine lactone lized by the catalytic calcium.
ring of AHL signals.The purified recombinant PONs are well known for their broad-
human PON2 efficiently hydrolyzes several spectrum enzyme activities. The most-charac-
tested AHL compounds (16).The recombinant terized PON1 shows at least three types of
animal CHO cells expressing mouse PON1, enzyme activities, i.e., organophosphatase,
PON2, and PON3, respectively, display strong arylesterase, and lactonase, and hydrolyzes a
AHL degradation activity (71).The hydrolytic wide range of substrates (2, 16). The chemical
activity of the PON1 from human against P. structures of these substrates are so different that
aeruginosa 3-O-C12-HSL signal has also been one may wonder whether the enzyme uses the
demonstrated (46), which explains the previ- same mechanism in catalysis. However, the
ously observed AHL inactivation activity in crystal structure analysis has revealed only one
human epithelial cells (8).These PON enzymes active site, and the essential role of the newly
seem to be most active with long-chain AHL identified His115-His134 dyad in enzyme cataly-
signals, such as 3-O-C12-HSL, but less effi- sis has been confirmed by site-directed mutage-
cient with short-chain AHL signals (8, 71). In nesis and assay using two classes of substrates,
contrast to bacterial AHL-lactonase, PON including phenyl acetate and paraoxon (18).
enzymes do not contain the typical “His-X- The above results suggest that PON1 may also
Asp-His~His~Asp” motif required for AHL- rely on the same catalytic mechanisms in
lactonase enzyme activity, indicating that hydrolysis of AHL signals.
the two quorum-quenching enzymes likely The three PON family members, i.e.,
use different catalytic mechanisms for AHL- PON1, PON2, and PON3, share about 60%
degradation. homology at the peptide level. By sequence
The crystal structure of a PON1 variant alignment, we found that the secondary struc-
(designated as rePON1), obtained by directed ture features and all the catalytically important
evolution, has been solved recently (18). The residues identified in PON1, including Glu53,
results show that rePON1 is a six-bladed - His115, His134, Asn168, Asn224, Asp269, and Asn270,
propeller with two Ca2 ions in its central tun- are also perfectly conserved in PON2 and
nel. One Ca2 ion lies at the bottom of the PON3, indicating that these enzymes most
active site and is postulated to play a role in likely share the same catalytic mechanism.These
catalysis, while the other calcium ion is largely three enzymes share overlapping substrates but
buried and appears to have a structural func- also display distinct substrate specificities (16), as
tion. The putative catalytic Ca2 ion seems to well as different catalytic efficiencies against
interact with five amino acid residues, i.e., AHL signals (16, 71). Future investigations on
Asn224,Asn270,Asn168,Asp269, and Glu53, and one these enzymes may reveal the intriguing mech-
water molecule and one of the oxygens of a anisms that influence the substrate specificity
phosphate ion. Based on the structural similar- and enzyme activity.
ity to secreted phospholipase A2 (58), and the
pH-rate profiles of rePON1 using 2-naththyl AHL-Acylase
acetate and paraoxon as substrates, a general- AHL-acylases, which inactivate AHL signals by
base catalytic model has been proposed (18). In breaking the amide bond of AHLs to produce
this model, the catalysis involves three critical corresponding fatty acids and homoserine lac-
steps, i.e., (i) deprotonating a water molecule by tone (Fig. 2), may also be widely conserved.
the His115-His134 dyad to generate a hydroxide Several bacterial species, including Variovorax
anion, then (ii) generating an oxyanionic inter- paradoxus, a Ralstonia isolate, P. aeruginosa PAO1,
mediate by nucleophilic attack at the ester car- a Streptomyces sp., and Rhodococcus erythropolis
bonyl center of the substrates, followed by (iii) W2, have been reported to encode AHL-acy-
lases (22, 23, 30, 32, 48, 61, 68).The four identi- the AHL signals with the corresponding substi-
fied AHL-acylases, i.e., AiiD from Ralstonia sp. tution to generate corresponding 3-hydroxy
XJ12B (32), PvdQ and QuiP from P. aeruginosa derivatives. Depending on the specificity of
PAO1 (22, 23, 61), and AhlM from Streptomyces AHL receptors, this modification may or may
sp.(48),share relatively low homologies ranging not affect the signaling activity of AHLs.
from 20 to 35%. As expected, there are also Although the TraR-based quorum system of A.
notable differences in the substrate specificities tumefaciens responds more or less equally to 3-
among AHL-acylases.AiiD effectively degrades oxo-substituted AHL signals and their 3-
long-chain AHLs and also short-chain AHLs, hydroxy derivatives (76), the LasR system of P.
albeit less effectively (32). AhlM appears to be aeruginosa requires about sevenfold higher con-
more effective in degrading AHLs with an acyl centration of 3-hydroxy derivatives than its 3-
chain longer than six carbons,and is particularly oxo counterparts to reach a same level of
more active against unsubstituted rather than 3- activation (51).
oxo-substituted AHLs (48). PvdQ is unable to The AHL oxidoreductase activity was first
degrade AHLs with acyl chains shorter than observed from R. erythropolis, which reduces
eight carbons (22, 61). Similarly, QuiP is only AHLs with 3-oxo substituents to their corre-
able to degrade AHLs with side chains of seven sponding 3-hydroxy derivatives but exhibited
or more carbons in length (23). Furthermore, relatively poor activity against the unsubstituted
AiiD fails to degrade penicillin G and ampi- AHLs (68). This enzyme does not seem to be
cillin (32), whereas AhlM is able to catalyze the specific for only naturally occurring 3-oxo-
hydrolysis of penicillin G, suggesting a broad AHLs, since it can reduce compounds such as
substrate specificity. N-(3-oxo-6-phenylhexanoyl)-homoserine
These AHL-acylases are structurally similar lactone and 3-oxododecanamide. Further-
to the cephalosporin acylase from Pseudomonas more, it can also modify the D-isomers of
diminuta (32, 48). Crystal structure analysis of N-(3-oxododecanoyl)-L-homoserine lactones
cephalosporin acylase reveals a side-chain bind- (68). Although the gene encoding the AHL-
ing pocket, in which the residues Gln50 and oxidoreductase has not yet been cloned,
Arg57 have been proposed as the key compo- it is interesting to note that an oxidoreduc-
nents determining the substrate specificity tase has been purified and characterized from
(28). Interestingly, sequence alignment of the R. erythropolis (73, 74). This enzyme accepts a
four AHL-acylases with cephalosporin acylase broad range of aliphatic and aromatic ketones,
shows that these four acylases have different keto esters, and halogenated carbonyl com-
residues in the two corresponding positions pounds as substrates and, for example, reduces
(Ile50 and Ser57 in AiiD; Leu50 and Asp57 in methyl 3-oxobutanoate and ethyl 4-chloro-3-
PvdQ;Ile50 andVal57 in QuiP;Leu50 and Ser57 in oxobutanoate to the corresponding 3-hydroxy
AhlM). Further mutagenesis and crystal struc- compounds. However, this carbonyl reductase
ture analysis of these AHL-acylases would be purified from cell-free extract does not seem to
critical for elucidating the molecular mecha- like the reported AHL oxidoreductase, whose
nisms implicated in substrate specificity and activity is associated with the whole cell but is
catalysis. not in cell crude extract. Further work is
required to identify the AHL-oxidoreductase
AHL-Oxidoreductase gene from R. erythropolis for determination of
In contrast to the above three groups of the substrate specificity and catalytic mecha-
enzymes, i.e., AHL-lactonases, PONs, and nisms of the encoded enzyme at the purified
AHL-acylases, which degrade AHL signals by protein level.
breaking the bond in either the lactone ring or
in the junction connecting fatty acid moiety NADH-Oxidase
and homoserine lactone component, AHL- The phagocyte NADPH oxidase is a critical
oxidoreductase modifies the 3-oxo group of component of innate immunity, responsible for
388 ■ WANG ET AL.
generation of microbicidal reactive oxygen CONCLUSIONS AND FUTURE

species, which participate in host defense by PERSPECTIVES
killing or damaging invading microbes. The The biological importance of quorum-sensing
NADPH oxidase is a multicomponent enzyme communication in bacterial pathogens and fair
system, composed of membrane proteins understanding of the general molecular mecha-
(gp91phox, p22phox, and the small G-protein nisms have significantly propelled our effort in
Rap1A) and cytosolic proteins (p47phox, searching for effective quorum-quenching
p67phox, p40phox, and the small G-proteins Rac2 mechanisms. The past decade has witnessed
and Cdc42) and other components (where tremendous progress in this field with numer-
phox denotes phagocyte oxidase) (59). The ous previously unknown mechanisms and
membrane-bound subunits gp91phox and quorum-quenching molecules being identi-
p22phox together form the heterodimeric fied. Due to space limitations and the general
cytochrome b558. Upon oxidase activation, the scheme of this book, this chapter has listed only
cytosolic subunits p47phox, p67phox, and p40phox a portion of reported quorum-quenching mol-
are translocated to the plasma membrane and ecules as representatives to illustrate the modes
bind to the cytochrome b558 complex. Addi- of action, the molecular mechanisms, and the
tionally, the small GTPase proteins Rac2, potentials of various quorum-quenching
Cdc42,and Rap1A are involved in the assembly strategies. It is delightful to realize that nature
and activation of the NADPH oxidase.This sys- has generously evolved various mechanisms
tem is arranged vectorially in the phagosome against the critical processes of bacterial quo-
membrane so that electrons pass through it rum sensing. While the potentials of most
from the NADPH-oxidizing site to the O2- quorum-quenching molecules have been eval-
reducing site, resulting in the production of uated under in vitro conditions, several in vivo
superoxide anion and, consequently, the gener- tests have been conducted with certain potent
ation of reactive oxygen species such as H2O2, quorum-quenching molecules such as expres-
ONOO–, and HOCl. sion of AHL-lactonase in transgenic plants
The phagocyte NADPH oxidase represents and application of furanone compound in
another dimension of quorum-quenching mouse pulmonary infection model (11, 19).
mechanisms. Rather than interacting directly The outcome of these tests highlights the
with quorum-sensing signals,this enzyme inac- promising potential of quorum-quenching
tivates AIP signals through its enzymatic prod- strategies in the control and prevention of plant
ucts—the above-mentioned reactive oxygen and human pathogens. It is expected that more
and nitrogen intermediates (56). The inactiva- molecules of different mechanisms or a combi-
tion of the AIP-I signal of S. aureus is caused by nation of mechanisms will be evaluated in vivo
oxidation of the C-terminal methionine of the in our effort to develop therapeutic quorum-
signal and can be prevented by addition of oxi- quenching drugs and to increase the host resist-
dant scavenger N-acetyl methionine.The struc- ance to bacterial pathogens. In addition, the
tural analysis shows that the oxidants produced available knowledge on structure-activity rela-
by NADPH oxidase oxidize the methionine tionship and corresponding molecular mecha-
sulfanyl group of the signal to the correspon- nisms of quorum-quenching molecules may
ding sulfoxide form, causing the loss of AIP significantly facilitate development of superac-
activity (Fig. 2).The results agree well with the tive quorum-sensing inhibitors. For example,
previous finding that the methionine sulfoxide substitution of the original zinc ion in the cat-
form of this AIP does not have quorum-sensing alytic center of AHL-lactonase with Cd2
signaling activity (37). It is likely that oxidation results in a superenzyme that can hydrolyze
may alter either the conformation or polarity of AHLs drastically faster than the parental
AIP-I, which prevents its proper interaction enzyme (40).Furthermore,it is amazing to note
with the bacterial receptor AgrC. the quorum-quenching mechanisms exist
widely in microorganisms as well as in the host 10. Dong, Y. H., A. R. Gusti, Q. Zhang, J. L.
organisms of bacterial pathogens. Although it Xu, and L. H. Zhang. 2002. Identification of
quorum-quenching N-acyl homoserine lac-
is debatable as to whether they have been tonases from Bacillus species. Appl. Environ. Micro-
evolved specifically for counteracting bacterial biol. 68:1754–1759.
quorum sensing, the impact and influence of 11. Dong, Y. H., L. H. Wang, J. L. Xu, H. B.
these quorum-quenching molecules on micro- Zhang, X. F. Zhang, and L. H. Zhang. 2001.
bial ecology and pathogen-host interactions Quenching quorum sensing-dependent bacterial
should be closely monitored in the future. infection by an N-acyl homoserine lactonase.
Nature 411:813–817.
12. Dong, Y. H., J. L. Xu, X. Z. Li, and L. H.
REFERENCES Zhang. 2000.AiiA, an enzyme that inactivates the
1. Aendekerk, S., B. Ghysels, P. Cornelis, and C. acylhomoserine lactone quorum-sensing signal
Baysse. 2002. Characterization of a new efflux and attenuates the virulence of Erwinia carotovora.
pump, MexGHI-OpmD, from Pseudomonas aerugi- Proc. Natl.Acad. Sci. USA 97:3526–3531.
nosa that confers resistance to vanadium.Microbiology 13. Dong,Y. H., and L. H. Zhang. 2005. Quorum
148:2371–2381. sensing and quorum-quenching enzymes. J. Micro-
2. Billecke, S., D. Draganov, R. Counsell, P. biol. 43:101–109.
Stetson, C. Watson, C. Hsu, and B. N. La Du. 14. Dong,Y. H., X. F. Zhang, J. L. Xu, and L. H.
2000. Human serum paraoxonase (PON1) Zhang. 2004. Insecticidal Bacillus thuringiensis
isozymes Q and R hydrolyze lactones and cyclic silences Erwinia carotovora virulence by a new form
carbonate esters. Drug Metab. Dispos. 28:1335–1342. of microbial antagonism—signal interference.
3. Calfee, M. W., J. P. Coleman, and E. C. Pesci. Appl. Environ. Microbiol. 70:954–960.
2001. Interference with Pseudomonas quinolone sig- 15. Draganov, D. I., and B. N. La Du. 2004. Phar-
nal synthesis inhibits virulence factor expression by macogenetics of paraoxonases: a brief review.
Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA Naunyn Schmiedebergs Arch. Pharmacol. 369:78–88.
98:11633–11637. 16. Draganov, D. I., J. F. Teiber, A. Speelman, Y.
4. Cameron, A. D., M. Ridderstrom, B. Olin, Osawa, R. Sunahara, and B. N. La Du. 2005.
and B. Mannervik. 1999. Crystal structure of Human paraoxonases (PON1, PON2, and PON3)
human glyoxalase II and its complex with a glu- are lactonases with overlapping and distinct sub-
tathione thiolester substrate analogue. Structure strate specificities. J. Lipid Res. 46:1239–1247.
7:1067–1078. 17. Guide, S.V., F. Stock,V. J. Gill,V. L.Anderson,
5. Carlier, A., S. Uroz, B. Smadja, R. Fray, X. H. L. Malech, J. I. Gallin, and S. M. Holland.
Latour,Y. Dessaux, and D. Faure. 2003.The Ti 2003.Reinfection,rather than persistent infection,
plasmid of Agrobacterium tumefaciens harbors an attM- in patients with chronic granulomatous disease. J.
paralogous gene, aiiB, also encoding N-acyl Infect. Dis. 187:845–853.
homoserine lactonase activity. Appl. Environ. Micro- 18. Harel, M., A. Aharoni, L. Gaidukov, B.
biol. 69:4989–4993. Brumshtein, O. Khersonsky, R. Meged, H.
6. Chan,Y.Y., and K. L. Chua. 2005.The Burkholde- Dvir, R. Ravelli, A. McCarthy, L. Toker, I.
ria pseudomallei BpeAB-OprB efflux pump: expres- Silman, J. L. Sussman, and D. S.Tawfik. 2004.
sion and impact on quorum sensing and virulence.J. Structure and evolution of the serum paraoxonase
Bacteriol. 187:4707–4719. family of detoxifying and anti-atherosclerotic
7. Cheng, Z.,Y. Kumagai, M. Lin, C. Zhang, and enzymes. Nat. Struct. Mol. Biol. 11:412–419.
Y. Rikihisa. 2006. Intra-leukocyte expression of 19. Hentzer, M., H.Wu, J. B.Andersen, K. Riedel,
two-component systems in Ehrlichia chaffeensis and T. B. Rasmussen, N. Bagge, N. Kumar, M. A.
Anaplasma phagocytophilum and effects of the histi- Schembri, Z. Song, P. Kristoffersen, M.
dine kinase inhibitor closantel. Cell Microbiol. Manefield, J.W. Costerton, S. Molin, L. Eberl,
8:1241–1252. P. Steinberg, S. Kjelleberg, N. Hoiby, and M.
8. Chun, C. K., E.A. Ozer, M. J.Welsh, J. Zabner, Givskov. 2003.Attenuation of Pseudomonas aerugi-
and E. P. Greenberg. 2004. Inactivation of a nosa virulence by quorum sensing inhibitors.
Pseudomonas aeruginosa quorum-sensing signal by EMBO J. 22:3803–3815.
human airway epithelia. Proc. Natl. Acad. Sci. USA 20. Hoang,T.T., and H. P. Schweizer. 1999. Char-
101:3587–3590. acterization of Pseudomonas aeruginosa enoyl-acyl
9. de la Sierra-Gallay, I. L., O. Pellegrini, and C. carrier protein reductase (FabI): a target for the
Condon. 2005. Structural basis for substrate bind- antimicrobial triclosan and its role in acylated
ing, cleavage and allostery in the tRNA maturase homoserine lactone synthesis. J. Bacteriol. 181:
RNase Z. Nature 433:657–661. 5489–5497.
390 ■ WANG ET AL.
21. Hu, J.Y., F.Yang,Y. H. Lin, H. B. Zhang, S. L. Acyl-homoserine lactone acylase from Ralstonia
Ong, N. Dong, J. L. Xu, W. J. Ng, and L. H. strain XJ12B represents a novel and potent class of
Zhang. 2003. Microbial diversity and prevalence quorum-quenching enzymes. Mol. Microbiol.
of virulent pathogens in biofilms developed in a 47:849–860.
water reclamation system. Res. Microbiol. 154: 33. Liu, D. L., B. W. Lepore, G. A. Petsko,
623–629. P. W. Thomas, E. M. Stone, W. Fast, and
22. Huang, J. J., J. I. Han, L. H. Zhang, and J. R. D. Ringe. 2005. Three-dimensional structure of
Leadbetter. 2003.Utilization of acyl-homoserine the quorum-quenching N-acyl homoserine lac-
lactone quorum signals for growth by a soil tone hydrolase from Bacillus thuringiensis. Proc. Natl.
pseudomonad and Pseudomonas aeruginosa PAO1. Acad. Sci. USA 102:11882–11887.
Appl. Environ. Microbiol. 69:5941–5949. 34. Lu, X.,Y.Yuan, X. L. Xue, G. P. Zhang, and S.
23. Huang, J. J., A. Petersen, M. Whiteley, and N. Zhou. 2006. Identification of the critical role
J. R. Leadbetter. 2006. Identification of QuiP, of Tyr-194 in the catalytic activity of a novel N-
the product of gene PA1032, as the second acyl-homoserine lactonase from marine Bacillus
acyl-homoserine lactone acylase of Pseudomonas cereus strain Y2. Curr. Microbiol. 53:346–350.
aeruginosa PAO1. Appl. Environ. Microbiol. 72: 35. Lyon, G. J., P. Mayville, T. W. Muir, and R. P.
1190–1197. Novick. 2000. Rational design of a global
24. Jackson, S. H., J. I. Gallin, and S. M. Holland. inhibitor of the virulence response in Staphylo-
1995. The p47phox mouse knock-out model of coccus aureus, based in part on localization of
chronic granulomatous disease. J. Exp. Med. the site of inhibition to the receptor-histidine
182:751–758. kinase,AgrC. Proc. Natl.Acad. Sci. USA 97:13330–
25. Kato, N., T. Morohoshi, T. Nozawa, H. 13335.
Matsumoto, and T. Ikeda. 2006. Control of 36. Manefield, M.,T. B. Rasmussen, M. Henzter,
gram-negative bacterial quorum sensing with J. B.Andersen, P. Steinberg, S. Kjelleberg, and
cyclodextrin immobilized cellulose ether gel. J. M. Givskov. 2002. Halogenated furanones inhibit
Incl. Phenom. Macro. 56:55–59. quorum sensing through accelerated LuxR
26. Kaufmann, G. F., R. Sartorio, S. H. Lee, J. M. turnover. Microbiology 148:1119–1127.
Mee, L. J.Altobell, D. P. Kujawa, E. Jeffries, B. 37. MDowell, P., Z. Affas, C. Reynolds, M. T.
Clapham, M. M. Meijler, and K. D. Janda. Holden, S. J. Wood, S. Saint, A. Cockayne,
2006. Antibody interference with N-acyl P. J. Hill, C. E. Dodd, B. M. Bycroft, W. C.
homoserine lactone-mediated bacterial quorum Chan, and P. Williams. 2001. Structure, activity
sensing. J.Am. Chem. Soc. 128:2802–2803. and evolution of the group I thiolactone peptide
27. Kim, M. H., W. C. Choi, H. O. Kang, J. S. quorum-sensing system of Staphylococcus aureus.
Lee, B. S. Kang, K. J. Kim, Z. S. Derewenda, Mol. Microbiol. 41:503–512.
T. K. Oh, C. H. Lee, and J. K. Lee. 2005. 38. Miyairi, S., K. Tateda, E. T. Fuse, C. Ueda,
The molecular structure and catalytic mechanism H. Saito,T.Takabatake,Y. Ishii, M. Horikawa,
of a quorum-quenching N-acyl-L-homoserine M. Ishiguro, T. J. Standiford, and K.
lactone hydrolase. Proc. Natl. Acad. Sci. USA Yamaguchi. 2006. Immunization with 3-
102:17606–17611. oxododecanoyl-L-homoserine lactone-protein
28. Kim, Y., Y. H. Yoon, Y. Khang, S. Turley, and conjugate protects mice from lethal Pseudomonas
W. G. J. Hol. 2000.The 2.0 Å crystal structure of aeruginosa lung infection. J. Med. Microbiol.
cephalosporin acylase. Structure 8:1059–1068. 55:1381–1387.
29. Kohler,T., C. van Delden, L. K. Curty, M. M. 39. Molina, L., F. Constantinescu, L. Michel, C.
Hamzehpour, and J. C. Pechere. 2001. Over- Reimmann, B. Duffy, and G. Défago. 2003.
expression of the MexEF-OprN multidrug efflux Degradation of pathogen quorum-sensing mole-
system affects cell-to-cell signaling in Pseudomonas cules by soil bacteria:a preventive and curative bio-
aeruginosa. J. Bacteriol. 183:5213–5222. logical control mechanism. FEMS Microbiol. Ecol.
30. Leadbetter, J. R., and E. P. Greenberg. 2000. 45:71–81.
Metabolism of acyl-homoserine lactone quorum- 40. Momb, J., P. W. Thomas, R. M. Breece, D. L.
sensing signals by Variovorax paradoxus. J. Bacteriol. Tierney, and W. Fast. 2006. The quorum-
182:6921–6926. quenching metallo--lactonase from Bacillus
31. Lee, S. J., S.Y. Park, J. J. Lee, D.Y.Yum, B. T. thuringiensis exhibits a leaving group thio effect.
Koo, and J. K. Lee. 2002. Genes encoding the N- Biochemistry 45:13385–13393.
acyl homoserine lactone-degrading enzyme are 41. Moré, M. I., L. D. Finger, J. L. Stryker, C.
widespread in many subspecies of Bacillus Fuqua, A. Eberhard, and S. C. Winans. 1996.
thuringiensis.Appl.Environ.Microbiol.68:3919–3924. Enzymatic synthesis of a quorum-sensing autoin-
32. Lin, Y. H., J. L. Xu, J. Hu, L. H. Wang, S. L. ducer through use of defined substrates. Science
Ong, J. R. Leadbetter, and L. H. Zhang. 2003. 272:1655–1658.
42. Müh, U., B. J. Hare, B. A. Duerkop, M. model of X-linked chronic granulomatous disease,
Schuster, B. L. Hanzelka, R. Heim, E. R. an inherited defect in phagocyte superoxide pro-
Olson, and E. P. Greenberg. 2006. A struc- duction. Nat. Genet. 9:202–209.
turally unrelated mimic of a Pseudomonas aeruginosa 54. Rasmussen, T. B., M. E. Skindersoe, T.
acyl-homoserine lactone quorum-sensing signal. Bjarnsholt, R. K. Phipps, K. B. Christensen,
Proc. Natl.Acad. Sci. USA 103:16948–16952. P. O. Jensen, J. B. Andersen, B. Koch, T. O.
43. Nakayama, K., Y. Ishida, M. Ohtsuka, H. Larsen, M. Hentzer, L. Eberl, N. Hoiby,
Kawato, K.Yoshida,Y.Yokomizo, S. Hosono, and M. Givskov. 2005. Identity and effects of
T. Ohta, K. Hoshino, H. Ishida, K.Yoshida,T. quorum-sensing inhibitors produced by Penicil-
E. Renau, R. Leger, J. Z. Zhang,V. J. Lee, and lium species. Microbiology 151:1325–1340.
W. J. Watkins. 2003. MexAB-OprM-specific 55. Reimmann, C., N. Ginet, L. Michel, C. Keel,
efflux pump inhibitors in Pseudomonas aeruginosa. P. Michaux, V. Krishnapillai, M. Zala, K.
Part 1: discovery and early strategies for lead opti- Heurlier, K. Triandafillu, H. Harms, G.
mization. Bioorg. Med. Chem. Lett. 13:4201–4204. Defago, and D. Haas. 2002. Genetically pro-
44. Ng, C. J., D. M. Shih, S.Y. Hama, N.Villa, M. grammed autoinducer destruction reduces viru-
Navab, and S.T. Reddy. 2005.The paraoxonase lence gene expression and swarming motility in
gene family and atherosclerosis. Free Radic. Biol. Pseudomonas aeruginosa PAO1. Microbiology
Med. 38:153–163. 148:923–932.
45. Novick, R. P. 2003. Autoinduction and signal 56. Rothfork, J. M., G. S.Timmins, M. N. Harris,
transduction in the regulation of staphylococcal X. Chen, A. J. Lusis, M. Otto, A. L. Cheung,
virulence. Mol. Microbiol. 48:1429–1449. and H. D. Gresham. 2004. Inactivation of a bac-
46. Ozer, E. A., A. Pezzulo, D. M. Shih, C. Chun, terial virulence pheromone by phagocyte-derived
C. Furlong, A. J. Lusis, E. P. Greenberg, and oxidants: new role for the NADPH oxidase in host
J. Zabner. 2005. Human and murine paraoxonase defense. Proc. Natl. Acad. Sci. USA 101:13867–
1 are host modulators of Pseudomonas aeruginosa 13872.
quorum-sensing. FEMS Microbiol. Lett. 253: 57. Schaefer, A. L., D. L.Val, B. L. Hanzelka, J. E.
29–37. Cronan, Jr., and E. P. Greenberg. 1996. Gener-
47. Park, S. Y., B. J. Hwang, M. H. Shin, J. A. ation of cell-to-cell signals in quorum sensing: acyl
Kim, H. K. Kim, and J. K. Lee. 2006. N- homoserine lactone synthase activity of a purified
acylhomoserine lactonase producing Rhodococcus Vibrio fischeri LuxI protein. Proc. Natl. Acad. Sci.
spp. with different AHL-degrading activities. USA 93:9505–9509.
FEMS Microbiol. Lett. 261:102–108. 58. Sekar, K., B. Z.Yu, J. Rogers, J. Lutton, X. H.
48. Park, S.Y., H. O. Kang, H. S. Jang, J. K. Lee, B. Liu, X. Chen, M. D. Tsai, M. K. Jain, and M.
T. Koo, and D. Y. Yum. 2005. Identification Sundaralingam. 1997. Phospholipase A2 engi-
of extracellular N-acylhomoserine lactone acylase neering. Structural and functional roles of the
from a Streptomyces sp. and its application to highly conserved active site residue aspartate-99.
quorum quenching. Appl. Environ. Microbiol. 71: Biochemistry 36:3104–3114.
2632–2641. 59. Sheppard, F. R., M. R. Kelher, E. E. Moore,
49. Park, S.Y., S. J. Lee, T. K. Oh, J. W. Oh, B. T. N. J. D. McLaughlin, A. Banerjee, and C. C.
Koo, D.Y.Yum, and J. K. Lee. 2003.AhlD,an N- Silliman. 2005. Structural organization of the
acylhomoserine lactonase in Arthrobacter sp., and neutrophil NADPH oxidase: phosphorylation and
predicted homologues in other bacteria. Microbiol- translocation during priming and activation. J.
ogy 149:1541–1550. Leukoc. Biol. 78:1025–1042.
50. Parsek, M. R., D. L.Val, B. L. Hanzelka, J. E. 60. Singh,V., W. Shi, S. C. Almo, G. B. Evans, R.
Cronan Jr., and E. P. Greenberg. 1999. Acyl H. Furneaux, P. C. Tyler, G. F. Painter, D. H.
homoserine-lactone quorum-sensing signal gen- Lenz, S. Mee, R. Zheng, and V. L. Schramm.
eration. Proc. Natl.Acad. Sci. USA 96:4360–4365. 2006. Structure and inhibition of a quorum sens-
51. Passador, L., K. D.Tucker, K. R. Guertin, M. ing target from Streptococcus pneumoniae. Biochem-
P. Journet, A. S. Kende, and B. H. Iglewski. istry 45:12929–12941.
1996. Functional analysis of the Pseudomonas aerug- 61. Sio, C. F., L. G. Otten, R. H. Cool, S. P. Diggle,
inosa autoinducer PAI. J. Bacteriol. 178:5995–6000. P. G. Braun, R. Bos, M. Daykin, M. Camara,
52. Pearson, J. P., C. Van Delden, and B. H. P. Williams, and W. J. Quax. 2006. Quorum
Iglewski. 1999. Active efflux and diffusion are quenching by an N-acyl-homoserine lactone acy-
involved in transport of Pseudomonas aeruginosa lase from Pseudomonas aeruginosa PAO1. Infect.
cell-to-cell signals. J. Bacteriol. 181:1203–1210. Immun. 74:1673–1682.
53. Pollock, J. D., D.A.Williams, M.A. Gifford, L. 62. Smith, R. S., S. G. Harris, R. Phipps, and B.
L. Li, X. Du, J. Fisherman, S. H. Orkin, C. M. Iglewski. 2002. The Pseudomonas aeruginosa
Doerschuk, and M. C. Dinauer. 1995. Mouse quorum-sensing molecule N-(3-oxododecanoyl)
392 ■ WANG ET AL.
homoserine lactone contributes to virulence and 69. Wang, L. H., L. X. Weng,Y. H. Dong, and L.
induces inflammation in vivo. J. Bacteriol. 184: H. Zhang. 2004. Specificity and enzyme kinetics
1132–1139. of the quorum-quenching N-acyl homoserine
63. Stephenson, K., Y. Yamaguchi, and J. A. lactone lactonase (AHL-lactonase). J. Biol. Chem.
Hoch. 2000. The mechanism of action of 14:13645–13651.
inhibitors of bacterial two-component signal 70. Waters, C. M., and B. L. Bassler. 2005. Quo-
transduction systems. J. Biol. Chem. 275:38900– rum sensing: cell-to-cell communication in bacte-
38904. ria. Annu. Rev. Cell Dev. Biol. 21:319–346.
64. Tateda, K., Y. Ishii, M. Horikawa, T. 71. Yang, F., L. H. Wang, J. Wang, Y. H. Dong,
Matsumoto, S. Miyairi, J. C. Pechere, T. J. J. Y. Hu, and L. H. Zhang. 2005. Quorum
Standiford, M. Ishiguro, and K. Yamaguchi. quenching enzyme activity is widely conserved in
2003.The Pseudomonas aeruginosa autoinducer N- the sera of mammalian species. FEBS Lett.
3-oxododecanoyl homoserine lactone accelerates 579:3713–3717.
apoptosis in macrophages and neutrophils. Infect. 72. Yoshida, K., K. Nakayama, N. Kuru, S.
Immun. 71:5785–5793. Kobayashi, M. Ohtsuka, M. Takemura, K.
65. Telford, G., D. Wheeler, P. Williams, P. T. Hoshino, H. Kanda, J. Z. Zhang,V. J. Lee, and
Tomkins, P.Appleby, H. Sewell, G. S. Stewart, W. J. Watkins. 2006. MexAB-OprM specific
B. W. Bycroft, and D. I. Pritchard. 1998.The efflux pump inhibitors in Pseudomonas aeruginosa.
Pseudomonas aeruginosa quorum-sensing signal Part 5: Carbon-substituted analogues at the C-2
molecule N-(3-oxododecanoyl)-L-homoserine position. Bioorg. Med. Chem. 14:1993–2004.
lactone has immunomodulatory activity. Infect. 73. Zelinski, T., and M. R. Kula. 1994. A kinetic
Immun. 66:36–42. study and application of a novel carbonyl reductase
66. Thomas, P.W., E. M. Stone, A. L. Costello, D. isolated from Rhodococcus erythroplis. Bioorg. Med.
L. Tierney, and W. Fast. 2005. The quorum- Chem. 2:421–428.
quenching lactonase from Bacillus thuringiensis is a 74. Zelinski, T., J. Peters, and M. R. Kula. 1994.
metalloprotein. Biochemistry 44:7559–7569. Purification and characterization of a novel car-
67. Ulrich, R. L. 2004. Quorum quenching: bonyl reductase isolated from Rhodococcus erythro-
enzymatic disruption of N-acylhomoserine polis. J. Biotechnol. 33:283–292.
lactone-mediated bacterial communication in 75. Zhang, H. B., L. H. Wang, and L. H. Zhang.
Burkholderia thailandensis. Appl. Environ. Microbiol. 2002. Genetic control of quorum-sensing signal
70:6173–6180. turnover in Agrobacterium tumefaciens. Proc. Natl.
68. Uroz, S., S. R. Chhabra, M. Camara, P. Acad. Sci. USA 99:4638–4643.
Williams, P. Oger, and Y. Dessaux. 2005. 76. Zhu, J., J.W. Beaber, M. I. More, C. Fuqua,A.
N-acylhomoserine lactone quorum-sensing Eberhard, and S. C. Winans. 1998. Analogs
molecules are modified and degraded by Rhodococ- of the autoinducer 3-oxooctanoyl-homoserine
cus erythropolis W2 by both amidolytic and lactone strongly inhibit activity of the TraR pro-
novel oxidoreductase activities. Microbiology 151: tein of Agrobacterium tumefaciens. J. Bacteriol. 180:
3313–3322. 5398–5405.
QUORUM-SENSING INHIBITION
Staffan Kjelleberg, Diane McDougald, Thomas Bovbjerg Rasmussen,
and Michael Givskov
25
The last 10 years have witnessed a remarkable bacterial lifestyle: surface-associated bacteria
change in our understanding of how bacteria living in matrix-encased aggregates or micro-
live and function. Following more than 200 colonies (5, 19, 20, 55, 58, 81–83, 108, 127). It is
years of primarily studying bacteria as single generally accepted that this lifestyle confers
cells dwelling as free-living organisms, the high levels of tolerance against a number of
advent of molecular biology-based approaches environmental stresses, including grazing, as
to study bacteria in high-density matrix- well as the activity of host immune systems. It is
embedded assemblages on surfaces, i.e., bio- also conceivable that, well within the biofilm,
films, and the involvement of cell-cell signaling the bacteria use a range of regulatory circuits to
in various aspects of the biofilm life cycle (61) adjust gene expression to the sessile mode of
led to a surge in interest to study bacterial growth and coordinate cell-to-cell interactions.
biofilms in a range of habitats. With our now These circuits include density-dependent cell-
improved understanding of biofilm biology, cell communication systems to pool single-cell
research also focuses on the development of activities, such as the secretion of bioactives and
biofilm control measures in medical as well as virulence factors, at the strategically right
industrial and environmental settings. moment to successfully compete or collaborate
High-resolution microscopic investigations with other microbial species for space and
of surfaces in several habitats, including chroni- nutrients, and to coordinate the progressive
cally infected tissues, medical implants, terres- attack on host surfaces.
trial and marine plant surfaces, and inanimate Bacteria have evolved several mechanisms to
marine and freshwater surfaces, suggest com- facilitate the interaction among themselves.
mon structural features indicative of a preferred These interactions extend beyond the species
level. One such strategy is the ability to coordi-
Staffan Kjelleberg and Diane McDougald Centre for Marine
Biofouling and Bio-Innovation,The University of New South nate gene expression in accordance with popu-
Wales, New South Wales, Australia. Thomas Bovbjerg Ras- lation density and hence act as a group in a
mussen Chr. Hansen A/S, Bøge Allé 10-12, 2970 Hørsholm, process termed quorum sensing (QS), as
Denmark. Michael Givskov BioScience and Technology,
Biocentrum, Bldg. 277, Technical University of Denmark, described in previous chapters of this book. It
Copenhagen, Denmark. has been proposed by several investigators that
393
394 ■ KJELLEBERG ET AL.
this capability is a fundamental prerequisite for neutrophil phagocytosis than biofilms formed
the development of and organization into more by a wild-type counterpart (1, 8, 24, 48, 53,
complex structural communities. Indeed, QS 54,73,94,96,98).Furthermore,in animal mod-
signals can be detected in biofilms across a els these QS mutants appear less virulent and
range of environments. For example, biofilms can be more readily cleared by the host organ-
grown on rocks in the San Marcos River in ism (7, 8).
Texas have been shown to produce acyl- Such observations on impaired biofim resist-
homoserine lactone (AHL) signals (80), as have ance and the fact that the QS systems function
biofilm communities on marine snow particles by means of small, extracellular signal mole-
(46) and sponge surfaces (119). Moreover, in cules have fueled the interest in exploring
different settings, signal molecules have also bacterial communication systems as an attrac-
been found in the lungs of patients with cystic tive target for novel antimicrobial control
fibrosis (CF) (17, 112) and to effect biofilm chemotherapy. Moreover, in the protected
architecture and mutualistic interactions by oral biofilm environment, bacteria produce and
bacteria (9, 65, 105). secrete a large number of extracellular com-
Currently several model organisms are pounds. In several bacterial species, many of
employed for pursuing detailed studies of how these molecules are virulence factors that are
biofilms form (42, 61). Of these organisms, controlled by QS (see other chapters in this
Pseudomonas aeruginosa has received particular book). In vivo, such virulence factors, in con-
attention for exploring the variety of factors junction with immune complexes and phago-
required for the life cycle of biofilm develop- cytic enzymes released by cells of the immune
ment and dispersal. This organism is one of system, may cause extensive tissue destruction.
the most widespread opportunistic bacterial For example, tissue destruction contributes
pathogens and causes a range of respiratory and significantly to loss of pulmonary function in
nosocomial infections (123). After initial CF patients (21,36,37,91).Hence,the QS con-
attachment of P. aeruginosa to a surface, micro- trol of both biofilm resistance and production
colonies are formed, which in turn can grow to of virulence factors identifies QS systems as
larger structures such as towers and mush- highly attractive targets for chemotherapy
rooms. Recent analysis based on transcrip- against biofilm-based, chronic infections. More
tomics revealed that the bulk of biofilm cells, generally, the presence of QS signaling systems
even at the early stages, express genes in a pat- in a range of bacteria in different medical,
tern that is reminiscent of gene expression seen industrial, and environmental settings, and their
in early stationary phase of planktonic cells, role in bacterial biofilm development and
including expression of QS-controlled genes maturation, speaks to the interest and utility in
(50, 54), showing that QS in principle is neither more broadly developing QS inhibitory tech-
restricted to the biofilm nor the planktonic nologies.
lifestyle. However, the same studies also identi-
fied a large subset of QS-controlled genes that THE ADVENT OF QS INHIBITORS
appear to be uniquely induced during the As the elucidation of the molecular mechanisms
biofilm mode of growth, and may be involved behind QS gained momentum,researchers con-
in generating biofilm-specific phenotypes such ceived several strategies to block QS systems.
as those involved in adaptation and likely One of the major problems with conventional
increased resistance and/or tolerance. For biofilm-control measures is the development
example,in vitro biofilms grown by P.aeruginosa of resistance.The use of methods targeting QS
cells in which the QS communication systems circuits constitutes a significant difference in
have been disabled by mutations are more approach in that such systems regulate nones-
prone to killing by various antibiotics, antimi- sential phenotypes. Consequently, a QS inhibi-
crobials, shear forces, protozoan grazing, and tor (QSI) is not expected to interfere with the
25. QUORUM-SENSING INHIBITION ■ 395
growth of the bacteria.Accordingly,it is believed

that the selective pressure for development of
resistance is significantly lower compared to that
of conventional antimicrobial measures.
On the basis of the AHL signal structure,
Eberhard et al. designed the first synthetic
inhibitors of the LuxR system in the 1980s
(39).However,the first natural,broad-spectrum
QSI compounds emerged from a marine
library of secondary metabolites. Further
research in this area has proven to be in line
with the general accepted view that nature is
superior to combinatorial chemistry in capabil- FIGURE 1 Halogenated furanone compounds. (A)
Compound 1 and (B) compound 2 produced by the
ity to produce “disease relevant” molecular algae D. pulchra. Synthetic furanone compounds 30 (C)
diversity (78, 86).The Australian red macro alga and 56 (D) (43, 53, 54, 68).
Delisea pulchra attracted the attention of marine
biologists because it was devoid of extensive
surface colonization, i.e., biofouling, unlike of furanones is stored in specialized vesicles
other plants in the same environment. Mature from which they become released at the surface
biofouling communities develop following the of the thallus at concentrations ranging from 1
settlement primarily of algal spores and inverte- to 100 ng/cm2 (38, 104). Field experiments
brate larvae, but bacteria are the first colonizers have demonstrated that the surface concentra-
of submerged surfaces, providing an initial con- tion of furanones is inversely correlated to the
ditioning biofilm to which other marine degree of colonization by marine bacteria (62).
organisms attach (104).The red seaweed D. pul- In 1995, a visual and conceptual model for
chra produces a range of halogenated furanone the rapid formation of biofilms on moist sur-
compounds (31) that display antifouling and faces was offered by the QS-controlled Serratia
antimicrobial properties (29, 30, 99), altering liquefaciens (now S. marcescens) surface transloca-
the abundance and composition of the bacterial tion and colonization (40). Several of the
community on the surface and hence the sub- furanone compounds that exhibit structural
sequent development of a biofouling commu- similarity to the short-chain AHL molecules
nity (6). Generally, it is now understood that (Color Plate 14) inhibited swarming motility of
primitive eukaryotes, devoid of sophisticated S. marcescens MG1 (43). From the observation
immune systems, have developed chemical that biofilm formation on submerged surfaces
defense mechanisms (25, 30, 126), which, in precedes the attraction of greater fouling
several cases, are secondary metabolites that organisms and the development of detrimental
inhibit bacterial colonization phenotypes (62, biofouling, our prevailing hypothesis was that
74, 114).The genus Delisea produces a range of furanones of D. pulchra constitute a specific
unique natural products—halogenated fura- means of eukaryotic interference with bacterial
nones or “fimbrolides” (Fig. 1)—that have multicellular behavior (43). Neither swimming
interesting biological activities in both a natural motility (a non-QS-regulated process involving
and applied context. (For a recent review, see planktonic bacteria), flagellar protein content
reference 28). The D. pulchra furanone com- per cell mass, nor growth rates (in planktonic
pounds consist of a furan ring structure with a cultures) were affected by furanones (43). The
substituted acyl chain at the C-3 position and a effect was reversible by addition of external
bromine substitution at the C-4 position. The signal molecules. Addition of pure surfactant
substitution at the C-5 position may vary in to the swarm medium also relieved the fura-
terms of side chain structure.The full cocktail none effect, which suggested that swarming
inhibition was caused by a reduction in the protect, the AHL-dependent transcriptional

extracellular serrawettin W2 production as activator. The furanone-dependent reduction
later demonstrated by Rasmussen et al. (97). In in the cellular concentration of the LuxR pro-
other words, the furanone compounds enabled tein was associated with a reduction in expres-
growth and swimming but disabled expansion sion of a plasmid-encoded PluxI-gfp(ASV)
and therefore abolished surface colonization. fusion, suggesting that the reduction in LuxR
All experimental data then supported a model concentration is the mechanism by which fura-
by which the furanone compounds specifically nones control expression of AHL-dependent
block the QS system. Furanones have since phenotypes (71).
been shown to prevent the expression of spe- Several observations indicate a competitive
cific colonization and invasion traits in several type of inhibition by furanones but do not
marine bacterial epiphytic isolates at non- reveal further information on the molecular
growth-inhibitory concentrations (our unpub- mechanisms. A recent mutational analysis that
lished observations). Several of these isolates involves exchange of selected amino acids in
were shown to produce AHLs. More recently, it the putative binding pocket of LuxR shed some
has also been established that these furanone light on this. For example, a mutant LuxR pro-
compounds are effective in inhibiting QS sys- tein containing an amino acid substitution
tems that use non-AHL signal molecules, for (either L42A, L42S, or M135A) was still
example, those that use the autoinducer 2 (AI- responsive to AHLs, but the synthetic QSI
2) QS system (27, 78, 101, 102). compound N-(propylsulfanylacetyl)-L-HSL
was no longer able to block QS (64).This indi-
FURANONE INHIBITORS cates that such an agonist-based QSI design
AND THEIR MODE OF ACTION blocks QS by entering the AHL-binding
One of the important pieces of evidence in pocket. Conversely, the mutant LuxR was still
favor of the proposed mode action of furanones sensitive to inhibition by halogenated fura-
in AHL QS systems was the furanone-mediated nones, indicating that they bind at a different
displacement of 3H-labeled signal molecules location on the LuxR protein (64).The above-
from LuxR (68). These studies suggested that mentioned observations on accelerated turno-
the furanone compounds competed with the ver of the LuxR suggest that the receptor folds
cognate OHHL signal for the LuxR receptor incorrectly, thereby making the receptor pro-
site. It is difficult to ignore the structural simi- tein prone to degradation by household pro-
larities (Color Plate 14) between AHLs and teases.As the production of the receptor protein
furanones when formulating a model for the is considered to be constitutive, the concentra-
mode of action of the inhibitory activity, but at tion of available LuxR homologues that can
the same time it is puzzling that similar experi- bind AHL is reduced. Consequently, the con-
ments revealed no substantial affinity of the centration of active LuxR multimers is reduced
furanones for Escherichia coli cells overproduc- and, hence, QS-controlled genes are not acti-
ing LuxR (70). This apparent paradox was vated (64, 71).
resolved with the finding that halogenated Does the model of furanone inhibition in
furanones accelerate the degradation of the QS signal systems apply also to other AHL-
LuxR protein (71). While unable to demon- based regulating systems such as the LasR and
strate a stable interaction between furanones RhlR proteins of the P. aeruginosa QS systems?
and the QS regulator LuxR, Manefield et al. The strongest support so far for the interaction
(71) discovered, by Western analysis, that the between furanones with these receptor proteins
half-life of the protein is reduced up to 100-fold is based on circumstantial evidence. By using
in the presence of halogenated furanones.This microarrays for transcriptomic analysis, one can
suggests that halogenated furanones modulate simultaneously analyze the expression of all
LuxR activity but act to destabilize, rather than genes, including known virulence genes. This
technique offers an, until now, unmatched pos- although this LuxR is not homologous to AHL
sibility to monitor the target specificity of QSI system regulators. Microarray analysis of Vibrio
compounds (54, 96, 98). In the present context, cholerae grown in the presence of furanones also
the strategy has been to characterize the entire indicates an overlap with the QS regulon,
QS regulon by transcriptomic analysis per- although to a lesser extent (78). Bioassays with
formed on defined QS mutants in the presence different reporter strains of Vibrio harveyi (e.g.,
or absence of externally added AHL signals. If AHL,AI-2, etc.) indicated that furanones affect
subsequent treatment with a putative QS all of the QS systems (79), which led us to
inhibitor results in a significant, global repres- hypothesize that they acted at a point at or
sion of the identified QS target genes, it below the convergence of the signaling cascade.
strongly suggests that the LasR and RhlR pro- Recent evidence has shown that the presence
teins are in fact targeted by the test compound. of furanones, while not affecting luxRVh
In addition, the number of unrelated genes mRNA or LuxRVh protein levels in V. harveyi,
affected by the treatment reports on the speci- rendered LuxRVh unable to bind to promoter
ficity of the test compound. Comparative sequences of QS-controlled genes (W. Ver-
analysis of the synthetic furanone 30 target straete, personal communication). It is not yet
genes and the QS regulon showed that approx- clear how the furanone mediates this effect, but
imately 80% of the furanone-repressed genes because furanone binding does not lower
are also QS controlled, using a fivefold cut-off LuxRVh levels, it is unlikely to do so by causing
limit for furanone repression and QS induc- degradation of LuxRVh, as is observed for the
tion. Likewise, 46% of AHL-induced genes AHL receptor proteins.The fact that furanones
were more than fivefold repressed by furanone are effective at QS inhibition in a range of QS
30, and another 39% were two- to fivefold systems highlights their potential as therapies
repressed (54). In general, there was a strong where QS inhibition is important.
correlation between genes strongly induced
and repressed by AHLs and furanone 30, QSI Compounds Produced by
respectively. Microorganisms and Plants
Similar evidence exists for the mode of Bacteria also produce compounds that are able
action of furanones in Autoinducer-2 (AI-2) to interfere with QS systems. Members of
QS systems. Microarray experiments with E. the filamentous fungus Streptomyces produce
coli demonstrated that 79% of the genes furanone compounds that are intermediates
repressed by furanone were induced by AI-2, in butanolide production. Inhibitors of the
which indicated that furanone compound 2 QS-controlled purple pigment production
inhibited AI-2 signaling (101). Furthermore, in Chromobacterium violaceum have been identi-
biofilm formation was repressed by the fura- fied by screening a library of furanone com-
none compound. AI-2 bioassays indicated that pounds (and their analogues) produced by S.
the furanone compound decreased the extra- antibioticus (72).
cellular concentration of AI-2, but luxS and pfs, Since AHL-producing bacteria are ubiq-
the enzymes involved in synthesis of the AI-2 uitous in nature where they coexist with a
signal, were not significantly altered, indicating multitude of higher organisms, the latter, not
that furanones alter AI-2 signaling posttran- surprisingly,have evolved the ability to interfere
scriptionally. Vibrio species utilize a variety of with QS by means of QSIs. Indeed, some fungi
QS systems (AHL, AI-2, Cqs, LuxM/N) that of the genus Penicillium produce effective QSI
act as coordinate regulators of a wide range of compounds. Among a group of 50 Penicillium
phenotypes at different cell densities; i.e., some isolates, 66% were found to produce QSIs; P.
are active at low and others at high cell density radicicola produces penicillic acid and P. copro-
(75).These QS systems converge to regulate the bium produces patulin (Fig. 2), both of which
response regulator, LuxRVh (and homologues), are inhibitors of both the lux system and the QS
ing X-Gal; the lux gene cluster from Vibrio fis-

cheri allowing expression of bioluminescence;
and production of the green fluorescent protein
(GFP) (or its variants with different fluorescent
colors) obtained from the jelly fish Aequorea vic-
toria. As these reporter systems are fused to a
QS-controlled promoter, they become acti-
FIGURE 2 The two QSIs, penicillic acid (A) and vated when the host bacterial cells sense the
patulin, (B) produced by certain fungi (98). presence of exogenous signal molecules. Con-
versely, when the bacteria are challenged with
QSI compounds (along with the inducing
systems found in P. aeruginosa (98).A transcrip- AHLs), the signal from the reporter systems is
tomic analysis (performed on planktonic cells reduced. Bacteria using GFP-based screening
at an optical density at 600nm [OD600] 2) systems can be grown in liquid media in
revealed that the two compounds were able to microtiter dishes where many different com-
down-regulate 45% (patulin) and 60% (penicil- pounds and/or concentrations can be probed at
lic acid), respectively, of the QS-controlled a time. Usually both growth (OD) and expres-
genes in P.aeruginosa, indicating that these com- sion of the reporter systems are monitored over
pounds are indeed QSI compounds targeting time (Fig. 3). The disadvantage to this type of
the las and rhl systems (98). Treatment with screening system is that compounds that either
patulin also caused turnover of the LuxR pro- inhibit or slow growth of the bacterial host
tein (64). inevitably reduce reporter expression and con-
Several plants produce compounds capable sequently may lead to the scoring of false posi-
of interfering with bacterial QS systems.They tives. Hence, growth of the bacterial screen has
include crown vetch, carrot, soybean, water lily, to be carefully monitored to ensure that the test
tomato, pea seedlings (Pisum sativum), habanero compounds are not interfering with growth
chili, and garlic (96, 121). Upon closer exami- and thereby general protein synthesis.
nation, some garlic subspecies produce more To circumvent the problem of separating
than three different QSIs, whereas others pro- specific QS-targeted effects from pleiotrophic
duce none.We recently found that a relatively effects caused by growth inhibition, another
crude extract of garlic, where most of the type of screen, termed the QSI selector, has
toxic allicin-derived compounds have been been developed (96). In this system the QS-
removed, was able to down-regulate about 34% controlled promoter is fused to a gene causing
of the QS-regulated genes in P. aeruginosa.The growth arrest when expressed. The bacterial
extract primarily targeted genes regulated by host does not produce any QS signals by itself,
RhlR and genes that are regulated by both so in the absence of AHL molecules, growth is
LasR and RhlR (96). unrestrained. If the growth medium is supple-
mented with AHL molecules, the QS-
Screens and Procedures for controlled killing system becomes activated,
Identification of QSI Compounds leading to growth arrest (Fig. 4). Addition of a
Several convenient screens for QSI activity QSI compound inhibits expression of the QS-
of AHL systems have been developed in controlled killing system and the bacteria are
recent years.The basic principle includes QS- allowed to grow. Hence, the presence of QSI
controlled gene expression of an easily recog- molecules in a sample is indicated by growth.
nizable reporter, such as the violacein genes Furthermore, the bacteria are expressing phe-
from C. violaceum producing a purple pigment notypes that ease the identification of growth,
when expressed; the lacZ gene from E. coli giv- such as -galactosidase or bioluminescence.
ing rise to blue-green color when it is hydrolyz- Briefly,the bacteria are cast into agar along with
FIGURE 3 Concentration-dependent inhibition of QS by furanone 30. P. aeruginosa harboring a lasB-gfp fusion is

treated with different concentrations of the furanone (as indicated). Specific activity of gene expression fluorescence
per OD600 is monitored over time.
signal molecules that will activate the killing activity will rescue the bacteria and allow a pos-
system (Color Plate 15). Wells are punched in itive outcome of the screen (96).
the agar, and compounds or mixtures to be
tested are added to the wells. From the wells, Molecular Design of QSI Compounds
compounds diffuse into the semisolid agar, A highly investigated strategy to interfere with
establishing a concentration gradient with the QS is blocking the LuxR homologue signal
highest concentration closest to the well. This receptor by means of small molecules; in the
enables the researchers to test the effect of following paragraphs we summarize the results
numerous concentrations in just one assay. If of previous studies indicating the molecular
the compound has no QSI activity, the killing requirements of such inhibitors.
system in the bacteria is active due to the AHLs
present in the agar; hence, the bacteria are SUBSTITUTIONS IN THE SIDE CHAIN
killed, no growth is observed, and a negative AHL analogues can be substituted in either the
screen is observed. If the compound is toxic, no side chain or the ring moiety.Analogues of the
growth is observed.This is also scored as a neg- 3-oxo-C6-HSL molecule with different sub-
ative outcome of the screen. Only test com- stituents in the side chain (Fig. 5A) are able
pounds with nontoxic properties and QSI to displace the native signal from the LuxR
FIGURE 4 Principle of the QSIS system. (A) The screening bacteria are grown without AHL.
The LuxR homologue is not activated; hence, there is no expression from the QS-controlled
promoter (PQS) and the killing gene is not expressed.This condition is used when the bacteria are
grown for purposes other than screening.(B) Exogenously added AHL molecules activate the PQS
promoter, and the killing gene is thereby expressed, causing growth arrest of the bacteria. (C)
Presence of an exogenously added QSI compound blocks QS, and the killing gene is not
expressed, which rescues the host bacteria (96).
receptor. However, most of these compounds both the lux and las systems is generated (93).
also exhibit agonist effects that limit their use as Likewise, if the C-1 atom is replaced by a sul-
QSIs (109). If the C-3 atom in the side chain is fonyl group, QSI activity is also achieved (14).
replaced by a sulfur atom, a potent inhibitor of Another strategy to modify the AHL signal
FIGURE 5 AHL analogues with changes in the side chain. (A) 3-oxo-C6-HSL signal
molecule. (B to D) Analogues with agonist effect. (E to H) Analogues with antagonistic
effect. (I and J) Analogues without any effect (14, 41, 93, 103, 109).
molecules is to place atoms or groups at the end (109). One exception to this is when the sub-
of the side chain. Substituting secondary alkyl stituents are placed on the C-3 carbon atom of
groups at the C-6 atom of 3-oxo- C6-HSL (Fig. the ring (Fig. 6). Compounds having acyl alco-
5C to H) gives rise to agonists, whereas posi- hols or acyl amides attached at this position are
tioning of a secondary aryl group on that loca- still able to function as agonists of LuxR. Con-
tion gives rise to an antagonist. As the size versely, if the substituents are placed on the C-4
difference between the two types of molecules atom, the compounds are not able to interact
is negligible,the differences in activity are prob- with the LuxR receptor. This indicates that
ably due to the ability of the aryl compounds to there is more “free space” around the C-3 atom
interact hydrophobically with aromatic amino of the ring inside the AHL-binding pocket
acids in the protein.If the size of the substituents (87). Instead of making single substitutions, the
is increased to include tertiary alkyl derivatives entire ring can be exchanged with another
or even larger alkyl and aryl moieties, the cyclic structure. In exploring compounds able
agonistic activity of the molecules is lost, indi- to interfere with QS in P. aeruginosa, the side
cating that they are too bulky to enter the AHL- chains of 3-oxo-C12-HSL and C4-HSL have
binding site in the receptor protein (41, 103). been attached to amino-cyclo-alcohol and
amino-cyclo-ketone with either five or six car-
SUBSTITUTIONS ON THE HSL RING bons in the ring (Fig. 7).The C12 amino-cyclo-
Most compounds with a keto-oxygen in the hexanol compound was found to be a strong
ring or an extra carbon expanding the ring activator of the LasR protein whereas the C4
exhibit little or no binding to the LuxR protein keto compounds were the most potent agonists
FIGURE 6 AHL analogues with changes in the ring. (A) 3-oxo- C6 HSL signal
molecule. (B) Analogue with agonist effect. (C) Analogue with antagonistic effect,
(D to G) Analogues without any effect (87, 109).
of the RlhR protein.This indicates that the two the HSL ring is replaced by a benzyl group, an
QS receptors do not perceive the HSL moiety inhibitor of LuxR is generated (103).
of the AHL signal in the same manner.Another Muh et al. (85) developed an ultrahigh-
molecule, 3-oxo- C12-(2-aminocyclohexa- throughput cell-based assay to screen a library
none) (Fig. 7), is an inhibitor of the LasR-based of approximately 200,000 compounds for the
QS system. This molecule is able to down- presence of inhibitors of LasR-dependent gene
regulate the LasR-dependent expression of the expression. The rsaL promoter was chosen to
LasI AHL synthetase. When applied at a con- drive the reporter-screen as one of the most
centration of 100 M (which is relatively high sensitive LasR-controlled genes, being induced
for a QSI), the compound significantly reduced several hundredfold in the presence of 3-oxo-
the production of exo-proteins (115). A more C12-HSL. After screening the synthetic library,
potent inhibitor of LasR is 3-oxo-C12-(2- two QS inhibitors were selected: V-06-018
aminophenol), which is able to abolish produc- (Fig. 7E) exhibiting a 50% inhibitory concen-
tion of pyocyanin and elastase and, in addition, tration (IC50) of 10 M and PD12 (Fig. 7F)
has been claimed to disrupt normal biofilm for- with an IC50 of 30 nM. Both inhibitors carry a
mation by P. aeruginosa (115). Interestingly, the 12-carbon aliphatic tail, which is attached to a
very similar molecule 3-oxo-C12-(2-aminocy- tetrazole in the case of PD12 and a phenyl in
clohexanol), having the phenol replaced by a the case of V-06-018, thereby resembling the
hexane ring, is a potent agonist of the las system native AHL signal 3-oxo-C12-HSL. Although
(116).A similar situation is seen with analogues the library screen included approximately
of 3-oxo-C6-HSL from the lux system. If the 200,000 compounds selected to cover a wide
HSL ring is replaced by a hexane ring, the abil- range of chemical scaffolds, the two most
ity to activate LuxR is retained. Contrarily, if potent inhibitors resembled the AHL molecule
FIGURE 7 AHL analogues with

exchanged ring part. (A) 3-oxo-C12-
HSL signal molecule and (B) analogue
showing agonist effects. (C and D) Ana-
logues with antagonistic effects from
Smith et al. and (E and F) from Muh
et al. (84, 85, 115, 116).
that normally binds to LasR. However, the NON-AHL-BASED INHIBITORS

authors could not detect any LasR degradation QSIs have also been identified by screening
by means of Western blotting (85). Similar to random libraries of synthetic chemical com-
some of the furanone compounds, interaction pounds. Among the inhibitors identified as
with the receptor is also based on circumstantial able to interfere with both the lux system from
evidence: A transcriptomic analysis of the V. fischeri and the las and rhl QS systems in P.
effects by the inhibitors on global gene expres- aeruginosa are para-benzoquinone, 2,4,5-
sion showed that both compounds were gen- tri-bromo-imidazole, indole, 3-nitrobenzene-
eral inhibitors of LasR controlled QS, i.e., the sulfonamide,and 4-nitro-pyridine-N-oxide (4-
expression levels of most LasR-dependent NPO) (Fig. 8B to D), the latter being the most
genes were affected. effective in this group (96). Transcriptomic
analysis of gene expression shows that 4-NPO

mainly affects genes that are regulated by either
RhlR alone or RhlR and LasR in concert.This
suggests that 4-NPO interacts most readily
with the RhlR receptor. In total, 37% of the
QS-regulated genes in P. aeruginosa (when the
planktonic cells are investigated at OD600 2)
are significantly down-regulated by treatment
with 100 M 4-NPO (96).
Riedel et al. (106) conducted a combined in
silico-based iterative process to develop antago-
nists to the Burkholderia cenocepacia QS system.
Despite the fact that the end product, denoted
compound 3 (Fig. 8E), showed no structural
similarities to AHL molecules, this hydrazide-
derived compound (in 0.5 to 1 mM range) was
able to specifically block QS in a competitive
manner as judged from two-dimensional poly-
acrylamide gel electrophoresis analysis, biosen-
sor and virulence assays, and attenuation of FIGURE 8 Molecules with QSI properties.
(A) 2,4,5-tri-bromo-imidazole, (B) indole, (C) 3-
virulence in the Caenorhabditis elegans patho- nitrobenzene-sulfonamide, (D) 4-nitro-pyridine-
genesis model. This is interesting and promis- N-oxide (96), and (E) compound 3 (106).
ing, since to date no QS inhibitor for
Burkholderia has been described. In fact, the
first-generation furanone-based compounds produce AHLs with side chains shorter than C4
assessed to date are ineffective due to the (12, 134). Some plants use this pH-dependent
rapid inactivation of the molecules by the instability against invading bacteria. For exam-
Burkholderia spp. ple,when the plant pathogen E.carotovora subsp.
carotovora attacks its host, the plant may respond
DESTRUCTION OF THE AHL SIGNAL by increasing the pH to 8.2, thereby achieving
The AHLs are inherently unstable at pH levels destruction of the signal molecules that direct
above neutral where they undergo lactonolysis, expression of virulence factors (12).
or ring opening. Some bacteria such as Erwinia Another mean of removing the QS signal
carotovora, P. aeruginosa, and Yersinia pseudotuber- molecule is employed by the marine alga Lami-
colosis raise the pH when they enter stationary narta digitata. This seaweed secretes oxidized
phase (at least when growing in shake flask cul- halogen compounds such as hypobromous
tures), leading to almost complete destruction and hypochlorous acids that react with 3-oxo-
of the signal molecules (134). Even exoge- substituted AHL compounds such as 3-oxo-
nously added AHLs are rapidly inactivated if C6-HSL produced by many marine Vibrio
added to cell-free supernatants of the above- strains. By secreting these compounds, the alga
mentioned bacteria.The lactonolysis reaction is is able to control formation of biofilm and bio-
also influenced by increasing temperature (Fig. fouling on its leaves (10).
9). In addition, the reaction is also dependent A different strategy to interfere with QS sys-
on the length of the side chain of the AHLs. tems is the enzymatic destruction of the signal
Taken together, this means that AHLs with a molecule, referred to as quorum quenching.
side chain shorter than 4 carbon atoms cannot This can be achieved by the action of the AiiA
retain their activity at physiological conditions. AHL lactonase (Fig. 9). This enzyme and its
Interestingly, no bacteria have been found to analogues are produced by many Bacillus species
FIGURE 9 Proposed pathway of C4-HSL degradation (66).The molecules of the side chain are channeled into
cell material, whereas the ring part is converted into a waste product.
such as B. cereus, B. mycoides, and several sub- some soil bacteria to compete with the AHL-
species of B.thurengiensis (32,34).The enzyme is producing strains in natural ecosystems. The
highly specific for acylated HSLs as it does not AHL-degrading enzymes have potential com-
degrade non-acyl lactones and noncyclic esters. mercial interest as they are likely to be used in
Conversely, the enzyme can degrade AHLs agriculture and food manufacturing, but for
with a range of lengths of the side chain and dif- treatment of human patients they have limited
ferent substitutions at the C-3 position (128). use.Because of the obstacles involved in deliver-
When expressed in P. aeruginosa, this enzyme ing proteinaceous agents systemically,AHL lac-
not only reduces the concentration of AHL in tonases can, at best, be applied topically. One
the surrounding media, it also lowers expres- aspect of the lactonolysis reaction should be
sion of major virulence factors such as elastase, kept in mind;it is reversible at acidic pH.A ring-
rhamnolipids,hydrogen cyanide,and pyocyanin opened AHL molecule will spontaneously
(100). When AiiA is expressed in E. carotovora, undergo ring formation if the environment is
the AHL-controlled production of extracellular not alkaline. Hence, if lactonases are applied,
pectolytic enzymes is abolished. This leads to steps to prevent reversal of the AHL molecules
lowered virulence against potato, eggplant, car- to the active form should be taken.
rots, and many other vegetables where the bac- As may be expected for the cells to be able to
terium causes soft rot disease (34, 35). To respond rapidly to changes in the environment
demonstrate the feasibility of the approach in and requirements for optimal adaptation to
vivo, the aiiA gene was inserted in tobacco and prevailing conditions,P.aeruginosa produces two
potato plants.These plants, which are normally enzymes, PvdQ and QuiP, that degrade AHLs.
prone to attack by E.carotovora, became resistant The two enzymes have different substrate
against the bacterium, which in turn was atten- specificities; PvdQ enables the bacterium to
uated by the quorum-quenching enzymes (33). grow on short-chain AHLs, whereas QuiP is
A variety of other bacteria,including Agrobac- responsible for growth on long-chain AHLs as a
terium tumefaciens,Arthrobacter sp., Klebsiella pneu- carbon and energy source (56, 57, 113). Some
moniae, Commomamonas sp., Rhodococcus sp., and soil bacteria such as Variovorax paradoxus and P.
Streptomyces sp., also produce enzymes that are aeruginosa PAI-A are able to grow and prolifer-
able to interfere with rhizosphere pathogens by ate using AHLs as the sole source of carbon,
producing AHL-degrading enzymes (13,89,90, energy,and nitrogen.The molar growth yield of
122). Expression of AHL-degrading enzymes the bacteria correlates directly with the length
has probably evolved as a survival strategy of of the side chain, indicating that this part of the
signal molecule is used as a carbon source.The well on, for example, a nonpathogenic E. coli
bacteria produce an aminoacylase that cleaves lawn on top of an agar plate.If the E.coli strain is
the amide bond of the signal molecule into an replaced by a pathogenic bacterium such as P.
organic acid (from the side chain) and the HSL aeruginosa, the worms are killed by cyanide and
moiety (from the ring). The acid undergoes phenazines that are secreted by the bacterium.
beta-oxidation and is used as an energy source A QS-deficient mutant only kills 10% (com-
and for biomass formation.By the action of lac- pared to 100% when the worms are feeding on
tonases, ammonium is released from the HSL the wild-type strain).This indicates that QS is
molecule and is subsequently used as a nitrogen indeed important for the infection process in
source by the bacteria (56, 66). C. elegans (22, 67, 96, 118). In a similar way, a
Differentiated epithelial cells are also able to functional RhlIR QS system has been found to
produce enzymes that degrade AHL signal be important for establishing P. aeruginosa infec-
molecules. Such cells produce the enzymes tions in the amoeba Dictyostelium discoideum.
paraoxynase 1, 2, and 3 (PON1 to 3), which are RhlR controls the production of rhamnolipid,
able to degrade 3-oxo-C12-HSL. PON1 to 3 which induces lysis of the amoeba (18).
act as lactonases degrading the AHL by hydrol- A model with more relevance for humans is
ysis of the ring part of the molecule. The the pulmonary infection model in mice. The
enzymes produced by the epithelial cells are initial stages of a chronic lung infection can be
highly specific for the AHLs they degrade. In mimicked by casting P. aeruginosa into seaweed
addition to 3-oxo-C12-HSL,these enzymes can alginate beads and surgically installing them
degrade C6-HSL but intriguingly not 3-oxo- through the trachea into the mouse lung.
C6-HSL and C4-HSL molecules.This indicates Under normal circumstances,the activity of the
that it is both the length of the side chain and its mucociliary escalator clears the lungs of foreign
oxidation state that determine whether the particular matter such as dust and bacteria.The
AHL molecule can be targeted by the enzymes. alginate beads partly impair the function of the
Treatment with purified PON1 enzyme from escalator. Consequently, neutrophils (polymor-
humans was able to abolish biofilm formation phonuclear leukocytes [PMNs]) are then
of P. aeruginosa. Treatment with 0.25% serum recruited to the sites of infection (92). For a
from wild-type mice was also able to abolish short time, this is reminiscent of the situation in
biofilm formation by P. aeruginosa, while serum the CF lung. When the mice are infected with a
from PON1 knockout mice did not have this QS mutant, both mortality of the mice and
ability (16, 49, 88). horizontal spread and dissemination are signifi-
cantly reduced compared to the situation with
QS AS DRUG TARGETS the wild type (7, 107). If rodents are infected
Roughly one-third of the QS-controlled genes with a QS-deficient mutant, the immune
in P. aeruginosa encode known virulence factors response is faster, the PMNs respond with a
such as elastase, alkaline protease, chitinases, stronger oxidative burst, and antibodies accu-
cyanide, phenazines, lectins, and rhamnolipids mulate faster in the infected lungs (117, 132).A
(54, 110, 124, 125) and genes involved in adap- recent transcriptomic analysis strongly suggests
tation to particular environmental conditions, that the wild-type P. aeruginosa further activates
including iron limitation in biofilms (50). Fur- a QS-controlled strategic defense system,
thermore, the importance of QS-dependent which reacts upon the encounter with PMNs
gene expression for bacterial virulence has been (Givskov et al., unpublished data) and sup-
established in several animal models. presses the powerful cellular immune response
One of the simplest infection models is by paralyzing the PMNs (7). Similar to the case
based on the nematode C. elegans.This 1-mm- for P. aeruginosa-amoeba interactions, the main
long worm feeds on bacteria from its surround- factor here is a rhamnolipid that causes necrosis
ings; in a laboratory setting, the worms feed of the incoming PMNs within minutes (59).
Interestingly, only cells carrying a functional may not be feasible for the physician to offer an
QS system are able to mount this aggressive effective treatment of biofilm-based infections
response. (2,3,23,120).As discussed in a previous section,
The relationship between QS and infection the tolerance of P. aeruginosa biofilms is, at least
has been established for several opportunistic in part, dependent on functional QS systems. In
pathogens, including S. liquefaciens (now S. concert with this dependency, it is possible to
marcescens) (47), C. violaceum (11), Burkholderia lower the tolerance of biofilms by treatment
cepacia (131), and Yersinia species (4), all of with QSI compounds. The genes involved
which cause infections in humans. Also, A. remain to be identified (54, 96, 98).
tumefaciens (111) and E. carotovora (129) employ
QS to control infection and virulence. Bacteria QSI Effect in Animal Infection Models
such as Serratia proteamaculans B5a and Enter- Only a few of many QSI compounds have been
obacter agglomerans B6a, which cause food reported to be tested in various infectious in
quality deterioration, utilize QS to control vivo models. The halogenated furanone com-
expression of exoenzymes that are involved in pound 4 is able to interfere with QS-dependent
decay (15, 45). virulence of the marine bioluminescent bac-
Bacterial species have also been demon- terium V. harveyi (69). The bacterium is a pri-
strated to utilize AI-2 QS as regulators of viru- mary pathogen in many aquaculture systems
lence. For example, the locus of enterocyte where invertebrates are raised.V.harveyi uses QS
effacement,a chromosomal pathogenicity island to control both bioluminescence and
that encodes proteins involved in the formation production of a 100-KDa toxin termed T1.
of attaching and effacing lesions by enterohem- Grown in the presence of furanone compound
orrhagic E. coli and enteropathogen E. coli, is 4, both bioluminescence and T1 production by
under QS control. Similarly, it has been shown V. harveyi were reduced in a concentration-
that QS through the integration of several sig- dependent manner. If specimens of P. monodon
naling systems regulates virulence in V. cholerae (black tiger prawn) were injected with cell-free
(76, 135), Vibrio vulnificus (60, 77), V. harveyi, and supernatant of cultures grown with and with-
Vibrio anguillarum (26). Expression of virulence out the furanone compound, a significant
factors by both V. vulnificus and V. cholerae, and reduction in mortality was recorded in the
pathogenicity of V.cholerae to tissue culture cells, group of prawns injected with the treated cul-
was prevented by furanones (78). V. harveyi viru- ture compared to the untreated group. Like-
lence against the black tiger prawn Penaeus mon- wise, if supernatant from the untreated culture
odon (69) and the brine shrimp Artemia was injected into mice, the survival rate was
franciscana (27) was prevented by treatment with 20%, whereas it was 90% in the case of super-
furanones. In other words, inhibition of QS in natant from the furanone-treated culture. This
bacterial pathogens not only will be beneficial indicates that the furanone compound indeed
in a clinical context but also will be applied for interferes with virulence of V. harveyi (69). Fur-
disease control in aquaculture, agriculture, and ther support of the ability of furanones to pre-
food preservation. vent infection by V. harveyi is the fact that
survival of the brine shrimp,Artemia,after expo-
sure to V. harveyi was significantly increased by
IMPACT OF QSIs
treatment with furanones (27).
QSI Effect on Biofilm Persistence Furanone compounds are also able to attenu-
Biofilms are inherently tolerant to the effects of ate P.aeruginosa in a mouse model.Encapsulating
most antimicrobial measures. The concentra- P. aeruginosa harboring a lasB-gfp fusion into
tions that would be expected to eradicate med- alginate beads and installing these beads in the
ically relevant biofilms often exceed the highest lungs of mice allow the study of QSI efficacy in
deliverable doses of antibiotics, and hence, it vivo (Color Plate 16).When the QS systems are
activated, the infecting bacteria express green The halogenated furanones are not the only
fluorescence. About 5 h after injection of fura- QSIs that are functional in animal models. In
none compound 30 or 56 into the tail vein of the nematode worm C. elegans model, 100% of
the mice, the GFP signal was reduced to the the worms are killed when feeding on wild-
level of the noninduced state, indicating that the type P. aeruginosa. If the growth medium is sup-
QS compound has traveled through the blood- plemented with 100 M 4-NPO or 2% QSI
stream,entered the lungs,and blocked QS in the containing garlic extract, the mortality of the
infecting bacteria (51,54).The 5 h is the approx- worms is reduced to 5% and 40%, respectively
imate turnover time of the unstable GFP repor- (96). In a similar fashion in the lung infection
ter used.As the bacteria were also equipped with model in mice, garlic extract reduces mortality
a constitutive red fluorescent tag, it was possible from 72% when mice are untreated to 32%
to determine that the bacteria were not killed or when they are treated with 1.5% of the extract
cleared.At 8 to 10 h after the injection, de novo (96).The clearance of bacteria is also enhanced
GFP expression commenced,indicating that the when the mice are treated with garlic. Mice
furanone compounds have been turned over by subcutaneously treated with 2% QSI contain-
the host organism.This system therefore offers a ing garlic extract (v/w) displayed a bacterial
rough pharmacokinetic estimate of the test load in the lungs of three orders of magnitude
compound. lower than their untreated counterparts already
The model described above also offers an at the second day after start of the infection and
estimate of the effect on bacterial attenuation.If treatment (8). On day 5 postinfection, the
the bacterial load of the infected lungs in mice treated mice had cleared their lungs, whereas
is followed over time, mice treated with fura- the mice in the untreated group harbored 104
none compound (0.7 g/g of body weight) to 105 CFU/lung (8).The QSI compound pat-
(three injections per day with 8-h intervals) had ulin isolated from fungi was also able to lower
a bacterial content 1,000-fold lower than that virulence of P. aeruginosa in the described
of an untreated group on day 5 postinfection. mouse model.When treated with 2.5 g/g of
Also, mortality of infected mice could be body weight once a day, the mice in the treated
reduced by treatment with the furanone com- group showed a 20-fold lower bacterial load in
pounds. In the untreated half of the animals, the lungs on day 3 postinfection.Also, mortality
88% of the mice died, whereas only 55% in the was less than half in the patulin-treated com-
treated group died (53, 54). Taken together, pared to the untreated group (98).
these results indicate that the halogenated fura-
none compounds are able to attenuate P. aerugi- QSI Effects on Components
nosa in vivo. More recently, based on the natural of the Host Immune System
furanones produced by D. pulchra, a structure- A histopathological investigation of mice lungs
function-derived synthetic library of com- revealed that lungs of mice receiving treatment
pounds with strong efficacy as QS blockers in with QSI containing garlic extract were
both in vitro and in vivo animal trials has been inflamed to a much higher degree on day 5
developed by Biosignal Ltd.The in vivo models postinfection (8). In concert with the inflam-
employed included subcutaneous, eye, and pul- mation, a high number of PMNs and mono-
monary lung infection models in which QS- cytes were detected in the lungs of garlic
dependent biofilms were removed by novel extract-treated mice, indicating that the
subclasses of furanone-derived nontoxic QSIs. immune system is actively removing the bacte-
Subclasses of compounds of high QS blocking ria.This fits well with the finding that P. aerugi-
and in vivo efficacy included the series of 1,5- nosa produces rhamnolipid in order to actively
dihydropyrrol-2-ones of the general formula engage and kill the PMNs and possibly other
are shown in Fig. 10. cells from the host immune system.This can be
Perhaps surprisingly, none of the QSIs tested in

this fashion was able to significantly down-
regulate all QS-controlled genes in P.aeruginosa.
Does this disqualify them as true QSI com-
pounds? Rather than answering this question,it
may seem more pertinent to first address which
genes are in fact QS regulated. Surprisingly,
there does not appear to be a simple answer to
this question. Several reports have sought to
identify the QS regulon in P. aeruginosa PAO1
(50, 54, 98, 110, 125, 130). DNA array-based
transcriptomics were employed in all of these
studies, including the comparison of gene
FIGURE 10 The general structure of the series 1.5- expression profiles of the wild-type parent with
dihydropyrrol-2-ones of a novel class of furanone- either a lasRrhlR or lasIrhlI mutant or the lasIrhlI
derived nontoxic QSI. The functional groups are as in the presence or absence of exogenous signal
follows: R1 H; R2 alkyl, aryl; R3 R4 H. Br; molecules. The combined outcome of these
R5 H, aryl.
studies is that a core set of QS-regulated genes
visualized with a P. aeruginosa biofilm present in has been identified. However, many additional
a flow cell.When PMNs are inoculated on top QS genes appear conditionally controlled, i.e.,
of a P. aeruginosa biofilm, the PMNs fail to they are only responsive to QS control under
become activated and develop oxidative burst. certain growth conditions.
In contrast,when the biofilm was formed in the Furthermore, in the emerging picture of P.
presence of QSIs including garlic extract, the aeruginosa QS control biology, target genes are
PMNs developed an oxidative burst (7, 8). Sim- not activated at a certain threshold concentra-
ilarly, the PMNs were unable to graze on the tion. Rather, each gene has its own individual
communicating biofilms, while the QSIs threshold concentration (133). The “AHL-
including garlic extract-treated biofilms were LuxR homologue complex concentration” at
readily phagocytosed by the neutrophiles (7, 8). which the genes are turned on merely depends
Furanone compound 30 and QSI compounds on the DNA sequence,hence the binding affin-
present in garlic extracts block rhamnolipid ity for the “complex”at the promoter site of the
synthesis, which is responsible for the elimina- QS-controlled genes. This means that some
tion of the PMNs (59).If this is true also in vivo, genes are activated at lower “AHL-LuxR
administration of a QSI drug will be expected homologue complex” concentrations and oth-
to tip the balance between biofilm growth and ers,yet at higher concentrations.On the basis of
immune system in favor of the host defense, this, we propose the following: In a normal sit-
which is likely to eliminate the biofilm. uation, the “complex” concentration is simply
dependent on the amount of AHL present, but
DOES A POSITIVE SCREEN OF QSI when the bacteria are treated with QSI com-
BIOACTIVITY EQUAL INHIBITION pounds that prevent activation of the LuxR
OF ALL QS-REGULATED GENES? homologue, the true “complex” concentration
The majority of QSI compounds reported in is lower than indicated by the AHL concentra-
the literature have been identified by their tion. Consequently, those genes that require a
ability to inhibit expression of a single QS- higher “receptor complex” concentration for
controlled promoter. Only a few have been activation are more readily inhibited by QSIs.
investigated by means of DNA array-based Hence, genes that only require a minor “com-
transcriptomics to identify their global effects. plex concentration” for activation might be
more difficult to inhibit. The furanone com- human pathogens (54).The very real need for
pound 30 study by Hentzer et al. (54) supports non-cidal, environmentally benign rather than
this since genes highly activated by the AHL current toxin biofilm-control technologies in,
signals were readily inhibited by the furanone for example, water, sewage, and oil distribution
compound. systems should strongly encourage the devel-
opment of stand-alone or combination-based
CONCLUDING REMARKS
QSI measures.In the medical context,whatever
The biofilm lifestyle where bacteria live in the QSI treatments may be, the development of
close proximity to each other presents ideal pharmacologically relevant molecules must be
conditions for QS control of gene expression. aggressively pursued by clinical trials.For this to
Many opportunistic pathogens utilize QS to happen, serious involvement of the pharma-
control expression of a battery of extracellular ceutical industry is required. Unfortunately, the
virulence factors. In addition, QS is involved in industry focus still seems to be in favor of the
control of tolerance of biofilms to antimicrobial development of broad-range antibiotics with
treatments and to cells of the host immune block-buster capabilities.Interestingly,since the
system,whereas the failure to eradicate biofilm- transcriptomic studies of in vitro biofilms sug-
forming bacteria in medical, domestic, indus- gest the existence of multiple pathways by
trial, and environmental settings calls for the which a biofilm can be built, the use and
development of novel strategies. One such may administration of QS blocking technologies
rely on controlling QS-regulated gene expres- may undergo further development and be
sion. Several strategies to inhibit QS systems combined with treatments directed at targets in
have been investigated; the two most promising additional pathways instrumental for biofilm
strategies are the inhibition by enzymatic development.
destruction of the signal molecule and inhibi-
tion by blocking the signal receptor with small ACKNOWLEDGMENTS
molecules or QSIs. Moreover, the synergistic Our work has received financial support from the Aus-
effect of QSI treatment and conventional tralian Research Council, Centre for Marine Biofoul-
antimicrobials is particularly appealing and may ing and Bio-Innovation, Danish Technical Research
be a first priority in industrial and other non- Council, the Villum Rann Rasmussen Foundation, and
the German Mukoviscidose E.V.
medical settings. However, it remains to be
demonstrated whether this is also achievable in REFERENCES
animal models.An additional interesting feature 1. Allesen-Holm, M., K. B. Barken, L. Yang, M.
of QSI compounds is that they appear to ham- Klausen, J. S.Webb, S. Kjelleberg, S. Molin, M.
per the ability of pathogens to inactivate the Givskov, and T.Tolker-Nielsen. 2006.A charac-
cellular immune system. Experimental docu- terization of DNA release in Pseudomonas aeruginosa
cultures and biofilms. Mol. Microbiol. 59:
mentation for such an effect in vivo remains to 1114–1128.
be obtained, but the inactivation of the cellular 2. Allison, C., H. C. Lai, D. Gygi, and C. Hughes.
immune system may explain why QSI treat- 1993. Cell differentiation of Proteus mirabilis is
ment promotes clearing in the mouse models, initiated by glutamine, a specific chemoattractant
discussed in this chapter (54, 102). for swarming cells. Mol. Microbiol. 8:53–60.
3. Anwar, H., M. K. Dasgupta, and J. W.
The many striking observations reported to Costerton. 1990. Testing the susceptibility of
date in the literature should encourage us to bacteria in biofilms to antibacterial agents. Antimi-
develop future QSI-based antimicrobial and crob.Agents Chemother. 34:2043–2046.
biofilm-control measures. Reports to date 4. Atkinson, S., R. E. Sockett, M. Camara, and P.
speak to the success in employing such Williams. 2006. Quorum sensing and the lifestyle
of Yersinia. Curr. Issues Mol. Biol. 8:1–10.
approaches in the biofilm control of solid sur- 5. Auschill, T. M., N. B. Arweiler, L. Netuschil,
faces and prevention of biofilm and/or disease M. Brecx, E. Reich, and A. Sculean. 2001.
caused by prawn (69), fish (95), plant, and Spatial distribution of vital and dead microor-
ganisms in dental biofilms. Arch. Oral Biol. 46: tein expression in Serratia proteamaculans B5a.
471–476. Microbiology 149:471–483.
6. Belas, R., R. Schneider, and M. Melch. 1998. 16. Chun, C. K., E.A. Ozer, M. J.Welsh, J. Zabner,
Characterization of Proteus mirabilis precocious and E. P. Greenberg. 2004. Inactivation of a
swarming mutants:identification of rsbA,encoding Pseudomonas aeruginosa quorum-sensing signal by
a regulator of swarming behavior. J. Bacteriol. human airway epithelia. Proc. Natl. Acad. Sci. USA
180:6126–6139. 101:3587–3590.
7. Bjarnsholt, T., P. O. Jensen, M. Burmolle, M. 17. Collier, D. N., L.Anderson, S. L. McKnight,T.
Hentzer, J. A. Haagensen, H. P. Hougen, H. L. Noah, M. Knowles, R. Boucher, U.
Calum, K. G. Madsen, C. Moser, S. Molin, N. Schwab, P. Gilligan, and E. C. Pesci. 2002. A
Hoiby, and M. Givskov. 2005. Pseudomonas bacterial cell to cell signal in the lungs of cystic
aeruginosa tolerance to tobramycin, hydrogen fibrosis patients. FEMS Microbiol. Lett. 215:41–46.
peroxide and polymorphonuclear leukocytes 18. Cosson, P., L. Zulianello, O. Join-Lambert, F.
is quorum-sensing dependent. Microbiology 151: Faurisson, L. Gebbie, M. Benghezal, C. Van
373–383. Delden, L. K. Curty, and T. Kohler. 2002.
8. Bjarnsholt, T., P. O. Jensen, T. B. Rasmussen, Pseudomonas aeruginosa virulence analyzed in a Dic-
L. Christophersen, H. Calum, M. Hentzer, H. tyostelium discoideum host system. J. Bacteriol.
P. Hougen, J. Rygaard, C. Moser, L. Eberl, N. 184:3027–3033.
Hoiby, and M. Givskov. 2005. Garlic blocks 19. Costerton, J.W., K. J. Cheng, G. G. Geesey,T.
quorum sensing and promotes rapid clearing of I. Ladd, J. C. Nickel, M. Dasgupta, and T. J.
pulmonary Pseudomonas aeruginosa infections. Marrie. 1987. Bacterial biofilms in nature
Microbiology 151:3873–3880. and disease. Annu. Rev. Microbiol. 41:435–464.
9. Blehert, D. S., R. J. Palmer, Jr., J. B. Xavier, J. 20. Costerton, J. W., Z. Lewandowski, D. E.
S. Almeida, and P. E. Kolenbrander. 2003. Caldwell, D. R. Korber, and H. M.
Autoinducer 2 production by Streptococcus gordonii Lappin-Scott. 1995. Microbial biofilms. Annu.
DL1 and the biofilm phenotype of a luxS Rev. Microbiol. 49:711–745.
mutant are influenced by nutritional conditions. J. 21. Costerton, J. W., P. S. Stewart, and E. P.
Bacteriol. 185:4851–4860. Greenberg. 1999. Bacterial biofilms: a common
10. Borchardt, S.A., E. J.Allain, J. J. Michels, G.W. cause of persistent infections. Science 284:
Stearns, R. F. Kelly, and W. F. McCoy. 2001. 1318–1322.
Reaction of acylated homoserine lactone bacterial 22. Darby, C., C. L. Cosma, J. H.Thomas, and C.
signaling molecules with oxidized halogen antimi- Manoil. 1999. Lethal paralysis of Caenorhabditis
crobials. Appl. Environ. Microbiol. 67:3174–3179. elegans by Pseudomonas aeruginosa. Proc. Natl. Acad.
11. Brito, C. F., C. B. Carvalho, F. Santos, R. T. Sci. USA 96:15202–15207.
Gazzinelli, S. C. Oliveira,V. Azevedo, and S. 23. Davies, D. 2003. Understanding biofilm resist-
M. Teixeira. 2004. Chromobacterium violaceum ance to antibacterial agents. Nat. Rev. Drug Disc.
genome: molecular mechanisms associated with 2:114–122.
pathogenicity. Genet. Mol. Res. 3:148–161. 24. Davies, D. G., M. R. Parsek, J. P. Pearson,
12. Byers, J. T., C. Lucas, G. P. Salmond, and M. B. H. Iglewski, J. W. Costerton, and E. P.
Welch. 2002. Nonenzymatic turnover of an Greenberg. 1998.The involvement of cell-to-cell
Erwinia carotovora quorum-sensing signaling mole- signals in the development of a bacterial biofilm.
cule. J. Bacteriol. 184:1163–1171. Science 280:295–298.
13. Carlier, A., S. Uroz, B. Smadja, R. Fray, X. 25. Davis, A. R., N. M.Targett, O. J. McConnell,
Latour,Y. Dessaux, and D. Faure. 2003.The Ti and C. M.Young. 1989.Epibiosis of marine algae
plasmid of Agrobacterium tumefaciens harbors an and benthic invertebrates: natural products chem-
attM-paralogous gene, aiiB, also encoding N-acyl istry and other mechanisms inhibiting settlement
homoserine lactonase activity. Appl. Environ. and overgrowth. Bioorg. Mar. Chem. 3:86–114.
Microbiol. 69:4989–4993. 26. Defoirdt, T., P. Bossier, P. Sorgeloos, and W.
14. Castang, S., B. Chantegrel, C. Deshayes, R. Verstraete. 2005.The impact of mutations in the
Dolmazon, P. Gouet, R. Haser, S. Reverchon, quorum sensing systems of Aeromonas hydrophila,
W. Nasser, N. Hugouvieux-Cotte-Pattat, and Vibrio anguillarum and Vibrio harveyi on their viru-
A. Doutheau. 2004. N-sulfonyl homoserine lac- lence towards gnotobiotically cultured Artemia
tones as antagonists of bacterial quorum sensing. franciscana. Environ. Microbiol. 7:1239–1247.
Bioorg. Med. Chem. Lett. 14:5145–5149. 27. Defoirdt, T., R. Crab, T. K. Wood, P.
15. Christensen, A. B., K. Riedel, L. Eberl, L. R. Sorgeloos,W.Verstraete, and P. Bossier. 2006.
Flodgaard, S. Molin, L. Gram, and M. Quorum sensing-disrupting brominated fura-
Givskov. 2003. Quorum-sensing-directed pro- nones protect the gnotobiotic brine shrimp
Artemia franciscana from pathogenic Vibrio harveyi, ciens into swarm cells is controlled by the expres-
Vibrio campbellii,and Vibrio parahaemolyticus isolates. sion of the flhD master operon. J. Bacteriol.
Appl. Environ. Microbiol. 72:6419–6423. 178:554–559.
28. de Nys, R., M. Givskov, N. Kumar, S. 41. Geske, G. D., R. J.Wezeman, A. P. Siegel, and
Kjelleberg, and P. Steinberg. 2006. Furanones. H. E. Blackwell. 2005. Small molecule inhibitors
Progr. Mol. Subcell. Biol. 42:55–86. of bacterial quorum sensing and biofilm forma-
29. de Nys, R., P. Steinberg, C. N. Rogers, T. S. tion. J.Am. Chem. Soc. 127:12762–12763.
Charlton, and M. W. Duncan. 1996. Quantita- 42. Ghannoum, M., and G. A. O′Toole. 2004.
tive variation of secondary metabolites in the sea Microbial Biofilms. ASM Press,Washington, DC.
hare Apylsia parvula and its host plant, Delisea 43. Givskov, M., R. de Nys, M. Manefield,
pulchra. Mar. Ecol. Prog. Ser. 130:135–146. L. Gram, R. Maximilien, L. Eberl, S. Molin, P.
30. de Nys, R., P. Steinberg, P. Willemsen, S. A. D. Steinberg, and S. Kjelleberg. 1996. Eukary-
Dworjanyn, C. L. Gabelish, and R. J. King. otic interference with homoserine lactone-
1995. Broad spectrum effects of secondary mediated prokaryotic signalling. J. Bacteriol. 178:
metabolites from the red alga Delisea pulchra in 6618–6622.
antifouling assays. Biofouling 8:259–271. 44. Reference deleted.
31. de Nys, R.,A. D.Wright, G. M. König, and O. 45. Gram, L., A. B. Christensen, L. Ravn, S.
Sticher. 1993. New halogenated furanones from Molin, and M. Givskov. 1999. Production of
the marine alga Delisea pulchra. Tetrahedron 49: acylated homoserine lactones by psychrotrophic
11213–11220. members of the Enterobacteriaceae isolated from
32. Dong, Y. H., A. R. Gusti, Q. Zhang, J. L. foods. Appl. Environ. Microbiol. 65:3458–3463.
Xu, and L. H. Zhang. 2002. Identification of 46. Gram, L., H. P. Grossart, A. Schlingloff, and
quorum-quenching N-acyl homoserine lactonases T. Kiorboe. 2002. Possible quorum sensing in
from bacillus species. Appl. Environ. Microbiol. marine snow bacteria: production of acylated
68:1754–1759. homoserine lactones by Roseobacter strains isolated
33. Dong, Y. H., L. H. Wang, J. L. Xu, H. B. from marine snow. Appl. Environ. Microbiol.
Zhang, X. F. Zhang, and L. H. Zhang. 2001. 68:4111–4116.
Quenching quorum-sensing-dependent bacterial 47. Grimont P.A.D., and F. Grimont. 1978. The
infection by an N-acyl homoserine lactonase. genus Serratia. Ann. Rev. Microbiol. 32:248.
Nature 411:813–817. 48. Hassett, D. J., J. F. Ma, J. G. Elkins, T. R.
34. Dong, Y. H., J. L. Xu, X. Z. Li, and L. H. McDermott, U. A. Ochsner, S. E. West, C. T.
Zhang. 2000.AiiA, an enzyme that inactivates the Huang, J. Fredericks, S. Burnett, P. S.
acylhomoserine lactone quorum-sensing signal Stewart, G. McFeters, L. Passador, and B. H.
and attenuates the virulence of Erwinia carotovora. Iglewski. 1999. Quorum sensing in Pseudomonas
Proc. Natl.Acad. Sci. USA 97:3526–3531. aeruginosa controls expression of catalase and
35. Dong,Y. H., X. F. Zhang, J. L. Xu, and L. H. superoxide dismutase genes and mediates biofilm
Zhang. 2004. Insecticidal Bacillus thuringiensis susceptibility to hydrogen peroxide. Mol. Microbiol.
silences Erwinia carotovora virulence by a new form 34:1082–1093.
of microbial antagonism, signal interference. Appl. 49. Hastings, J. W. 2004. Bacterial quorum-sensing
Environ. Microbiol. 70:954–960. signals are inactivated by mammalian cells. Proc.
36. Donlan, R. M. 2002. Biofilms: microbial life on Natl.Acad. Sci. USA 101:3993–3994.
surfaces. Emerg. Infect. Dis. 8:881–890. 50. Hentzer, M., L. Eberl, and M. Givskov. 2005.
37. Donlan, R. M., and J. W. Costerton. 2002. Transcriptome analysis of Pseudomonas aeruginosa
Biofilms: survival mechanisms of clinically biofilm development: anaerobic respiratoin and
relevant microorganisms. Clin. Microbiol. Rev. 15: iron limitation. Biofilms 2:37–61.
167–193. 51. Hentzer, M., and M. Givskov. 2003.Pharmaco-
38. Dworjanyn, S. A., and Steinberg, P. 1999. logical inhibition of quorum sensing for the treat-
Localisation and surface quantification of second- ment of chronic bacterial infections. J. Clin. Invest.
ary metabolites in the red alga Delisea pulchra. 112:1300–1307.
Marine Biol. 133:727–736. 52. Reference deleted.
39. Eberhard,A.,A. L. Burlingame, C. Eberhard, 53. Hentzer, M., K. Riedel, T. B. Rasmussen, A.
G. L. Kenyon, K. H. Nealson, and N. J. Heydorn, J. B. Andersen, M. R. Parsek, S. A.
Oppenheimer. 1981. Structural identification of Rice, L. Eberl, S. Molin, N. Høiby, S.
autoinducer of Photobacterium fischeri luciferase. Kjelleberg, and M. Givskov. 2002. Inhibition of
Biochemistry 20:2444–2449. quorum sensing in Pseudomonas aeruginosa biofilm
40. Eberl, L., G. Christiansen, S. Molin, and M. bacteria by a halogenated furanone compound.
Givskov. 1996. Differentiation of Serratia liquefa- Microbiology 148:87–102.
54. Hentzer, M., H.Wu, J. B.Andersen, K. Riedel, oral multi-species biofilms: bacterial communities
T. B. Rasmussen, N. Bagge, N. Kumar, M. A. in health and disease, p. 175–193. In S. Kjelleberg
Schembri, Z. Song, P. Kristoffersen, M. and M. Givskov (ed.), The Biofilm Mode of Life.
Manefield, J.W. Costerton, S. Molin, L. Eberl, Horizon Scientific Press, Norwich, United King-
P. Steinberg, S. Kjelleberg, N. Hoiby, and M. dom.
Givskov. 2003.Attenuation of Pseudomonas aerugi- 66. Leadbetter, J. R., and E. P. Greenberg. 2000.
nosa virulence by quorum sensing inhibitors. Metabolism of acyl-homoserine lactone quorum-
EMBO J. 22:3803–3815. sensing signals by Variovorax paradoxus. J. Bacteriol.
55. Hope, C. K., D. Clements, and M. Wilson. 182:6921–6926.
2002. Determining the spatial distribution of 67. Mahajan-Miklos, S., M.W.Tan, L. G. Rahme,
viable and nonviable bacteria in hydrated micro- and F. M.Ausubel. 1999. Molecular mechanisms
cosm dental plaques by viability profiling. J. Appl. of bacterial virulence elucidated using a
Microbiol. 93:448–455. Pseudomonas aeruginosa-Caenorhabditis elegans path-
56. Huang, J. J., J. I. Han, L. H. Zhang, and J. R. ogenesis model. Cell 96:47–56.
Leadbetter. 2003.Utilization of acyl-homoserine 68. Manefield, M., R. de Nys, N. Kumar,
lactone quorum signals for growth by a soil R. Read, M. Givskov, P. Steinberg, and S.
pseudomonad and Pseudomonas aeruginosa PAO1. Kjelleberg. 1999. Evidence that halogenated
Appl. Environ. Microbiol. 69:5941–5949. furanones from Delisea pulchra inhibit acylated
57. Huang, J. J.,A. Petersen, M.Whiteley, and J. R. homoserine lactone (AHL)-mediated gene
Leadbetter. 2006. Identification of QuiP, the expression by displacing the AHL signal from its
product of gene PA1032, as the second acyl- receptor protein. Microbiology 145:283–291.
homoserine lactone acylase of Pseudomonas aerugi- 69. Manefield, M., L. Harris, S. A. Rice, R. de
nosa PAO1.Appl.Environ.Microbiol. 72:1190–1197. Nys, and S. Kjelleberg. 2000. Inhibition of
58. Jacques, M. A., K. Josi, A. Darrasse, and luminescence and virulence in the black tiger
R. Samson. 2005. Xanthomonas axonopodis pv. prawn (Penaeus monodon) pathogen Vibrio harveyi
phaseoli var. fuscans is aggregated in stable biofilm by intercellular signal antagonists. Appl. Environ.
population sizes in the phyllosphere of field- Microbiol. 66:2079–2084.
grown beans. Appl. Environ. Microbiol. 71: 2008– 70. Manefield, M., and S. L.Turner. 2002.Quorum
2015. sensing in context: out of molecular biology and
59. Jensen, P. O., T. Bjarnsholt, R. Phipps, T. B. into microbial ecology. Microbiology 148:3762–
Rasmussen, H. Calum, C. Christoffersen, C. 3764.
Moser, P. Williams, T. Pressler, M. Givskov, 71. Manefield, M.,T. B. Rasmussen, M. Henzter,
and N. Hoiby. 2007. Rapid necrotic killing of J. B.Andersen, P. Steinberg, S. Kjelleberg, and
PMNs is caused by quorum sensing controlled M. Givskov. 2002. Halogenated furanones inhibit
production of rhamnolipid by Pseudomonas aerugi- quorum sensing through accelerated LuxR
nosa. Microbiology 151:1329–1338. turnover. Microbiology 148:1119–1127.
60. Kim, S.Y., S. E. Lee,Y. R. Kim, C. M. Kim, P. 72. Martinelli, D., G. Grossmann, U. Sequin, H.
Y. Ryu, H. E. Choy, S. S. Chung, and J. H. Brandl, and R. Bachofen. 2004. Effects of natu-
Rhee. 2003. Regulation of Vibrio vulnificus viru- ral and chemically synthesized furanones on quo-
lence by the LuxS quorum-sensing system. Mol. rum sensing in Chromobacterium violaceum. BMC
Microbiol. 48:1647–1664. Microbiol. 4:25.
61. Kjelleberg, S., and M. Givskov. 2007. The 73. Matz, C., T. Bergfeld, S. A. Rice, and S.
Biofilm Mode of Life. Horizon Scientific Press, Kjelleberg. 2004.Microcolonies,quorum sensing
Norwich, United Kingdom. and cytotoxicity determine the survival of
62. Kjelleberg, S., P. Steinberg, M. Givskov, L. Pseudomonas aeruginosa biofilms exposed to proto-
Gram, M. Manefield, and R. de Nys. 1997. Do zoan grazing. Environ. Microbiol. 6:218–226.
marine natural products interfere with prokaryotic 74. Maximilien, R., R. de Nys, C. Holmstrom, L.
AHL regulatory systems? Aquat. Microb. Ecol. Gram, M. Givskov, S. Kjelleberg, and P.
13:85–93. Steinberg. 1998.Chemical mediation of bacterial
63. Reference deleted. surface colonisation by secondary metabolites
64. Koch, B.,T. Liljefors,T. Persson, J. Nielsen, S. from the red alga Delisea pulchra. Aquat. Microb.
Kjelleberg, and M. Givskov. 2005. The LuxR Ecol. 15:233–246.
receptor: the sites of interaction with quorum 75. McDougald, D.,W. H. Lin, S. A. Rice, and S.
sensing signals and inhibitors. Microbiology 151: Kjelleberg. 2006.The role of quorum sensing and
3589–3602. the effect of environmental conditions on biofilm
65. Kolenbrander, P. E., N. S. Jakubovics, N. I. formation by strains of Vibrio vulnificus. Biofouling
Chalmers, and R. J. Palmer, Jr. 2007. Human 22:133–144.
76. McDougald, D., S.A. Rice, and S. Kjelleberg. Zabner. 2005. Human and murine paraoxonase 1
2000. The marine pathogen Vibrio vulnificus are host modulators of Pseudomonas aeruginosa quo-
encodes a putative homologue of the Vibrio harveyi rum-sensing. FEMS Microbiol. Lett. 253:29–37.
regulatory gene, luxR: a genetic and phylogenetic 89. Park, S.Y., H. O. Kang, H. S. Jang, J. K. Lee, B.
comparison. Gene 248:213–221. T. Koo, and D. Y. Yum. 2005. Identification of
77. McDougald, D., S.A. Rice, and S. Kjelleberg. extracellular N-acylhomoserine lactone acylase
2001. SmcR-dependent regulation of adaptive from a Streptomyces sp. and its application to quo-
phenotypes in Vibrio vulnificus. J. Bacteriol. 183: rum quenching. Appl. Environ. Microbiol. 71:
758–762. 2632–2641.
78. McDougald, D., S.A. Rice, and S. Kjelleberg. 90. Park, S.Y., S. J. Lee, T. K. Oh, J. W. Oh, B. T.
2007. Bacterial quorum sensing and interference Koo, D.Y.Yum, and J. K. Lee. 2003.AhlD,an N-
by naturally occurring biomimics. Anal. Bioanal. acylhomoserine lactonase in Arthrobacter sp., and
Chem. 387:445–453. predicted homologues in other bacteria. Microbiol-
79. McDougald, D., S. Srinivasan, S.A. Rice, and ogy 149:1541–1550.
S. Kjelleberg. 2003. Signal-mediated cross-talk 91. Parsek, M. R., and P. K. Singh. 2003. Bacterial
regulates stress adaptation in Vibrio species. Microbi- biofilms: an emerging link to disease pathogenesis.
ology 149:1923–1933. Annu. Rev. Microbiol. 57:677–701.
80. McLean, R. J., M.Whiteley, D. J. Stickler, and 92. Pedersen, S. S., G. H. Shand, B. L. Hansen,
W. C. Fuqua. 1997. Evidence of autoinducer and G. N. Hansen. 1990. Induction of experi-
activity in naturally occurring biofilms. FEMS mental chronic Pseudomonas aeruginosa lung infec-
Microbiol. Lett. 154:259–263. tion with P. aeruginosa entrapped in alginate
81. Monier, J. M., and S. E. Lindow. 2005.Aggre- microspheres. APMIS 98:203–211.
gates of resident bacteria facilitate survival of 93. Persson,T.,T. H. Hansen,T. B. Rasmussen, M.
immigrant bacteria on leaf surfaces. Microb. Ecol. E. Skinderso, M. Givskov, and J. Nielsen.
49:343–352. 2005. Rational design and synthesis of new
82. Monier, J. M., and S. E. Lindow. 2005. Spatial quorum-sensing inhibitors derived from acylated
organization of dual-species bacterial aggregates homoserine lactones and natural products from
on leaf surfaces. Appl. Environ. Microbiol. 71: garlic. Org. Biomol. Chem. 3:253–262.
5484–5493. 94. Queck, S.Y., M.Weitere, A. M. Moreno, S. A.
83. Morris, C. E., J. M. Monier, and M.A. Jacques. Rice, and S. Kjelleberg. 2006.The role of quo-
1998. A technique to quantify the population size rum sensing mediated developmental traits in the
and composition of the biofilm component in resistance of Serratia marcescens biofilms against pro-
communities of bacteria in the phyllosphere. Appl. tozoan grazing. Environ. Microbiol. 8:1017-1025.
Environ. Microbiol. 64:4789–4795. 95. Rasch, M., C. Buch, B. Austin, W. J.
84. Muh, U., B. J. Hare, B. A. Duerkop, M. Slierendrecht, K. S. Ekmann, J. L. Larsen, C.
Schuster, B. L. Hanzelka, R. Heim, E. R. Johansen, K. Riedel, L. Eberl, M. Givskov, and
Olson, and E. P. Greenberg. 2006. A struc- L. Gram. 2004.An inhibitor of bacterial quorum
turally unrelated mimic of a Pseudomonas aeruginosa sensing reduces mortalities caused by vibriosis in
acyl-homoserine lactone quorum-sensing signal. rainbow trout (Oncorhynchus mykiss, Walbaum).
Proc. Natl.Acad. Sci. USA 103:16948–16952. Syst.Appl. Microbiol. 27:350–359.
85. Muh, U., M. Schuster, R. Heim, A. Singh, E. 96. Rasmussen, T. B., T. Bjarnsholt, M. E.
R. Olson, and E. P. Greenberg. 2006. Novel Skindersoe, M. Hentzer, P. Kristoffersen, M.
Pseudomonas aeruginosa quorum-sensing inhibitors Kote, J. Nielsen, L. Eberl, and M. Givskov.
identified in an ultra-high-throughput screen. 2005. Screening for quorum-sensing inhibitors
Antimicrob.Agents Chemother. 50:3674–3679. (QSI) by use of a novel genetic system, the QSI
86. Newman, D. J., G. M. Cragg, and K. M. selector. J. Bacteriol. 187:1799–1814.
Snader. 2003. Natural products as sources of new 97. Rasmussen, T. B., M. Manefield, J. B.
drugs over the period 1981–2002. J. Nat. Prod. Andersen, L. Eberl, U. Anthoni, C.
66:1022–1037. Christophersen, P. Steinberg, S. Kjelleberg,
87. Olsen, J. A., R. Severinsen, T. B. Rasmussen, and M. Givskov. 2000. How Delisea pulchra fura-
M. Hentzer, M. Givskov, and J. Nielsen. nones affect quorum sensing and swarming motil-
2002. Synthesis of new 3- and 4-substituted ana- ity in Serratia liquefaciens MG1. Microbiology
logues of acyl homoserine lactone quorum sensing 146:3237–3244.
autoinducers. Bioorg. Med. Chem. Lett. 12: 98. Rasmussen,T. B., M. E. Skindersoe,T. Bjarn-
325–328. sholt, R. K. Phipps, K. B. Christensen, P. O.
88. Ozer, E. A., A. Pezzulo, D. M. Shih, C. Chun, Jensen, J. B. Andersen, B. Koch,T. O. Larsen,
C. Furlong,A. J. Lusis, E. P. Greenberg, and J. M. Hentzer, L. Eberl, N. Hoiby, and M.
Givskov. 2005. Identity and effects of quorum- predation in an experimental microbial commu-
sensing inhibitors produced by Penicillium nity. Aquat. Microb. Ecol. 38:215–229.
species. Microbiology 151:1325–1340. 109. Schaefer,A. L., B. L. Hanzelka,A. Eberhard,
99. Reichelt, J. L., and M. A. Borowitzka. 1984. and E. P. Greenberg. 1996. Quorum sensing in
Antimicrobial activity from marine algae: results Vibrio fischeri: probing autoinducer-LuxR inter-
of a large-scale screening programme. Hydrobiol- actions with autoinducer analogs. J. Bacteriol.
ogy 116/117:158–168. 178:2897–2901.
100. Reimmann, C., N. Ginet, L. Michel, C. Keel, 110. Schuster, M., C. P. Lostroh,T. Ogi, and E. P.
P. Michaux, V. Krishnapillai, M. Zala, K. Greenberg. 2003. Identification, timing, and
Heurlier, K. Triandafillu, H. Harms, G. signal specificity of Pseudomonas aeruginosa
Defago, and D. Haas. 2002. Genetically pro- quorum-controlled genes: a transcriptome analy-
grammed autoinducer destruction reduces viru- sis. J. Bacteriol. 185:2066–2079.
lence gene expression and swarming motility in 111. Sheng, J., and V. Citovsky. 1996. Agrobac-
Pseudomonas aeruginosa PAO1. Microbiology 148: terium-plant cell DNA transport: have virulence
923–932. proteins, will travel. Plant Cell 8:1699–1710.
101. Ren, D., L. A. Bedzyk, R. W. Ye, S. M. 112. Singh, P. K., A. L. Schaefer, M. R. Parsek,T.
Thomas, and T. K. Wood. 2004. Differential O. Moninger, M. J. Welsh, and E. P.
gene expression shows natural brominated fura- Greenberg. 2000.Quorum-sensing signals indi-
nones interfere with the autoinducer-2 bacterial cate that cystic fibrosis lungs are infected with
signaling system of Escherichia coli. Biotechnol. Bio- bacterial biofilms. Nature 407:762–764.
eng. 88:630–642. 113. Sio, C. F., L. G. Otten, R. H. Cool, S. P.
102. Ren, D., J. J. Sims, and T. K. Wood. 2001. Diggle, P. G. Braun, R. Bos, M. Daykin, M.
Inhibition of biofilm formation and swarming of Camara, P. Williams, and W. J. Quax. 2006.
Escherichia coli by (5Z)-4-bromo-5-(bromometh- Quorum quenching by an N-acyl-homoserine
ylene)-3-butyl-2(5H)-furanone. Environ. Micro- lactone acylase from Pseudomonas aeruginosa
biol. 3:731–736. PAO1. Infect. Immun. 74:1673–1682.
103. Reverchon, S., B. Chantegrel, C. Deshayes, 114. Slattery, M., J. B. McClintocj, and J. N.
A. Doutheau, and N. Cotte-Pattat. 2002. Heine. 1995.Chemical defences in Antarctic soft
New synthetic analogues of N-acyl homoserine corals: evidence for antifouling compounds. J.
lactones as agonists or antagonists of transcrip- Exp. Mar. Biol. Ecol. 190:61–77.
tional regulators involved in bacterial quorum 115. Smith, K. M., Y. Bu, and H. Suga. 2003.
sensing. Bioorg. Med. Chem. Lett. 12:1153–1157. Induction and inhibition of Pseudomonas aerugi-
104. Rice, S. A., M. Givskov, P. Steinberg, and S. nosa quorum sensing by synthetic autoinducer
Kjelleberg. 1999. Bacterial signals and anta- analogs. Chem. Biol. 10:81–89.
gonists: the interaction between bacteria and 116. Smith, K. M., Y. Bu, and H. Suga. 2003.
higher organisms. J. Mol. Microbiol. Biotechnol. Library screening for synthetic agonists and
1:23–31. antagonists of a Pseudomonas aeruginosa autoin-
105. Rickard, A. H., R. J. Palmer, Jr., D. S. ducer. Chem. Biol. 10:563–571.
Blehert, S. R. Campagna, M. F. Semmel- 117. Smith, R. S., S. G. Harris, R. Phipps, and B.
hack, P. G. Egland, B. L. Bassler, and P. E. Iglewski. 2002.The Pseudomonas aeruginosa quo-
Kolenbrander. 2006.Autoinducer 2: a concen- rum-sensing molecule N-(3- oxododecanoyl)
tration-dependent signal for mutualistic bacterial homoserine lactone contributes to virulence and
biofilm growth. Mol.Microbiol. 60:1446–1456. induces inflammation in vivo. J. Bacteriol. 184:
106. Riedel, K., M. Kothe, B. Kramer,W. Saeb,A. 1132–1139.
Gotschlich, A. Ammendola, and L. Eberl. 118. Tan, M. W., S. Mahajan-Miklos, and F. M.
2006. Computer-aided design of agents that Ausubel. 1999. Killing of Caenorhabditis elegans
inhibit the cep quorum-sensing system of Burk- by Pseudomonas aeruginosa used to model mam-
holderia cenocepacia. Antimicrob. Agents Chemother. malian bacterial pathogenesis. Proc. Natl.Acad. Sci.
50:318–323. USA 96:715–720.
107. Rumbaugh, K. P., J. A. Griswold, B. H. 119. Taylor, M.W., P. J. Schupp, H. J. Baillie,T. S.
Iglewski, and A. N. Hamood. 1999. Contri- Charlton, R. de Nys, S. Kjelleberg, and P. D.
bution of quorum sensing to the virulence of Steinberg. 2004. Evidence for acyl homoserine
Pseudomonas aeruginosa in burn wound infections. lactone signal production in bacteria associated
Infect. Immun. 67:5854–5862. with marine sponges. Appl. Environ. Microbiol.
108. Salcher, M. M., R. Psenner, J. Pernthaler, 70:4387–4389.
and T. Posch. 2005. Succession of bacterial 120. Teitzel, G. M., and M. R. Parsek. 2003.Heavy
grazing defense mechanisms against protistan metal resistance of biofilm and planktonic
Pseudomonas aeruginosa. Appl. Environ. Microbiol. L. Everson, A. K. Harris, C. L. Pemberton,

69:2313–2320. N. J. Simpson, H. Slater, D. S. Smith, M.
121. Teplitski, M., J. B. Robinson, and W. D. Welch, N. Williamson, and G. P. Salmond.
Bauer. 2000. Plants secrete substances that 2002.The regulation of virulence in phytopatho-
mimic bacterial N-acyl homoserine lactone sig- genic Erwinia species: quorum sensing, antibiotics
nal activities and affect population density- and ecological considerations. Antonie Van
dependent behaviors in associated bacteria. Mol. Leeuwenhoek 81:223–231.
Plant-Microbe Interact. 13:637–648. 130. Whiteley, M., K. M. Lee, and E. P.
122. Uroz, S., C. D’Angelo-Picard,A. Carlier, M. Greenberg. 1999. Identification of genes con-
Elasri, C. Sicot, A. Petit, P. Oger, D. Faure, trolled by quorum sensing in Pseudomonas aerugi-
and Y. Dessaux. 2003.Novel bacteria degrading nosa. Proc. Natl.Acad. Sci. USA 96:13904–13909.
N-acylhomoserine lactones and their use as 131. Wopperer, J., S.T. Cardona, B. Huber, C. A.
quenchers of quorum-sensing-regulated func- Jacobi, M. A.Valvano, and L. Eberl. 2006. A
tions of plant-pathogenic bacteria. Microbiology quorum-quenching approach to investigate the
149:1981–1989. conservation of quorum-sensing-regulated func-
123. Van Delden, C., and B. H. Iglewski. 1998. tions within the Burkholderia cepacia complex.
Cell-to-cell signaling and Pseudomonas aeruginosa Appl. Environ. Microbiol. 72:1579–1587.
infections. Emerg. Infect. Dis. 4:551–560. 132. Wu, H., Z. Song, M. Givskov, G. Doring, D.
124. Vasil, M. L. 2003. DNA microarrays in analysis Worlitzsch, K. Mathee, J. Rygaard, and N.
of quorum sensing: strengths and limitations. J. Hoiby. 2001. Pseudomonas aeruginosa mutations
Bacteriol. 185:2061–2065. in lasI and rhlI quorum sensing systems result in
125. Wagner,V. E., D. Bushnell, L. Passador, A. I. milder chronic lung infection. Microbiology
Brooks, and B. H. Iglewski. 2003. Microarray 147:1105–1113.
analysis of Pseudomonas aeruginosa quorum- 133. Yarwood, J. M., E. M. Volper, and E. P.
sensing regulons: effects of growth phase and Greenberg. 2005. Delays in Pseudomonas aerugi-
environment. J. Bacteriol. 185:2080–2095. nosa quorum-controlled gene expression are
126. Wahl, M. 2003. Marine epibiosis. Fouling and conditional. Proc. Natl. Acad. Sci. USA 102:
antifouling: some basic aspects. Mar. Ecol. Prog. Ser. 9008–9013.
58:175–189. 134. Yates, E.A., B. Philipp, C. Buckley, S.Atkin-
127. Walker, T. S., H. P. Bais, E. Deziel, H. P. son, S. R. Chhabra, R. E. Sockett, M. Gold-
Schweizer, L. G. Rahme, R. Fall, and J. M. ner,Y. Dessaux, M. Camara, H. Smith, and
Vivanco. 2004.Pseudomonas aeruginosa-plant root P. Williams. 2002. N-acylhomoserine lactones
interactions. Pathogenicity, biofilm formation, undergo lactonolysis in a pH-, temperature-, and
and root exudation. Plant Physiol. 134:320–331. acyl chain length-dependent manner during
128. Wang, L. H., L. X.Weng,Y. H. Dong, and L. growth of Yersinia pseudotuberculosis and Pseudo-
H. Zhang. 2004. Specificity and enzyme kinet- monas aeruginosa. Infect. Immun. 70:5635–5646.
ics of the quorum-quenching N-acyl homoser- 135. Zhu, J., M. B. Miller, R. E. Vance, M.
ine lactone lactonase (AHL-lactonase). J. Biol. Dziejman, B. L. Bassler, and J. J. Mekalanos.
Chem. 279:13645–13651. 2002. Quorum-sensing regulators control
129. Whitehead, N. A., J. T. Byers, P. virulence gene expression in Vibrio cholerae. Proc.
Commander, M. J. Corbett, S. J. Coulthurst, Natl.Acad. Sci. USA 99:3129–3134.
EUKARYOTIC QUORUM
SENSING AND
INTERACTIONS WITH
QUORUM-SENSING
BACTERIA
IV
INTERDOMAIN CROSS TALK
Carla Cugini, Roberto Kolter, and Deborah A. Hogan
26
Diffusible signals are used to transmit informa- tionships that have been characterized, to vary-
tion between cells within a single organism and ing extents, at the molecular level. Some pre-
among organisms within a population. In mul- liminary findings that indicate the potential for
ticellular organisms, chemical signals allow signaling between bacteria and archaea are also
global responses to local stimuli as demon- discussed.
strated by the action of hormones or the multi- The identification of a bona fide signaling
faceted immune responses that occur in interaction is complicated by the potential of
reaction to injury or infection. In single-celled less-specific responses resulting from the physi-
organisms, extracellular signals coordinate dif- cal or chemical properties of the signaling
ferent population-level behaviors such as the molecules. For example, long-chain acyl-
formation of biofilms or sporulating fruiting homoserine lactones (AHLs) produced by
bodies (13, 36, 48, 73). It is now becoming gram-negative bacteria can have surfactant
increasingly apparent that some of the chemical activity (11), and the Pseudomonas aeruginosa
signals first identified for their roles within a quinolone signal (PQS) can also chelate iron (6,
species are also used to convey information 15). The identification of specific interspecies
across the domains of life.The study of interdo- signaling interactions will come as the receptors
main signaling has mainly focused on its role in and corresponding signal transduction path-
interactions between bacteria and eukarya.The ways are described in more detail.These molec-
processes found to involve interdomain signal- ular analyses remain one of the key challenges
ing range from proper eukaryotic organ devel- in the field.
opment to the formation of multispecies At the outset we wish to define four signal-
microbial communities (33, 44). This chapter ing categories that describe the different inter-
emphasizes those interdomain signaling rela- domain signaling mechanisms that we will
discuss in this chapter (Fig. 1). These mecha-
Carla Cugini and Deborah A. Hogan Department of Micro- nisms are: (i) one-way sensing—one organism
biology and Immunology, Dartmouth Medical School, senses and responds to a diffusible signal pro-
Hanover, New Hampshire 03755. Roberto Kolter
Department of Microbiology and Molecular Genetics, duced by a second organism; (ii) co-opting of a
Harvard Medical School, Boston, Massachusctts 02115. signal—one organism uses the signal produced
419
420 ■ CUGINI ET AL.
FIGURE 1 Different types of interdomain signaling interactions. (A) One-way communication describes
instances in which one organism responds to a signal produced by another organism.(B) Co-opting of a signal occurs
when one organism specifically uses the signaling molecule of another to regulate gene expression. (C) Modulation
of a signal by another organism can either stimulate or dampen the elicited response. (D) Two-way communication
describes a chemical conversation between two organisms.The signal producer is depicted as the shaded form.
by another to regulate its own gene expression; sensing. One-way sensing is exemplified by
(iii) modulation of a signal—one organism alters mammalian detection of microbial signaling
the production or stability of a signal from molecules and microbial chemotaxis either
another organism; and (iv) two-way communica- toward or away from signal-producing popula-
tion—multiple signals are exchanged between tions of bacteria (4, 60, 74).
organisms.While these classifications are useful A number of the one-way signaling interac-
in thinking about the different interactions that tions involve quorum-sensing molecules first
can occur between organisms, the categoriza- identified for their role in communication
tion of any one interaction is largely defined by within single species populations. For example,
the current state of knowledge. What might P. aeruginosa, an opportunistic bacterial
appear to be a one-way signaling interaction pathogen, regulates the expression of hundreds
today may, in fact, be revealed to be part of a of its genes via extracellular quorum-sensing
complex, two-way conversation upon further signals (58), and these same quorum-sensing
study. signals also participate in interdomain signaling
(62, 71). P. aeruginosa produces two AHL signals,
ONE-WAY SENSING 3-oxo-C12-homoserine lactone (3-O-C12-
OF CELLULAR SIGNALS HSL) and C4-homoserine lactone (C4-HSL)
In many instances, one organism responds to a (Fig. 2) (50). The 3-O-C12-HSL molecule has
signal produced by another organism (Fig. 1A) been shown to accumulate during P. aeruginosa
in a relationship that we refer to as one-way human infections and in biofilms formed in
26. INTERDOMAIN CROSS TALK ■ 421
FIGURE 2 Diffusible molecules that participate in interdomain signaling interactions.The compounds on the left
are bacterially derived; eukaryotic organisms produce the molecules on the right.The R group on the AHL and 3-
O-AHL represents acyl chains from C4 to C16.
vitro (8, 21).When exposed to 3-O-C12-HSL, cells and can sometimes cause alterations of cal-
but not to other AHLs, epithelial cells induce cium levels leading to apoptosis (59, 75).Taken
the expression of several molecules that partici- together these results suggest that 3-O-C12-
pate in the inflammatory response, such as HSL is able to elicit different effects on cells,
interleukin-8 (IL-8) (17, 60, 61, 63).Thus, the from specific immunomodulatory effects to the
production of 3-O-C12-HSL by P. aeruginosa alteration of endogenous signaling leading to
may contribute to the massive IL-8-mediated differing outcomes depending on the cell type.
neutrophil infiltration that is often observed Several interactions have been described
during chronic P. aeruginosa infections such as where one organism senses a quorum-sensing
those found in the lungs of patients with cystic signal from another organism and uses this info-
fibrosis (CF) (5).The mammalian host may be rmation to direct movement.When zoospores
sensing bacterially produced 3-O-C12-HSL as a produced by the seaweed Ulva intestinalis swim
signal indicating the presence of P. aeruginosa. away from the parent plant, they preferentially
Separate studies with different mammalian cells settle on surfaces that are already colonized by
have shown that 3-O-C12-HSL can enter the Vibrio anguillarum rather than on uncolonized
surfaces (32, 74). V. anguillarum produces several environment where it is in constant contact
AHLs including 3-oxo-C10-homoserine lac- with various microbes and where distinguish-
tone (3-O-C10-HSL), C6-homoserine lactone ing food source from pathogen is critical.
(C6-HSL), and 3-hydroxy- C6 homoserine lac-
tone (3-OH-C6-HSL). These AHLs cause a CO-OPTING OF A SIGNAL
reduction in zoospore swimming speed with- The following section focuses on two instances
out changing its orientation, thereby promot- where one organism uses, or co-opts, a secreted
ing zoospore colocalization with bacteria (33, cellular signal from another organism to regu-
45, 46). Non-AHL-producing strains did not late its own gene expression (Fig. 1B). In these
result in an alteration of swimming speed (33, examples, two bacteria, P. aeruginosa and entero-
66).Addition of the AHL lactonase,AiiA,which hemorrhagic E. coli (EHEC), respond to host
degrades AHLs, disrupted the AHL-induced immune factors by modulating the production
settling of the zoospores, further demonstrating of their own quorum signals and/or the
that AHLs play a role in zoospore and bacteria expression and production of quorum-sensing-
colocalization (18, 66). In addition, purified controlled factors.
AHLs did not alter U. intestinalis phototaxis, P. aeruginosa can cause a number of acute and
which also requires swimming motility. This chronic infections. In the course of these infec-
indicates that the effects of the AHLs are not tions,it interacts with different types of cells and
due to nonspecific effects on swimming (74). signaling molecules from the immune system.
The biological reason for why U. intestinalis has As a consequence, the bacterium appears to
a pathway that promotes colocalization with have developed ways to co-opt and utilize a
surface-associated bacteria remains unknown. number of eukaryotic factors by specifically
Caenorhabditis elegans, a free-living terrestrial recognizing them. Wu et al. found that T-cell
nematode that feeds on bacteria, can also direct culture supernatants specifically increase the
its movement in response to AHLs. In the labo- expression of the quorum-sensing-regulated
ratory setting, some strains of bacteria, among gene,lecA,which encodes a cytotoxic lectin and
them Escherichia coli, are innocuous to the adhesin (PA-1 lectin) (16, 76, 77).The increase
nematodes and serve as a food source. Other in lecA expression is due to the specific binding
bacteria, including P. aeruginosa, are pathogenic of a host cytokine, interferon- (INF-), to the
to the nematode (39).When exposed to P.aerug- outer membrane protein OprF, which results in
inosa for the first time, C. elegans exhibits stimulation of the quorum-sensing network
chemotaxis toward the bacteria (4). The without stimulating overall growth (1, 77).
chemoattractants sensed by the nematode INF-, a cytokine that is secreted by different
include the AHLs (3-O-C12-HSL and C4- immune cells, participates in a variety of signal
HSL) produced by P. aeruginosa. Upon repeated transduction pathways involved in activating
exposure to these molecules, however, C. ele- cells in response to potential pathogens (52,67).
gans displays a learned,aversive olfactory behav- The ability of P. aeruginosa to bind, sense, and
ior that allows the nematode to avoid this respond to this factor is remarkable and may
potential pathogen (4, 53, 81). Worms do not indicate a way that P. aeruginosa up-regulates
exhibit negative chemotaxis when presented virulence factors and destroys host cells over the
with P. aeruginosa mutants defective in AHL course of an infection. As lecA is a quorum-
production, nor are worm mutants with olfac- sensing-controlled gene, and other quorum-
tory defects repelled by 3-O-C12-HSL (4). C. sensing-controlled genes are important for P.
elegans is thus able to differentiate a pathogen aeruginosa virulence (58), it will be interesting
from a food source based on the detection of to learn whether there are increases in other
AHLs through adaptation. This must have quorum-sensing-regulated factors in response
important implications for the nematode in the to IFN-. A recent report by Zaborina et al.
shows that P. aeruginosa increases virulence fac- virulence genes in the context of mammalian
tor production in response to another host fac- hosts. Disruption of this cascade in EHEC may
tor, dynorphin A, a 17-amino-acid opioid that protect against disease.
can be found in the intestine upon stress or
injury (79). Immunostaining with antidynor- MODULATION AND ALTERATION
phin antibodies shows that dynorphin A enters OF CELLULAR SIGNALS
the cytoplasm of P. aeruginosa, and a variety of Interdomain signaling interactions can involve
virulence assays correlate the exposure to modulation of cell signals by stimulating signal
dynorphin with increased virulence factor pro- production or through signal interference
duction (79). The increased virulence in the (Fig. 1C). In many cases, quorum sensing has
presence of dynorphin is due to induction of been shown to regulate the expression of
the pqs biosynthetic operon, which leads to an virulence genes. Presumably to counteract
increased production of hydroxy-alkyl- antagonistic microbes, some eukaryotes pro-
quinolones, including the PQS quorum-sens- duce compounds that can inhibit bacterial
ing molecule, which regulates a number of quorum sensing (55, 80). A number of benefi-
cytotoxic factors (78, 79). Because P. aeruginosa cial plant-bacterial associations appear to pro-
increases its virulence factor production in mote quorum sensing for the proper regulation
response to host factors produced during a host of this specialized interdomain interaction (24).
response to damage or injury, it appears that the Algae have been shown to produce AHL-
bacterium has evolved ways to survive eradica- mimics that alter quorum-sensing-regulated
tion by the host immune system. gene expression in bacteria. Delisea pulchra, a
EHEC, which causes bloody diarrhea and benthic marine macro-alga, was found to inter-
hemolytic uremic syndrome, also up-regulates fere with AHL signaling in gram-negative bac-
virulence-related genes in response to host sig- teria (14, 23, 54) by producing furanones (Fig.
naling molecules. EHEC produces two signal- 2) that antagonize AHL-mediated quorum-
ing molecules, the furanone borate diester, sensing systems in multiple bacteria. Data indi-
autoinducer-2 (AI-2), produced by LuxS, and cate that furanones interfere with quorum
an as-of-yet uncharacterized AI-3, which also sensing by directly antagonizing the action of
requires LuxS for its synthesis (34, 35, 65). native quorum-sensing molecules and by
EHEC pathogenesis involves the production of promoting the turnover of quorum-sensing
a number of factors, including toxins and response proteins (40, 41).While D. pulchra and
adhesins, the assembly of secretion systems, and another alga, Chlamydomonas reinhardtii, both
the formation of attaching and effacing lesions produce molecules that antagonize quorum-
within the intestine (9, 20). Genes responsible sensing regulation,C.reinhardtii also produces at
for the formation of these structures and the least one uncharacterized compound that stim-
regulation of many virulence genes are con- ulates quorum-sensing-regulated transcription
trolled by AI-3, which activates the sensor histi- in multiple bacterial species (68). Presumably,
dine kinase QseBC (43, 64). Interestingly, the ability to modulate bacterial quorum sens-
Sperandio et al. demonstrated that, even in the ing promotes algal fitness in some aquatic com-
absence of AI-3, the pathogenicity genes were munities.
transcriptionally activated upon the addition of There are a variety of ways in which bacter-
epinephrine, a catecholamine neurotransmitter ial quorum sensing is impacted by plant factors
that can be found in the gastrointestinal tract during plant-microbe interactions. For exam-
(Fig. 2) (9, 72). Addition of antagonists of epi- ple, the legume Medicago truncatula has symbi-
nephrine prevents the transcriptional activation otic relationships with the AHL-producing,
by this hormone (9). These data suggest that nitrogen-fixing bacterium Sinorhizobium meliloti
EHEC has evolved a way to specifically induce (10). M. truncatula produces at least 12 products
that either stimulate or inhibit quorum sensing tion between A. muscaria and the spruce tree,
(22). At the same time, bacterially produced Picea abies (27, 56, 57).
AHLs affect gene expression and protein pro- In the case of the symbiotic relationship
duction in M. truncatula (42). Plant responses to between the Hawaiian bobtail squid, Euprymna
these AHLs differ depending on temporal fac- scolopes, and the gram-negative marine bac-
tors and concentrations. Furthermore, the M. terium, Vibrio fischeri, interdomain signaling
truncatula responses change depending on appears to be critical for proper light-organ
whether the AHL is one produced by its natural development. Luminescent V. fischeri colonizes
root symbiont, S. meliloti, or one produced by a the squid light-organ and provides lumines-
potential plant pathogen, P. aeruginosa, suggest- cence for the squid (38). The colonization of
ing some specificity in the plant responses (42). the light organ by V. fischeri stimulates profound
M.truncatula and other plants were also found to changes in the light-organ morphology, ulti-
produce quorum-sensing mimics, adding mately leading to macrophage-like cell infiltra-
another layer of complexity in plant-microbe tion and apoptosis (37, 70). Koropatnick et al.
signaling interactions (42, 69). determined that V. fischeri culture supernatants
contained tracheal cytotoxin, similar to that
TWO-WAY COMMUNICATION produced by Bordetella pertussis. In the squid,
Many symbiotic relationships have likely tracheal cytotoxin acts as a potent morphogen
evolved mechanisms for communication inducing hemocyte infiltration, epithelial field
between organisms from different domains of regression, and apoptosis in part due to the
life. The most widely characterized signaling activities of p53-family member proteins and
interactions within symbioses include those Toll/NF-B pathways (25,26,37).Nitric oxide
plant-bacteria and fungi-bacteria interactions (NO), which has been shown to be a signaling
involving the formation of specialized nitrogen- molecule in plant-bacterial interactions,may be
fixing structures that aid both organisms in another signal involved in establishing the
nutrient acquisition.The best-studied example squid-Vibrio symbiosis (12, 49). Both nitric
of this type of development is root nodule for- oxide synthases and NO were found in the
mation during the establishment of nitrogen- embryonic stages of light-organ development
fixing symbioses in the plant family Leguminosae and through the different stages of host-
with bacteria collectively known as rhizobia. symbiont association (12). Future studies on
The plant releases flavonoids (Fig. 2) that bind candidate bacterial NO-sensors will aid in
to LysR-family transcriptional regulators in determining whether NO signaling is impor-
the bacteria, leading to the production of Nod tant for establishing the squid-bacterium
factors, which are secreted lipo-chito-oligosac- symbiosis (12).
charides.The bacterially produced Nod factors P. aeruginosa and the dimorphic fungus Can-
facilitate bacterial entry into the plant root and dida albicans are two opportunistic pathogens
promote root structure alteration, possibly that are often coisolated from a variety of types
through the binding of plant transcription of chronic and acute infections (2, 3, 19, 31).
factors or activation of specific signaling cas- P. aeruginosa can form biofilms on and kill C.
cades (7, 51). albicans hyphae using some of the same viru-
There are indications that similar molecular lence determinants important in human dis-
conversations may be occurring in bacterial- ease, but P. aeruginosa can neither colonize nor
fungal associations. The mycorrhiza helper kill C. albicans yeast-form cells (28). In the pres-
bacterium Streptomyces strain AcH 505 was ence of P. aeruginosa, C. albicans grows as yeast-
recently shown to produce auxofuran (Fig.2) in form cells even under conditions that normally
response to being incubated with the fungus promote hyphal growth. Genetic and bio-
Amanita muscaria. In turn, auxofuran stimulates chemical experiments indicated that P. aerugi-
both fungal growth and promotes an interac- nosa 3-O-C12-HSL, described above for its role
FIGURE 3 P. aeruginosa and C. albicans responses to secreted chemical signals.
in quorum sensing, is sufficient to suppress C. regulation of P. aeruginosa pyocyanin produc-

albicans hypha formation (Fig. 3) (29).The fun- tion, a redox active virulence factor (C. Cugini
gal response facilitates its survival by altering its and D. A. Hogan, unpublished data). Although
morphology to a form that is resistant to P. farnesol does not affect P.aeruginosa growth rate,
aeruginosa colonization. it does lead to the decreased transcript levels of
C. albicans produces farnesol (Fig. 2), its own genes involved in the synthesis of PQS. This
quorum-sensing molecule that autoregulates C. repression translates into decreased production
albicans morphology at high culture densities of PQS and PQS-controlled factors including
(30). Interestingly, farnesol leads to the down- pyocyanin (C. Cugini and D.A. Hogan, unpub-
lished data). It is not yet known if P. aeruginosa Understanding the genetic regulation and
3-O-C12-HSL affects C. albicans farnesol pro- consequences of these signaling interactions
duction. If so, these two organisms may indeed may be crucial for developing new ways to
be participating in two-way communication.As modulate bacterial behaviors. With the emer-
it stands, the two distinct signaling interactions gence of antibiotic resistance, using signal
between P. aeruginosa and C. albicans (Fig. 3) mimics to specifically attenuate detrimental
decrease P. aeruginosa killing of the fungus and organisms, while leaving beneficial organisms
promote the coexistence of these two species. It unharmed, may be one answer to controlling
is possible that these interactions reflect an microorganisms in medically and environmen-
unknown benefit for these organisms upon tally important settings.
growth in coculture, or these interactions may
reflect relationships between P. aeruginosa and REFERENCES
other fungi in other environments. 1. Azghani, A. O., S. Idell, M. Bains, and R. E.
Hancock. 2002. Pseudomonas aeruginosa outer
CONCLUSIONS AND membrane protein F is an adhesin in bacterial
FUTURE DIRECTIONS binding to lung epithelial cells in culture. Microb.
Although this chapter has focused on signaling Pathog. 33:109–114.
interactions that involve molecules whose pri- 2. Bakare, N.,V. Rickerts, J. Bargon, and G. Just-
mary role appears to be the transmission of Nubling. 2003. Prevalence of Aspergillus fumigatus
information,there are likely many other signals, and other fungal species in the sputum of adult
patients with cystic fibrosis. Mycoses 46:19–23.
such as metabolites or pH and oxygen gradi- 3. Bauernfeind, A., R. M. Bertele, K. Harms,
ents, that may play equally important roles in G. Horl, R. Jungwirth, C. Petermuller, B.
the regulation of processes important during Przyklenk, and C. Weisslein-Pfister. 1987.
interdomain interactions. Physical interactions Qualitative and quantitative microbiological
between organisms may also be important analysis of sputa of 102 patients with cystic fibrosis.
Infection 15:270–277.
for communication within mixed species com- 4. Beale, E., G. Li, M. W. Tan, and K. P.
munities. Rumbaugh. 2006. Caenorhabditis elegans senses
All of the interactions that have been dis- bacterial autoinducers. Appl. Environ. Microbiol. 72:
cussed thus far have been those between the 5135–5137.
domains eukarya and bacteria. A recent report 5. Berger, M. 2002. Inflammatory mediators in cys-
tic fibrosis lung disease. Allergy Asthma Proc. 23:
suggests that archaea may also produce signal- 19–25.
ing molecules that could participate in interdo- 6. Bredenbruch, F., R. Geffers, M. Nimtz, J.
main signaling interactions. Ethyl acetate Buer, and S. Haussler. 2006. The Pseudomonas
extracts of supernatants from the haloalka- aeruginosa quinolone signal (PQS) has an iron-
liphilic archaeon, Natronococcus occultus, were chelating activity. Environ. Microbiol. 8:1318–1329.
7. Broughton, W. J., S. Jabbouri, and X. Perret.
found to possess a factor that increases the 2000. Keys to symbiotic harmony. J. Bacteriol.
expression of AHL-regulated genes in an 182:5641–5652.
Agrobacterium tumefaciens reporter assay, suggest- 8. Charlton, T. S., R. de Nys, A. Netting, N.
ing that this archeal species can produce AHLs Kumar, M. Hentzer, M. Givskov, and S.
or functional mimics (47). It remains to be seen Kjelleberg. 2000. A novel and sensitive method
for the quantification of N-3-oxoacyl homoserine
if this is a common occurrence and if the lactones using gas chromatography-mass spec-
archaea use these chemicals for interspecies trometry: application to a model bacterial biofilm.
signaling.The close association between bacte- Environ. Microbiol. 2:530–541.
ria and archea in myriad environments over 9. Clarke, M. B., D. T. Hughes, C. Zhu, E. C.
the course of millions of years has provided Boedeker, and V. Sperandio. 2006. The QseC
sensor kinase: a bacterial adrenergic receptor. Proc.
ample opportunity for bacteria-archea signal- Natl.Acad. Sci. USA 103:10420–10425.
ing interactions to evolve; we likely just need to 10. Cook, D. R. 1999. Medicago truncatula—a model
discover them. in the making! Curr. Opin. Plant Biol. 2:301–304.
11. Daniels, R., S. Reynaert, H. Hoekstra, C. patients with cystic fibrosis. Infect. Immun. 70:
Verreth, J. Janssens, K. Braeken, M. Fauvart, 1783–1790.
S. Beullens, C. Heusdens, I. Lambrichts, D. E. 22. Gao, M., M.Teplitski, J. B. Robinson, and W.
De Vos, J. Vanderleyden, J. Vermant, and D. Bauer. 2003. Production of substances by Med-
J. Michiels. 2006. Quorum signal molecules as icago truncatula that affect bacterial quorum sensing.
biosurfactants affecting swarming in Rhizobium Mol. Plant-Microbe Interact. 16:827–834.
etli. Proc. Natl.Acad. Sci. USA 103:14965–14970. 23. Givskov, M., R. de Nys, M. Manefield, L.
12. Davidson, S. K., T. A. Koropatnick, R. Koss- Gram, R. Maximilien, L. Eberl, S. Molin, P. D.
mehl, L. Sycuro, and M. J. McFall-Ngai. 2004. Steinberg, and S. Kjelleberg. 1996. Eukaryotic
NO means ‘yes’ in the squid-vibrio symbiosis: interference with homoserine lactone-mediated
nitric oxide (NO) during the initial stages of a prokaryotic signalling. J. Bacteriol. 178:6618–6622.
beneficial association.Cell Microbiol. 6:1139–1151. 24. Gonzalez, J. E., and M. M. Marketon. 2003.
13. De Kievit,T. R., R. Gillis, S. Marx, C. Brown, Quorum sensing in nitrogen-fixing rhizobia.
and B. H. Iglewski. 2001. Quorum-sensing Microbiol. Mol. Biol. Rev. 67:574–592.
genes in Pseudomonas aeruginosa biofilms: their role 25. Goodson, M. S.,W. J. Crookes-Goodson, J. R.
and expression patterns. Appl. Environ. Microbiol. Kimbell, and M. J. McFall-Ngai. 2006.Charac-
67:1865–1873. terization and role of p53 family members in the
14. de Nys, R., M. Givskov, N. Kumar, S. Kjelle- symbiont-induced morphogenesis of the
berg, and P. D. Steinberg. 2006. Furanones. Euprymna scolopes light organ. Biol. Bull. 211:7–17.
Prog. Mol. Subcell. Biol. 42:55–86. 26. Goodson, M. S., M. Kojadinovic, J.V.Troll,T.
15. Diggle, S. P., S. Matthijs, V. J. Wright, M. P. E. Scheetz, T. L. Casavant, M. B. Soares, and
Fletcher, S. R. Chhabra, I. L. Lamont, X. M. J. McFall-Ngai. 2005. Identifying compo-
Kong, R. C. Hider, P. Cornelis, M. Camara, nents of the NF-kappaB pathway in the beneficial
and P. Williams. 2007. The Pseudomonas aerugi- Euprymna scolopes-Vibrio fischeri light organ sym-
nosa 4-quinolone signal molecules HHQ and PQS biosis. Appl. Environ. Microbiol. 71:6934–6946.
play multifunctional roles in quorum sensing and 27. Gupte, M., P. Kulkarni, and B. N. Ganguli.
iron entrapment. Chem. Biol. 14:87–96. 2002. Antifungal antibiotics. Appl. Microbiol.
16. Diggle, S. P., R. E. Stacey, C. Dodd, M. Biotechnol. 58:46–57.
Camara, P.Williams, and K.Winzer. 2006.The 28. Hogan, D. A., and R. Kolter. 2002.
galactophilic lectin, LecA, contributes to biofilm Pseudomonas-Candida interactions: an ecological
development in Pseudomonas aeruginosa. Environ. role for virulence factors. Science 296:2229–2232.
Microbiol. 8:1095–1104. 29. Hogan, D. A., A. Vik, and R. Kolter. 2004.
17. DiMango, E., H. J. Zar, R. Bryan, and A. A Pseudomonas aeruginosa quorum-sensing mole-
Prince. 1995.Diverse Pseudomonas aeruginosa gene cule influences Candida albicans morphology. Mol.
products stimulate respiratory epithelial cells to Microbiol. 54:1212–1223.
produce interleukin-8. J. Clin. Invest. 96: 30. Hornby, J. M., E. C. Jensen, A. D. Lisec, J. J.
2204–2210. Tasto, B. Jahnke, R. Shoemaker, P. Dussault,
18. Dong, Y. H., J. L. Xu, X. Z. Li, and L. H. and K. W. Nickerson. 2001. Quorum sensing
Zhang. 2000.AiiA, an enzyme that inactivates the in the dimorphic fungus Candida albicans is medi-
acylhomoserine lactone quorum-sensing signal ated by farnesol. Appl. Environ. Microbiol. 67:
and attenuates the virulence of Erwinia carotovora. 2982–2992.
Proc. Natl.Acad. Sci. USA 97:3526–3531. 31. Hughes,W.T., and H. K. Kim. 1973. Mycoflora
19. El-Azizi, M.A., S. E. Starks, and N. Khardori. in cystic fibrosis: some ecologic aspects of
2004. Interactions of Candida albicans with other Pseudomonas aeruginosa and Candida albicans. Myco-
Candida spp. and bacteria in the biofilms. J. Appl. pathol. Mycol.Appl. 50:261–269.
Microbiol. 96:1067–1073. 32. Joint, I. 2006. Bacterial conversations: talking, lis-
20. Elliott, S. J., L.A.Wainwright,T. K. McDaniel, tening and eavesdropping. A NERC discussion
K. G. Jarvis,Y. K. Deng, L. C. Lai, B. P. McNa- meeting held at the Royal Society on 7 December
mara, M. S. Donnenberg, and J. B. Kaper. 2005. J. R. Soc. Interface 3:459–463.
1998.The complete sequence of the locus of ente- 33. Joint, I., K.Tait, M. E. Callow, J.A. Callow, D.
rocyte effacement (LEE) from enteropathogenic Milton, P. Williams, and M. Camara. 2002.
Escherichia coli E2348/69. Mol. Microbiol. 28:1–4. Cell-to-cell communication across the prokary-
21. Erickson, D. L., R. Endersby, A. Kirkham, ote-eukaryote boundary. Science 298:1207.
K. Stuber, D. D. Vollman, H. R. Rabin, 34. Kanamaru, K., K. Kanamaru, I. Tatsuno, T.
I. Mitchell, and D. G. Storey. 2002. Pseudomonas Tobe, and C. Sasakawa. 2000. SdiA, an
aeruginosa quorum-sensing systems may control Escherichia coli homologue of quorum-sensing
virulence factor expression in the lungs of regulators, controls the expression of virulence
factors in enterohaemorrhagic Escherichia coli Williams. 1997. Quorum sensing in Vibrio anguil-
O157:H7. Mol. Microbiol. 38:805–816. larum: characterization of the vanI/vanR locus and
35. Kaper, J. B., and V. Sperandio. 2005. Bacterial identification of the autoinducer N-(3-oxode-
cell-to-cell signaling in the gastrointestinal tract. canoyl)-L-homoserine lactone. J. Bacteriol.
Infect. Immun. 73:3197–3209. 179:3004–3012.
36. Kaplan, H. B. 2003. Multicellular development 47. Paggi, R.A., C. B. Martone, C. Fuqua, and R.
and gliding motility in Myxococcus xanthus. Curr. E. De Castro. 2003. Detection of quorum sens-
Opin. Microbiol. 6:572–577. ing signals in the haloalkaliphilic archaeon
37. Koropatnick,T.A., J.T. Engle, M.A.Apicella, Natronococcus occultus. FEMS Microbiol. Lett. 221:
E.V. Stabb,W. E. Goldman, and M. J. McFall- 49–52.
Ngai. 2004. Microbial factor-mediated develop- 48. Parsek, M. R., and E. P. Greenberg. 2005.
ment in a host-bacterial mutualism. Science 306: Sociomicrobiology: the connections between
1186–1188. quorum sensing and biofilms. Trends Microbiol.
38. Lupp, C., M. Urbanowski, E. P. Greenberg, 13:27–33.
and E. G. Ruby. 2003.The Vibrio fischeri quorum- 49. Pauly, N., C. Pucciariello, K. Mandon, G.
sensing systems ain and lux sequentially induce Innocenti, A. Jamet, E. Baudouin, D. Her-
luminescence gene expression and are important ouart, P. Frendo, and A. Puppo. 2006. Reactive
for persistence in the squid host. Mol. Microbiol. oxygen and nitrogen species and glutathione: key
50:319–331. players in the legume-Rhizobium symbiosis. J. Exp.
39. Mahajan-Miklos, S., M.W.Tan, L. G. Rahme, Bot. 57:1769–1776.
and F. M.Ausubel. 1999. Molecular mechanisms 50. Pearson, J. P., E. C. Pesci, and B. H. Iglewski.
of bacterial virulence elucidated using a 1997. Roles of Pseudomonas aeruginosa las and rhl
Pseudomonas aeruginosa-Caenorhabditis elegans path- quorum-sensing systems in control of elastase and
ogenesis model. Cell 96:47–56. rhamnolipid biosynthesis genes. J. Bacteriol.
40. Manefield, M., R. de Nys, N. Kumar, R. 179:5756–5767.
Read, M. Givskov, P. Steinberg, and S. Kjelle- 51. Perret, X., C. Staehelin, and W. J. Broughton.
berg. 1999. Evidence that halogenated furanones 2000. Molecular basis of symbiotic promiscuity.
from Delisea pulchra inhibit acylated homoserine Microbiol. Mol. Biol. Rev. 64:180–201.
lactone (AHL)-mediated gene expression by dis- 52. Pestka, S., C. D. Krause, and M. R. Walter.
placing the AHL signal from its receptor protein. 2004. Interferons, interferon-like cytokines, and
Microbiology 145:283–291. their receptors. Immunol. Rev. 202:8–32.
41. Manefield, M.,T. B. Rasmussen, M. Henzter, 53. Rankin, C. H. 2004. Invertebrate learning: what
J. B.Andersen, P. Steinberg, S. Kjelleberg, and can’t a worm learn? Curr. Biol. 14:R617-R618.
M. Givskov. 2002. Halogenated furanones inhibit 54. Rasmussen, T. B., M. Manefield, J. B. Ander-
quorum sensing through accelerated LuxR sen, L. Eberl, U. Anthoni, C. Christophersen,
turnover. Microbiology 148:1119–1127. P. Steinberg, S. Kjelleberg, and M. Givskov.
42. Mathesius, U., S. Mulders, M. Gao, M.Teplit- 2000.How Delisea pulchra furanones affect quorum
ski, G. Caetano-Anolles, B. G. Rolfe, and W. sensing and swarming motility in Serratia liquefa-
D. Bauer. 2003. Extensive and specific responses ciens MG1. Microbiology 146:3237–3244.
of a eukaryote to bacterial quorum-sensing signals. 55. Rice, S. A., M. Givskov, P. Steinberg, and S.
Proc. Natl.Acad. Sci. USA 100:1444–1449. Kjelleberg. 1999. Bacterial signals and antago-
43. McDaniel, T. K., K. G. Jarvis, M. S. Donnen- nists: the interaction between bacteria and higher
berg, and J. B. Kaper. 1995. A genetic locus of organisms. J. Mol. Microbiol. Biotechnol. 1:23–31.
enterocyte effacement conserved among diverse 56. Riedlinger, J., S. D. Schrey, M. T. Tarkka, R.
enterobacterial pathogens. Proc. Natl. Acad. Sci. Hampp, M. Kapur, and H. P. Fiedler. 2006.
USA 92:1664–1668. Auxofuran, a novel metabolite that stimulates the
44. McFall-Ngai, M. J., and E. G. Ruby. 2000. growth of fly agaric,is produced by the mycorrhiza
Developmental biology in marine invertebrate helper bacterium Streptomyces strain AcH 505.
symbioses. Curr. Opin. Microbiol. 3:603–607. Appl. Environ. Microbiol. 72:3550–3557.
45. Milton, D. L.,V. J. Chalker, D. Kirke,A. Hard- 57. Schrey, S. D., M. Schellhammer, M. Ecke,
man, M. Camara, and P. Williams. 2001.The R. Hampp, and M. T. Tarkka. 2005.
LuxM homologue VanM from Vibrio anguillarum Mycorrhiza helper bacterium Streptomyces AcH
directs the synthesis of N-(3-hydroxyhexanoyl) 505 induces differential gene expression in the
homoserine lactone and N-hexanoylhomoserine ectomycorrhizal fungus Amanita muscaria. New
lactone. J. Bacteriol. 183:3537–3547. Phytol. 168:205–216.
46. Milton, D. L., A. Hardman, M. Camara, S. R. 58. Schuster, M., and E. P. Greenberg. 2006. A
Chhabra, B.W. Bycroft, G. S. Stewart, and P. network of networks: quorum-sensing gene regu-
lation in Pseudomonas aeruginosa. Int. J. Med. Micro- associated bacteria. Mol. Plant-Microbe Interact.
biol. 296:73–81. 13:637–648.
59. Shiner, E. K., D. Terentyev, A. Bryan, S. 70. Visick, K. L., J. Foster, J. Doino, M. McFall-
Sennoune, R. Martinez-Zaguilan, G. Li, S. Ngai, and E. G. Ruby. 2000. Vibrio fischeri lux
Gyorke, S. C.Williams, and K. P. Rumbaugh. genes play an important role in colonization and
2006. Pseudomonas aeruginosa autoinducer modu- development of the host light organ. J. Bacteriol.
lates host cell responses through calcium signalling. 182:4578–4586.
Cell. Microbiol. 8:1601–1610. 71. Visick, K. L., and C. Fuqua. 2005. Decoding
60. Smith, R. S., E. R. Fedyk, T. A. Springer, N. microbial chatter: cell-cell communication in bac-
Mukaida, B. H. Iglewski, and R. P. Phipps. teria. J. Bacteriol. 187:5507–5519.
2001. IL-8 production in human lung fibroblasts 72. Walters, M., and V. Sperandio. 2006. Quorum
and epithelial cells activated by the Pseudomonas sensing in Escherichia coli and Salmonella. Int. J. Med.
autoinducer N-3-oxododecanoyl homoserine lac- Microbiol. 296:125–131.
tone is transcriptionally regulated by NF-kappa B 73. Weijer, C. J. 2004. Dictyostelium morphogenesis.
and activator protein-2. J. Immunol. 167:366–374. Curr. Opin. Genet. Dev. 14:392–398.
61. Smith, R. S., S. G. Harris, R. Phipps, and B. 74. Wheeler, G. L., K. Tait, A. Taylor, C. Brown-
Iglewski. 2002. The Pseudomonas aeruginosa lee, and I. Joint. 2006.Acyl-homoserine lactones
quorum-sensing molecule N-(3-oxodode- modulate the settlement rate of zoospores of the
canoyl)homoserine lactone contributes to viru- marine alga Ulva intestinalis via a novel chemoki-
lence and induces inflammation in vivo. J. Bacteriol. netic mechanism. Plant Cell. Environ. 29:
184:1132–1139. 608–618.
62. Smith, R. S., and B. H. Iglewski. 2003. P. aerug- 75. Williams, S. C., E. K. Patterson, N. L. Carty, J.
inosa quorum-sensing systems and virulence. Curr. A. Griswold, A. N. Hamood, and K. P. Rum-
Opin. Microbiol. 6:56–60. baugh. 2004. Pseudomonas aeruginosa autoinducer
63. Smith, R. S., R. Kelly, B. H. Iglewski, and R. enters and functions in mammalian cells. J. Bacte-
P. Phipps. 2002.The Pseudomonas autoinducer N- riol. 186:2281–2287.
(3-oxododecanoyl) homoserine lactone induces 76. Winzer, K., C. Falconer, N. C. Garber, S. P.
cyclooxygenase-2 and prostaglandin E2 produc- Diggle, M. Camara, and P.Williams. 2000.The
tion in human lung fibroblasts: implications for Pseudomonas aeruginosa lectins PA-IL and PA-IIL
inflammation. J. Immunol. 169:2636–2642. are controlled by quorum sensing and by RpoS.
64. Sperandio,V., J. L. Mellies, R. M. Delahay, G. J. Bacteriol. 182:6401–6411.
Frankel, J. A. Crawford,W. Nguyen, and J. B. 77. Wu, L., O. Estrada, O. Zaborina, M. Bains, L.
Kaper. 2000. Activation of enteropathogenic Shen, J. E. Kohler, N. Patel, M.W. Musch, E. B.
Escherichia coli (EPEC) LEE2 and LEE3 operons by Chang, Y. X. Fu, M. A. Jacobs, M. I.
Ler. Mol. Microbiol. 38:781–793. Nishimura, R. E. Hancock, J. R.Turner, and J.
65. Sperandio, V., A. G. Torres, B. Jarvis, J. P. C. Alverdy. 2005. Recognition of host immune
Nataro, and J. B. Kaper. 2003. Bacteria-host activation by Pseudomonas aeruginosa. Science
communication: the language of hormones. Proc. 309:774–777.
Natl.Acad. Sci. USA 100:8951–8956. 78. Xiao, G., E. Deziel, J. He, F. Lepine, B. Lesic,
66. Tait, K., I. Joint, M. Daykin, D. L. Milton, P. M. H. Castonguay, S. Milot, A. P.Tampakaki,
Williams, and M. Camara. 2005. Disruption of S. E. Stachel, and L. G. Rahme. 2006. MvfR, a
quorum sensing in seawater abolishes attraction of key Pseudomonas aeruginosa pathogenicity LTTR-
zoospores of the green alga Ulva to bacterial class regulatory protein, has dual ligands. Mol.
biofilms. Environ. Microbiol. 7:229–240. Microbiol. 62:1689–1699.
67. Takaoka,A., and H.Yanai. 2006. Interferon sig- 79. Zaborina, O., F. Lepine, G. Xiao,V.Valuckaite,
nalling network in innate defence. Cell. Microbiol. Y. Chen, T. Li, M. Ciancio, A. Zaborin,
8:907–922. E. Petroff, J. R. Turner, L. G. Rahme, E.
68. Teplitski, M., H. Chen, S. Rajamani, M. Gao, Chang, and J. C.Alverdy. 2007.Dynorphin acti-
M. Merighi, R.T. Sayre, J. B. Robinson, B. G. vates quorum sensing quinolone signaling in
Rolfe, and W. D. Bauer. 2004. Chlamydomonas Pseudomonas aeruginosa. PLoS Pathog. 3:e35.
reinhardtii secretes compounds that mimic bacterial 80. Zhang, L. H., and Y. H. Dong. 2004. Quorum
signals and interfere with quorum sensing regula- sensing and signal interference: diverse implica-
tion in bacteria. Plant Physiol. 134:137–146. tions. Mol. Microbiol. 53:1563–1571.
69. Teplitski, M., J. B. Robinson, and W. D. Bauer. 81. Zhang,Y., H. Lu, and C. I. Bargmann. 2005.
2000.Plants secrete substances that mimic bacterial Pathogenic bacteria induce aversive olfactory
N-acyl homoserine lactone signal activities and learning in Caenorhabditis elegans. Nature 438:
affect population density-dependent behaviors in 179–184.
INTERCELLULAR SIGNALING BY
RHOMBOIDS IN EUKARYOTES
AND PROKARYOTES
Matthew Freeman and Philip Rather
27
The rhomboid family of intramembrane serine of the above themes and provides some future
proteases controls a variety of functions in both directions for the analysis of this novel class of
eukaryotes and prokaryotes. The rhomboid membrane proteases.
proteins were originally identified in Drosophila,
where they are required for growth factor signal CONTROL OF GROWTH FACTOR
generation. However, in recent years, a number SIGNALING IN DROSOPHILA
of diverse functions for the rhomboid proteins Most of animal development is controlled by
have been identified.These functions include (i) intercellular signaling. Cells receive diffusible
the cleavage of TatA, a membrane-bound com- signals from their neighbors and interpret them
ponent of the twin arginine transport system to make fate decisions. Many physiological
that is required for cell-cell signaling in a responses in adult organisms are also regulated
prokaryote; (ii) regulating mitochondrial mem- by intercellular communication, and the dis-
brane fusion in Saccharomyces; and (iii) cleavage ruption of these signals can lead to disease.The
of cell surface adhesions in apicomplexan para- fruit fly Drosophila melanogaster is a genetically
sites. Recent biochemical analyses combined tractable organism that has been extensively
with crystallography studies have confirmed used for over 100 years to investigate the con-
these enzymes use a Ser-His catalytic dyad. trol of development; indeed, research into flies
Moreover, the active-site serine of these was instrumental in highlighting the signifi-
enzymes is embedded within the membrane cance of intercellular signaling (3). One impor-
bilayer, and access to water in the membrane is tant conclusion from Drosophila and many
mediated by a hydrophilic cavity that extends other systems is that rather few signaling path-
from the extracellular environment to the ways control a very large number of signaling
active-site serine.This chapter expands on each events. The response of a cell to activation of
one of these widely used pathways is dictated
Matthew Freeman MRC Laboratory of Molecular Biology, not by the signal itself but instead by the con-
Hills Rd., Cambridge CB2 0QH, United Kingdom. text and developmental history of the cell
Philip Rather Department of Microbiology and Immunol-
ogy, 3001 Rollins Research Bldg., Emory University School receiving the signal (16). A necessary conse-
of Medicine,Atlanta, Georgia 30322. quence of this is that the signal must be pre-
431
432 ■ FREEMAN AND RATHER
cisely regulated in time, space, and amplitude: serine proteases (61).This discovery contributed
inappropriate signaling leads to inaccurate and to a clear understanding of how EGFR signal-
potentially deleterious fate decisions. ing is controlled in Drosophila (30, 39, 58).The
One of the principal signaling pathways in active ligand for the EGFR, called Spitz,
Drosophila is controlled by the homologue of the homologue of mammalian transforming
the mammalian epidermal growth factor recep- growth factor (TGF-), needs to be released
tor (EGFR), a receptor tyrosine kinase. EGFR from a membrane-tethered precursor in order
activity regulates a wide variety of developmen- to diffuse to neighboring cells and activate the
tal decisions in many tissues (53). Extensive EGFR (Fig.1).The protease that releases Spitz is
genetic screening has been used to define rhomboid, primarily Drosophila Rhomboid-1
the physiologically significant regulators of (61). Regulation of this key proteolytic activa-
Drosophila EGFR activity,and these studies have tion is unusual in that Rhomboid-1 appears to
highlighted the significance of the rhomboid be constitutively active; instead of the protease
gene,so named because of the shape of the head activity being controlled, the enzyme and sub-
skeleton of mutant Drosophila embryos (2, 15, strate are kept physically separated until signal-
20, 21, 34, 49, 56, 65). Even after the rhomboid ing occurs. Spitz is retained in the endoplasmic
gene was sequenced, however, there was no reticulum (ER) until a third transmembrane
information about what the rhomboid protein protein, Star, is expressed. Star traffics Spitz out
does at the molecular level; it was a novel proof the ER, allowing it to move to the Golgi
tein with multiple transmembrane domains but apparatus, the location of Rhomboid-1. Here
no obvious homology to known proteins (6). Spitz is cleaved and the extracellular domain is
Eventually,a combination of genetic,cellular, released into the lumen,from where it leaves the
and biochemical evidence was used to build the cell as a secreted protein (30).
case that rhomboid was in fact the first member Although the rhomboid sequence did not
of a new class of proteases, the intramembrane provide the key insight into its function, bioin-
FIGURE 1 Drosophila Rhomboid-1 has seven transmembrane domains, and its

active site comprises residues within the plane of the lipid bilayer. It cleaves its sub-
strate, Spitz, within the TMD.This allows the extracellular domain of Spitz to be
released from the cell, so that it can activate the EGF receptor in neighboring cells.
Catalytic and other key residues are shown.
27. INTERCELLULAR SIGNALING BY RHOMBOIDS ■ 433
formatic analysis indicated that Drosophila Intriguingly, at the time that Drosophila
Rhomboid-1 belongs to a widespread family rhomboid function was first discovered, there
of conserved proteins that exist in all branches was only one other rhomboid that had received
of life (28, 65). To date, nearly all sequenced any attention, the AarA protein of the bac-
eukaryotic genomes have rhomboids. In fact, terium Providencia stuartii.
most have several in the secretory pathway and
at least one mitochondrial rhomboid. More- Identification of AarA in P. stuartii
over, a majority of prokaryotic organisms, P.stuartii is a gram-negative bacterium responsi-
both archaea and bacteria, have at least one ble for a variety of human infections, particu-
rhomboid. Those prokaryotes that do not larly within the urinary tract (44). The aarA
have rhomboids are scattered across evolution. locus was independently identified in two sepa-
Importantly, although the degree of identity rate genetic screens that were employed to
between distant rhomboids is low, the catalyti- identify regulatory mutations in P. stuartii.
cally essential residues are highly conserved, The first screen was used to identify mutations
implying that most rhomboid-like proteins are that altered transcription of the aac(2′)-Ia gene
indeed intramembrane serine proteases. (45).The aac(2′)-Ia gene encodes an acetyltrans-
The wide conservation of rhomboids in ferase capable of acetylating the aminoglyco-
organisms that have no EGFR receptors, nor sides gentamicin and tobramycin (44, 46).The
even any receptor tyrosine kinases, poses the aac(2′)-Ia gene represents a unique class of
question of the biological role of rhomboids aminoglycoside resistance genes because it is
beyond Drosophila. For example, are they all chromosomally encoded and universally pres-
involved in intercellular signaling? Or might ent in P. stuartii (44).The universal presence of
there be an ancient “core” function that has the aac(2′)-Ia gene suggested that it encoded a
been exploited in the Drosophila lineage to con- housekeeping enzyme required for some aspect
trol EGFR signaling? The full answers to these of cell physiology and that the ability to acety-
kinds of question are not yet clear, but there has late aminoglycosides was fortuitious. Work by
been enough work on rhomboids in a variety Payie et al.revealed that the AAC(2′)-Ia enzyme
of species to begin to address them. For exam- possessed O-acetyltransferase activity and was
ple, in the nematode Caenorhabditis elegans, capable of O-acetylating peptidoglycan at N-
ROM-1 has a related function to Drosophila acetylmuramyl residues (40).This modification
Rhomboid-1: it controls EGFR activity by is likely involved in regulating the activity of
releasing the LIN3 TGF--like growth factor autolysins, which are enzymes that degrade
(14). In this case, the developmental require- peptidoglycan (40).
ment for the rhomboid is less than that in flies, In wild-type P. stuartii, the levels of aac(2′)-Ia
but the basic function is extremely similar.So,at expression are not sufficient to confer resistance
least between nematodes and Drosophila, a dis- to aminoglycosides. Aminoglycoside-resistant
tance of approximately several hundred million mutants often result from regulatory mutations
years, there is a common connection between that increase aac(2′)-Ia expression (44, 46). An
rhomboids and EGFR signaling.We emphasize, additional aspect of aac(2′)-Ia regulation is that
however, that this evidence does not constitute expression is markedly decreased on agar plates
proof for a conserved function throughout relative to growth in liquid.This difference was
evolution: a rhomboid may have been recruited shown to be the result of a secreted molecule
to do the same job independently in the two produced by P. stuartii cells (47). This secreted
lineages by a process of convergent evolution. It factor was a small, heat-resistant, and protease-
will be necessary to determine how widespread sensitive factor that reduced aac(2′)-Ia mRNA
the relationship between rhomboids and accumulation when it accumulated in high-
EGFR signaling is before this issue can be density cultures (47). These findings provided
resolved. the first experimental evidence that P. stuartii
utilized a cell-cell signaling pathway to regulate were conducted by Gallio et al. in which the
gene expression. To identify genes that were functional interchangeability of AarA and the
involved in regulation of aac(2′)-Ia, a transpo- Rhomboid-1 protein from Drosophila was
son insertion library was screened for mutations examined (18). First, the ectopic expression of
that increased expression of an aac(2′)-Ia-lacZ AarA in wings of transgenic flies resulted in
fusion. Transposon insertions in a gene desig- increased vein tissue, vein thickening, and blis-
nated aarA resulted in a fourfold increase in tering of the wing.All of these phenotypes were
aac(2′)-Ia transcription based on both RNA similar to that previously reported when
analyses and reporter lacZ fusions (45). Interest- Rhomboid-1 was overexpressed (56).This phe-
ingly, the increase in aac(2′)-Ia expression in notype was strongly enhanced when the Spitz
aarA mutants was more pronounced in cells at or Gurken ligand was coexpressed with AarA
high density, suggesting an involvement in the (18).A more direct confimation was the finding
above-mentioned cell-to-cell signaling path- that AarA could rescue the phenotypes of vein-
way. In addition, aarA mutations resulted in two let alleles (18), which specifically act by reduc-
additional prominent phenotypes:(i) the inabil- ing Rhomboid-1 activity within the wing (56).
ity of cells to properly separate during the final In reciprocal experiments, the expression of
stages of cell division, resulting in chains of cells rhomboid-1 from the lac promoter in P. stuartii
connected together, and (ii) loss of a diffusible resulted in a restoration of extracellular signal
yellow pigment that is secreted into the sur- production to an aarA mutant (18). In addition,
rounding agar during growth. the phenotypes of cell chaining and lack of pig-
The second identification of aarA occurred ment production were restored to wild type
in a screen for the identification of genes that when the rhomboid-1 gene was expressed in a
were required for the expression of lacZ fusions P. stuartii aarA mutant.The ability to rescue sig-
regulated by cell-cell signaling (43). In this nal production to an aarA mutant was not
study, a lacZ fusion (cma37::lacZ) to an operon restricted to the Drosophila rhomboid.A human
that likely represents the D-methionine uptake rhomboid, RHBDL2, could also mediate this
locus for P.stuartii was used to identify mutations effect, albeit at much lower efficiency (9). It was
that increased expression. Insertions in aarA also shown that AarA could directly cleave the
were identified, and further characterization of Drosophila substrate Spitz (62).At the time,these
the aarA mutants revealed that they failed to experiments suggested that both AarA and
produce an extracellular activating signal.More- Rhomboid-1 generate a signaling molecule via
over,the aarA-dependent signal was required for intramembrane proteolysis of an unknown
activation of three additional lacZ fusions, sug- membrane-bound substrate. However, subse-
gesting that the AarA-dependent extracellular quent experiments revealed that the role of
signal participated in global gene regulation AarA in signal production was indirect, and sig-
(43). In summary, the AarA-dependent signal nal production required the twin-arginine
was required for repression of the aac(2′)-Ia gene translocase (Tat) protein export system as
and for activation of at least three additional described below.
genes in P. stuartii.
Role of AarA in Function of the
Conservation of AarA and Rhomboid Tat Protein Export System
Function across Species The twin-arginine-dependent translocation
The independent discovery of rhomboids in system exports prefolded, cofactor-containing
intercellular signaling roles in Drosophila and P. proteins with a twin-arginine motif within the
stuartii led to the proposal that eukaryotic and signal sequence (50, 51, 67). These exported
prokaryotic rhomboids may function in a simi- proteins participte in a variety of functions,
lar manner (17).To obtain experimental confir- including anaerobic respiration. In Escherichia
mation of this hypothesis, a set of experiments coli, five proteins,TatA to E,make up the Tat sys-
tem, although only three subunits, TatA to C, inability to grow on MacConkey agar, and the
are required for protein export (51, 68). Early loss of extracellular signal production were all
observations of aarA mutants in P. stuartii sug- restored to wild-type levels.This ability to sup-
gested a relationship with the Tat system. The press an aarA mutation was not unique to the P.
primary phenotype linking loss of aarA with a mirabilis tatA protein. The E. coli tatA or tatE
Tat defect was a prominent cell division pheno- proteins were also identified from an E.coli plas-
type where cells failed to separate during cell mid library in a screen for rhomboid-like genes.
division, resulting in chains of interconnected The ability of tatE to restore these phenotypes
cells. In previous studies, this phenotype was was likely due to the fact that it is highly similar
observed in E. coli tat mutants (54).The basis for to tatA and functionally interchangeable (51).
this chaining phenotype was revealed by studies Unexpectedly, through screens for restored pig-
that demonstrated that both the AmiA and ment or growth on MacConkey agar, the tatA
AmiC amidases were dependent on the Tat sys- gene from P. stuartii was never isolated from a
tem for transport into the periplasm (5, 27). genomic library that had yielded many other P.
Both of these amidases cleave the peptide moi- stuartii genes. However, the P. stuartii tat locus
ety from N-acetyl muramic acid and are was subsequently isolated with an approach that
involved in murein cleavage that facilitates the did not involve functional complementation of
separation of cells during division. In addition an aarA mutation. Surprisingly, when the native
to cell chaining and loss of extracellular signal TatA protein from P. stuartii was expressed in an
production, P. stuartii aarA mutants exhibit a aarA mutant, there was no restoration of the
number of additional phenotypes, including an aarA mutant phenotypes.The TatAPs was shown
inability to grow on MacConkey agar and loss to be functional as it complemented the pheno-
of a diffusible yellow pigment (9).The inability types of an E. coli tatA/tatE double mutant.
to grow on MacConkey agar, which contains Therefore, some feature of the TatAPs protein
the detergent sodium deoxycholate, can be prevented it from “complementing” the aarA
linked to the Tat phenotype because tat mutants mutation.When the TatAPs sequence was ana-
are sensitive to detergents (54). lyzed, it was found to exhibit an unusual feature
Experimental data that directly linked the when compared to the TatAPm and TatAEc pro-
AarA rhomboid protease with function of the teins in that it contained an extension of seven
Tat system were the observation that overex- amino acids at the N terminus (55).This exten-
pression of the TatA protein suppressed all the sion was atypical with respect to other bacterial
above phenotypes associated with loss of AarA. TatA proteins (Fig. 2). From this information, it
In a screen for rhomboid-like genes in Proteus was hypothesized that the P. stuartiiTatA protein
mirabilis,an organism closely related to P.stuartii, may need to be processed by the AarA rhom-
the tatA gene was repeatedly isolated as a high- boid protease in order to function normally.The
copy suppressor of the aarA mutation based on TatAPm and TatAEc proteins functioned in an
restored production of extracellular pigment AarA-independent manner because they were
(55). Further examination of a P. stuartii aarA both naturally missing this N-terminal exten-
mutant containing the P. mirabilis tatA gene sion (Fig. 2).This was experimentally tested in
indicated that the cell chaining phenotype, the vivo by the construction of a C-terminally
FIGURE 2 Alignment of the N-terminal region of TatA proteins from P. mirabilis, E. coli, and P. stuartii. The
arrowhead designates the site of AarA-dependent cleavage for the P. stuartiiTatA protein.
tagged TatA-His6 fusion protein. Western blot cleavage is the signaling molecule.This is based
analysis of cell extracts from wild-type and aarA on the observation that in a tatC mutant, there
mutant P. stuartii strains indicated the TatA-His is no signal activity, yet TatA is still processed
was processed in wild type, but not in an aarA normally (55). In addition, the TatAEc and Tat-
mutant (55). The processing site was between APm proteins are missing the extension that
the eighth and ninth amino acids.To investigate would generate the signal, yet they still restore
whether the processing of TatA by AarA was signal production to an aarA mutant (55).
direct,purified AarA and TatA-His were used to
examine proteolytic processing in vitro. The How Is the Tat System Required
processing of TatA-His6 was confirmed in vitro, for Signal Production?
and the processing site was identical to that To our knowledge, a role for the Tat export sys-
observed in vivo (55).The above data suggested tem in cell-to-cell signaling has not been previ-
that a TatAPs variant that was missing the first ously reported.A number of possibilities can be
eight amino acids should restore Tat function in considered for why Tat export is required; how-
an aarA mutant because the requirement for ever,at this point,all are highly speculative.First,
processing would be bypassed. A TatAPs 2-8 a gene product that has an essential role in signal
variant restored all Tat-dependent phenotypes production may be exported by the Tat system.
to an aarA mutant. A large number of Tat-exported proteins func-
The basis of the requirement for TatA pro- tion in anaerobic respiration,and signal produc-
cessing in P. stuartii is unclear. In E. coli, the TatA tion or activity may be coupled to this process
protein can exist in complexes with TatB, with (38).A second possibility is that processing of a
TatBC and as a separate TatA homo-oligomer periplasmic component by a Tat-dependent
(7, 10, 41, 52).TatA protein itself may form the enzyme may be required for signal production.
actual pore for protein secretion (19).The sim- As noted previously,the export of the AmiA and
plest explanation is that this N-terminal exten- AmiC enzymes is Tat-dependent (5, 27).There-
sion is disruptive to correct insertion within the fore, one intriguing possibility that has not been
membrane and/or interaction with other Tat experimentally tested is that one or more of
proteins.The processing event does not seem to these amidases have a role in signal production,
be required for insertion of TatA within the possibly via the breakdown of peptidoglycan.
membrane,as unprocessed TatA is found within The AmiA and AmiC enzymes remove peptides
membrane fractions of an aarA mutant (unpub- from the N-acetyl muramic acid moiety of the
lished data). glycan backbone (22,23).This activity will gen-
erate either a pentapeptide or a tetrapeptide
Role of the Tat System in fragment (42). A third possibility is that the
Production/Activity of a Cell- extracellular signal is a cofactor normally associ-
to-Cell Signaling Molecule ated with a protein exported by the Tat system.
The ability of TatA in high copy to restore sig- Fe-S, molydopterin, NADP, and nickel are
nal production to an aarA mutant indicated that some examples of cofactors associated with
it was likely the Tat export system that was proteins exported by the Tat system (38). Upon
required for activity or production of the extra- completion of export,the cofactor may become
cellular signal. The construction of a tatC null disassociated and act as a signaling molecule.
allele and a tatC/aarA double mutant indicated
that the tat locus was required for signal pro- RECENT ADVANCES IN RHOMBOID
duction and the requirement for aarA was due MECHANISM AND STRUCTURE
to loss of Tat function. The mechanism by Rhomboids are one of four known families of
which the Tat system is required for signal pro- intramembrane proteases.The first such enzyme
duction is unclear.One possibility that has been to be discovered was site-2 protease, a puta-
ruled out is that the peptide released by TatA tive metalloprotease that controls cholesterol
biosynthesis in mammals but which, like rhom- (iii) their activity may be modulated by the lipid
boids, has prokaryotic homologues (48). Per- composition of the membrane; and (iv) highly
haps the most famous intramembrane protease purified rhomboid protein retains its activity.
is presenilin, an aspartyl protease and the active This last property provided a major boost to the
subunit of the -secretase complex, implicated effort to generate a crystal structure of rhom-
in the generation of the amyloid protein that boids, and this has very recently born fruit as
constitutes the plaques in Alzheimer’s disease the structure of GlpG, the E. coli rhomboid,
(12, 70). -Secretase also controls signaling has been solved at high resolution (64), the first
through the Notch receptor in animals—a key such structure available for any intramembrane
developmental pathway. The fourth family protease (Color Plate 17).Very soon after this
member is signal peptide peptidase, an aspartyl first crystal structure, additional structures were
protease mechanistically similar to presenilin published (4, 71). Coupled with the nuclear
(66). Although the different families are unre- magnetic resonance structure of the soluble
lated by sequence, mechanism, and evolution, cytoplasmic domain of a Pseudomonas aeruginosa
all intramembrane proteases have multiple rhomboid (11), there is now a wealth of struc-
transmembrane domains (TMDs) and all share tural information available about these enzymes.
the property of apparently performing a prote- There are still numerous unsolved mechanis-
olytic cleavage reaction within the lipid bilayer tic questions, many of which will need the
of membranes. Proteolysis requires water, and structural solution of rhomboid in complex
since the lipid bilayer is hydrophobic, the enzy- with a substrate or inhibitors, but some key
mology of rhomboids and the other intramem- functional messages emerge from these recent
brane proteases has been mysterious (60, 69). advances. Perhaps the most important is the
Indeed, the whole notion of intramembrane structural support for the idea that rhomboid is,
proteolysis has been rather heretical and only as predicted, a serine protease that relies on a
slowly accepted in the protease field. However, catalytic dyad and whose active site is within the
genetic and biochemical evidence has become plane of the lipid bilayer (64, 71); intramem-
increasingly compelling over the last few years, brane proteolysis is not just the fantasy of mis-
and now recent structural work on rhomboids guided geneticists and cell biologists. The
has dispelled all doubt about the reality of problem of water and its access to the active site
intramembrane proteolysis and has started to is solved by the simple feature of the fourth
solve the mysteries of how these enzymes work. transmembrane domain of GlpG being shorter
The first important step toward understand- than predicted, leading to a hydrophilic inden-
ing the rhomboid mechanism was the demon- tation on the extracellular/luminal side of the
stration that rhomboid enzymatic activity could membrane (64, 71). The catalytic heart of the
be reconstituted in vitro with detergent- enzyme lies at the bottom of this cavity. In the
solubilized and purified proteins (32, 33, 63). absence of crystal structures for the other
Until that point, all work on rhomboids had intramembrane proteases,it is too early to know
been either genetic or reliant on cell-based whether they have solved the water problem in
assays,neither of which provided direct access to the same way, but recent biochemical evidence
enzymological questions. The development of and low-resolution structures certainly suggest
in vitro biochemical assays allowed the follow- that the presenilin active site is accessible to the
ing conclusions to be drawn: (i) rhomboids do aqueous environment (29, 37, 57).
not require cofactors for enzymatic activity
(note that this does not necessarily mean that Functional and Evolutionary
cofactors are never involved);(ii) the mechanism Implications of Rhomboid Structure
relies on a core catalytic dyad of serine and histi- One spin-off of this much clearer picture of
dine that is related but not identical to the classi- rhomboid function is the ability to define true
cal soluble serine proteases like chymotrypsin; rhomboids more accurately.This allows distant
rhomboids to be aligned with more confidence substrates can leave the membrane in either
than before,leading to some revision of the phy- direction. The longer transmembrane product
logenetic tree of rhomboids (31). One of the is released by mitochondrial rhomboids, while
outcomes of this is that the predicted transmem- the shorter product is released by the secretase
brane topologies of some of the rhomboids have rhomboids studied so far. This raises the
changed (Color Plate 18). Most prokaryotic intriguing possibility of rhomboid-induced
rhomboids have six TMDs. Drosophila rhom- bidirectional signaling, although no clear
boid-1 and most of the eukaryotic rhomboids example of this has yet been found.
that reside in the secretory pathway have seven It has been proposed that the phylogenetic
TMDs: an extra one has been added to the C tree of rhomboids supports an evolutionary
terminus of the “core” six-TMD unit. Mito- model where rhomboids arose in bacteria and
chondrial rhomboids (known as PARLs [prese- subsequently spread by multiple horizontal
nilin-associated rhomboid-like, now not gene transfer events (28).The argument for this
believed to be associated with presenilin], after somewhat counterintuitive suggestion was
the first to be discovered in mammals) also have based on analysis of rhomboid phylogeny at a
seven, but, in contrast to the secretase rhom- time when there was very rudimentary func-
boids, the extra TMD has been added to the N tional and no structural information available.A
terminus (Color Plate 18). Notably, however, more recent proposal is that, based on the
some prokaryotes, including Providencia, have a updated phylogeny, a six-TMD rhomboid was
seven-TMD rhomboid, and there is one sub- present in the last universal common ancestor
family of eukaryotic secretase rhomboids with and has been vertically spread through the
six TMDs.To summarize these variant topolo- branches of the tree of life (31). In this alterna-
gies, there is a six-TMD core rhomboid struc- tive proposal, the absence of a rhomboid from
ture that is present in most prokaryotes and a some prokaryotes is explained by occasional
small subfamily of eukaryotic rhomboids. Most loss in some lineages,presumably where an evo-
eukaryotic rhomboids and a few from prokary- lutionary niche has removed the otherwise
otes have acquired a seventh TMD; in secretase strong selective pressure for its maintenance.
rhomboids, this TMD is added to the C termi- Superimposed on this basic scheme is at least
nus, and in mitochondrial rhomboids it is added one presumed horizontal gene transfer event:
to the N terminus (28, 31). the acquisition of mitochondrial rhomboids
There are at least two important functional (and probably chloroplast rhomboids) when
consequences of this pattern of topologies (31). these organelles first formed by endosymbiosis.
First, now that we know that all mitochondrial
rhomboids have an extra N-terminal TMD, it is Rhomboids in Parasites and Yeast
clear that the main catalytic TMD (TMD5 in Although we have focused here on the role of
mitochondrial rhomboids, equivalent to rhomboids in animals and bacteria, the last few
TMD4 in the secretase and prokaryotic rhom- years have seen the discovery of rhomboid
boids) is in the opposite orientation to its coun- functions in other branches of life. The yeast
terpart in secretase rhomboids. Notably, the Saccharomyces cerevisiae has two rhomboids, one
known substrates for mitochondrial rhomboids in the secretory pathway whose function
are type II proteins, whereas all known sub- remains unknown, and a mitochondrial rhom-
strates for secretase rhomboids are type I.This boid that regulates mitochondrial membrane
striking correlation between the orientations of fusion (24,25,35).Another example is the work
the active site and the substrate strongly sup- implicating rhomboids in host cell invasion by
ports the idea that a given rhomboid enzyme apicomplexan parasites like the malaria parasite
can only cleave one substrate orientation. Sec- (Plasmodium) and Toxoplasma.It has been shown
ond, the cleavage site of mitochondrial sub- by a number of groups that rhomboids can
strates indicates that, once cleaved, rhomboid cleave the cell surface adhesion proteins that
mediate the physical interaction between the of rhomboid function could be identified in
parasite and host cells (1, 8, 13, 26, 36, 59). Since cell-based assays using an aarA mutant of P.
the cleavage of these proteins is essential for stuartii.
invasion, there is growing evidence for the As described above, when rhomboids were
exciting possibility that rhomboids might be first discovered to be proteases that control
valuable therapeutic targets for diseases, like growth factor signaling in Drosophila and the
malaria, caused by the apicomplexa. link with Providencia quorum sensing was made,
the question arose of whether all rhomboids
CONCLUSIONS would be involved with intercellular signaling.
More generally, although knowledge of the The answer is clearly no: there are already mul-
medical relevance of rhomboids is still in its tiple examples of rhomboids being involved in
infancy, there is considerable potential. We other kinds of signaling and control events.
already know that rhomboids are involved in Instead of being specialized proteases with a
multiple developmental and physiological con- single conserved biological role, rhomboids are
trol/signaling events, some of which, like para- versatile control enzymes that catalyze an irre-
site invasion, quorum sensing, and growth versible release of protein domains from mem-
factor signaling, have clear medical significance. branes in multiple contexts.This is a powerful
Unlike many signaling proteins that function by tool in the cellular kit, and although the nature
protein-protein interactions, rhomboids are of the reaction lends itself to controlling inter-
enzymes, and these are much easier to inhibit cellular signaling between membranes of
with small-molecules than interacting protein neighboring cells, it has also been exploited in
interfaces. Indeed, pilot studies demonstrate many different ways.
that small-molecule inhibitors against rhom- A number of unanswered questions remain
boids can be identified. Of course, major ques- that will provide numerous opportunities to
tions remain about the possibility of targeting study these proteins further. First, despite the
rhomboids for therapeutic use. For example, almost universal presence of rhomboids in
although there are many promising possibilities, prokaryotes,we only understand the function of
in no case has a clinical opportunity been fully a rhomboid in one system,which is the cleavage
validated. Furthermore, if and when such cases of TatA in P. stuartii, an event required for the
are made, it may be necessary to develop rhom- generation of a cell-cell signaling molecule via a
boid inhibitors that show specificity between functional Tat system.What are the functions of
rhomboids; this could be a formidable chemical rhomboids in other bacteria, and why do some
challenge. Nevertheless, it seems likely that this bacteria have multiple rhomboids? In eukary-
will become a major focus of the next few years. otes, a similar question arises about the roles of
An attractive system that will be useful for the multiple rhomboids; there are at least five true
identification and functional analysis of both rhomboids in mammals and as many as 15 in the
eukaryotic and prokaryotic rhomboid proteins plant Arabidopsis. By analogy to Drosophila and
is the aarA mutant of P.stuartii. A number of eas- C. elegans, there is a possibility that mammalian
ily scored phenotypes are present in aarA EGFR signaling might involve a rhomboid,
mutants, including an inability to grow on but there is no evidence to support this yet.
MacConkey plates and loss of a diffusible yel- Another key issue for all rhomboids is how they
low pigment. These phenotypes can be are regulated. As described above, Drosophila
exploited to select for new rhomboid genes by Rhomboid-1 activity is primarily regulated by
growth on MacConkey agar or in screens to limiting substrate access rather than by more
identify colonies with restored pigment pro- common posttranslational mechanisms. Is this a
duction. In a reciprocal manner, structure/ common theme, or is the enyme activity itself
function analyses of rhomboids can be assessed regulated in other contexts, for example, by
using this system.In addition,specific inhibitors phosphorylation, membrane lipid composition,
or cofactors? Finally, beyond the multiple bio- tional role in intramembrane proteolysis. J. Mol.
logical and mechanistic questions that need to Biol. 365:109–122.
12. De Strooper, B., W. Annaert, P. Cupers, P.
be resolved, current preliminary evidence Saftig, K. Craessaerts, J. S. Mumm, E. H.
makes it likely that there will be an increasing Schroeter,V. Schrijvers, M. S.Wolfe,W. J. Ray,
focus on the possible medical relevance of the A. Goate, and R. Kopan. 1999.A presenilin-1-
large and diverse family of rhomboid proteases. dependent gamma-secretase-like protease medi-
ates release of Notch intracellular domain. Nature
398:518–522.
REFERENCES 13. Dowse,T. J., J. C. Pascall, K. D. Brown, and D.
1. Baker, R. P., R. Wijetilaka, and S. Urban. Soldati. 2005. Apicomplexan rhomboids have
2006.Two Plasmodium rhomboid proteases prefer- a potential role in microneme protein cleavage dur-
entially cleave different adhesins implicated in all ing host cell invasion. Int. J. Parasitol. 35:747–756.
invasive stages of malaria. PLoS Pathog. 2:e113. 14. Dutt, A., S. Canevascini, E. Froehli-Hoier,
2. Bang, A. G., and C. Kintner. 2000. Rhomboid and A. Hajnal. 2004. EGF signal propagation
and Star facilitate presentation and processing of during C. elegans vulval development mediated by
the Drosophila TGF-alpha homolog Spitz. Genes ROM-1 rhomboid. PLoS Biol. 2:e334.
Dev. 14:177–186. 15. Freeman, M. 1994.The spitz gene is required for
3. Bate, M., and A. Martinez-Arias. 1993. The photoreceptor determination in the Drosophila eye
Development of Drosophila melanogaster. Cold where it interacts with the EGF receptor. Mech.
Spring Harbor Laboratory Press, Plainview, NY. Dev. 48:25–33.
4. Ben-Shem, A., D. Fass, and E. Bibi. 2006. 16. Freeman, M. 1997. Cell determination strategies
Structural basis for intramembrane proteolysis by in the Drosophila eye. Development 124:261–270.
rhomboid serine proteases. Proc. Natl. Acad. Sci 17. Gallio, M., and P. Kylsten. 2000. Providencia may
USA 104:462–466. help find a function for a novel widespread protein
5. Bernhardt, T. G., and P. A. J. de Boer. 2003. family. Curr. Biol. 10:693–694.
The Escherichia coli amidase AmiC is a periplasmic 18. Gallio, M., G. Sturgill, P. N. Rather, and P.
septal ring component exported via the twin- Kylsten. 2002. A common mechanism for extra-
arginine transport pathway. Mol. Microbiol. 48: cellular signaling in eukaryotes and prokaryotes.
1171–1182. Proc. Natl.Acad Sci. USA 99:12208–12213.
6. Bier, E., L.Y. Jan, and Y. N. Jan. 1990. Rhom- 19. Gohlke, U., L. Pullan, C. A. McDevitt, I.
boid,a gene required for dorsoventral axis establish- Porcelli, E. de Leeuw,T. Palmer, H. R. Saibil,
ment and peripheral nervous system development and B. C. Berks. 2005.The TatA component of
in Drosophila melanogaster. Genes Dev. 4:190–203. the twin-arginine protein transport system forms
7. Bolhuis, A., E. G. Bogsch, and C. Robinson. channel complexes of variable diameter. Proc. Natl.
2000. Subunit interactions in the twin arginine Acad. Sci. USA 102:10482–10486.
translocase complex of Escherichia coli. FEBS Lett. 20. Golembo, M., E. Raz, and B. Z. Shilo. 1996.
472:88–92. The Drosophila embryonic midline is the site of
8. Brossier, F., T. J. Jewett, L. D. Sibley, and S. Spitz processing,and induces activation of the EGF
Urban. 2005.A spatially localized rhomboid pro- receptor in the ventral ectoderm. Development
tease cleaves cell surface adhesins essential for inva- 122:3363–3370.
sion by Toxoplasma. Proc. Natl. Acad. Sci. USA 21. Guichard, A., B. Biehs, M. A. Sturtevant,
102:4146–4151. L. Wickline, J. Chacko, K. Howard, and E.
9. Clemmer, K. M., G. M. Sturgill, A.Veenstra, Bier. 1999. rhomboid and Star interact synergisti-
and P. N. Rather. 2006. Functional characteriza- cally to promote EGFR/MAPK signaling during
tion of Escherichia coli GlpG and additional rhom- Drosophila wing vein development. Development
boid proteins using an aarA mutant of Providencia 126:2663–2676.
stuartii. J. Bacteriol. 188:3415–3419. 22. Heidrich, C., M. F. Templin, A. Ursinus, M.
10. de Leeuw, E.,T. Granjon, I. Porcelli, M.Alami, Merdanovic, J. Berger, H. Schwarz, M. A. de
S. B. Carr, M. Mueller, F. Sargent, T. Palmer, Pedro, and J.V. Höltje. 2001. Involvement of N-
and B. C. Berks. 2002.Oligomeric properties and acetylmuramyl-L-alanine amidases in cell separa-
signal peptide binding by Escherichia coli Tat protein tion and antibiotic-induced autolysis of Escherichia
transport complexes. J. Mol. Biol. 322:1135–1146. coli. Mol. Microbiol. 41:167–178.
11. Del Rio, A., K. Dutta, J. Chavez, I. Ubarretx- 23. Heidrich C.,A. Ursinus, J. Berger, H. Schwarz,
ena-Belandia, and R. Ghose. 2006. Solution and J.V. Holtje. 2002. Effects of multiple deletions
structure and dynamics of the N-terminal cytoso- of murein hydrolases on viability, septum cleavage,
lic domain of rhomboid intramembrane protease and sensitivity to large toxic molecules in Escherichia
from Pseudomonas aeruginosa: insights into a func- coli. J. Bacteriol. 184:6093–6099.
24. Herlan, M., C. Bornhovd, K. Hell, W. 36. O’Donnell, R. A., F. Hackett, S. A. Howell,
Neupert, and A. S. Reichert. 2004.Alternative M. Treeck, N. Struck, Z. Krnajski, C.
topogenesis of Mgm1 and mitochondrial mor- Withers-Martinez, T. W. Gilberger, and M. J.
phology depend on ATP and a functional import Blackman. 2006. Intramembrane proteolysis
motor. J. Cell. Biol. 165:167–173. mediates shedding of a key adhesin during ery-
25. Herlan, M., F. Vogel, C. Bornhovd, W. throcyte invasion by the malaria parasite. J. Cell
Neupert, and A. S. Reichert. 2003. Processing Biol. 174:1023–1033.
of Mgm1 by the rhomboid-type protease Pcp1 is 37. Ogura, T., K. Mio, I. Hayashi, H. Miyashita,
required for maintenance of mitochondrial mor- R. Fukuda, R. Kopan, T. Kodama, T.
phology and of mitochondrial DNA. J. Biol. Chem. Hamakubo, T. Iwatsubo, T. Tomita, and C.
278:27781–27788. Sato. 2006. Three-dimensional structure of the
26. Howell, S. A., F. Hackett, A. M. Jongco, C. gamma-secretase complex. Biochem. Biophys. Res.
Withers-Martinez, K. Kim,V. B. Carruthers, Commun. 343:525–534.
and M. J. Blackman. 2005. Distinct mechanisms 38. Palmer,T., F. Sargent, and B. C. Berks. 2005.
govern proteolytic shedding of a key invasion pro- Export of complex cofactor-containing proteins
tein in apicomplexan pathogens. Mol. Microbiol. by the bacterial Tat pathway. Trends Microbiol.
57:1342–1356. 13:175–180.
27. Ize, B., N. R. Stanley, G. Buchanan, and T. 39. Pascall, J. C., J. E. Luck, and K. D. Brown.
Palmer. 2003. Role of the Escherichia coliTat path- 2002. Expression in mammalian cell cultures
way in outer membrane integrity. Mol. Microbiol. reveals interdependent, but distinct, functions for
48:1183–1193. Star and rhomboid proteins in the processing of
28. Koonin, E.V., K. S. Makarova, I. B. Rogozin, the Drosophila transforming-growth-factor-alpha
L. Davidovic, M. C. Letellier, and L. homologue Spitz. Biochem. J. 363:347–352.
Pellegrini. 2003.The rhomboids:a nearly ubiqui- 40. Payie, K. G., P. N. Rather, and A. J. Clarke.
tous family of intramembrane serine proteases that 1995. Contribution of gentamicin 2′-N-acetyl-
probably evolved by multiple ancient horizontal transferase to the O-acetylation of peptidoglycan
gene transfers. Genome Biol. 4:R19. in Providencia stuartii. J. Bacteriol. 177:4303–4310.
29. Lazarov, V. K., P. C. Fraering, W. Ye, M. S. 41. Porcelli, I., E. de Leeuw, R. Wallis, E. van
Wolfe, D. J. Selkoe, and H. Li. 2006. Electron den Brink-van der Laan, B. de Kruijff,
microscopic structure of purified, active gamma- B. A. Wallace, T. Palmer and B. C. Berks.
secretase reveals an aqueous intramembrane cham- 2002. Characterization and membrane assembly of
ber and two pores. Proc. Natl. Acad. Sci. USA the TatA component of the Escherichia coli twin-
103:6889–6894. arginine protein transport system. Biochemistry
30. Lee, J. R., S. Urban, C. F. Garvey, and M. 41:13690–13697.
Freeman. 2001. Regulated intracellular ligand 42. Priyadarshini, R., D. L. Popham, and K. D.
transport and proteolysis control EGF signal Young. 2006. Daughter cell separation by peni-
activation in Drosophila. Cell 107:161–171. cillin-binding proteins and peptidoglycan ami-
31. Lemberg, M. K., and M. Freeman. 2007.Func- dases in Escherichia coli. J. Bacteriol. 188:5345–5355.
tional and evolutionary implications of enhanced 43. Rather, P. N., X. Ding, R. R. Baca-DeLancey,
genomic analysis of rhomboid intramembrane and S. Sidduqui. 1999. Providencia stuartii genes
proteases, Genome Res. 17:1634–1646. activated by cell-cell signaling and identification of
32. Lemberg, M. K., J. Menendez, A. Misik, M. a gene required for the production or activity of an
Garcia, C.M. Koth, and M. Freeman. 2005. extracellular factor. J. Bacteriol. 181:7185–7191.
Mechanism of intramembrane proteolysis investi- 44. Rather, P. N., and D. R. Macinga. 1999. The
gated with purified rhomboid proteases. EMBO J. chromosomal 2′-N-acetyltransferase of Providencia
24:464–472. stuartii: physiological functions and genetic regula-
33. Maegawa, S., K. Ito, and Y. Akiyama. 2005. tion. Front. Biosci. 4:132–140.
Proteolytic action of GlpG,a rhomboid protease in 45. Rather, P. N., and E. Orosz. 1994. Characteri-
the Escherichia coli cytoplasmic membrane. Bio- zation of aarA, a pleiotrophic negative regulator of
chemistry 44:13543–13552. the 2′-N-acetyltransferase in Providencia stuartii. J.
34. Mayer, U., and C. Nüsslein-Volhard. 1988. A Bacteriol. 176:5140–5144.
group of genes required for pattern formation in 46. Rather, P. N., E. Orosz, R. Hare, G. Miller, and
the ventral ectoderm of the Drosophila embryo. K. Shaw. 1993. Characterization and transcrip-
Genes Dev. 2:1496–1511. tional regulation of the 2′-N-acetyltransferase gene
35. McQuibban, G. A., S. Saurya, and M. of Providencia stuartii. J. Bacteriol. 175: 6492–6498.
Freeman. 2003. Mitochondrial membrane 47. Rather, P. N., M. M. Parojcic, and M. R.
remodelling regulated by a conserved rhomboid Paradise. 1997.An extracellular factor regulating
protease. Nature 423:537–541. expression of the chromosomal aminoglycoside
2′-N-acetyltransferase in Providencia stuartii. 59. Urban, S., and M. Freeman. 2003. Substrate
Antimicrob.Agents Chemother. 41:1749–1754. specificity of rhomboid intramembrane proteases
48. Rawson, R. B., N. G. Zelenski, D. Nijhawan, J. is governed by helix-breaking residues in the
Ye, J. Sakai, M. T. Hasan, T. Y. Chang, M. S. substrate transmembrane domain. Mol. Cell 11:
Brown, and J. L. Goldstein. 1997.Complemen- 1425–1434.
tation cloning of S2P, a gene encoding a putative 60. Urban, S., and M. Freeman. 2002. Intramem-
metalloprotease required for intramembrane brane proteolysis controls diverse signaling path-
cleavage of SREBPs. Mol. Cell. 1:47–57. ways throughout evolution. Curr. Opin. Genet. Dev.
49. Ruohola-Baker, H., E. Grell, T. B. Chou, D. 12:512–518.
Baker, L. Y. Jan, and Y. N. Jan. 1993. Spatially 61. Urban, S., J. R. Lee, and M. Freeman. 2001.
localized rhomboid is required for establishment of Drosophila rhomboid-1 defines a family of putative
the dorsal-ventral axis in Drosophila oogenesis. Cell intramembrane serine proteases.Cell 107:173–182.
73:953–965. 62. Urban, S., D. Schlieper, and M. Freeman.
50. Santini, C. L., B. Ize, A. Chanal, M. Muller, 2002. Conservation of intramembrane proteolytic
G. Giordano, and L. F. Wu. 1998. A novel sec- activity and substrate specificity in prokaryotic and
independent periplasmic protein translocation eukaryotic rhomboids. Curr. Biol. 12:1507–1512.
pathway in Escherichia coli. EMBO J. 17:101–112. 63. Urban, S., and M. S.Wolfe. 2005.Reconstitution
51. Sargent, F., E. G. Bogsch, N. R. Stanley, M. of intramembrane proteolysis in vitro reveals that
Wexler, C. Robinson, B. C. Berks, and T. pure rhomboid is sufficient for catalysis and speci-
Palmer. 1998. Overlapping functions of compo- ficity. Proc. Natl.Acad. Sci. USA 102:1883–1888.
nents of a bacterial Sec-independent protein 64. Wang, Y., Y. Zhang, and Y. Ha. 2006. Crystal
export pathway. EMBO J. 17:3640–3650. structure of a rhomboid family intramembrane
52. Sargent, F., U. Gohlke, E. de Leeuw, N. R. protease. Nature 444:179–180.
Stanley, T. Palmer, H. R. Saibil, and B. C. 65. Wasserman, J. D., S. Urban, and M. Freeman.
Berks. 2001. Purified components of the 2000.A family of rhomboid-like genes: Drosophila
Escherichia coli Tat protein transport system form a rhomboid-1 and roughoid/rhomboid-3 cooper-
double layered ring structure. Eur. J. Biochem. ate to activate EGF receptor signalling. Genes Dev.
268:3361–3367. 14:1651–1663.
53. Shilo, B. Z. 2003. Signaling by the Drosophila epi- 66. Weihofen, A., K. Binns, M. K. Lemberg, K.
dermal growth factor receptor pathway during Ashman, and B. Martoglio. 2002.Identification
development. Exp. Cell. Res. 284:140–149. of signal peptide peptidase, a presenilin-type aspar-
54. Stanley, N. R., K. Findlay, B. C. Berks, and T. tic protease. Science 296:2215–2218.
Palmer. 2001. Escherichia coli strains blocked in 67. Weiner, J. H., P. T. Bilous, G. M. Shaw, S. P.
Tat-dependent protein export exhibit pleiotrophic Lubitz, L. Frost, G. H.Thomas, J.A. Cole, and
defects in the cell envelope. J. Bacteriol. 183: R. J.Turner. 1998 A novel and ubiquitous system
139–144. for membrane targeting and secretion of cofactor
55. Stevenson, L. G., K. Strisovsky, K. M. containing proteins. Cell. 93:93–101.
Clemmer, S. Bhatt, M. Freeman, and P. N. 68. Wexler, M., F. Sargent, R. L. Jack, N. R.
Rather. 2007. The rhomboid protease AarA Stanley, E. G. Bogsch, C. Robinson, B. C.
mediates quorum sensing in Providencia stuartii by Berks, and T. Palmer. 2000.TatD is a cytoplas-
activating TatA of the twin arginine translocase. mic protein with DNase activity. No requirement
Proc. Natl.Acad. Sci USA 104:1003–1008. for TatD family proteins in sec-independent pro-
56. Sturtevant, M. A., M. Roark, and E. Bier. tein export. J. Biol. Chem. 275:16717–16722.
1993.The Drosophila rhomboid gene mediates the 69. Wolfe, M. S., and R. Kopan. 2004. Intramem-
localized formation of wing veins and interacts brane proteolysis: theme and variations. Science
genetically with components of the EGF-R sig- 305:1119–1123.
naling pathway. Genes Dev. 7:961–973. 70. Wolfe, M. S., W. Xia, B. L. Ostaszewski, T. S.
57. Tolia, A., L. Chavez-Gutierrez, and B. De Diehl,W.T. Kimberly, and D. J. Selkoe. 1999.
Strooper. 2006. Contribution of presenilin trans- Two transmembrane aspartates in presenilin-1
membrane domains 6 and 7 to a water-containing required for presenilin endoproteolysis and
cavity in the gamma-secretase complex. J. Biol. gamma-secretase activity. Nature 398:513–517.
Chem. 281:27633–27642. 71. Wu, Z., N.Yan, L. Feng,A. Oberstein, H.Yan,
58. Tsruya, R., A. Schlesinger, A. Reich, L. R. P. Baker, L. Gu, P. D. Jeffrey, S. Urban, and
Gabay,A. Sapir, and B. Z. Shilo. 2002.Intracel- Y. Shi. 2006. Structural analysis of a rhomboid
lular trafficking by Star regulates cleavage of the family intramembrane protease reveals a gating
Drosophila EGF receptor ligand Spitz. Genes Dev. mechanism for substrate entry. Nat. Struct. Mol.
16:222–234. Biol. 13:1084–1091.
QUORUM SENSING IN FUNGI
Claire C.Tseng and Gerald R. Fink
28
The ubiquity of quorum-sensing molecules which single-celled organisms are capable. In
and recognition systems in bacteria might lead the presence of mating pheromones, two and
one to believe that such molecules and systems only two cells fuse to make a diploid. This
are also widespread in other single-celled monogamy is a precise counting mechanism
eukaryotes, such as fungi. However, sequence that prevents formation of higher ploidy cells
analysis of fungal genomes, such as those of Sac- that would disrupt the normal life cycle.
charomyces cerevisiae and Candida albicans, does Mating pheromones have been characterized in
not reveal any homologues to the genes com- both ascomycete (which form spores inside
mon in the gram-negative and gram-positive asci) and basidiomycete (which form spores on
bacterial quorum-sensing systems. In recent basidia) fungi, and also in the more phylogenet-
years, evidence of systems for cell-cell commu- ically distant zygomycete fungi (which form
nication unique to fungi has begun to emerge, zygospores).
and this chapter provides a synthesis of the The most well studied of the more recently
known examples of fungal quorum sensing. identified quorum-sensing molecules in fungi
One family of systems for cell-cell com- are small primary alcohols, thus chemically dif-
munication in fungi has been known for ferent from the acyl-homoserine lactones and
some time, namely the secretion and recog- modified peptides preferred by bacteria.These
nition of mating pheromones. The most primary alcohols include farnesol and tyrosol in
well-characterized systems are the peptide C.albicans,and phenylethanol and tryptophol in
pheromone pathways in S. cerevisiae and S. cerevisiae.These molecules are involved in the
Schizosaccharomyces pombe.These are among the regulation of dimorphism and biofilm forma-
first described systems for intercellular commu- tion in their respective organisms in response to
nication among single-celled organisms and are environmental signals.
beautiful examples of the exquisite level of In addition to the emerging primary alco-
recognition and response to one another of hol family of intercellular signaling mole-
cules, several other molecules have also been
Claire C. Tseng and Gerald R. Fink Whitehead Institute for described that are important for intercellular
Biomedical Research, Cambridge, Massachusetts 02142. communication among fungi. This chapter
443
444 ■ TSENG AND FINK
touches on these molecules after a review of characterized. In each system, generally one
the mating pheromones and a more detailed pheromone is quite hydrophobic due to car-
discussion of the primary alcohol quorum- boxymethylation and S-farnesylation on a C-
sensing molecules. terminal cysteine, while the other pheromone
is unmodified, and therefore more hydrophilic.
MATING PHEROMONES The mating pheromones used by zygomycete
S. cerevisiae haploid cells secrete mating fungi, such as Mucor mucedo, are not peptide-
pheromones that initiate at low nanomolar based, but rather carotene-derived compounds,
concentrations a mating response pathway in such as trisporic acid (42) (Fig. 1).
cells of the opposite mating type, resulting in The secretion and recognition of the fungal
the fusion of two cells of opposite mating types peptide pheromones, as well as their structures,
to form stable diploid a/ cells (reviewed in are reminiscent of the quorum-sensing systems
reference 26). Haploid cells of the a mating used by gram-positive bacteria (reviewed in ref-
type produce a-factor and respond to -factor, erence 52). In both the bacterial and fungal sys-
whereas cells of the mating type produce - tems, the peptide signals are secreted out of the
factor and respond to a-factor.The structures of producer cell, as they are not diffusible across
the two pheromones were determined to be the membrane, and are recognized by dedicated
peptides:a-factor is a dodecapeptide with a car- receptors that propagate the signal into the tar-
boxymethylated and S-farnesylated C-terminal get cell. In gram-positive bacteria such as
cysteine (2), while -factor is a tridecapeptide Staphylococcus aureas, the receptor is a two-
(48) (Fig. 1). component sensor kinase,whereas fungi such as
Mating peptide pheromones of several other S. cerevisiae use G-protein-coupled receptors.
ascomycete and basidiomycete fungi, including Detailed study of the peptide pheromone
those of S. pombe (6), Ustilago maydis (47), and pathway in S. cerevisiae has provided unique
Rhodosporidium toruloides (19), have also been insight into an elegantly fine-tuned cell-cell sig-
FIGURE 1 Structures of various molecules used by fungi for cell-cell communication.

28. QUORUM SENSING IN FUNGI ■ 445
naling system. Numerous biological responses (15) and aromatic alcohol tyrosol (5) in C. albi-
occur in response to stimulation by mating cans, and the aromatic alcohols phenylethanol
pheromones, including synthesis of cell surface and tryptophol in S. cerevisiae (4) (Fig. 1). The
molecules for agglutination with the mating latter three are derived from the aromatic
partner, arrest in the G1 phase of the cell cycle amino acids tyrosine, phenylalanine, and tryp-
to achieve synchrony for mating, and formation tophan, respectively (Fig. 2). Each molecule
of mating projections involved in the fusion increases in the growth medium in proportion
process (reviewed in reference 26). The exact to cell density and initiates a change in cell
concentration of pheromone is important for physiology upon reaching a critical concentra-
the morphological events that lead to cell tion.Tryptophol also autostimulates its produc-
fusion.For example,cells exhibit a half-maximal tion levels.These characteristics are all requisites
response for agglutination induction and cell of bacterial quorum-sensing molecules.
division arrest at 1010 M levels of pheromone, All four primary alcohol quorum-sensing
but the half-maximal response for projection molecules influence the morphological transi-
formation occurs only at pheromone concen- tion of their respective organisms between the
trations 2 orders of magnitude greater (29).The yeast form and the filamentous, mycelial form,
cells’ graded response provides checkpoints to in response to environmental cues. Effects of
return to vegetative growth if mating becomes the molecules on biofilm formation have also
no longer feasible. S. cerevisiae obtains this been observed, which is expected, as genes
remarkable level of response regulation through required for the filamentous form are impor-
the modulation of two parallel mitogen- tant in biofilm formation in both C.albicans and
activated protein kinase (MAPK) pathways (10). S. cerevisiae (39, 40). Dimorphism and biofilm
A final interesting aspect of the fungal formation have been shown to be important
mating pathways is their connection to inva- for pathogenicity and increased drug resistance
sive growth and virulence. In S. cerevisiae, in fungi (reviewed in references 18 and 27), and
members of the same MAPK signaling cas- thus the study of the biosynthesis and signal
cade are involved in both pheromone response transduction pathways of these quorum-
and agar invasion in haploid cells (41), and in sensing molecules could lead to promising new
U. maydis, the loci expressed in response to antifungal targets.
mating pheromone contain genes involved
in filamentous growth and pathogenicity Farnesol
(11). This coordination between mating and Under otherwise identical conditions, C. albi-
morphogenesis in fungi is analogous to the con- cans grows in the yeast form when inoculated at
nection between competence and virulence in 106 cells/ml, but as mycelia and germ tubes
Streptococcus pneumoniae and other gram-positive when inoculated at 106 cells/ml. Farnesol,
bacteria, which is also regulated by quorum- specifically the E,E-isomer, was identified as
sensing peptides (reviewed in reference 49).This the component of conditioned medium that
link also demonstrates the use by fungi of the prevents the conversion of C. albicans from the
mating pheromones to propagate environmen- yeast form to the mycelial form at higher cell
tal signals, which is also a function of the pri- densities, where conditioned medium is the fil-
mary alcohol family of fungal quorum-sensing trate from a stationary-phase culture (15,34,46)
molecules, as discussed in the next section. (Fig. 1).The molecule has no effect on growth
rate, and its production is not dependent on the
PRIMARY ALCOHOL QUORUM- nature of the components of the growth
SENSING MOLECULES medium, reaching a concentration of 10 to 50
Four primary alcohols have recently been iden- M in the medium of stationary-phase cultures
tified and characterized as quorum-sensing regardless of the carbon or nitrogen source (15).
molecules in fungi: the sesquiterpene farnesol In addition, farnesol reversibly inhibits biofilm
drugs, such as fluconazole and terbinafine, that

target ergosterol biosynthesis and therefore have
clinical relevance. For example, C. albicans cells
exposed to subinhibitory concentrations of flu-
conazole so that they secrete 10 times more far-
nesol than do untreated cells are 4 to 8 times
more lethal to mice than are untreated cells (32).
At present, little is known about farnesol’s
mode of action at the molecular level or the
effector proteins associated with the propaga-
tion of its effect. Structure-activity analysis of
farnesol reveals that all parts of the molecule,
including the head group, chain length, double
bonds, and hydrophobic tail, are critical for full
activity, as none of 50 characterized analogs has
greater than 12% of the activity of farnesol (45,
46). However, the fluorescent analogs may be
useful in elucidating the binding partners for
the molecule and for pharmacokinetic studies.
Genetic evidence indicates that the histidine
kinase Chk1p is important for mediating the
effects of farnesol, as a chk1/chk1 mutant is
unresponsive to farnesol’s inhibitory effects on
filamentation and biofilm formation (20). As
Chk1p is a cytoplasmic protein without an
apparent binding motif for farnesol, it is not
likely to be a receptor, but rather an intermedi-
ate component in the signaling pathway initi-
ated by farnesol. Microarray expression analysis
on C. albicans cells exposed to farnesol upon
dilution from stationary phase (9) and in devel-
FIGURE 2 Biosynthetic pathways of tyrosol, tryp- oping biofilms (3) has also been performed.
tophol, and phenylethanol in S. cerevisiae. Both analyses indicate decreased expression lev-
els of genes associated with hyphal formation,
formation by C. albicans, although it does not and increased expression levels of genes related
affect the elongation of preexisting hyphae to drug resistance, in the presence of farnesol.
(38),and the molecule also plays a role in oxida- Although farnesol does not affect the
tive stress resistance (53). growth rate of C. albicans, it does have a potent
C. albicans likely synthesizes farnesol by growth inhibitory effect on several other fungi,
dephosphorylation of farnesyl pyrophosphate, including Aspergillus nidulans (43) and S. cere-
an intermediate in sterol biosynthesis, although visiae (25). Similarly, farnesol shows antibacter-
the phosphatase responsible has not yet been ial activity against certain bacteria, such as
identified (16). By inhibiting the ergosterol Propionibacterium acnes (21) and Haloferax volcanii
biosynthetic pathway downstream of farnesyl (50). In addition, farnesol induces apoptosis in
pyrophosphate, the intracellular and extracellu- tobacco cells (13) and inhibits the proliferation
lar levels of farnesol can be increased up to 45- of human acute leukemia cells (28).The broad
fold (16, 17). The increased levels of farnesol inhibitory effects of farnesol indicate that C.
may affect the efficacy of the many antifungal albicans may also use the molecule to gain selec-
tive advantage against other microorganisms cultures at high density and biofilm formation
and host cells in its natural environment. (15,38).However,tyrosol does not directly alter
the activity of farnesol, even at 16-fold molar
Tyrosol excess, indicating that the effect of farnesol is
A second quorum-sensing molecule produced dominant (1, 33). The balance between the
by C. albicans is tyrosol (Fig. 1). When a high- effects of the two molecules in the culture
density culture of C. albicans is diluted into fresh medium over time was examined regarding
medium at low density, there is a long lag phase biofilm formation, indicating that the effects of
before growth resumes and a morphological tyrosol prevail in the early stages of biofilm for-
conversion occurs—the yeast cells form long mation, while those of farnesol are foremost in
protuberances called germ tubes. This lag can the later stages (1). Further studies on the cellu-
be abolished by the addition of conditioned lar processes responsible for the biological
medium, and tyrosol was found to be the active activity of these compounds, including the
component responsible for this (5). When receptors and downstream signaling pathways,
added to C. albicans cultures at low density, would provide insights into the mechanisms by
tyrosol specifically shortens the lag phase, with which these two molecules dynamically modu-
no effect on the rate of exponential growth (5). late the response of C. albicans to environmental
It also promotes germ tube formation in yeast signals. It would be interesting to learn whether
form cells (5) and stimulates production of such regulation is analogous to the parallel
hyphae during the early stages of biofilm devel- MAPK signaling cascades used in the S. cere-
opment (1).Tyrosol is present in the medium of visiae pheromone pathways to adjust response
stationary-phase cultures at a concentration of based on pheromone concentration.
3 to 10 M (1, 5).
Tyrosol is synthesized through the transami- Phenylethanol and Tryptophol
nation, followed by decarboxylation and reduc- Two quorum-sensing molecules in S. cerevisiae
tion, of tyrosine (44) (Fig. 2). This amino acid have recently been identified, phenylethanol
catabolism route is known as the Ehrlich path- and tryptophol (Fig. 1). Either molecule stimu-
way (8). It has been known for some time that lates pseudohyphal growth in diploid S.cerevisiae
tyrosol is present in the culture medium of C. under nitrogen-poor conditions,and their com-
albicans, but its quorum-sensing function was bined effect is synergistic (4).Phenylethanol also
not initially uncovered (31).It is also secreted by increases invasive growth in haploids, and the
S. cerevisiae, in amounts comparable to those further addition of tryptophol results in even
produced by C. albicans, although its function in greater invasive growth, but tryptophol alone
S. cerevisiae is not yet known (4). Tyrosol is a has no effect, again indicating synergy between
component of virgin olive oil and has been the two molecules (4). In stationary-phase cul-
studied extensively in that context due to its tures grown in ammonium-free medium,
antioxidant properties (30). It does not, how- phenylethanol reaches concentrations of 5 to 8
ever, appear to play a role in oxidative stress M,and tryptophol reaches concentrations of 1
resistance in C. albicans as farnesol does (53). to 2 M (4).
As with farnesol, very little is known about Analogously to tyrosol, phenylethanol and
the molecular basis for tyrosol’s mode of action. tryptophol are synthesized through the
Microarray expression analysis indicates that transamination, followed by decarboxylation
the presence of tyrosol prevents the dramatic and reduction, of phenylalanine and trypto-
decrease in the expression levels of genes phan, respectively (7) (Fig. 2). In S. cerevisiae, the
involved in DNA synthesis and cell cycle regu- transamination step of these three aromatic
lation that occurs upon culture dilution (5). amino acids is catalyzed by two transaminases,
Tyrosol promotes the reciprocal effects of Aro8p and Aro9p (51), and the decarboxylation
farnesol, which inhibits both filamentation in step by four decarboxylases, Aro10p, Pdc1p,
Pdc5p, and Pdc6p (7). The aro8aro9 double under nitrogen-poor conditions (4, 23). In C.
mutant is impaired in the production of the albicans, either molecule at relatively high con-
three alcohols and is also defective in both centration (500 M) reduces both filamenta-
diploid pseudohyphal formation and haploid tion and biofilm formation, suggesting that
invasive growth,with the defects suppressible by their roles in C. albicans are different than in S.
exogenous supplementation of the appropriate cerevisiae and also different from the role of
alcohol(s) (4). The expression of ARO8 and tyrosol in C. albicans (4).
ARO9, as well as ARO10 and PDC6, is repres-
sed by ammonium,consistent with the observed OTHER FUNGAL QUORUM-
production of tryptophol and phenylethanol SENSING MOLECULES
only under nitrogen-poor conditions (4). In addition to the mating pheromones and four
Flo11p is a gene essential for filamentous primary alcohols discussed in the previous sec-
growth in S. cerevisiae (24), and as expected, tions, there are various examples of quorum-
exogenous addition of phenylethanol and tryp- sensing-like phenomena and molecules in the
tophol could not suppress the diploid pseudo- literature, indicating that quorum sensing may
hyphal and haploid invasive defects of a flo11 indeed be as ubiquitous among fungi as it is
mutant (4). Flo11p expression is regulated by among bacteria. A few of these examples are
both protein kinase A (PKA) and MAPK path- summarized here, although in all cases, little is
ways (reviewed in reference 36), and deletion known about the pathways, receptors, or signal
analysis indicates that phenylethanol and tryp- transduction mechanisms involved.
tophol exert their effects primarily through the Upon starvation for both nitrogen and fer-
PKA pathway (4).Specifically,responsiveness to mentable carbon sources, diploid S. cerevisiae
phenylethanol and tryptophol requires Tpk2p, cells undergo meiosis and sporulation. Sporu-
the catalytic subunit of PKA, and Flo8p, the lation at low densities (106 cells/ml) is inef-
transcription factor downstream of Tpk2p, but ficient, but it proceeds readily at high densities
not Gpa2p and Gpr1p, the G-protein and G- (106 cells/ml), suggesting the presence of a
protein-coupled receptor upstream of Tpk2p, quorum-sensing molecule that promotes
indicating that Tpk2p may play a role in sensing meiosis and sporulation. Careful studies with
these quorum-sensing molecules (4). sporulation mutants and conditioned medium
Exogenous addition of tryptophol, but not indicate that alkalization of the medium via
phenylethanol or tyrosol, stimulates the pro- secretion of bicarbonate is at least in part
duction of all three aromatic alcohols, with the responsible for this density-dependent effect
autoregulation dependent on the presence of (12, 35).
the transcription activator Aro80p (4). The When grown on solid medium, colonies of
addition of tryptophol to wild-type cells upreg- various fungal species, including S. cerevisiae,
ulates the transcript levels of ARO9 and Candida mogii, and Kluyveromyces lactis, can
ARO10 25- to 30-fold but has no effect on the communicate with each other using pulses of
transcripts in aro80 mutants (4). Microarray ex- the volatile compound ammonia (37). The
pression analysis intriguingly indicates that the pulses result in growth inhibition of the parts of
presence of tryptophol and phenylethanol, colonies facing other colonies, allowing for
either individually or together, upregulates overall coordinated growth of colonies away
largely the same subset of genes that are upreg- from one another and toward potential fresh
ulated upon entry of growing cells into station- nutrient sources instead. The ammonia pulses
ary phase (4).These results sketch out the first thus appear to be an example of interspecies
molecular details of a quorum-sensing pathway communication among fungi. The availability
in fungi (Fig. 3). of amino acids in the medium is necessary for
Both phenylethanol and tryptophol are also ammonia production, as S. cerevisiae mutants
secreted in significant quantities by C. albicans unable to internalize external amino acids nei-
FIGURE 3 Quorum-signaling pathway

involving tryptophol and phenylethanol in S.
cerevisiae. Adapted from Chen and Fink (4).
See text for details.
ther produce ammonia nor show asymmetric sensing molecules. Without answers to these
growth inhibition, suggesting that the produc- basic physiological questions, the principles
tion of ammonia is connected to the uptake of of quorum sensing unique to single-celled
amino acids (37). eukaryotes will remain obscure. Promising new
In addition, there is evidence for the exis- insights into mechanisms of fungal pathogene-
tence of quorum-sensing molecules that regu- sis and targets for the development of antifungal
late dimorphism in Histoplasma capsulatum (22), drugs also depend on the unraveling of these
Ceratocystis ulmi (14), and several other fungi puzzles.
(33), although no chemical identities have been
determined.The molecule in C.ulmi has similar REFERENCES
effects on the morphogenesis of that organism 1. Alem, M. A. S., M. D.Y. Oteef, T. H. Flowers,
as farnesol in C. albicans, but farnesol cannot and L. J. Douglas. 2006. Production of tyrosol by
substitute for it in C. ulmi, and conversely, con- Candida albicans biofilms and its role in quorum
sensing and biofilm development. Eukaryot. Cell
ditioned medium from C. ulmi has no effect on 5:1770–1779.
C. albicans (14).This indicates the specificity of 2. Anderegg, R. J., R. Betz, S.A. Carr, J.W. Crabb,
these molecules for their respective organisms, and W. Duntze. 1988. Structure of Saccharomyces
similar to the species specificity of quorum- cerevisiae mating hormone a-factor. J. Biol. Chem.
sensing molecules in bacteria. 263:18236–18240.
3. Cao, Y.-Y., Y.-B. Cao, Z. Xu, K. Ying, Y. Li,
Y. Xie, Z.-Y. Zhu, W.-S. Chen, and Y.-Y.
CONCLUSIONS Jiang. 2005. cDNA microarray analysis of differen-
From the available data, it appears that quorum tial gene expression in Candida albicans biofilm
sensing is indeed a prevalent mechanism used exposed to farnesol. Antimicrob. Agents Chemother.
by fungi to modulate response to each other 49:548–549.
4. Chen, H., and G. R. Fink. 2006. Feedback con-
and their environment, through systems unique trol of morphogenesis in fungi by aromatic alcohols.
from those used in bacteria. If the S. cerevisiae Genes Dev. 20:1150–1161.
peptide pheromone response pathway is repre- 5. Chen, H., M. Fujita, Q. Feng, J. Clardy, and G.
sentative, then fungal quorum-sensing path- R. Fink. 2004.Tyrosol is a quorum-sensing mole-
ways are well-developed systems that modulate cule in Candida albicans. Proc. Natl. Acad. Sci. USA
101:5048–5052.
response based on the integration of quorum- 6. Davey, J. 1992. Mating pheromones of the fission
sensing molecule concentration and environ- yeast Schizosaccharomyces pombe: purification and
mental signals. However, this area of research is structural characterization of M-factor and isolation
still in its infancy, with a great deal of work and analysis of two genes encoding the pheromone.
remaining to be done in the elucidation of the EMBO J. 11:951–960.
7. Dickinson, J. R., L. E. J. Salgado, and M. J. E.
receptors and signal transduction pathways Hewlins. 2003. The catabolism of amino acids to
involved in response to signals, as well as the long chain and complex alcohols in Saccharomyces
chemical identification of additional quorum- cerevisiae. J. Biol. Chem. 278:8028–8034.
8. Ehrlich, F. 1907. Über die Bedingungen der protein Chk1p regulates quorum sensing in Can-
Fuselölbildung und über ihren Zusammenhang dida albicans. Eukaryot. Cell 3:1062–1065.
mit dem Eiweiaufbau der Hefe. Ber. Dtsch. Chem. 21. Kubo, I., H. Muroi, and A. Kubo. 1994. Natu-
Ges. 40:1027–1047. rally occurring antiacne agents. J. Nat. Prod.
9. Enjalbert, B., and M.Whiteway. 2005. Release 57:9–17.
from quorum-sensing molecules triggers hyphal 22. Kügler, S., T. S. Sebghati, L. G. Eissenberg,
formation during Candida albicans resumption of and W. E. Goldman. 2000. Phenotypic variation
growth. Eukaryot. Cell 4:1203–1210. and intracellular parasitism by Histoplasma capsula-
10. Farley, F. W., B. Satterberg, E. J. Goldsmith, tum. Proc. Natl.Acad. Sci. USA 97:8794–8798.
and E. A. Elion. 1999. Relative dependence of 23. Lingappa, B.T., M. Prasad,Y. Lingappa, D. F.
different outputs of the Saccharomyces cerevisiae Hunt, and K. Biemann. 1969. Phenethyl alco-
pheromone response pathway on the MAP kinase hol and tryptophol: autoantibiotics produced by
Fus3p. Genetics 151:1425–1444. the fungus Candida albicans. Science 163:192–194.
11. Hartmann, H. A., R. Kahmann, and M. 24. Lo,W.-S., and A. M. Dranginis. 1998.The cell
Bölker. 1996. The pheromone response factor surface flocculin Flo11 is required for pseudohy-
coordinates filamentous growth and pathogenicity phae formation and invasion by Saccharomyces cere-
in Ustilago maydis. EMBO J. 15:1632–1641. visiae. Mol. Biol. Cell 9:161–171.
12. Hayashi, M., K. Ohkuni, and I. Yamashita. 25. Machida, K.,T.Tanaka,Y.Yano, S. Otani, and
1998.An extracellular meiosis-promoting factor in M. Taniguchi. 1999. Farnesol-induced growth
Saccharomyces cerevisiae.Yeast 14:617–622. inhibition in Saccharomyces cerevisiae by a cell cycle
13. Hemmerlin,A., and T. J. Bach. 2000. Farnesol- mechanism. Microbiology 145:293–299.
induced cell death and stimulation of 3-hydroxy- 26. Madhani, H. D. 2007.From a to :Yeast as a Model
3-methylglutaryl-coenzyme A reductase activity for Cellular Differentiation. Cold Spring Harbor
in tobacco cv Bright Yellow-2 cells. Plant Physiol. Laboratory Press, Cold Spring Harbor, NY.
123:1257–1268. 27. Madhani, H. D., and G. R. Fink. 1998. The
14. Hornby, J. M., S. M. Jacobitz-Kizzier, D. J. control of filamentous differentiation and viru-
McNeel, E. C. Jensen, D. S.Treves, and K.W. lence in fungi. Trends Cell Biol. 8:348–353.
Nickerson. 2004. Inoculum size effect in dimor- 28. Melnykovych, G., J. S. Haug, and C. M.
phic fungi: extracellular control of yeast-mycelium Goldner. 1992. Growth inhibition of leukemia
dimorphism in Ceratocystis ulmi. Appl. Environ. cell line CEM-C1 by farnesol: effets of phos-
Microbiol. 70:1356–1359. phatidylcholine and diacylglycerol. Biochem. Bio-
15. Hornby, J. M., E. C. Jensen, A. D. Lisec, J. J. phys. Res. Commun. 186:543–548.
Tasto, B. Jahnke, R. Shoemaker, P. Dussault, 29. Moore, S.A. 1983. Comparison of dose-response
and K. W. Nickerson. 2001. Quorum sensing in curves for factor-induced cell division arrest,
the dimorphic fungus Candida albicans is mediated agglutination, and projection formation of yeast
by farnesol.Appl.Environ.Microbiol. 67:2982–2992. cells. J. Biol. Chem. 258:13849–13856.
16. Hornby, J. M., B. W. Kebaara, and K. W. 30. Moreno, J. J. 2003. Effect of olive oil minor com-
Nickerson. 2003. Farnesol biosynthesis in Can- ponents on oxidative stress and arachidonic acid
dida albicans: cellular response to sterol inhibition mobilization and metabolism by macrophages
by zaragozic acid B. Antimicrob. Agents Chemother. RAW 264.7. Free Radic. Biol. Med. 35:1073–1081.
47:2366–2369. 31. Narayanan, T. K., and G. R. Rao. 1976. Pro-
17. Hornby, J. M., and K. W. Nickerson. 2004. duction of -(4-hydroxyphenyl)ethanol and -(4-
Enhanced production of farnesol by Candida albi- hydroxyphenyl)lactic acid by Candida species.Can.
cans treated with four azoles. Antimicrob. Agents J. Microbiol. 22:384–389.
Chemother. 48:2305–2307. 32. Navarathna, D. H. M. L. P., J. M. Hornby,
18. Jabra-Rizk, M. A., W. A. Falkler, and T. F. N. Hoerrmann, A. M. Parkhurst, G. E.
Meiller. 2004. Fungal biofilms and drug resist- Dunhamel, and K. W. Nickerson. 2005.
ance. Emerg. Infect. Dis. 10:14–19. Enhanced pathogenicity of Candida albicans pre-
19. Kamiya, Y., A. Sakurai, S. Tamura, N. treated with subinhibitory concentrations of flu-
Takahashi, K. Abe, E. Tsuchiya, S. Fukui, C. conazole in a mouse model of disseminated
Kitada, and M. Fujino. 1978. Structure of candidiasis.J.Antimicrob.Chemother. 56:1156–1159.
rhodoturucine A, a novel lipopeptide, inducing 33. Nickerson, K. W., A. L. Atkin, and J. M.
mating tube formation in Rhodosporidium toruloides. Hornby. 2006. Quorum sensing in dimorphic
Biochem. Biophys. Res. Commun. 83:1077–1083. fungi:farnesol and beyond.Appl.Environ.Microbiol.
20. Kruppa, M., B. P. Krom, N. Chauhan, A. V. 72:3805–3813.
Bambach, R. L. Cihlar, and R. A. Calderone. 34. Oh, K.-B., H. Miyazawa, T. Naito, and H.
2004. The two-component signal transduction Matsuoka. 2001. Purification and characteriza-
tion of an autoregulatory substance capable of reg- tic interactions between fungi. Mol. Microbiol.
ulating the morphological transition in Candida 59:753–764.
albicans. Proc. Natl.Acad. Sci. USA 98:4664–4668. 44. Sentheshanmuganathan, S., and S. R. Elsden.
35. Ohkuni, K., M. Hayashi, and I. Yamashita. 1958.The mechanism of the formation of tyrosol
1998. Bicarbonate-mediated social communica- by Saccharomyces cerevisiae. Biochem. J. 69:210–218.
tion stimulates meiosis and sporulation of Saccha- 45. Shchepin, R., R. Dumitru, K. W. Nickerson,
romyces cerevisiae.Yeast 14:623–631. M. Lund, and P. H. Dussault. 2005.Biologically
36. Palecek, S. P., A. S. Parikh, and S. J. Kron. active fluorescent farnesol analogs. Chem. Biol.
2002. Sensing, signalling and integrating physical 12:639–641.
processes during Saccharomyces cerevisiae invasive and 46. Shchepin, R., J. M. Hornby, E. Burger, T.
filamentous growth. Microbiology 148:893–907. Niessen, P. Dussault, and K. W. Nickerson.
37. Palková, Z., B. Janderová, J. Gabriel, B. 2003.Quorum sensing in Candida albicans:probing
Zikánová, M. Pospíšek, and J. Forstová. 1997. farnesol’s mode of action with 40 natural and syn-
Ammonia medicated communication between thetic farnesol analogs. Chem. Biol. 10:743–750.
yeast colonies. Nature 390:532–536. 47. Spellig, T., M. Bölker, F. Lottspeich, R. W.
38. Ramage, G., S. P. Saville, B. L.Wickes, and J. Frank, and R. Kahmann. 1994. Pheromones
L. López-Ribot. 2002. Inhibition of Candida trigger filamentous growth in Ustilago maydis
albicans biofilm formation by farnesol, a quorum- EMBO J. 13:1620–1627.
sensing molecule. Appl. Environ. Microbiol. 68: 48. Stötzler, D., H.-H. Kiltz, and W. Duntze.
5459–5463. 1976. Primary structure of -factor peptides
39. Ramage, G., K. VandeWalle, J. L. López- from Saccharomyces cerevisiae. Eur. J. Biochem.
Ribot, and B. L. Wickes. 2002.The filamenta- 69:397–400.
tion pathway controlled by the Efg1 regulator 49. Suntharalingam, P., and D. G. Cvitkovitch.
protein is required for normal biofilm formation 2005. Quorum sensing in streptococcal biofilm
and development in Candida albicans. FEMS Micro- formation. Trends Microbiol. 13:3–6.
biol. Lett. 214:95–100. 50. Tachibana,A.,T.Tanaka, M.Taniguchi, and S.
40. Reynolds,T. B., and G. R. Fink. 2001. Bakers’ Oi. 1996. Evidence for farnesol-mediated iso-
yeast, a model for fungal biofilm formation. Science prenoid synthesis regulation in a halophilic
291:806–807. archaeon, Haloferax volcanii. FEBS Lett. 379:43–46.
41. Roberts, R. L., and G. R. Fink. 1994. Elements 51. Urrestarazu, A., S. Vissers, I. Iraqui, and
of a single MAP kinase cascade in Saccharomyces M. Grenson. 1998. Phenylalanine- and tyrosine-
cerevisiae mediate two developmental programs in auxotrophic mutants of Saccharomyces cerevisiae
the same cell type: mating and invasive growth. impaired in transamination. Mol. Gen. Genet.
Genes Dev. 8:2974–2985. 257:230–237.
42. Schachtschabel, D., C. Schimek, J. Wöste- 52. Waters, C. M., and B. L. Bassler. 2005. Quo-
meyer, and W. Boland. 2005. Biological activity rum sensing: cell-to-cell communication in bacte-
of trisporoids and trisporoid analogues in Mucor ria. Annu. Rev. Cell Dev. Biol. 21:319–346.
mucedo (). Phytochemistry 66:1358–1365. 53. Westwater, C., E. Balish, and D. A. Schofield.
43. Semighini, C. P., J. M. Hornby, R. Dumitru, 2005. Candida albicans-conditioned medium pro-
K. W. Nickerson, and S. D. Harris. 2006. tects yeast cells from oxidative stress: a possible link
Farnesol-induced apoptosis in Aspergillus nidu- between quorum sensing and oxidative stress
lans reveals a possible mechanism for antagonis- resistance. Eukaryot. Cell 4:1654–1661.
QUORUM SENSING IN ROTIFERS
Julia Kubanek and Terry W. Snell
29
Induction of sexual reproduction (mixis) that quorum sensing, as a mechanism for coor-
among monogonont rotifers appears to be the dinating behavior, is at least as common in ani-
first described example of quorum sensing mals as in bacteria. However, there are few
among aquatic animals. Rotifer quorum sens- examples in animals and none where the chem-
ing involves a mixis-inducing protein (MIP) ical signals involved in synchronizing behav-
released and sensed by female rotifers that trig- ioral interactions have been fully described.
gers a life cycle change to produce males, lead- Aquatic animals may be especially likely candi-
ing to copulation between males and mictic dates for using quorum sensing, since their
females, and the production of diapausing eggs, watery medium should permit the reliable
which can lay dormant for months to years.To transmission of dissolved chemical signals.
date, the MIP of only one rotifer, Brachionus pli- Many aquatic invertebrates live in close prox-
catilis, has been partially characterized, although imity to conspecifics, sometimes as colonies of
other MIPs are suspected based on the species genetically related or even clonal individuals,
and occasional population specificity of sex suggesting that intraspecific communication
induction. Rotifer quorum-sensing is similar in could be important. It therefore seems reason-
many respects to the more intensely studied able to hypothesize that quorum sensing is used
bacterial quorum-sensing processes, suggesting by aquatic animals for the regulation of various
that chemosensory processes for assessing con- life cycle processes that require coordination
specific population density may have ancient with conspecifics, such as reproduction, devel-
origins. opment, or entry into dormancy. In addition,
other types of chemically mediated interactions
IS THERE QUORUM SENSING have been described among a wide variety of
AMONG AQUATIC ANIMALS? aquatic invertebrates, including chemical cues
Coordination of behavior is ubiquitous within to attract and recognize mates, inhibit competi-
the animal kingdom, and so we might expect tors, deter predators, and suppress colonization
by potential parasites (19). Many aquatic ani-
Julia Kubanek School of Biology and School of Chemistry mals have keen chemosensory abilities that
and Biochemistry, Georgia Institute of Technology, Atlanta,
Georgia 30332. Terry W. Snell School of Biology, enable them to “smell” predator activity and to
Georgia Institute of Technology,Atlanta, Georgia 30332. differentiate, using chemical cues, between
453
454 ■ KUBANEK AND SNELL
predator attack of conspecifics versus het- the first clear demonstration of a quorum-
erospecifics (14). Thus, it seems likely that the sensing system in an aquatic animal.
reason why we have so few examples of quo-
rum sensing in aquatic animals is not because it ROTIFER MIXIS:
is rare, but that we have not looked for it. HISTORICAL PERSPECTIVE
A few behavioral and physiological processes Monogonont rotifers in freshwater and pro-
in aquatic invertebrates that appear to be coor- tected coastal marine environments are cyclical
dinated among conspecifics are likely candi- parthenogens, reproducing asexually with
dates for regulation by quorum sensing. For episodic bouts of sexual reproduction (38).
example, it has long been recognized that many Early observation of their life history suggested
corals and other sessile invertebrates living in that, as with daphnids, rotifer cyclical partheno-
dense colonies synchronize the release of genesis is adaptive, particularly among popula-
gametes and zygotes, probably to minimize tions in seasonal environments (9,24,39).When
predatory loss (13). However, evidence to date rotifer population densities are low and condi-
suggests that reproductive synchronization is tions for growth are favorable,asexual reproduc-
directly controlled by environmental signals tion predominates and the population consists
like solar insolation and lunar cycles,rather than of females that produce clonal diploid eggs that
by diffusible chemical cues from the corals develop into females. Sometime during the
themselves (13, 35). In contrast, freshwater growing season,typically when population den-
hydrozoans of the genus Hydra undergo sexual sity is high,diapausing eggs are produced;other-
differentiation at a rate controlled at least partly wise the rotifer population will go locally
by the concentration of dissolved carbon diox- extinct during the harsh conditions of winter.
ide, which correlates with high population Because resting (diapausing) eggs are solely the
density (16). Given that CO2 is a nonspecific product of sexual reproduction among mono-
signal indicating the presence of any organism gonont rotifers, induction of sexuality (mixis)
engaged in respiration, decomposition, or fer- must precede resting egg production (37).Upon
mentation, it is not clear how such a signal pro- reaching a population density threshold, female
vides information about the presence of rotifers produce mictic daughters that produce
conspecific Hydra, nor whether there is a CO2 haploid eggs that, if unfertilized, develop into
concentration threshold (quorum) above haploid males.When these males fertilize other
which sexual forms develop. Similarly, crowd- mictic females, the resulting zygote is a resting
ing in the cladoceran Daphnia has been associ- egg (cyst) that is deposited in sediments or dis-
ated with the induction of sexual reproduction persed to another habitat, diapausing until the
and the production of diapausing eggs, and the following growing season (38). Resting eggs
combination of an unidentified dissolved hatch when exposed to increased temperature,
chemical signal, food limitation, and shortened light, and moisture, typically at the beginning of
photoperiod have been implicated in this a new growing season (6, 7, 12, 21).
process (15, 33). For a convincing example of Gilbert (8) showed that conspecific crowd-
quorum sensing in aquatic animals to emerge,it ing causes mixis among brachionid rotifers,
must satisfy several criteria: the chemical signal without the need for other environmental sig-
must be produced by the animals themselves; nals such as altered temperature, light, or food
there must be some degree of species specificity availability (10, 31).The effect of crowding on
in the signal to communicate information freshwater rotifer mixis appears to be generally
about the density of conspecifics; and a coordi- species specific, and even population specific,
nated response of conspecifics must occur once for Brachionus calyciflorus (8, 10). However, for
density has exceeded a certain threshold. None the salt-tolerant B. plicatilis species complex,
of the examples cited above meet these criteria, specificity of the mixis cue is low: members of
but the example of sex induction in rotifers this species complex can induce mixis in each
described below is an excellent candidate for other (32). Among conspecific B. plicatilis, the
29. QUORUM SENSING IN ROTIFERS ■ 455
protease; (ii) protected by treatment with pro-

tease inhibitors; (iii) retained on an anion
exchange chromatography column; and (iv) of
high molecular weight, passing through a 100-
kDa filter but retained by a 10-kDa filter (28).
The ready loss of sex-inducing activity from
conditioned medium or its chromatographed
fractions when not protected by protease
inhibitors suggests that the MIP is unstable
and/or easily denatured.The nonlinear density-
dependent response of female rotifers (Fig. 1)
may indicate a positive feedback effect on the
FIGURE 1 Evidence for a population density-
dependent, nonlinear response on rotifer sex induction production of MIP as with quorum sensing in
(mixis).The x axis label refers to the volume of water in gram-negative bacteria (17), or that chemore-
which a single female B. plicatilis rotifer was cultured for ception itself may be subject to positive feed-
48 h,and the y axis reports the percentage of her daugh- back, such that females become more sensitive
ters that became mictic (thus, produced males). to the MIP with increased exposure to the MIP.
Redrawn from Marine Biology (28) with kind permis-
sion of Springer Science and Business Media. Several structural characteristics of the MIP
have been elucidated, although the full
sequence and three-dimensional structure of
mixis response to crowding is nonlinear with a the protein have yet to be determined (28).
clear threshold (Fig. 1) (28), thus suggesting a The molecular weight of the MIP is clearly
“quorum”-type process. between 10 and 100 kDa,based on the results of
Medium conditioned by B. plicatilis rotifers the rotifer mixis assay following ultrafiltration
caused increased mixis rates in conspecific and dialysis. Analysis by Sodium dodecyl sul-
females exposed in a laboratory assay, strongly fate-polyacrylamide gel electrophoresis sug-
suggesting accumulation of a density- gested a molecular weight of 39 kDa; however,
dependent chemical cue (3, 31). Only develop- on occasion other protein bands representing
ing embryos exposed to conditioned medium 16, 23, and 42 kDa have been observed in
before egg extrusion were responsive to the mixis-inducing fractions (Stout, Snell, and
mixis-inducing cue, indicating that this cue acts Kubanek, unpublished data).The strong reten-
on female rotifer embryos early in their devel- tion of the mixis-inducing fraction by anion
opment (11, 28). The fact that sex in mono- exchange chromatography suggests overall
gonont rotifers is induced by a chemical cue anionic character, whereas strong retention by
released by conspecifics in a density-dependent C3 reversed-phase high-performance liquid
manner, with induction occurring in a coordi- chromatography (HPLC) suggests a fairly
nated fashion above a threshold concentration, hydrophobic protein (Fig. 3). HPLC purifica-
fulfills the traditional criteria for a quorum- tion of adequate quantities of MIP to enable
sensing process (Fig. 2). Evidence that the mixis crystallization for X-ray diffraction analysis,
signal is a protein secreted by rotifers is nuclear magnetic resonance solution structure
described below. determination, or even mass spectral analysis
have been impeded by the very low concentra-
Experimental Evidence for tion of MIP produced by rotifers and the sus-
Mixis-Inducing Protein (MIP) ceptibility of the MIP to decomposition under
as a Quorum-Sensing Signal laboratory conditions.
The chemical cue causing mixis in B. plicatilis N-terminal sequencing by Edman degrada-
rotifers appears to be proteinaceous, since the tion of a 39-kDa candidate MIP yielded a
sex-inducing activity of conditioned medium 17-amino-acid sequence of dVNGGGAT-
was (i) abolished by treatment with a general LPQpLYQTA, with lowercase letters indicat-
FIGURE 2 Conceptual model for quorum sensing in the rotifer B. plicatilis, in which the MIP acts as autoinducer, and interaction of the MIP with
a chemoreceptor located on the body surface of female conspecifics induces mixis in 10 to 30% of the population, once a threshold of approximately
71 females per liter has been reached.
FIGURE 3 (Top) High-performance liquid chromatogram of mixis-inducing fraction from conditioned

medium of the rotifer B.plicatilis,using a C3 reversed-phase column with gradient elution of aqueous methanol (AU,
absorbance units at 280 nm; numbers above peaks refer to retention times). (Bottom) Mixis activity of HPLC frac-
tions (11 to 25 min retention time).NC,negative control;PC,positive control;numbers inside of bars indicate num-
ber of females scored;asterisks indicate data significantly greater than negative control data by G test.Reprinted from
Marine Biology (28) with kind permission of Springer Science and Business Media.
ing tentatively assigned amino acids (28). This taxa such as cholesterol and its derivatives have
sequence was 100% identical to the N terminus been identified from rotifers (34). If the rotifer
of a partially characterized human steroidogen- MIP induces mixis via steroid production,these
esis-inducing protein (SIP) (Genbank accession steroids are likely to be unusual in their molec-
number P83897), with no strong homology to ular structures, since steroids like cholesterol
any other previously identified protein. It involved in essential metabolism could not act
should be noted that the gene for the human as specific signals.The possible involvement of
SIP has not been found in human genome unique steroids in rotifer mixis is consistent
datasets, suggesting that the SIP is a product of with the fact that aquatic invertebrates are
substantial posttranslational modification. Like known to synthesize a variety of species-
the rotifer MIP, the SIP is a secreted protein. specific steroids of unusual structure (1). Identi-
The apparent involvement of the human SIP in fication of steroids from mictic rotifer cultures
steroid production suggests a similar role for the is ongoing,toward a better understanding of the
rotifer MIP, such that MIP quorum sensing pathway for sex induction (Stout,Kubanek,and
could trigger steroid hormone production in Snell, unpublished data).
female rotifers, leading to meiotic oogenesis.To Considerable experimental work remains to
date, only steroids common to many animal be done toward full characterization of the
mixis pathway in rotifers.The complete struc- relates to its function in inducing sex. Popula-
ture of the B. plicatilis MIP is being pursued by tion genetics approaches will be used to deter-
protein purification techniques followed by mine when the mixis pathways of different
mass spectral analysis of intact protein and groups of rotifers diverged and to test hypothe-
digested peptide fragments. Given the variable ses related to the importance of rotifer mixis in
molecular weight of MIPs isolated from B. pli- evolutionary processes.
catilis, it seems possible that a larger-molecular-
weight parent protein yields active MIPs of Ecological Consequences
varying size via nonspecific cleavage. So far of Rotifer Quorum Sensing
there is no indication that multiple MIPs are Among cyclical parthenogens, the timing of
required to work synergistically to coordinate periodic sexual reproduction has a large impact
mixis, but rather that several MIPs may exist on fitness (24). Successful sexual reproduction
with molecular weights of 16 to 42 kDa, any of requires that mixis be synchronized with high
which are sufficient to cause conspecific rotifer population density so that there are sufficient
mixis (Stout, Snell, and Kubanek, unpublished male-female encounters to effect high rates of
data). Sequencing of a B. plicatilis cDNA library fertilization (27). Additional environmental
is in progress (M. Welch, unpublished data), conditions that promote mixis and resting egg
which should enable identification of the production include adequate food of good
MIP gene and resolution of whether a larger- quality, moderate temperatures and salinities,
molecular-weight parent protein is synthesized and an absence of stressors (26). Under these
before cleavage to active MIPs. Once the MIP conditions, brachionid rotifers are responsive to
gene is identified, it may be possible to knock mixis signals and are able to complete resting
out MIP transcription using RNAi to confirm egg formation. To maximize resting egg pro-
the key role of the MIP in sex induction. A duction in a given season, an optimal pattern of
genetically transformable rotifer model is also induction of sexuality, including delayed mixis,
under development to facilitate further knock- should be favored (23). Field observations indi-
out studies involving the MIP and its putative cate that mixis rates in natural Brachionus popu-
chemoreceptor. lations range from 0 to 29%, with peaks
Preliminary data support the hypothesis that occurring in late autumn (2), and phenotypes
the MIP acts via a G-protein-coupled receptor with reduced sensitivity to mixis cues exist
accessible to the secreted MIP on the surface of (22). This is consistent with predicted optimal
female rotifer bodies. The nontoxic abolish- mixis rates based on simulation models, which
ment of mixis by addition of a G-protein antag- suggest that delayed mixis enables rotifer popu-
onist, GP-Ant-2, to live rotifers (Grubbs, Snell, lations to grow quickly by asexual reproduc-
and Kubanek, unpublished data) suggested the tion, and then produce a maximum number of
involvement of a Gi, Go, or Gs protein in resting eggs when mixis occurs at high popula-
the mixis pathway (18). However, further tion densities (23).A mixis ratio of 14% with a
experiments are required to confirm this result. mixis threshold of 70 females per liter, includ-
Future identification of the MIP gene will ing a mixis delay of 8 to 10 days, was predicted
enable the testing of various hypotheses regard- to be optimal for avoiding invasion by rotifer
ing the species and population specificity of populations with different mixis strategies (23).
rotifer mixis. Comparison of MIP genes from This predicted optimal threshold is remarkably
different rotifer populations and species com- similar to the mixis threshold of 71 females per
bined with experimental manipulation of MIP liter, observed in laboratory cultures of B. pli-
structure (by laboratory synthesis of MIP- catilis (28), suggesting that rotifer sensitivity to
related peptides or by cloning and expression of the mixis-inducing cue is close to optimal.
the MIP in a heterologous host) could result in The MIP produced by rotifers appears to be
further understanding of how MIP structure chemically unstable and produced at low
(nanomolar to picomolar) concentrations (28). see reference 32). Gilbert (10) proposed that
A very dilute cue to which rotifers are highly two allopatric rotifer populations could diverge
sensitive represents a low-cost, species-specific as different mutations accumulate in their
communication system for coordinating sexual mixis-inducing cue and/or chemoreceptor,
reproduction. The sensitivity of the MIP to eventually making these two populations fail to
decomposition by proteases allows the MIP to recognize each other’s mixis signals. Alterna-
function as a “real-time” estimate of population tively, two overlapping populations whose
density.The trade-offs between accurate popu- hybrid offspring have reduced fitness could be
lation assessment and the energy costs associ- selected for divergence in their mixis-inducing
ated with loss of a complex signaling molecule cue or chemoreceptor structures. However,
to the environment have been recognized in because some rotifers of the B. plicatilis species
bacterial quorum sensing (36) and may apply complex can induce mixis in each other (32), it
similarly to rotifers.An additional advantage of seems likely that species boundaries are prima-
using a secreted protein as a quorum-sensing rily maintained by other mechanisms, such as
signal is that its complex three-dimensional the mate recognition pheromone (29).The dual
structure allows for the communication of con- chemical signaling systems for rotifer mixis (via
siderable information. A signal transduction the secreted MIP) and male-female recognition
pathway initiated by interaction between the (via contact chemoreception of the mate
MIP and a G-protein-coupled chemoreceptor recognition pheromone) share some character-
could lead to amplification of the signal, consis- istics with a dual pheromone system found in
tent with the nonlinear (threshold) response of abalone (20). These authors proposed that
rotifers to the MIP (Fig. 1). species integrity is mainly controlled by a con-
A receptor-ligand interaction involving two tact pheromone,whereas a secreted pheromone
proteins, as a first step for mixis, should act as a serves to increase the probability that gametes
sensitive conspecific recognition system and at locate each other, thereby playing more of an
the same time as a powerful barrier to mixis ecological role. For rotifers that first induce
induction by heterospecifics. Barriers to production of males under favorable environ-
hybridization by the mixis induction system mental conditions and then ensure that conspe-
could prevent the production of large numbers cific males and females recognize each other, a
of low-fitness offspring in habitats in which similar separation of ecological and evolution-
closely related rotifer species co-occur. Species ary roles may exist with the mixis and mate
boundaries are further reinforced in brachionid recognition signaling systems.
rotifers by a chemically mediated mate recogni- There also appear to be some similarities
tion system,whereby males recognize a species- between rotifer and bacterial quorum-sensing
specific glycoprotein signal on the body surface processes; for example, both involve the secre-
of females before initiating mating (25, 29, 30). tion of autoinducers into extracellular space.
The rotifer MIP,as a proteinaceous autoinducer,
EVOLUTIONARY IMPLICATIONS is more structurally similar to gram-positive
OF ROTIFER QUORUM SENSING bacterial peptide autoinducers than to acylated
It is possible that chemical signals for inducing homoserine lactones (AHLs) or to the boron-
sex played a role in promoting speciation containing autoinducer 2 of gram-negative bac-
among monogonont rotifers and may continue teria, although the MIP has a substantially larger
to play a role in maintaining species integrity molecular weight (10 kDa) than gram-
(10). Differences probably exist among bra- positive peptides. Gram-negative AHLs diffuse
chionid rotifers in the molecular structure of into bacterial cells to regulate transcription,
their mixis-inducing cues, since mixis induc- whereas gram-positive peptide autoinducers
tion has been shown to be species specific, and interact with membrane-bound chemorecep-
sometimes even population specific (8, 10, but tors, launching two-component signal trans-
duction or diffuse into cells as do AHLs (4). ing, including the production and release of an
Rotifer quorum sensing, in contrast, likely autoinducing chemical cue (MIP) by rotifers in
involves interaction of the MIP and an external a density-dependent manner, the response by
(probably membrane-bound) chemoreceptor, conspecifics upon reaching a threshold concen-
since a 10-kDa protein would not be tration, and the coordination of this response,
expected to readily enter cells. Given the appar- which includes both physiological and behav-
ent differences among these systems, it seems ioral changes. Future research is required to
likely that these mechanisms of quorum sensing fully characterize rotifer mixis inducing pro-
evolved independently in bacteria and animals. teins and their chemoreceptors, and to under-
The use by rotifers of a structurally complex stand the importance of rotifer quorum sensing
peptide-based pheromone has the advantage, in speciation, the maintenance of species
over bacterial AHLs or small peptides, of reduc- boundaries, and the evolution of sex.
ing the chance that another species will inde-
pendently evolve the same signaling system. At ACKNOWLEDGMENTS
the same time,since the amino acid sequences of Our work on rotifer reproduction and plank-
proteins are encoded directly in genes,the use of ton chemical ecology has been supported by
protein pheromones should allow rapid evolu- NSF grants BE/GenEn MCB-0412674 (to T.
tion via mutation of the gene encoding the MIP. W. S.) and OCE-0134843 (to J. K.).We thank
Since the rotifer MIP appears to share sub- M. Serra for comments that improved the man-
stantial sequence identity with the human SIP uscript and the journal Marine Biology for per-
(28), it may be that the evolution of chemical mission to reproduce the data in Fig. 1 and 3.
signaling involving these types of proteins
occurred very early in animal evolution and has REFERENCES
been conserved in diverse taxa. That both the 1. Baker, B. J., and R. G. Kerr. 1993. Biosynthesis
MIP and SIP control sexual processes seems to of marine sterols. Topics Curr. Chem. 167:1–31.
indicate an ancestral function involving sexual- 2. Carmona, M. J., A. Gomez, and M. Serra.
1995. Mictic patterns of the rotifer Brachionus pli-
ity and possibly the involvement of steroids. catilis Muller in small ponds. Hydrobiologia 313/
Quorum sensing among rotifers, therefore, 314:365–371.
would be analogous to cell-cell communica- 3. Carmona, M. J., M. Serra, and M. R. Miracle.
tion between tissues of a complex multicellular 1993. Relationships between mixis in Brachionus
organism—both require coordination,whether plicatilis and preconditioning of culture medium
by crowding. Hydrobiologia 255/256:145–152.
of animal behavior or cell physiology. Support- 4. Dunny, G. M., and S. C.Winans. 1999. Bacterial
ing this hypothesis and involving an even more life: neither lonely nor boring, p. 1–5. In G. M.
distant ancestor,Gallio and coworkers (5) found Dunny and S. C.Winans (ed.), Cell-Cell Signaling in
that bacterial and Drosophila genes share some Bacteria.ASM Press,Washington, DC.
common function, such that Drosophila RHO- 5. Gallio, M., G. Sturgill, P. Rather, and P.
Kylsten. 2002. A conserved mechanism for extra-
1 gene expressed in a gram-negative bacterium cellular signaling in eukaryotes and prokaryotes.
induced bacterial quorum sensing, whereas Proc. Natl.Acad. Sci. USA 99:12208–12213.
bacterial AarA gene expressed in Drosophila 6. Garcia-Roger, E. M., M. J. Carmona, and
activated epidermal growth factor signaling, M. Serra. 2006. Hatching and viability of rotifer
which is involved in wing development.These diapausing eggs collected from pond sediments.
Freshwater Biol. 51:1351–1365.
authors argue that the signals themselves may 7. Garcia-Roger, E. M., M. J. Carmona, and
not be conserved, but that cell-cell communi- M. Serra. 2006. Patterns in rotifer diapausing egg
cation may be based on some common banks: density and viability. J. Exp. Mar. Biol. Ecol.
processes that evolved long ago. 336:198–210.
8. Gilbert, J. J. 1963. Mictic female production in the
rotifer Brachionus calyciflorus. J. Exp. Zool. 153:
CONCLUSIONS 113–123.
Induction of sexual reproduction in rotifers sat- 9. Gilbert, J. J. 1993. Rotifera, p. 231–263. In K. G.
isfies the conditions that define quorum sens- Adiyodi and T. G.Adiyodi (ed.), Reproductive Biology
of Invertebrates, volume VI, part A. Asexual Propaga- 25. Snell, T.W. 1998. Chemical ecology of rotifers.
tion and Reproductive Strategies. Oxford and IBH, Hydrobiologia 387/388:267–276.
New Delhi, India. 26. Snell, T.W., and E. M. Boyer. 1988.Threshold
10. Gilbert, J. J. 2003. Specificity of crowding for mictic female production in the rotifer Bra-
response that induces sexuality in the rotifer Bra- chionus plicatilis. J. Exp. Mar. Biol. Ecol. 124:73–85.
chionus. Limnol. Oceanog. 48:1297–1303. 27. Snell, T. W., and B. L. Garman. 1986.
11. Gilbert, J. J. 2007. Induction of mictic females in Encounter probabilities between male and female
the rotifer Brachionus: oocytes of amictic females rotifers. J. Exp. Mar. Biol. Ecol. 97:221–230.
respond individually to population-density signal 28. Snell,T.W., J. Kubanek,W. Carter,A. B. Payne,
only during oogenesis shortly before oviposition. J. Kim, M. K. Hicks, and C. P. Stelzer. 2006.
Freshwater Biol. 52:1417–1426. A protein signal triggers sexual reproduction in
12. Gilbert, J. J. and T. Schroder. 2004. Rotifers Brachionus plicatilis (Rotifera). Mar. Biol. 149:
from diapausing, fertilized eggs: unique features 763–773.
and emergence. Limnol. Oceanogr. 49:1341–1354. 29. Snell, T. W., R. Rico-Martinez, L. N. Kelly,
13. Harrison, P. L., R. C. Babcock, G. D. Bull, J. and T. E. Battle. 1995. Identification of a sex
K. Oliver, C. C.Wallace, and B. L.Willis. 1984. pheromone from a rotifer. Mar. Biol. 123:347–353.
Mass spawning in tropical reef corals. Science 223: 30. Snell,T.W. and C. P. Stelzer. 2005. Removal of
1186–1189. surface glycoproteins and transfer among Bra-
14. Kats, L. B., and L. M. Dill. 1998.The scent of chionus species. Hydrobiologia 546:267–274.
death: chemosensory assessment of predation risk 31. Stelzer, C. P., and T.W. Snell. 2003.Induction of
by animals. EcoScience 5:361–394. sexual reproduction in Brachionus plicatilis (Mono-
15. Kleiven, O. T., P. Larsson, and A. Hobaek. gononta, Rotifera) by a density-dependent chemi-
1992. Sexual reproduction in Daphnia magna cal cue. Limnol. Oceanog. 48:939–943.
requires three stimuli. OIKOS 65:197–206. 32. Stelzer, C. P., and T.W. Snell. 2006. Specificity
16. Loomis, W. F. 1957. Sexual differentiation in of the crowding response in the Brachionus plicatilis
Hydra. Science 126:735–739. species complex. Limnol. Oceanog. 51:125–130.
17. Miller, M. B., and B. L. Bassler. 2003. Quorum 33. Stross, R. G., and J. C. Hill. 1965. Diapause
sensing in bacteria. Annu. Rev. Microbiol. 55: induction in Daphnia requires two stimuli. Science
165–199. 150:1462–1464.
18. Mukai, H., E. Munekata, and T. Higashijima. 34. Teshima, S. I., A. Kanazawa, N. Kamezakin,
1992. G protein antagonists. J. Biol. Chem. 267: and H. Hirata. 1979. Sterols of the rotifer. Bull.
16237–16243. Jap. Soc. Sci. Fisheries 42:1495–1501.
19. Paul, V. J., M. P. Puglisi, and R. Ritson- 35. van Woesik, R., F. Lacharmoise, and S.
Williams. 2006. Marine chemical ecology. Nat. Koksal. 2006. Annual cycles of solar insolation
Prod. Rep. 23:153–180. predict spawning times of Caribbean corals. Ecol-
20. Riffell, J. A., P. J. Krug, and R. K. Zimmer. ogy Lett. 9:390–398.
2004. The ecological and evolutionary conse- 36. Visick, K. L., and E. G. Ruby. 1999.The emer-
quences of sperm chemoattraction. Proc. Natl. gent properties of quorum sensing: consequences
Acad. Sci. USA 101:4501–4506. to bacteria of autoinducer signaling in their natural
21. Schroder, T. 2005. Diapause in monogonont environment,p.333–352.In G.M.Dunny and S.C.
rotifers. Hydrobiologia 546:291–306. Winans (ed.), Cell-Cell Signaling in Bacteria. ASM
22. Schroder,T., and J. J. Gilbert. 2004.Transgener- Press,Washington, DC.
ational plasticity for sexual reproduction and dia- 37. Wallace, R. L., and T. W. Snell. 2001. Rotifera,
pause in the life cycle of monogonont rotifers: p. 195–254. In J. Thorpe and A. Covich (ed.),
intraclonal, intraspecific and interspecific variation Ecology and Classification of North American
in the response to crowding. Functional Ecol. 18: Freshwater Invertebrates, 2nd ed. Academic Press,
458–466. New York, NY.
23. Serra, M., T. W. Snell, and J. J. Gilbert. 2005. 38. Wallace, R. L., T. W. Snell, C. Ricci, and T.
Delayed mixis in rotifers: an adaptive response to Nogrady. 2006.Biology,Ecology,and Systematics,
the effects of density-dependent sex on population p. 1–299. In H. Segers (ed.), Rotifera, 2nd ed. Back-
growth. J. Plankton Res. 27:37–45. huys Publisher, Leiden,The Netherlands.
24. Serra, M., T. W. Snell, and C. E. King. 2004. 39. Walz, N. 1997. Rotifer life history strategies and
The timing of sex in monogonont rotifers, p. evolution in freshwater plankton communities, p.
135–156. In A. Moya and E. Font (ed.), Evolution: 119–149. In B. Streit, T. Stadler, and C. M. Lively
from Molecules to Ecosystems. Oxford University (ed.), Evolutionary Ecology of Freshwater Animals.
Press, New York, NY. Birkhauser Verlag, Basel, Switzerland.
“QUORUM SENSING” IN HONEYBEES:
PHEROMONE REGULATION OF
DIVISION OF LABOR
Yves Le Conte, Zachary Huang, and Gene E. Robinson
30
Social insects live in colonies; each colony, learned about the mechanisms by which it is
composed of up to hundreds of thousands or governed (4, 5, 32). Several different types of
even millions of individuals, called workers, group decision-making activities in insect soci-
behaves as a single integrated entity. A colony eties appear to be regulated by processes similar
reproduces as a unit, and its properties are to quorum sensing in bacteria, including deci-
believed to result from selection mainly at the sions on nest sites (32a) and the allocation of
level of the whole colony (25).Although social labor to different activities. Identifying com-
insects possess brains capable of sophisticated monalities in the social behavior of social
cognitive functions (22), it is unlikely that an insects and microbes might provide ideas on
individual colony member has the capacity social regulation in general. Studies of some
to acquire and integrate information on the species of microorganisms have revealed that
global state of its colony and how that changes the core elements of sociality—altruism and
as a result of changing internal and external division of labor—are possible without a brain
conditions (28). One of the central problems in at all (7, 29). This chapter reviews our under-
social insect biology, therefore, is that of colony standing of the regulation of division of labor
integration: how the activities of thousands of in honeybee colonies from the perspective of
individual workers are integrated to form a quorum sensing.
productive colony (37).
Colony integration in the insect societies has REGULATION OF DIVISION OF
been studied intensively, and much has been LABOR IN HONEYBEE COLONIES
Age-related division of labor among workers
is a central feature of many insect societies.
Yves Le Conte INRA,UMR406 INRA/UAPV Ecologie des The highly structured worker force is generally
Invertébrés, Laboratoire Biologie et Protection de l’Abeille,
Site Agroparc, Domaine Saint-Paul, 84914 Avignon Cedex 9, seen as a consequence of having evolved two
France. Zachary Huang Department of Entomology, female castes—queens and workers—which is
Michigan State University, E. Lansing, Michigan 48824. characteristic of many insect societies (37).
Gene E. Robinson Department of Entomology, Neuroscience
Program, and Institute for Genomic Biology, University of Illi- Once workers were limited to serve mostly as
nois at Urbana-Champaign, Urbana, Illinois 61801. nonreproductive helpers to the queen, their
463
464 ■ LE CONTE ET AL.
characteristics could be shaped further by natu- SOCIAL REGULATION OF

ral selection, acting at the level of the whole DIVISION OF LABOR
colony (34). Colony age structure is an important variable in
In virtually all species of social insects, age- the regulation of division of labor in honey-
related division of labor among workers is bees. Colony age structure changes throughout
based on a pattern of behavioral maturation the year, due to changing individual birth and
(28). Individuals perform tasks in the nest death rates, reproductive colony fission, preda-
such as brood care (“nursing”) and nest mainte- tion on foragers, and brood diseases (8). The
nance when they are young.They then venture following research findings demonstrate that
outside to collect food from flowers and other social interactions provide bees with informa-
materials and defend the nest when they get tion on this important demographic variable.
older. In honeybee colonies, adult workers Huang and Robinson reported (11) results
spend the first 2 to 3 weeks of adult life working that led them to hypothesize that the age at
in the hive and the remaining 1 to 3 weeks of onset of foraging in honeybee colonies is regu-
life mostly as foragers (28). Foraging is arguably lated by worker-worker interactions, with old
the most complex task performed by a bee, bees inhibiting young bees. There is a strong
because the skills required to navigate in negative relationship between the proportion
the environment and efficiently collect floral of old bees in a colony and the proportion of
resources are thought to be greater than those precocious foragers: the more old bees present,
required to perform tasks inside the hive (6). It the slower is the maturation of the younger bees
is therefore appropriate to view the bee’s (12). Experimental manipulations revealed that
behavioral transitions as part of a process of old bees inhibit the maturation of younger
maturation. bees. For example, when a portion of a colony’s
Division of labor in honeybee colonies is foragers is removed to simulate predation,
not rigid, because bees are sensitive to changes young bees matured faster than those in a con-
in their environment (11).One response of bees trol colony in which the same number of indi-
to changing conditions is an alteration in the viduals are depleted, but evenly across different
typical pattern of behavioral maturation.Deter- age classes (12).Conversely,when foragers were
mining the age at onset of foraging most easily confined to their hive by artificial rain, young
reveals such changes, because the transition bees delayed, rather than accelerated, their mat-
from hive tasks to foraging is the most striking uration (12). In a “transplant” assay (11), when
and best understood aspect of honeybee behav- foragers were used as the transplants, behavioral
ioral maturation. For example, in experimental maturation in young bees was inhibited, but
“single-cohort” colonies (31) composed ini- when young bees were used as the transplant
tially of 1-day-old adult bees, some bees initiate instead, there was no inhibition. Was the
foraging when they are as young as 5 days of inhibitory effect of foragers mediated by their
age—more than 2 weeks earlier than under interaction with resident bees or by changes in
more typical conditions—while other bees act the hive environment caused by the transplants’
as normal-age nurses. If necessary, bees also can foraging activity? To test these alternatives, the
delay or even reverse their behavioral develop- transplant experiments were performed with
ment and become overage or reverted nurses, the colony entrance closed, thereby eliminating
respectively. A flexible system of division of any change of the nest due to foraging. Inhibi-
labor presumably is very important to colony tion of behavioral maturation was still observed
fitness because a bee colony must develop and (11). In addition to the above empirical find-
produce reproductive individuals despite con- ings, the feasibility of this social inhibition con-
stant changes in external and colony condi- cept also has been supported by theoretical
tions. models (1–3, 23).
30. “QUORUM SENSING”AND HONEYBEE DIVISION OF LABOR ■ 465
Social regulation of the rate of behavioral trophallaxis, a form of food exchange that also
maturation in honeybees requires physical con- serves as a prominent communication channel
tact among bees. Older bees separated from in insect societies (21). EO is synthesized de
younger bees with a screen that permits some novo by honeybees, but the process is not yet
forms of physical contact—food transfer, completely elucidated. Because behavioral
antennal contact, and licking—are able to maturation is socially regulated, it is likely that
inhibit behavioral maturation, but not when the activities of at least some EO synthesis
they are separated with a double screen that enzymes are under social control. Consistent
prevents these interactions (10, 20). These with these findings, Pankiw (26) reported that a
results suggested that the worker inhibitory hexane extract of foragers delays age at onset of
factor was either a nonvolatile “contact” phero- foraging. It is possible that EO acts with other,
mone, a behavior, or both, with old bees having still unidentified, compounds produced by the
greater inhibitory potency than younger bees. workers; multicomponent pheromones are
This led to research ultimately culminating common in insects, especially social insects
in the discovery of the worker inhibitory (33). The other two known honeybee primer
pheromone (21). In other words, the “quorum” pheromones, queen mandibular pheromone
in this context represents the number of old (QMP) and brood pheromone (BP), are multi-
bees, or the ratio of young to old bees, and component pheromone blends. EO itself is part
“quorum sensing” occurs via a pheromone of BP (13, 15), but the other components of BP
produced by older bees. are either not found on foragers or found in
As will be discussed in the following sec- higher quantity in foragers (21).
tions, three pheromones have been identified The identification of EO as the worker
that regulate honeybee division of labor: a inhibitor provides important validation for a
worker inhibitory pheromone, queen man- model that explains how social interactions
dibular pheromone, and brood pheromone. can regulate a key aspect of colony division of
These pheromones act directly or indirectly on labor,the age at onset of foraging (21),and helps
physiological factors including juvenile hor- us understand how the regulation of the size
mone and molecular pathways associated with of the colony foraging force can be controlled
the foraging, malvolio, and vitellogenin genes (24, by a self-organizing mechanism of social
29). However, microarray analyses indicate that integration.
these are but a few of the presumably many
genes that play a causal role in honeybee behav- REGULATION OF DIVISION OF
ioral maturation (9, 35, 36). LABOR BY BROOD PHEROMONE
The amount of brood (eggs, larvae, and pupae)
REGULATION OF DIVISION present in a honeybee colony depends on
OF LABOR BY WORKER extrinsic factors such as season, weather, and
INHIBITORY PHEROMONE food availability, and intrinsic factors such as
Ethyl oleate (EO), a substance produced by colony size, age demography, and food stores.
adult forager honeybees,was recently identified The amount of brood in a colony in turn affects
as a chemical inhibitory factor, delaying the age division of labor—more brood requires more
of onset of foraging in field experiments (21). “nurse” bees and stimulates foraging for pollen,
EO production is related to behavioral matura- the bees’ sole source of protein (38). It is thus
tion, as predicted (11); foragers have the highest not surprising that pheromones produced by
amounts of EO,and 14-day-old bees have more brood regulate these processes, with complex
than 7-day-old bees. EO is present in highest dose-dependent effects. The best understood
concentrations in the “crop” (foregut) of for- effect concerns the inhibition of behavioral
agers, suggesting that it is transmitted via maturation by BP.
BP is a blend of five ethyl and five methyl influenced by QMP. It is unlikely that the
esters of the common fatty acids palmitic, queen’s effects can also be interpreted in a “quo-
linoleic, linolenic, stearic, and oleic acids that is rum-sensing” context, because there is usually
secreted by the salivary glands of honeybee lar- only one queen present in a honeybee colony.
vae (15). It was first identified as a kairomone
that attracts the parasitic mite, Varroa jacobsoni CONCLUSIONS
(14). Later, the components of this blend were Unlike quorum sensing in bacteria, the regula-
found to have releaser effects on various aspects tion of age-related division of labor in honey-
of brood recognition and care (13, 17–19).The bee colonies involves a “decision” that unfolds
larval blend of these 10 fatty-acid esters acts as a over a relatively long period of time. In this
pheromone in the regulation of division of sense, the regulation of division of labor differs
labor among workers. Field tests showed that from quorum-sensing-like decisions that bees
bees in colonies receiving supplementary BP make concerning nest sites (32a). Behavioral
initiated foraging at older ages than did bees in maturation involves extensive changes in brain
control colonies (16).These results indicate that chemistry and structure (6), as well as changes
brood-adult communication also involves a in various physiological systems (30). However,
quorum-sensing-like mechanism, which serves the basic outlines of the process are intriguingly
to coordinate division of labor with the needs similar. Bees detect molecules (pheromones)
of the brood. Greater amounts of brood that reflect the abundance of a particularly rele-
inhibit maturation,and this leads to a later onset vant category of conspecifics and respond
age of foraging and a prolonged period of accordingly. With the identification of these
brood care. pheromones and the ability to detect their
effects on brain gene expression (9),it should be
REGULATION OF DIVISION OF possible to trace the flow of information from
LABOR BY QUEEN MANDIBULAR the environment ultimately to neurons in the
PHEROMONE
brain to determine whether quorum sensing in
There is typically only one queen in a honeybee bacteria and pheromone regulation of division
colony. She releases a mixture of substances that of labor in the honey bee share any features at
are attractive to workers, thus triggering “ret- the molecular level.
inue behavior.” This behavior is characterized
by workers surrounding the queen antennating ACKNOWLEDGMENTS
and licking her.In this way workers get exposed Research discussed here was funded by grants from
to a complex blend of pheromones that have a National Institutes of Health, National Science Foun-
variety of effects, especially inhibition of repro- dation, and U. S. Department of Agriculture to G. E. R.
ductive physiology and behavior.Another effect
relates to the regulation of age-related division REFERENCES
of labor by QMP,a five-component pheromone 1. Amdam, G. V., and S. W. Omholt. 2003.
produced by the mandibular glands of the The hive bee to forager transition in honeybee
queen (27). QMP is composed of the fatty colonies: the double repressor hypothesis. J. Theor.
acids 9-keto-(E)2-decenoic acid (9-0DA), Biol. 223:451–464.
2. Beshers, S. N., and J. H. Fewell. 2001. Models of
R-9-hydroxy-(E)2-decenoic acid (R-9-HDA), division of labor in social insects. Annu. Rev. Ento-
and S-9-HDA, and two aromatics, methyl p- mol. 46:413–440.
hydroxybenzoate and 4-hydroxy-3-meth- 3. Beshers, S. N., Z. Y. Huang, Y. Oono, and G.
oxyphenylethanol (39). Bees from colonies in E. Robinson. 2001. Social inhibition and the
the field treated with supplemental doses of regulation of temporal polyethism in honey bees. J.
Theor. Biol. 213:461–479.
QMP begin to forage at older ages than bees 4. Camazine, S. 1991. Self-organizing pattern for-
from control colonies (27). These results show mation on the combs of honey bee colonies. Behav.
that honeybee behavioral maturation also is Ecol. Sociobiol. 28:61–76.
30. “QUORUM SENSING”AND HONEYBEE DIVISION OF LABOR ■ 467
5. Detrain, C., J. L. Deneubourg, and J. M. 18. Le Conte,Y., L. Sreng, N. Sacher, J.Trouiller,

Pasteels. 1999. Decision-making in foraging by G. Dusticier, and S. H. Poitout. 1995. Chemi-
social insects, p. 331–354. In C. Detrain, J. L. cal recognition of queen cells by honey bee
Deneubourg, and J. M. Pasteels (ed.), Information workers Apis mellifera (Hymenoptera: Apidae).
Processing in Social Insects. Birkhauser Verlag, Basel, Chemoecology 5/6:6–12.
Switzerland. 19. Le Conte,Y., L. Sreng, and J. Trouiller. 1994.
6. Fahrbach, S. E.,T. Giray, and G. E. Robinson. The recognition of larvae by worker honeybees.
1995.Volume changes in the mushroom bodies of Naturwissenschaften 81:462–465.
adult honey bee queens. Neurobiol. Learn. Memory 20. Leoncini, I., D. Crauser, G. E. Robinson, and
63:181–191. Y. Le Conte. 2004.Worker-worker inhibition of
7. Foster, K. R., K. Parkinson, and C. R. L. honey bee behavioural development independent
Thompson. 2007. What can microbial genetics of queen and brood. Insect. Soc. 51:392–394.
teach sociobiology? Trends Genet. 23:74–80. 21. Leoncini, I., Y. Le Conte, G. Costagliola, E.
8. Fukuda, H. 1983.The relationship between work Plettner,A. L.Toth, M.Wang, Z. Huang, J. M.
efficiency and population size in a honeybee Bécard, D. Crauser, K. N. Slessor, and G. E.
colony. Res. Popul. Ecol. 25:249–263. Robinson. 2004. Regulation of behavioral matu-
9. Grozinger, C. M., N. M. Sharabash, C. W. ration by a primer pheromone produced by adult
Whitfield, and G. E. Robinson. 2003. worker honey bees. Proc. Natl. Acad. Sci. USA
Pheromone-mediated gene expression in the 101:17559–17564.
honey bee brain. Proc. Natl. Acad. Sci. USA 22. Menzel, R., and M. Giurfa. 2006. Dimensions
100:14519–14525. of cognition in an insect, the honeybee. Behav.
10. Huang, Z. Y., E. Plettner, and G. E. Robinson. Cogn. Neurosci. Rev. 5:24–40.
1998. Effects of social environment and worker 23. Naug, D., and R. Gadagkar. 1999. Flexible
mandibular glands on endocrine-mediated behav- division of labor mediated by social interactions in
ioral development in honey bees. J. Comp. Physiol. an insect colony—a simulation model.J.Theor.Biol.
A-Sens. Neural Behav. Physiol. 183:143–152. 197:123–133.
11. Huang, Z.Y., and G. E. Robinson. 1992. Hon- 24. Nelson, C. M., K. E. Ihle, M. K. Fondrk, J. R.
eybee colony integration: worker-worker interac- E. Page, and G. V. Amdam. 2007. The gene
tions mediate hormonally regulated plasticity in vitellogenin has multiple coordinating effects on
division of labor. Proc. Natl. Acad. Sci. USA social organization. PLoS Biol. 5:e62.
89:11726–117269. 25. Oster, G. F., and E. O. Wilson. 1978. Caste and
12. Huang, Z.Y., and G. E. Robinson. 1996. Reg- Ecology in the Social Insects. Princeton University
ulation of honey bee division of labor by colony Press, Princeton, NJ.
age demography. Behav. Ecol. Sociobiol. 39:147–158. 26. Pankiw, T. 2004. Worker honey bee pheromone
13. Le Conte, Y., G. Arnold, J. Trouiller, C. regulation of foraging ontogeny. Naturwis-
Masson, and B. Chappe. 1990. Identification of senschaften 91:178–181.
a brood pheromone in honeybees. Naturwis- 27. Pankiw, T., Z.Y. Huang, M. L. Winston, and
senschaften 77:334–336. G. E. Robinson. 1998. Queen mandibular gland
14. Le Conte, Y., G. Arnold, J. Trouiller, C. pheromone influences worker honey bee (Apis
Masson, B. Chappe, and G. Ourisson. 1989. mellifera L.) foraging ontogeny and juvenile hor-
Attraction of the parasitic mite Varroa to the drone mone titers. J. Insect Physiol. 44:685–692.
larvae of honey bees by simple aliphatic esters. 28. Robinson, G. E. 1992. Regulation of division
Science 245:638–639. of labor in insect societies. Annu. Rev. Entomol.
15. Le Conte,Y., J. M. Bécard, G. Costagliola, G. 37:637–665.
de Vaublanc, M. El Maataoui, D. Crauser, E. 29. Robinson, G. E., C. M. Grozinger, and C.W.
Plettner, and K. N. Slessor. 2006. Larval salivary Whitfield. 2005. Sociogenomics: social life in
glands are a source of primer and releaser molecular terms. Nat. Rev. Genet. 6:257–270.
pheromone in honey bee (Apis mellifera L.). Natur- 30. Robinson, G. E., R. E. Page, C. Strambi, and
wissenschaften 93:237–241. A. Strambi. 1992. Colony integration in honey
16. Le Conte, Y., A. Mohammedi, and G. E. bees: mechanisms of behavioral reversion. Ethology
Robinson. 2001.Primer effects of a brood phero- 90:336–348.
mone on honeybee behavioural development. 31. Robinson, G. E., R. E. J. Page, C. Strambi,
Proc. R. Soc. Lond. Ser. B-Biol. Sci. 268:163–168. and A. Strambi. 1989. Hormonal and genetic
17. Le Conte, Y., L. Sreng, and S. H. Poitout. control of behavioral integration in honey bee
1995.Brood pheromone can modulate the feeding colonies. Science 246:109–112.
behavior of Apis mellifera workers (Hymenoptera: 32. Seeley, T. D. 1995. The Wisdom of the Hive. Har-
Apidae). J. Econ. Entomol. 88:798–804. vard University Press, Cambridge, MA.
32a. Seeley, T. D., P. K. Visscher, and K. M. the honey bee. Proc. Natl. Acad. Sci. USA
Passino. 2006. Group decision making in 1033:16068–16075.
honeybee swarms. Am. Sci. 94:220–230. 36. Whitfield, C. W., A. M. Cziko, and G. E.
33. Slessor, K. N., M. L. Winston, and Y. Le Robinson. 2003.Gene expression profiles in the
Conte. 2005. Pheromone communication in the brain predict behavior in individual honey bees.
honeybee (Apis mellifera L.). J. Chem. Ecol. 31: Science 302:296–299.
2731–2745. 37. Wilson, E. O. 1971. The Insect Societies. Harvard
34. West-Eberhard, M. J. 1986.Alternative adapta- University Press, Cambridge, MA.
tions, speciation, and phylogeny. Proc. Natl. Acad. 38. Winston, M. L. 1987. The Biology of the
Sci. USA 83:1388–1392. Honey Bee. Harvard University Press,Cambridge,
35. Whitfield, C.W.,Y. Ben-Shahar, C. Brillet, I. MA.
Leoncini, D. Crauser, Y. LeConte, S. 39. Winston, M. L., and K. N. Slessor. 1992.The
Rodriguez-Zas, and G. E. Robinson. 2006. essence of royalty—honey bee queen pheromone.
Genomic dissection of behavioral maturation in Am. Sci. 80:374–385.
COLOR PLATE 1 (chapter 1) Stage- and compartment-specific gene expression during sporu-
lation. Sporulating cells harboring fluorescent fusions to forespore- and mother-cell-specific pro-
moters were visualized by fluorescence microscopy. In the top panels, the cells (from hour 2 of
sporulation) contained F- and E-responsive promoters fused to cfp (false-colored green) and yfp
(false-colored red), respectively. Note that all sporangia that have F-dependent expression of yfp
also have E-dependent expression of cfp.In the middle panels,the cells (from hour 3 of sporulation)
contained E- and G-responsive promoters fused to cfp (blue) and yfp (yellow), respectively. Note
that some sporangia that have E-dependent expression of cfp do not yet have detectable G-
dependent expression of yfp. In the bottom panel, the cells (from hour 4 of sporulation) contained
G- and K-responsive promoters fused to cfp (false-colored purple) and yfp (false-colored green),
respectively.Note that some sporangia that have G-dependent expression of cfp do not yet have K-
dependent expression of yfp. Schematic diagrams to the right show the three signal transduction
pathways that comprise the conversation between the mother cell and forespore.The membrane
dye (TMA-DPH) stains the double membrane that surrounds the forespore more intensely prior to
the completion of engulfment.
COLOR PLATE 2 (chapter 1) Activation of E in the mother cell is controlled by a signal transduc-
tion pathway emanating from the forespore under the control of F. Schematic diagram of the
signal transduction pathway. Proteolytic activation of pro-E in the mother cell requires the putative
membrane-tethered aspartyl protease GA. GA localizes to the septal membrane and is inactive in its
default state.The signaling protein R is produced in the forespore compartment under the control of E
and is secreted into the space between the forespore and mother cell membranes, where it activates the
GA processing enzyme. Both pro-E and GA are synthesized in the predivisional cell and are therefore
present in both the mother cell and forespore compartments. Pro-E is degraded by an unknown mech-
anism in the forespore. It is hypothesized that the signaling protein R activates GA by promoting
its dimerization. Fluorescence micrographs of wild-type cells harboring a functional pro-E-green
fluorescent protein (GFP) fusion. Early during sporulation pro-E localizes to all the membranes of the
sporangia. Cell-cell signaling results in proteolytic processing and release of E-GFP into the mother cell
cytosol, where it localizes to the nucleoid and directs gene transcription. Pro-E processing requires
forespore gene expression under the control of F. Pro-E and E were visualized by immunoblot
analysis of whole cell lysates from wild-type and F mutant (sigF) cells in a sporulation time course.
Time (in hours) is indicated.
COLOR PLATE 3 (chapter 1) Two models for the activation of G in the forespore by E in the
mother cell. E in the mother cell directs the synthesis of several engulfment proteins as well as the
IIIA locus. In the first model, the IIIA proteins monitor the engulfment process. Upon completion
of engulfment, they transduce a signal that triggers G activation. In the second model, the IIIA
complex is involved in transducing an unknown mother cell signal. In this second model, the
requirement for engulfment must also be satisfied to activate G,and this need not require IIIA.The
engulfment protein IIQ localizes to the septal membrane on the forespore side and anchors IIIAH
in the septal membrane on the mother cell side.The interaction between these two proteins has
been established biochemically and cytologically. Fluorescence micrographs of a sporulating cell
harboring cfp-IIIAH and yfp-IIQ fusions.
COLOR PLATE 4 (chapter 1) The activation of K in the mother cell is controlled by a signal
transduction pathway emanating from the forespore under the control of G.Schematic diagram of the
signaling pathway. Proteolytic activation of pro-K in the mother cell requires the putative membrane-
embedded metalloprotease B. B resides in a multimeric membrane complex with A and BofA. A
anchors the complex in the mother cell membranes that surround the forespore and serves as a plat-
form to bring B and BofA into close proximity, wherein BofA holds B inactive.Two signals from the
forespore under the control of G trigger pro-K processing. Both signaling proteins (IVB and CtpB)
are serine proteases, and both target the regulatory protein A. Cleavage of A triggers activation of the B
metalloprotease and pro-K processing. It is hypothesized that cleavage of A results in a conformational
change in the complex that allows pro-K access to the caged interior of the membrane-embedded
protease.The putative pro-K processing enzyme B localizes to the engulfing septal membranes and is
anchored there by the regulatory protein A.Wild-type and A mutant cells harboring a B-GFP fusion
were visualized during sporulation by fluorescence microscopy. Pro-K processing requires forespore
gene expression under the control of G. Pro-K and K were visualized by immunoblot analysis of
whole cell lysates from wild-type and G mutant (sigG) cells in a sporulation time course.Time (in
hours) is indicated.
COLOR PLATE 5 (chapter 7) Extracellular complementation of S. coelicolor HU261 (bldJ) by conditioned spent
medium.When grown on medium to which its own spent medium has been added, HU261 displays a bald pheno-
type after 4 days of growth (A). However, when grown on R5YE medium to which the conditioned spent medium
of S. coelicolor J660 (bldC) has been added, HU261 makes both aerial hyphae and pigment, as noted by the white
periphery and darkening of the surrounding medium, respectively (B).
lasR lasI
3OC12-HSL
COLOR PLATE 6 (chapter 9) The
rhlI rhlR qscR acyl-homoserine lactone signaling cir-
cuitry in P. aeruginosa. The schematic
emphasizes the acyl-HSL signals, genes,
and corresponding protein components
for each quorum-sensing system. Small
C4-HSL circles designate acyl-HSL signals.
LasR, RhlR, and QscR control the
LasR expression of overlapping regulons.
regulon
RhlR QscR
regulon regulon
COLOR PLATE 7 (chapter 9) A Class 1
Model of the interaction of LuxR-type

proteins with their cognate acyl-HSL
signals. (A) Class 1 QS receptors: acyl-
HSL (diamonds) associates with the nas-
cent polypeptide during ribosomal
translation. The mature protein dimer B Class 2
with tightly bound signal activates tran-
scription of target genes. (B) Class 2 QS
receptors: acyl-HSL (circles) is also
required for proper folding, but once
folded,the mature protein binds its ligand
Class 3
reversibly. (C) Class 3 QS receptors: acyl- C
HSL (triangles) is not required for folding
of apo-protein, and it reversibly binds to
the mature protein.
COLOR PLATE 8 (chapter 11) Structure and activity of the AIP.At the right is a computer-predicted structure
for AIP-II determined by an energy-minimizing algorithm, kindly provided by G. Lyon (personal communication).
Residues belonging to the macrocycle are shown in green, those to the linear tail portion in magenta.At the left are
diagrammatic representations of the four known S. aureus AIP structures.The conserved essential cysteine is in red,
and the S-CO of the thiolactone bond is shown. Ring residues whose conserved hydrophobicity is thought to
mediate generalized receptor binding are enclosed by gray circles.Residues that are critical for receptor activation are
enclosed by colored circles; their replacement by alanine generates universal agr inhibitors for S. aureus. The N ter-
minus of AIP-III is marked with an asterisk to signify its importance for this peptide’s activity, as additional amino
acids on the N terminus abolish receptor activation, while the same does not hold true for AIP-I.
COLOR PLATE 9 (chapter 11) RNAIII secondary structure (adapted from Y. Benito, F.A. Kolb, P.
Romby, G. Lina, J. Etienne, and F. Vandenesch, RNA 6:668-679, 2000). Computer-predicted structure
was confirmed by enzymatic and chemical analyses. Numbers 1 to14 refer to hairpins;A, B, and C indi-
cate long-distance interactions (see Y. Benito, F.A. Kolb, P. Romby, G. Lina, J. Etienne, and F. Vandenesch,
RNA 6:668-679, 2000, for details). Regions of complementarity with the hla mRNA leader are high-
lighted in red (E. Morfeldt, EMBO J. 14:4569-4577, 1995; R. P. Novick, H. F. Ross, S. J. Projan, J. Korn-
blum, B. Kreiswirth, and S. Moghazeh, EMBO J. 12:3967-3975, 1993.) and the spa mRNA translation
initiation region is in green (unpublished data); the region of complementarity to the SA1000 mRNA,
which also centers on hairpin 13, overlaps with that to the spa mRNA (S. Boisset, T. Geissmann, E.
Huntzinger, P. Fechter, N. Bendridi, M. Possedko, C. Chevalier, A. C. Helfer,Y. Benito, A. Jacquier, C.
Gaspin,F.Vandenesch,and P.Romby,Genes Dev.21:1353-1366,2007).Staphylococcus aureus RNAIII coor-
dinately represses the synthesis of virulence factors and the transcription regulator Rot by an antisense
mechanism.).The hld coding sequence and potentially translatable regions upstream are outlined in blue,
solid and dashed respectively;its SD sequence (70–4) and start (85–7) and stop (163–5) codons are in bold.
C-rich loop sequences in hairpins 7, 13, and 14 complementary to the rot mRNA are also noted in bold.
COLOR PLATE 10 (chapter 11) Global regulation of the staphylococcal virulon—black-box model. See text
for description.
COLOR PLATE 11 (chapter 18) (A) A TraR dimer complexed with OOHL
bound to tra box DNA.Alpha helices involved in dimerization are shown in red,
while those required for decoding tra box DNA are shown in orange. Residues
that contact RNAP are shown in yellow. One molecule of OOHL, shown in
CPK colors,is bound to the N-terminal domain of each monomer.(B) Hydrogen
bonding between OOHL and four TraR residues: Trp57 bonds with the ring
oxo group of OOHL,Asp70 bonds with the amine of OOHL,Tyr53 bonds with
the 1-oxo group of OOHL, and Thr129 makes a water-mediated hydrogen bond
with the 3-oxo group of OOHL (A.Vannini, C.Volpari, C. Gargioli, E. Muraglia,
R. Cortese, R. De Francesco, P. Neddermann, and S. D. Marco, EMBO J.
21:4393–4401,2002;R.G.Zhang,T.Pappas, J.L.Brace,P.C.Miller,T.Oulmassov,
J. M. Molyneaux, J. C. Anderson, J. K. Bashkin, S. C.Winans, and A. Joachimiak,
Nature 417:971–974, 2002).
DNA TRANSPORT
R comX2 R comW purA R comC comD comE
EARLY
RECOMBINATION
R comX2 R comM R 1110 R comA comB
STRESS RESPONSE
1549 1548 R 1717 1716 1918 2156
REGULATION
METHYLATION
LYTIC ATTACK
c cinA recA dinF lytA c cbpD c cibA cibB cibC
LYTIC MMUNITY
c coiA c dalA 1264 c ssbB c cclA
c cglA cglB cglC cglD cglE 2048 2047 2046 c celA celB
LATE
c 0021 0022 radA 0024 0025 0026 c dpnA dpnB c cflA cflB
c 0031 0030 c 0200 0201 dnaG rpoD 1074 malM malP
c 0782 c 1088 c 1480 1479 1478 1714 1715
c 1981 1980 c SP2019

DELAYED
clpL hrcA grpE dnaK dnaJ 0785 0786 0787
ciaR ciaH 1027 1029 1380 htrA spoJ groES groEL
COLOR PLATE 12 (chapter 22) Organization of gene clusters in the early,late,and delayed regulons of the CSP
response. Pentagons and triangles indicate orientation of each open reading frame. Border colors indicate phenotype
of deletion mutant: red transformation defective; green transformation proficient; blacknot known.
Fill colors indicate functional class of protein, as indicated to the right. Narrow triangles, genes possibly subject
to transcriptional readthrough; R, CAxTT-16-CAxTT direct repeats; C, combox (TACGAATA). Bent hollow
arrow, apparent promoter; lollipops, stem-loop terminators at early and late operons. Open reading frames are
identified by common name or by designation in the genomic sequence of strain TIGR4 (H.Tettelin et al., Science
293:498–506, 2001).
COLOR PLATE 13 (chapter 23) Overall structure of CprB, an ArpA homolog. CprB constitutes a dimer, each
subunit of which contains a ligand-binding pocket in the C-terminal portion and a helix-turn-helix DNA-bind-
ing domain in the N-terminal portion. The receptor dimer binds the same face of the DNA by inserting the
DNA-binding helices in the major groove.The A-factor molecule in the pocket is illustrated with a ball model.
The binding of A factor so that it is embedded completely in the pocket relocates the DNA-binding domains
(DBD) outside the molecule via the long helix-4, thus dissociating ArpA from the DNA.This computer-modeled
structure was provided by R. Natsume (R. Natsume, Y. Ohnishi, T. Senda, and S. Horinuchi, J. Mol. Biol.
336: 409–419, 2004).
COLOR PLATE 14 (chapter 25) Three-dimensional molecular simulations
of OHHL (A) and of furanone compound 4 (B) of the furan ring structure anno-
tated with a selection of bond angles. (C) Furanones accelerate the QS receptor
degradation. Panel shows Western blots following the decay of overexpressed
LuxR protein (in an E. coli background) over time in the absence and presence of
furanone 56.
COLOR PLATE 15 (chapter 25)
The QSIS1 screening system in
action. The screening bacteria are
cast into an agar plate along with
AHL, X-Gal, and growth medium.
Wells are punched in the plate in
which test compounds are added.
The test compounds diffuse into the
agar, and where in appropriate con-
centration, QS is blocked, allowing
growth of the bacteria. This is
indicated by a blue rescue ring as
the growing bacteria produce -
galactosidase which turns over the
X-Gal.
COLOR PLATE 16 (chapter 25) (A) A QS monitor: Rfp-tagged P. aeruginosa harboring the lasB-gfp fusion.
(B) The ability to suppress P. aeruginosa QS in vivo was tested by infecting mouse lungs with alginate beads con-
taining 2 107 CFU of P. aeruginosa per lung equipped with the QS monitor (M. Hentzer, K. Riedel, T. B.
Rasmussen, A. Heydorn, J. B. Andersen, M. R. Parsek, S. A. Rice, L. Eberl, S. Molin, and M. Givskov, Microbiology
148:87-102, 2002). Mice were administered 2g of furanone 30 per g of body weight or saline via injection. (C)
Infected animals were sacrificed in groups of three; the lung tissue samples were examined by SCLM. Expression
of green fluorescence was used for detection of cell-cell signaling; for detection of bacteria in tissue samples, red
fluorescence was used.At the left (with inhibitor), cell-cell communication appears blocked, since no or very little
green fluorescence can be recorded.At the right (saline without inhibitor), cells are communicating.
COLOR PLATE 17 (chapter 27) The three-dimensional structure of the E. coli rhomboid GlpG.The catalytic serine lies at the top of
the fourth transmembrane helix, which is central and shorter than the others, making a hydrophilic indentation in the extracellular face
of the enzyme.This allows water access to the active site.The loops L1 and L5 are thought to participate in substrate access/gating, although
the precise mechanism remains uncertain (Y.Wang,Y. Zhang, and Y. Ha, Nature 444:179–180, 2006; Z.Wu, N.Yan, L. Feng,A. Oberstein, H.
Yan, R. P. Baker, L. Gu, P. D. Jeffrey, S. Urban, andY. Shi, Nat. Struct. Mol. Biol. 13:1084–1091, 2006).There is also uncertainty about the exact
position of the fifth transmembrane helix, which in other structures is tilted away from the core, providing a possible substrate access route
(Z.Wu, N.Yan, L. Feng,A. Oberstein, H.Yan, R. P. Baker, L. Gu, P. D. Jeffrey, S. Urban, and Y. Shi, Nat. Struct. Mol. Biol. 13:1084–1091, 2006).
This diagram is based on the structure of Wang et al. (Y.Wang,Y. Zhang, and Y. Ha, Nature 444:179–180, 2006).
COLOR PLATE 18 (chapter 27) Sec-
retase rhomboids (left) and mitochondrial
rhomboids (right) have opposite mem-
brane orientations. Drosophila rhomboid-1
typifies the secretase-type rhomboids.The
helices containing the active site serine and
histidine are oriented in an out-in direc-
tion, and it cleaves type I substrates. The
released fragment carries the short trans-
membrane remnant. S. cerevisiae Pcp1/
Rbd1 is the best studied of the PARL-type
mitochondrial rhomboids. The helices
containing the active site serine and histi-
dine are orientated in an in-out direction,
and Rbd1 cleaves mitochondrial substrates
that correspond to type II orientation
(the C terminus is in the intermembrane
space, which is topologically equivalent to
the luminal/extracellular compartment
[Schatz and Dobberstein, Science 271:
1519–1526, 1996]).The released fragment
carries the long transmembrane remnant.
N-terminal to C-terminal orientations
of key helices are indicated with white
arrowheads.
INDEX
A factor, Streptomyces, 93, 363–376 Agrobacterium tumefaciens, 291, 294–299

in ArpA dissociation, 371 Vibrio harveyi, 323
biological activities of, 364–365 structures of, 278–279
biosynthesis of, 367–370 Acyl-homoserine lactones, 275–289
homologues of, 365–366 acyl chain lengths of, 252
in quorum sensing, 313, 444 analogs of, molecular design of, 399–404
receptor for, 370–371 antibiotic resistance and, 318
regulation of, 370–373 degradation of, 285
in streptomycin production, 373 destruction of, 404–406
structure of, 368–370 furanones and, 396–406
A factor receptor protein (ArpA), in streptomycin in one-way sensing, 420–426
production, 366–367, 370–371 in quorum sensing, 312–313
AarA protein, Pantoea stewartii, 433–436 Aeromonas, 260
AbrB protein, in ComK regulation, 22 Agrobacterium tumefaciens, 254
Accessory genes, Staphylococcus aureus, see Antarctobacter, 255
Staphylococcus aureus, virulon regulation of antibiotic-producing organisms, 314–316
AccR protein, in quorum sensing, 293 Chromobacterium violaceum, 254
O-Acetylserine, 115–116 Delisea pulchra, 265
Acg proteins, in quorum sensing, 155 Dinoroseobacter shibae, 254, 257, 259
Acidification, of seawater, AHL signaling and, 253 Erwinia, 185–195
Acr proteins, in biofilm formation, 113 Erythrobacter, 256
Actinobacteria, see Streptomyces fish pathogens, 260–262
Actinorhodin, 95 general models for, 251–252
Acyl-homoserine lactone acylases, in quorum sensing Glaciecoloa polaris, 254
quenching, 386–387 inhibitors of, 381–388
Acyl-homoserine lactone lactonases, in quorum sens- Jannaschia, 254, 257, 259
ing quenching, 383–385 Laminaria digitata, 267
Acyl-homoserine lactone oxidoreductases, in quorum Lokanella, 255, 257, 260
sensing quenching, 387 macroalgae, 262–268
Acyl-homoserine lactone synthases marine bacterial systems, 251–272
active sites of, 279, 281 Marinobacter, 254
families of, 275–277 milky sea phenomenon, 265
list of, 278 Oceanibulbus, 255
LuxI type, 277–279 Oceanicola, 257, 259, 260
mechanisms of action of, 281 Octadecabacter, 255
in quorum sensing Pantoea stewartii, 202–210
469
470 ■ INDEX
Acyl-homoserine lactones (continued) in vivo studies of, 170–171

parameters for, 252–253 pathogenicity of, 170–171
Phaeobacter, 254 RNAIII, 170, 173, 175–176
Pseudoalteromonas atlantica, 254 specificity groups, 165, 167
Pseudomonas aeruginosa, 133–140, 333 Agrobacterium radiobacter, 292, 293
Rhizobia, 216–226 Agrobacterium rhizogenes, 292
Rhodobacteriales bacterium, 257 Agrobacterium rubi, 292
Roseobacter, 254–260 Agrobacterium tumefaciens
Roseovarius, 254, 257, 259 acyl-homoserine lactone synthase of, 276
Roseovivax, 255 in biofilms, 114
Ruegeria, 255, 256 cell-to-cell signaling in, 291–306
Sagittula, 255, 257 Ti plasmids and, 292–293
Salipiger, 255 TraR in, 293–302
Silicibacter, 255–257, 259 characteristics of, 292
Staleya, 255, 256 genome of, 292
Sulfitobacter, 254, 255, 257, 259 horizontal gene transfer in, 22
Ulva, 262–265 quorum sensing in, 138, 254, 291–306
Vibrio, 256, 260–264 inhibitors of, 405–407
Vibrio fischeri, 233–244, 251–252, 254 quenching of, 384, 385, 387
Vibrio harveyi, 252, 265 Agrobacterium vitis, 292
screening for, 398–399 Agrocinopines, in crown gall tumors, 291, 293
structures of, 275–280 AhlM protein, in quorum quenching, 387
synthase active sites in, 279, 281 AHLs, see Acyl-homoserine lactones
synthesis of, 251–252, 277–278 AhyRI protein, in quorum sensing, 260
enzymatic mechanism in, 281 AI-3, in co-opting, 423
intrinsic specificity of, 281–282 AI-1 and AI-2
LuxI synthases in, 277–279 in quorum sensing, 313
modulation of, 282–286 furanones and, 397
species differences and, 275–277 Vibrio cholerae, 148, 152–153, 325–329
specificity and, 282–286 Vibrio fischeri, 237–240, 244
S-Adenosyl-L-methionine Vibrio harveyi, 323–324, 327–329
in acyl-homoserine lactone synthesis, 276, 279, synthesis of, 327–329
281, 285 AiiA protein, in quorum quenching, 383–385
in AI-1 synthesis, 327 Ain proteins, in quorum sensing
inhibitors of, 381 acyl-homoserine lactone synthesis and, 276
Adhesins, in biofilm formation, 107 Vibrio fischeri, 237–241, 244
AdpA protein Akinetes, versus heterocysts, 84–85
in aerial hypha formation and secondary metabo- Algae
lism, 93 quorum sensing inhibitors in, 423
in A factor regulation, 366, 371–373 signaling of, acyl-homoserine lactones in, 262–268
AdsA protein, in aerial hypha formation and second- Alpha factor, in quorum sensing, 444
ary metabolism, 93–94 Alteration, in interdomain signaling, 423–424
Aeromonas hydrophila, quorum sensing in, 260 Amanita muscara, Streptomyces communication with,
Aeromonas salmonicida, quorum sensing in, 260 424
AfsA protein, in butyrolactone synthesis, 364–370 Amf proteins, in aerial hypha formation and second-
Aggregatibacter actinomycetemcomitans, quorum sensing ary metabolism, 93, 96–97
in, 338–339 Ami proteins,Tat protein export system and, 435,
Aggregation, Myxococcus xanthus, 58–59 436
Aggregation substance, in pheromone binding, Amino acids, metabolites of, 115–116
33, 45 Aminoglycosides, in biofilm formation, 107
Agr proteins, in virulon regulation, 167, 169–170 Ammonia, in quorum sensing, 449
agr system, in virulon regulation, Staphylococcus aureus, Ammonium, for nitrogen fixation, 76
161–171 Anabaena, heterocysts of, see Heterocysts
AgrA protein, 169–170 Anaerobic regulator, in quorum sensing, 136
AgrB protein, 167 ANR anaerobic regulator, in quorum sensing, 136
AgrC protein, 169 Antarctobacter, quorum sensing in, 255
AgrD protein, 167 Anthranilate, in quorum quenching, 381
autoinducing peptide, 165–169 Antibiotics, 310–311
INDEX ■ 471
in biofilm formation, 112–115 forespore response in, 8–9

definition of, 310 forespore signaling in, 4–6
economic importance of, 310 mother cell response in, 6–7
Pseudomonas aeruginosa, 336–337 quorum sensing and, 14–16
resistance to, 317–318 surfactin of, 97
signaling activity of, 314–316 Bacillus thuringiensis, quorum sensing in
Streptomyces, 363–377 inhibitors of, 405
subinhibitory, 106–109, 315–316 quenching of, 382, 384–385
synthesis of, 308–310 Bacteroids, in nodulation, 215
AphA protein, in quorum sensing, 152, 154–155 BarA protein, in butyrolactone synthesis, 370
AphD protein, in A factor regulation, 371 Bees, see Honeybees
App protein, in sporulation regulation, 16 Biofilms
Aquatic animals, quorum sensing in, 453–462 Agrobacterium tumefaciens, 114
Arabidopsis, salicylic acid effects on, 119 Bacillus subtilis, 15
ArcA/ArcB system, in quorum sensing, 244 Citrobacter koseri, 114
ArgC protein, in quorum sensing, 380 control of, see Quorum-sensing inhibitors
arlRS system, in virulon regulation, 164, 171–172 Escherichia coli, 107, 115
Aro proteins, in quorum sensing, 448 examples of, 394
ArpA protein formation of, 393–394
in aerial hypha formation and secondary metabo- agr system in, 171
lism, 93, 95 amino acid metabolites in, 115–116
in streptomycin production, 366–367, 370–372 antibiotics in, 112–115
Arr protein, in biofilm formation, 107 indole in, 112–115
Arthrobacter, quorum sensing in quorum sensing in, 133–134
inhibitors of, 405–406 Klebsiella oxytoca, 114
quenching of, 382, 384, 385 membrane vesicles in, 341
AsaRI protein, in quorum sensing, 260 Morganella morganii, 114
Asc10 protein, function of, 38–39, 44–46 Myxococcus xanthus, 66–67
Aspergillus nidulans, quorum sensing in, 447 Pantoea stewartii, 207–308
AstD protein, in biofilm formation, 113 persistence of, 407
AttN protein, in quorum quenching, 384 Providencia stuartii, 114, 115
Autoinducers, see also N-3-(Oxo-hexanoyl)-homoser- Pseudomonas aeruginosa, 107, 117–118, 394, 396,
ine lactone 398, 401–403, 405–410
in co-opting, 423 Pseudomonas aureofaciens, 114
inhibitors of, 380–381 Pseudomonas fluorescens, 114
Vibrio cholerae, 147–148 regulation of, Spo0 proteins in, 15
Vibrio harveyi, 147 resistance to environmental effects, 393–394
Autoinducing peptide, in virulon regulation, 165–169 Staphylococcus aureus, 171
Autoregulation, of PrgX protein, 38–40 Staphylococcus epidermidis, 107
Auxin, in dimorphic transition, 118 Streptococcus, 355–356
Auxofurans, in two-way communication, 424 Streptococcus pneumoniae, 354–356
Vibrio cholerae, 110, 146, 152–157
B factor, in secondary metabolite regulation, 373 Yersinia pestis, 109–110
Bacillus, quorum sensing in, quenching of, 382, Bioluminescence, Vibrio fischeri, 134, 138, 235,
383–385 241–244, 424
Bacillus cereus, quorum-sensing inhibitors of, 405 BisR protein, in quorum sensing, 216–217
Bacillus mycoides, quorum-sensing inhibitors of, 405 BlcC protein, in TraA regulation, 299
Bacillus subtilis, 13–30 BofA protein, in sporulation, 8–9
cell density phenomena in, 13–14 Brachionus calyciflorus, quorum sensing in, 454
competence development in, 13–22 Brachionus plicatilis, quorum sensing in, 453–462
gene transfer in ecological consequences of, 458–459
horizontal, 22–23 evolutionary implications of, 459–460
regulation of, 23–24 historical perspective of, 454–459
population density signals of, 17–21 Bradyoxetin, in quorum sensing, 224–225
quorum sensing in, 14–17, 313 Bradyrhizobium japonicum, quorum sensing in, 220,
quenching of, 382 224–225
sporulation in, 3–16 Brood pheromone, in regulation of labor in honey-
cell density and, 13–14 bees, 465–466
472 ■ INDEX
Burkholderia cenocepacia, quorum-sensing inhibitors of, CinR protein, in quorum sensing, 219–222
404 Citrobacter koseri, in biofilms, 114
Burkholderia cepacia, quorum-sensing inhibitors of, 407 Closantel, in quorum quenching, 382
Burkholderia pseudomallei, quorum sensing in, 381 ClyA toxin, 339
N-Butyryl homoserine lactone CodY protein, in ComK regulation, 22
analogues of, for quorum-sensing inhibition, Coi proteins, in competence-stimulating peptide
401–402 synthesis, 350, 351
in interdomain signaling, 420 Com proteins
in quorum sensing, 134, 136, 139 in cell population regulation, 17–22
in competence-stimulating peptide synthesis,
C signaling, in fruiting body development, 57–58 346–351, 354–357
Caenorhabditis elegans population regulation, 21–22
quorum-sensing inhibitors of, 408 Commomamonas, quorum-sensing inhibitors of,
rhomboid proteases of, 433 405–406
signaling mechanisms of, 422 Competence
CAI-1 protein, in quorum sensing biofilms and, 354–356
Vibrio anguillarum, 262 versus cell density, 13–14
Vibrio cholerae, 148–150, 152–153, 325–327 Com proteins in, 17–23
Vibrio harveyi, 323–324 pheromones in, 346–353
Calcium, in heterocyst development, 79–80 phosphorelay in, 14–15
Candida albicans versus quorum sensing, 352–353
Pseudomonas aeruginosa communication with, Rap-Phr signaling in, 15–17, 19–21
424–426 Streptococcus pneumoniae, 14, 345–362
quorum sensing in, 443, 445–449 Competence-stimulating peptide, Streptococcus pneu-
Candida mogii, quorum sensing in, 449 moniae
Capsular polysaccharides in biofilm formation, 354–356
in quorum sensing, 201, 206 gene clusters in, 348–349
in virulon regulation, 162 in infections, 356–357
Car proteins, production of, 186–189 regulation of, 349–352
Carbapenem, production of, 186–189 synthesis of, 346
CcbP protein, in heterocyst development, 80 transduction pathway for, 346–348
cCF10 pheromone Conjugal opines, in crown gall tumors, 291–292
naming of, 32 Conjugal transfer, in rhizobia, 215–225, see also
PrgX binding to, 39, 40–43 Rhizobia, quorum sensing in
regulation of, 32 Co-opting of signals, 422–423
synthesis of, 33–36 Corn flea beetle, as Pantoea stewartii vector, 201
CcfA peptide, in pheromone regulation, 33 CprB protein
Cdc42 protein, in quorum quenching, 387 in aerial hypha formation and secondary metabo-
Cel protein, in quorum sensing, 192 lism, 95
Cell density factor, in quorum sensing, 224–225 structure of, 371
Ceratocystis ulmi, quorum sensing in, 449 Cpx protein, in swarming, 112
Cfp protein, in sporulation, 4–5 Cqs proteins
CglB protein, in motility, 55 in CAI-1 synthesis, 148–140
Chaplins, 96 in quorum sensing
Che proteins, in motility, 69, 71–72 Vibrio anguillarum, 262
CheA-like histidine kinase, in motility, 66 Vibrio cholerae, 325–327
Chemosensory system, Dif, see Dif chemosensory Vibrio fischeri, 240
system Vibrio harveyi, 323–324
Chemotaxis, 56–57, 67–70, 420–422 Crown gall tumors, cell-cell signaling within,
Chk1p protein, in quorum sensing, 447 291–306
Chlamydomonas reinhardtii, quorum sensing inhibitors Ti plasmids and, 292–293
in, 423 TraR in, 293–302
Cholera, see Vibrio cholerae Crs proteins, in quorum sensing, 150, 325
Cholera toxin, 325–327 Csr proteins
Chromobacterium violaceum homologs of, in quorum sensing, 239
quorum sensing in, 254 in quorum sensing
quorum-sensing inhibitors of, 397, 398 Escherichia coli, 193
CinI protein, in quorum sensing, 217, 219–222 Vibrio cholerae, 152
INDEX ■ 473
CtpB protein, in sporulation, 8–9 Enterotoxins, in virulon regulation, 162, 173–175

Cyanobacteria, heterocysts of, see Heterocysts Epidermal growth factor receptor, homologue of,
Cyanophycin, in heterocysts, 82 in Drosophila, 432–433
Cyclodextrins, in quorum quenching, 382 Epinephrine, in co-opting, 423
Cys proteins, in biofilms formation, 113, 115–116 Erwinia amylovora, quorum sensing in, 195
Cysteine, metabolites of, 115 Erwinia carotovora
Cystic fibrosis, Pseudomonas aeruginosa infections in, acyl-homoserine lactone synthase of, 282
333, 337 quorum sensing in, 185–199
Cytolysin, pCF10 synergism with, 45 inhibitors of, 404–405
Cytotoxins quenching of, 382
in quorum sensing, 243 virulence factors of, 189–193
in virulon regulation, 162 Erwinia chrysanthemi, quorum sensing in, 138, 194,
195
Daughter cells, in sporulation, 3 Erwinia stewartii, see Pantoea stewartii
DegU protein, in ComK regulation, 22 Erythrobacter, quorum sensing in, 256
Delisea pulchra esaI/esaR system
quorum-sensing inhibitors of, 408, 423 in acyl-homoserine lactone synthesis, 282, 284
signaling mechanisms of, 265 in quorum sensing, 202–208
Desferrioxamine, in aerial hypha formation and sec- esaR/RcsA system, in quorum sensing, 206–207
ondary metabolism, 99 Escherichia coli
Dev proteins acyl-homoserine lactone synthase of, 276
in heterocyst development, 79 in biofilms, 107, 112–115
in heterocyst envelope synthesis, 82 chemotaxis of, 69, 71–72
Dickeya dadantii (Erinia chrysanthemi), quorum sensing ComX pheromone of, 17–19
in, 138, 194, 195 enterohemorrhagic, co-opting by, 422–423
Dictyostelium, fruiting body development in, 70 enterotoxigenic, membrane vesicles of, 339
Dif chemosensory system, Myxococcus xanthus, 65–74 A factor homolog of, 368
in fruiting body development, 70 microcin of, 97
function of, 70–72 nitrogen fixation in, 76
in lipid chemotaxis, 67–70 polyamine transport system of, 109, 110
in S motility, 65–66 quorum sensing in, 193
4,5-Dihydroxy-2,3-pentanedione, in AI-1 synthesis, inhibitors of, 396, 406, 407
327–328 membrane vesicles in, 335, 338
Dinoroseobacter shibae, quorum sensing in, 254, 257 Tat protein export system of, 435, 436
Division of labor in, honeybees, pheromones in, Euprymna scolopes,Vibrio fischeri luminescence in, 235,
465–466 241–244, 424
Drosophila, rhomboid proteases of, 431–433 Exfoliatins, in virulon regulation, 162
DsbA protein, in quorum sensing, 194 Exopolysaccharides
Dsp protein, see Dif chemosensory system in pili, 65–66
Dynorphin A, in co-opting, 423 in quorum sensing
Pantoea stewartii, 201–202, 205–208
ECA proteins, in quorum sensing, 190, 194 rhizobia, 215
Eep protein, in pheromone regulation, 34–35 ExpR proteins, in quorum sensing
Efflux pumps, amino acid, 115–116 Erwinia carotovora, 189–192
Ef-Ts protein, in quorum sensing, 221 Erwinia chrysanthemi, 195
Ehrlichia chaffeensis, quorum sensing in, quenching of, rhizobia, 224
382 Extracellular matrix, Dif chemosensory system and,
Endogalacturonase, in quorum sensing, 185 65–68–72
Engulfment, in sporulation, 6–7 Extracellular peptide signaling
Enhancer-binding activator protein, Myxococcus xan- cell density and, 13–14
thus, 59–60 in competence development, 17–22
Enterobacter agglomerans, quorum-sensing inhibitors of, in gene transfer, 22–24
407 in quorum sensing, 14–17
Enterococcus faecalis
autoinducing peptide of, 168
horizontal gene transfer in, 22–23 Factor C, in Streptomyces, 93, 98
pCF10 plasmid of, see PCF10 plasmid FarA protein, in butyrolactone synthesis, 370
quorum sensing in, 313 Farnesol, in quorum sensing, 425–426, 445–447
474 ■ INDEX
FibA protein Halobacterium salinarum, phototaxis of, 72

in motility, 70 Haloferax volcanii, quorum sensing in, 447
Myxococcus xanthus, 68 Hap proteins, in quorum sensing, 149–157, 325–327,
Fibronectins, in virulon regulation, 163 329–330
Fimbrolides, Delisea pulchra, 395 HdtS proteins, in quorum sensing, acyl-homoserine
Fis protein, in quorum sensing, 149–150, 327 lactone synthesis and, 276–277
Flavonoids, rhizobial recognition of, 213–214 Helicobacter pylori, membrane vesicles of, 339
Flo11 protein Hemolysins
in dimorphic transition, 118 pCF10 synergism with, 45
in quorum sensing, 448 in virulon regulation, 162
Forespores Hep proteins, in heterocyst envelope synthesis, 81–82
mother cell response to, 6–7 2-Heptyl-3-hydroxy-4-quinolone, in quorum sens-
response of, 8–9 ing, 136, 334
signal initiation by, 4–6 Het proteins, in heterocyst development
Frizzilator, in motility, 55–56 HetC, 80
FruAP protein, in motility, 58–60 HetL, 80–81
Fruit flies, rhomboid proteases of, 431–433 HetN, 84
Fruiting bodies, Myxococcus xanthus HetR, 78–79, 83–84
aggregation of, 58–59 Heterocysts, 75–90
appearance of, 51, 52 development of
development of, 56–60, 70 envelope synthesis in, 81–82
solid surface for, 51 metabolic changes in, 82
Frz proteins, in motility, 55–56, 67–68, 72–73 regulation of, 76–81
Fungi, see also specific fungi metabolic changes in, 82
quorum sensing in, 443–452 nitrogen fixation process in, 82
alcohols in, 445–449 versus other developmental alternatives, 84–85
ammonia, 449 pattern formation of, 82–84
mating pheromones in, 444–445 Hgl proteins, in heterocyst envelope synthesis, 81–82
Furanones Histoplasma capsulatum, quorum sensing in, 449
for animal infections, 407–408 H-NOX proteins, in nitric oxide signaling, 117
bacteria producing, 397–398 Honeybees, division of labor in, 463–468
Delisea pulchra, 395, 423 description of, 463–464
discovery of, 395 social factors in, 464–465
identification of, 398–399 Horizontal gene transfer, description of, 22–23
mode of action of, 396–397 Hormesis, in secondary metabolites, 309, 314
molecular design of, 399–404 Hormogonia, versus heterocysts, 85
plant producing, 398 HrpN protein, in quorum sensing, 194
Hydrogen cyanide, in quorum sensing, 136
GabT protein, in biofilm formation, 113 Hydrophobins, 97
GacA/GacS system, in quorum sensing, 137 Hydroxyalkanoic acid core, of rhamnolipids, 111–112
Garlic, quorum-sensing inhibitors of, 398, 408–409
-Butyrolactones, in Streptomyces, 93–96, 363–365, ICEBs1 element, regulation of, 23–24
see also A factor iCF10 peptide
Gene transfer, in Bacillus subtilis, 22–24 function of, 32
GerE protein, 297 PrgX binding to, 39, 40–43
Glaciecoloa polaris, quorum sensing in, 254 synthesis of, 33–36
Gln proteins, in nitrogen limitation detection, 76–77 IM-2 butyrolactone, 365
GlpG protein, structure of, 437 ImmR protein, in ICEBs1 regulation, 24
Glutamine synthase, in nitrogen limitation detection, Indole, in biofilm formation, 112–115
76 Indole-3-acetic acid (auxin), in dimorphic transition,
Glycolipids, in heterocyst envelope synthesis, 81–82 118
Goadsporin, 96, 97 Integrative and conjugative elements, regulation of,
Growth factor signaling, in Drosophila, 431–433 23–24
Intercellular adhesins, in biofilm formation, 107
Haemophilus influenzae, competence development in, Interdomain signaling, 419–429
14 co-opting in, 422–423
HAI-1 protein, in quorum sensing, 323–324 modulation in, 423–424
INDEX ■ 475
one-way sensing, 420–422 Vibrio cholerae, 148–155, 325–327

two-way communication in, 424–426 Vibrio fischeri, 185, 234–244, 312–313
types of, 419–420 Vibrio harveyi, 323–325
Interferon-, in co-opting, 422–423 regulation of, 396
Intramembrane serine proteases, see Rhomboid pro- LysR proteins, in quorum sensing, 214
teases
IntS protein, in quorum sensing, 219 Macroalgae, signaling of, acyl-homoserine lactones
Isoflavonoids, rhizobial recognition of, 213–214 in, 262–268
Mannopine, in crown gall tumors, 296
Jannaschia, quorum sensing in, 254, 257 Marine bacterial systems, acyl-homoserine lactone
signaling in, 251–272
KasO protein, in aerial hypha formation and second- diversity of, 253
ary metabolism, 95 examples of, 253–262
Klebsiella oxytoca, in biofilms, 114 fish pathogens, 260–262
Klebsiella pneumoniae, quorum sensing in LuxI-LuxR homologues in, 259–260
inhibitors of, 405–406 macroalgae zoospores, 262–268, 421–422
quenching of, 384, 385 in milky sea phenomenon, 265
Kluyveromyces lactis, quorum sensing in, 449 models for, 251–252
parameters for, 252–253
Lactococcus plantarum, autoinducing peptide of, 168 Marinobacter, quorum sensing in, 254
Lactonases, in quorum sensing quenching, 383–385 Mating pair formation apparatus, of pCF10, 43–44
LacZ protein, in biofilm formation, 113 Mating pheromones, fungal, 444–445
LadS protein, in quorum sensing, 137 MbaA protein, in biofilm formation, 110
Laminaria digitata, signaling mechanisms of, 267, 404 MecA protein, in ComK regulation, 22
Lantibiotics, 314 Medicago truncatula, quorum sensing inhibitors in,
Las protein 423–424
furanone effects on, 396–397 Membrane vesicles, signal trafficking with, 333–344
in quorum sensing, 333 antimicrobial factors in, 336–337
LasR-LasI system, in quorum sensing in biofilms, 341
acyl-homoserine lactone synthesis and, 275–276, characteristics of, 335
278–279, 282, 285 delivery of cargo in, 335–339
Pseudomonas aeruginosa, 134–140 DNA of, 338
Lectins, in co-opting, 422 formation of, 339–341
Legionella pneumophila, nitric oxide signaling in, 117 protective factors in, 337–338
Legumes, rhizobial communication with, see Pseudomonas aeruginosa, 333–337
Rhizobia toxin production in, 338–339
Leukotoxins, in membrane vesicles, 338–339 Mesorhizobium, quorum sensing in, 225
Light organs, Vibrio fischeri in, see Vibrio fischeri Mesorhizobium loti, quorum sensing in, 216, 218–219
Lipopolysaccharide Mesorhizobium tianshanense, quorum sensing in, 221
of membrane vesicles, 335, 339–341 Metabolites, 105–129, 307–322
in quorum sensing, 243 of amino acids, 115–116
LitR protein, in quorum sensing, 239–241 antibiotics, 106–109, 310–311, 314–318
Loktanella, quorum sensing in, 255, 257, 260 auxin, 118
Lsr proteins, in quorum sensing, 329 indole, 112–115
Ltx proteins, in membrane vesicles, 338–339 polyamines, 109–111
Luminescence in quorum sensing, 311–314
in milky sea phenomenon, 265 receptors, 316–317
Vibrio fischeri, 134, 138, 235, 241–244, 424 resistance to, 317–318
Lux proteins rhamnolipids, 111–112
in biofilm formation, 114–115 salicylic acid, 118–119
in quorum sensing, 379–380 secondary, see Secondary metabolites
acyl-homoserine lactone synthesis and, signaling with
275–276, 278–279 criteria for, 120–123
Erwinia carotovora, 186, 189–193 versus current paradigm, 123–124
marine bacterial systems, 252, 254, 257–262 evolution of, 119–120
Pantoea stewartii, 195, 205 examples of, 116–119
Pseudomonas aeruginosa, 134–140 Metalloproteases, in virulon regulation, 162
476 ■ INDEX
Methionine, acetylated derivatives of, 115 Nitric oxide signaling

Methyl-accepting chemosensory protein, 55, 66, Pseudomonas aeruginosa, 116–118
70–72 Vibrio fisheri–Euprymna scolopes, 424
Methylenenomycin, production of, 366 Nitrogen-fixing bacteria, heterocysts of, see
Mex proteins, inhibition of, 381 Heterocysts
Mgl proteins, in motility, 55 Nitrosomonas europaea, acyl-homoserine lactone syn-
Microcin, 97–98 thase of, 276–277
Milky sea phenomenon, 265 Nocardia mediterranei, in rifamycin synthesis, 373
Mitochondrial rhomboids, 438 Nod proteins, in quorum sensing, 213–215, 424
Mixis-inducing protein, in rotifers, see Rotifers, Nodule formation, rhizobial, see Rhizobia
quorum sensing in Nopaline-type Ti plasmids, 291–293
Modulation, in interdomain signaling, 423–424 Norfloxacin, in biofilm formation, 107
Morganella morganii, in biofilms, 114 Norspermidine, in biofilm formation, 110
Mother cells Nostoc, heterocysts of, see Heterocysts
forespore response to, 8–9 Nostoc punctiforme, nitric oxide signaling in, 117
forespore signaling to, 4–6 Nozzles, for slime, 53–54, 65
response of, 6–7 NrrA protein, in heterocyst development, 79
Mrt proteins, in quorum sensing, 225 NspS protein, in biofilm formation, 110
Mucor mucedo, quorum sensing in, 444 Ntr proteins, in nitrogen limitation detection, 76
Multicellularity, heterocysts as, see Heterocysts
Multimerization, in quorum sensing, 138 OccR protein, in quorum sensing, 293
MvfR protein, in quorum sensing, 136 Oceanibulbus, quorum sensing in, 255
MXAN4899 enhancer-binding protein, Myxococcus Oceanicola batensis, quorum sensing in, 257, 259, 260
xanthus, 60 Oceanicola granulosus, quorum sensing in, 257, 260
Myxococcus xanthus, 51–63 Octadecaacter, quorum sensing in, 255
aggregation of, 58–59 Octopines, in crown gall tumors, 291–293
in biofilms, 66–67 One-way sensing, 420–422
chemotaxis of, 67–70 Opines, in crown gall tumors, 291–296
C-signaling in, 57–58 Opp protein, in sporulation regulation, 16
Dif chemosensory system of, 65–74 OprF protein, in co-opting, 422
evolution of, 51 Organ of transfer, of pCF10, 44
fruiting body development and, 56–57 Outer membrane vesicles, see Membrane vesicles
gene alterations due to, 59–60 Oxidoreductases, in quorum sensing quenching, 387
life cycle of, 51, 52 N-3-(Oxo-decanoyl)-homoserine lactone
motility of, 51, 65–66 in interdomain signaling, 422
pilus engine of in quorum sensing, 140
description of, 51, 53, 65–66 N-3-(Oxo-dodecanoyl)-homoserine lactone
reversal of, 53–57 analogues of, for quorum-sensing inhibition,
slime engine of 401–404
description of, 53, 66 destruction of, 406
reversal of, 53–57 in interdomain signaling, 422
in quorum sensing, 134–140
NADPH oxidase, in quorum sensing quenching, quenching of, 382–383
387–388 2-Oxoglutarate, in nitrogen fixation, 76–77
NafA protein, in motility, 71 N-3-(Oxo-hexanoyl)-homoserine lactone
NarL protein, 297 analogues of, for quorum-sensing inhibition, 399,
NctA protein, in heterocyst development, 77–80 401, 402
Necrosis-inducing protein, in quorum sensing, destruction of, 404–406
194 in interdomain signaling, 422
Neisseria gonorrhoeae in quorum sensing
competence development in, 14 Erwinia carotovora, 186–193
membrane vesicles of, 338 Pantoea stewartii, 202–203
pilus fibers of, 53 Pseudomonas aeruginosa, 134, 137–139
Nif proteins, in heterocysts, 82 Vibrio fischeri, 235–238, 241–244
Nip protein, in quorum sensing, 194 N-3-(Oxo-octanoyl)-homoserine lactone
NirS protein, in nitric oxide signaling, 117 in interdomain signaling, 420–422, 426
Nitrate, for nitrogen fixation, 76 in quorum sensing
INDEX ■ 477
Agrobacterium tumefaciens, 291, 294–299 Escherichia coli, 17–19

Erwinia carotovora, 192 fungal, 444–445
Pantoea stewartii, 202 in regulation of labor in honeybees, 465–466
Pseudomonas aeruginosa, 138–140 Streptococcus pneumoniae
rhizobia, 216, 218, 220–221, 223 gene clusters in, 348–349
Vibrio fischeri, 236–244 regulation of, 349–352
synthesis of, 346
PA-1 lectin, in co-opting, 422 transduction pathway for, 346–348
pAD1 plasmid, pheromones and, 31–32, 35, 46–47 Phosphatidylethanolamine, in chemotaxis, 67–69
pAM373 plasmid, pheromones and, 32 Phosphorelay, in sporulation regulation, 14–15
Pamamycins, in aerial hypha formation and second- Phr peptides
ary metabolism, 99 in ICEBs1 regulation, 24
Pantoea stewartii, 201–212 in sporulation, 14–17, 19–21
acyl-homoserine lactone synthase of, 278 PI factor, in secondary metabolite regulation,
in biofilms, 207–308 373–374
characteristics of, 201–202 PII protein, in nitrogen fixation, 76–78
pathogenicity of, 201–202 Pil proteins, 53–55, 70
quorum sensing in, 138, 195, 202–210 Pilus engine, Myxococcus xanthus
esaI/esaR system in, 202–208 description of, 51, 53, 65–66
esaR/RcsA system in, 206–207 in fruiting body development, 70
Paraoxonases, in quorum sensing quenching, reversal of, 53–57
385–386 Pimaricin, production of, 373–374
Parasites, rhomboids in, 438–439 Plant cell wall-degrading enzymes, in quorum sens-
PARLs (mitochondrial rhomboids), 438 ing, 185, 190–194
Pat proteins, in heterocyst development, 80, 82–84 Plasmodium, rhomboids in, 438
Patulin Pneumococci, see Streptococcus pneumoniae
production of, 397–398 Polarity, Myxococcus xanthus, reversal of, 53–57
in quorum quenching, 381 Polyamines
pCF10 plasmid, 31–49 in biofilm formation, 109–110
conjugative transfer mechanism of, 32–33, 44–45 in swarming, 109
evolution of, 46–47 Polygalacturonases, in quorum sensing, 185
genes of, 32–33 Polysaccharide intercellular adhesin, in biofilm for-
induction mechanism of, 36–40 mation, 107
overview of, 31–33 Porphyromonas gingivalis, membrane vesicle effects on,
in pheromone synthesis and control, 33–36 338
pheromone-inducible functions of, 43–46 pPD1 plasmid, pheromones and, 32, 35, 46–47
pheromone-sensing machinery of, 46–47 PQa promoter, in PrgQ, 37
virulence traits of, 45–46 PqsR protein, in quorum sensing, 334
Pdc proteins, in quorum sensing, 448 Presenilin, 437
PecS protein, in quorum sensing, 195 Prevotella loescheii, membrane vesicle effects on, 338
Pectinases, in quorum sensing, 185 PrgQ protein, in pheromone regulation, 36–38
Penicillic acid, production of, 397–398 PrgX protein
Penicillin, resistance to, 317 autoregulation of, 38–40
Penicillium, quorum-sensing inhibitors of, in pheromone binding, 33
397–398 in pheromone regulation, 36, 38–43
Peptides, extracellular signaling with, see Extracellular C-terminal arm in, 40–43
peptide signaling negative, 39–40
Peptidoglycan, in quorum sensing, 243 structure of, 38, 40–44
Periodontal disease, membrane vesicle involvement as target of cDF10 and iCF10, 39
in, 338–339 tetramerization of, 40
pH, of seawater, AHL signaling and, 253 PrgY protein, in pheromone regulation, 33,
Phaeobacter, quorum sensing in, 254 35–36
Phenazines, in biofilm formation, 107–108 PrgZ protein, in pheromone binding, 32–33
Phenylethanol, in quorum sensing, 446, 448–449 Pristinamycin, production of, 366
Pheromones Proheterocysts, 75
designation of, 32 Propionibacterium acnes, quorum sensing in, 447
Enterococcus faecalis, 31–49 Proteases, in virulon regulation, 162
478 ■ INDEX
Proteus mirabilis Quinuprisin-dalfopristin, in biofilm formation, 107

swarming in, 109 QuiP protein, in quorum quenching, 387
Tat protein export system of, 435 Quorum quenching, see also Quorum-sensing
Providencia, rhomboids of, 438 inhibitors
Providencia stuartii mode of action of, 380–383
in biofilms, 114, 115 blocking signal generation, 381
rhomboid proteases of, 433–436 disturbing signal exchange, 381
Pseudoalteromonas atlantica, quorum sensing in, 254 enzymes in, 383–388
Pseudomonas aeruginosa inactivating signals, 382
acyl-homoserine lactone synthase of, 276, 278, preventing signal recognition, 381–382
282, 283, 285 signal trapping, 382
antibiotics isolated from, 336–337 Quorum sensing
in biofilms, 107, 117–118, 394, 396, 398, 401–403, acyl-homoserine lactones in, see Acyl-homoserine
405–410 lactones
Candida albicans communication with, 424–426 advantages of, 380
characteristics of, 133 Aggregatibacter actinomycetemcomitans, 338–339
chemotaxis of, 67 Agrobacterium tumefaciens, 291–306
co-opting by, 422–423 quenching of, 384, 385, 387
membrane vesicles of, 333–338 Arthrobacter, quenching of, 382, 384, 385
nitric oxide signaling in, 116–118 Aspergillus nidulans, 447
one-way sensing by, 420–422 Bacillus, quenching of, 382–385
pilus fibers of, 53–54 Bacillus subtilis, 14–17, 313
plant signaling with, 120 quenching of, 382
quorum sensing in, 133–144, 150–151 Bacillus thuringiensis, quenching of, 382, 384–385
membrane vesicles in, 333–338 Brachionus plicatilis, 453–462
quenching of, 381, 382, 386–387 Bradyrhizobium japonicum, 220, 224–225
rhamnolipids of, 111–112 Burkholderia pseudomallei, 381
salicylic acid interactions with, 118–119 Candida albicans, 443, 445–449
secondary metabolites of, 314–315 Candida mogii, 449
virulence factors of, 133 Ceratocystis ulmi, 449
Pseudomonas aureofaciens, in biofilms, 114 description of, 105
Pseudomonas diminuta, quorum sensing in, quenching as drug targets, 406–407
of, 387 Ehrlichia chaffeensis, quenching of, 382
Pseudomonas fluorescens Enterococcus faecalis, 313
acyl-homoserine lactone synthase of, 276–277 Erwinia amylovora, 195
in biofilms, 114 Erwinia carotovora, 185–199
quorum sensing in, quenching of, 382 quenching of, 382
Pseudomonas quinolone signal Erwinia chrysanthemi, 138, 194, 195
in co-opting, 423 Escherichia coli, 193, 335, 338
in quorum sensing, 334–336, 425–426 fungi, 443–452
Putrescine Haloferax volcanii, 447
in biofilm formation, 109–110 Histoplasma capsulatum, 449
in swarming, 109 in honeybees, 463–468
PvdQ protein, in quorum quenching, 387 inhibitors of, see Quorum quenching
Pyocyanin interdomain signaling and, 419–429
in biofilm formation, 107–108 interference with, see Quorum quenching
in quorum quenching, 382 Klebsiella pneumoniae, quenching of, 384, 385
Kluyveromyces lactis, 449
Qa RNA, in pheromone binding, 37, 40, 42–43 mechanisms of, 379–380
Qrr proteins, in quorum sensing, 324–325 Mesorhizobium, 225
Vibrio cholerae, 325, 329–330 Mesorhizobium loti, 216, 218–219
Vibrio harveyi, 329–330 Mucor mucedo, 444
QS-box, in quorum sensing, 135 Pantoea stewartii, 138, 195, 202–210
QscR protein, in quorum sensing, 134–136, 139–140 Propionibacterium acnes, 447
Queen mandibular pheromone, in regulation of labor Pseudomonas aeruginosa, 133–144, 150–151
in honeybees, 466 membrane vesicles in, 333–338
Quinolone signaling pathway, in quorum sensing, quenching of, 381, 382, 386–387
136, 336 Pseudomonas diminuta, quenching of, 387
INDEX ■ 479
Pseudomonas fluorescens, quenching of, 382 in quorum quenching, 387

Ralstonia, quenching of, 386–387 in sporulation, 14–17, 19–21
Rhizobia, 215–226 RcsA protein, in quorum sensing, 205–207
Rhizobium etli, 217, 220, 222 RdfS protein, in quorum sensing, 219
Rhizobium leguminosarum, 219–222 Receptors, for secondary metabolites, 316–317
Rhodococcus, quenching of, 385 Rep proteins, in crown gall disease, 300–302
Rhodococcus erythropolis, quenching of, 387 Resistance
Rhodosporidium toruloides, 444 antibiotic, 317–318
rotifers, 453–462 to beta-lactam antibiotics, 337–338
Saccharomyces cerevisiae, 443–445, 447–449 biofilm, 393–394
Saccharomyces pombe, 443, 444 to secondary metabolite inhibitors, 317–318
secondary metabolites, 311–314 RetS protein, in quorum sensing, 137
Serratia liquefaciens, 224 RghR protein, in ComA regulation, 20–21
Serratia marcescens, quenching of, 382 Rhamnolipids, in swarming, 111–112
Sinorhizobium meliloti, 215–216, 218, 220, 223–224 Rhi proteins, in quorum sensing, 221, 222
in sporulation control, 14–17 Rhizobia, 213–232, 424
Staphylococcus aureus, 150–151, 313, 380, 444 characteristics of, 292
quenching of, 382 conjugal transfer in, 215–225, see also Rhizobia,
Streptococcus pneumoniae, 313 quorum sensing in
versus competence, 352–353 mechanisms of, 216
quenching of, 381 legume recognition of, 214–215
Streptomyces, 313 legume signals to, recognition of, 213–214
quenching of, 386–387 nodulation induction in, 214
Ustilago maydis, 444 quorum sensing in, 215–226
Variovorax paradoxus, 224–225 Bradyrhizobium japonicum, 220, 224–225
quenching of, 386 mechanisms of, 216–219
Vibrio, 397 Mesorhizobium, 225
Vibrio anguillarum, 239–240 Mesorhizobium loti, 216, 218–219
Vibrio cholerae, 147–158, 323, 325–330 Rhizobium etli, 217, 220, 222
Vibrio fischeri, 185, 233–250, 312–314 Rhizobium leguminosarum, 219–222
Vibrio harveyi, 147–150, 233, 239–240, 323–325, Sinorhizobium meliloti, 215–216, 218, 220,
327–330, 397 223–224
Vibrio parahaemolyticus, 323 soil ecology of, 225–226
Quorum-sensing inhibitors, 393–416, see also two-way communication in, 424
Furanones Rhizobium etli, quorum sensing in, 217, 220, 222
AHL destruction by, 404–406 Rhizobium leguminosarum
for animal infections, 407–408 quorum sensing in, 219–222
bacteria producing, 397–398 secondary metabolites of, 314
biofilm persistence and, 407 Rhl proteins
global effects of, 409–410 in acyl-homoserine lactone synthesis, 285
historical review of, 394–396 furanone effects on, 396–397
host immune system and, 408–409 in quorum sensing, 259–260, 333–334
identification of, 398–400 in swarming, 112
inhibitors of, in algae, 423 RhlR-RhlI system, in quorum sensing, 134–140
molecular design of, 399–404 Rhodobacteriales bacterium, quorum sensing in, 257
homoserine lactone ring substitutions in, Rhodococcus, quorum sensing in
401–402 inhibitors of, 405–406
non-AHL-based, 403–404 quenching of, 385
side chain substitutions, 399, 401 Rhodococcus erythropolis, quorum sensing in, quenching
plants producing, 398 of, 387
Rhodosporidium toruloides, quorum sensing in, 444
Rac2 protein, in quorum quenching, 387 Rhomboid-1
Rai proteins, in quorum sensing, 221, 222 Drosophila, 432–434
Ralstonia, quorum sensing in, quenching of, 386–387 structure of, 438
RamR protein, in aerial hypha formation and sec- Rhomboid proteases, 431–442
ondary metabolism, 97 conservation of, 433
Rap proteins in Drosophila, 431–433
in ICEBs1 regulation, 24 evolutionary implications of, 437–438
480 ■ INDEX
Rhomboid proteases (continued) importance of, 310

functions of, 431, 436–438 in quorum sensing, 311–314
in P. stuartii, 433–436 receptors, 316–317
in parasites, 438–439 resistance to, 317–318
structures of, 437–438 Streptomyces, 363–377
TatA system, 436 synthesis of, 308–310
in yeasts, 438–439 -Secretase, 437
Rht proteins, 116 Serine proteases, rhomboid, see Rhomboid proteases
Rifamycin, production of, 373 Serratia liquefaciens, see Serratia marcescens
Rms proteins, in quorum sensing, 190–194 Serratia marcescens, quorum sensing in, 224
RNAIII, in virulon regulation, 170, 173, 175–176 inhibitors of, 395, 407
Rok protein, in ComK regulation, 22 quenching of, 382
Roots, of legumes, rhizobial communication with, see Serratia proteamaculans, quorum-sensing inhibitors of,
Rhizobia 407
Roseobacter, quorum sensing in, 254, 255, 260 Sexual reproduction, in rotifers, see Rotifers, quorum
Roseobacter denitrificans, quorum sensing in, 256, 257 sensing in
Roseobacter litoralis, quorum sensing in, 256 SgmA protein, in aerial hypha formation and second-
Roseovarius, quorum sensing in, 257, 259 ary metabolism, 94
Roseovarius mucosus, quorum sensing in, 254 Shigella, antibiotic resistance in, 317
Roseovivax, quorum sensing in, 255 Siderophores, in aerial hypha formation and second-
rot system, in virulon regulation, 162–164, 173 ary metabolism, 99
Rotifers, quorum sensing in, 453–462 Sigma factor(s)
ecological consequences of, 458–459 in aerial hypha formation and secondary metabo-
evolutionary implications of, 459–460 lism, 94
historical perspective of, 454–459 in competence-stimulating peptide synthesis,
Rpo proteins 346–347, 351
in nitrogen limitation detection, 77 in heterocyst development, 81
in quorum sensing, 136–137 precursors of, 8–9
Rsm proteins, in quorum sensing, 137 in sporulation, 4–9
Ruegeria, quorum sensing in, 255, 256 in virulon regulation, 172–173
RWJ-49815, in quorum quenching, 382 Sigma factor-54 activator proteins, Myxococcus xan-
thus, 59
Saccharomyces cerevisiae Signal peptide peptidase, 437
dimorphic transition of, 118 SilI proteins, in quorum sensing, 257, 259
quorum sensing in, 443–445, 447–449 Silicibacter, quorum sensing in, 255, 256, 259
rhomboids in, 438 Silicibacter pomeroyi, quorum sensing in, 257, 259, 260
Saccharomyces pombe, quorum sensing in, 443, 444 Sin proteins, in quorum sensing, 224
SadB protein, in swarming, 112 Sinorhizobium meliloti
saeRS system, in virulon regulation, 162–163, 171 acyl-homoserine lactone synthase of, 282
Sagittula, quorum sensing in, 255, 257 quorum sensing in, 215–216, 218, 220, 223–224
Salicylic acid, as virulence factor, 118–119 quorum sensing inhibitors in, 423–424
Salipiger, quorum sensing in, 255 Site-2 protease
Salmonella enterica serovar Typhimurium, subinhibitory discovery of, 436–437
antibiotic effects on, 108 in sporulation, 8
Sap proteins, in aerial hypha formation and second- Slime engine, Myxococcus xanthus
ary metabolism, 92, 96–98 description of, 53, 66
Sar proteins, in virulon regulation, 162–164, 173 reversal of, 53–57
SCB-1 protein, in aerial hypha formation and sec- Small bacteriocin, in quorum sensing, 219
ondary metabolism, 95 SmaR protein, in quorum sensing, 189
ScbR protein Soft-rot erwinias, quorum sensing in, 185–199
in aerial hypha formation and secondary metabo- SoxR protein, in biofilm formation, 108
lism, 95 Spe proteins, in biofilm formation, 109–110
in butyrolactone synthesis, 370 Spitz ligand, Drosophila, 432
SdiA protein, in biofilm formation, 114–115 Spo0A protein
Secondary metabolites, 307–322, see also Antibiotics in ICEBs1 regulation, 24
definition of, 308 in sporulation, 4, 14–15
A factor and, see A factor Spo0AP protein, in sporulation regulation, 15
functions of, 308–309 Spo0B protein, in sporulation regulation, 15
INDEX ■ 481
Spo0F protein, in sporulation regulation, 15 Streptococcus

Spo0FP protein, in sporulation regulation, 15–17 in biofilms, 355–356
SpoIIAB protein, in sporulation, 7 competence in, 355–356
SpoIIGA protein, in sporulation, 5–6 horizontal gene transfer in, 22
SpoIIIA protein, in sporulation, 6–8 Streptococcus gordonii
SpoIIQ protein, in sporulation, 7 in biofilms, 355–356
SpoIIR protein, in sporulation, 5–6 secondary metabolites of, 316
SpoIVB protein, in sporulation, 8–9 Streptococcus intermedius, in biofilms, 355
SpoIVFA protein, in sporulation, 8–9 Streptococcus mutans, in biofilms, 356
SpoIVFB protein, in sporulation, 8–9 Streptococcus oralis, in biofilms, 355
Sporulation, 3–10, see also Fruiting bodies Streptococcus pneumoniae
in Bacillus subtilis, 3–16 in biofilms, 354–356
benefits of, 14 competence in, 445
cell density and, 13–14 in biofilms, 354–356
costs of, 14 development of, 14, 345–362
Factor C effects on, 98 in infection, 353–354, 356–357
forespore response in, 8–9 pheromones in, 346–353
forespore signaling in, 4–6 versus quorum sensing, 352–353
mother cell response in, 6–7 quorum sensing in, 313
quorum sensing and, 14–16 versus competence, 352–353
Streptomyces, 91, 98 quenching of, 381
Spr proteins, in aerial hypha formation and secondary Streptococcus pyogenes, competence of, 347
metabolism, 94 Streptococcus sanguis, in biofilms, 355
srrAB system, in virulon regulation, 172 Streptomyces, 91–104
SsgA protein, in aerial hypha formation and second- bald (bld) mutants of, 91–92, 97
ary metabolism, 93 cell-cell communication in, 92–93
Staleya, quorum sensing in, 255 density factors in, 98
Staleya guttiformis, quorum sensing in, 256 desferrioxamine in, 99
Staphylococcus aureus factor C in, 98
antibiotic resistance in, 318 -butyrolactones in, 93–96
biofilms of, 171 hydrophobic peptides in, 96–98
Pseudomonas aeruginosa antibiotic effects on, pamamycins in, 99
336–337 characteristics of, 91–92, 363
quorum sensing in, 150–151, 313, 380, 444 A factor of, see A factor
quenching of, 382 fungal communication with, 424
virulon regulation of, 161–183 -butyrolactones of, 363–365, see also A factor
agr system in, 161–171 programmed death of, 98
alternative sigma factors in, 172–173 quorum sensing in, 313
arlRS system in, 164, 171–172 inhibitors of, 397, 405–406
proteins in, 161–164 quenching of, 386–387
regulatory organization and, 175–176 white (whi) mutants of, 91
rot system in, 162–164, 173 Streptomyces albidoflavus, 92
saeRS system in, 162–163, 171 Streptomyces alboniger, 99
sar system in, 162–164, 173 Streptomyces antibioticus, 368, 397
srrAB system in, 172 Streptomyces avermitilis, 96
superantigens in, 162–163, 173–175 Streptomyces bikiniensis, A factor homolog of, 365
transcription factors in, 164, 173–175 Streptomyces coelicolor, 92–93, 95–97
tst system in, 162–163 A factor homolog of, 365–366, 368, 371
Staphylococcus epidermidis Streptomyces cyanofuscatus, A factor homolog of, 365
autoinducing peptide of, 167 Streptomyces flavofungini, 92, 98
in biofilms, 107 Streptomyces fradiae, 93–94
Staphylococcus intermedius, autoinducing peptide of, A factor homolog of, 366
167 Streptomyces griseus, 93–94, 96–99
Staphylococcus lugdunensis, autoinducing peptide of, A factor of, see A factor
167 Streptomyces lavendulae, 95
Staphylococcus warneri, autoinducing peptide of, 167 A factor homolog of, 365
Star protein, Drosophila, 432 Streptomyces lividans, 97
Stewart’s wilt, 201 Streptomyces natalensis, in pimaricin production, 374
482 ■ INDEX
Streptomyces pristinaespiralis, A factor homolog of, 366 in pheromone binding, 33

Streptomyces scabies, 96 in pheromone regulation, 37
Streptomyces tanashiensis, 99 in quorum sensing, 216–224
Streptomyces tendae, 96 TraB, in pheromone regulation, 35
Streptomyces venezuelae, A factor homolog of, 366 TraI, in quorum sensing, 291–292
Streptomyces virginiae, 93 TraM, in TraA regulation, 298
A factor homolog of, 365–366, 368 TraR, 293–302
Streptomyces viridochromogenes, A factor homolog of, activity of, posttranscriptional control of,
365 293–296
Streptomycin antiactivators of, 297–299
production of, A factor in, see A factor gene expression of, regulation of, 293, 300–302
resistance to, 317 in quorum sensing, 137–140, 299–300
Streptomycin-6-phosphotransferase, in A factor regu- structure of, 296–297
lation, 371 in transcription activation, 299–300
StrR protein Tracheal cytotoxin, in quorum sensing, 243
in aerial hypha formation and secondary metabo- Transcription factors, in virulon regulation, 164,
lism, 93 173–175
in A factor regulation, 371, 373 Triclosan, in quorum quenching, 381
Succinoglycan, in quorum sensing, 215 TrlR protein, in TraA regulation, 298–299, 301–302
Sulfitobacter, quorum sensing in, 254, 255, 257, Tryptophanase, 112–114
259–260 Tryptophol, in quorum sensing, 446, 448–449
Superantigens, in virulon regulation, 162–163, Twin-arginine-dependent translocation system (Tat
173–175 protein system), AarA protein and, 434–436
Surfactants, rhamnolipids as, 111–112 Two-way communication, in interdomain signaling,
Surfactin, 17, 97 424–426
Svx protein, in quorum sensing, 194 Tyl proteins, in aerial hypha formation and secondary
Swarming metabolism, 94–95
Myxococcus xanthus, 53 Tylosin
Proteus mirabilis, 109 in aerial hypha formation and secondary metabo-
rhamnolipids in, 111–112 lism, 94–95
Symbiosis production of, 366
two-way communication in, 424–426 Tyrosol, in quorum sensing, 446–448
Vibrio fischeri, 241–243
Synechocystis, pilus fibers of, 53 Ulva zoospores, signaling mechanisms of, 262–265,
421–422
Tat protein export system Undecylprodigiosin, 95
AarA protein and, 434–436 Ustilago maydis, quorum sensing in, 444
signaling function of, 436
Tcp proteins, in quorum sensing, 151–152 Van proteins, in quorum sensing
Tendrils, in swarming, 111 acyl-homoserine lactone synthesis and, 276
Tet proteins, 95 Vibrio anguillarum, 261–262
Tetracycline, in biofilm formation, 107 Vibrio fischeri, 240
Tgl protein, in pili, 53–55 Var proteins, in quorum sensing, 150, 152, 325–327
Ti plasmids, in crown gall tumors, 292–293 Variovorax paradoxus, quorum sensing in, 224–225
TnaB protein, in biofilm formation, 113 inhibitors of, 405–406
Tobramycin, in biofilm formation, 107 quenching of, 386
Tox proteins, in quorum sensing, 151–152 Veillonella atypica, secondary metabolites of, 316
Toxic shock toxin-1, in virulon regulation, 162, Vesicles, membrane, see Membrane vesicles
173–175 Vfr protein, in quorum sensing, 136
Toxin-coregulated pilus, Vibrio cholerae, 325–327 Vibrio, quorum sensing in, 254, 256, 397
Toxins Vibrio anguillarum
in membrane vesicles, 338–339 quorum sensing in, 239–240, 261–264, 421–422
Vibrio cholerae, 146–147 quorum-sensing inhibitors of, 407
Toxoplasma, rhomboids in, 438 Vibrio cholerae, 145–160
Tpk2p protein, in quorum sensing, 448 in aquatic environment, 145–146, 155–157
Tra proteins in biofilms, 110, 146, 152–157
TraA characteristics of, 145
folding of, 293–295 as human pathogen, 145–146, 156–157
INDEX ■ 483
life cycle of, 155–157 Virginiamycin, production of, 366

phenotypes of, 150–155 VirR protein, in quorum sensing, 190–192
quorum sensing in, 323, 325–330 Virulence factors
inhibitors of, 407 in biofilms, 394
life cycle and, 155–157 Erwinia carotovora, 189–193
mechanisms of, 147–150 pCF10 plasmid, 45–46
phenotypes and, 150–155 Pseudomonas aeruginosa, 118–119, 133, 135
Vibrio fischeri salicylic acid, 118–119
acyl-homoserine lactone of, 276 Staphylococcus aureus, 161–183
Euprymna scolopes communication with, 235, Vibrio cholerae, 146–147
241–244, 424 Vpr proteins, in quorum sensing, 152–154
luminescence of, 134, 138, 235, 241–244, 424 Vps proteins, in quorum sensing, 327
quorum sensing in, 185, 233–250, 261, 312– VqmA protein, in quorum sensing, 150
313, 314
environmental control of, 243–244 Wce proteins, in quorum sensing, 205–207
in ES114 strain, 234–237 Worker inhibitory pheromone, in regulation of labor
inhibitors of, 403–404 in honeybees, 465
in light-organ symbiosis, 241–243 Wts proteins, in quorum sensing, 201
mechanistic model for, 237–241 Wyx proteins, in quorum sensing, 206–207
in MJ1 strain, 234–237 Wyz proteins, in quorum sensing, 206–207
models for, 251–252 Wzx proteins, in quorum sensing, 207
parameters for, 252–253
Vibrio harveyi Yap1 protein, in dimorphic transition, 118
acyl-homoserine lactone synthase of, 276 YceP protein, in biofilm formation, 114
quorum sensing in, 147–150, 233, 239–240, Yeasts, rhomboids in, 438–439
261–262, 323–325, 327–330, 397 Yersinia pestis, in biofilms, 109–110
inhibitors of, 407 Yersinia pseudotuberculosis, quorum-sensing inhibitors
models for, 252 of, 404
Vibrio parahaemolyticus, quorum sensing in, 323 Yfp protein, in sporulation, 4–5
Vibrio polysaccharide, in quorum sensing, 152–154 YliH protein, in biofilm formation, 114
Vibrio vulnificus, quorum sensing in, 261
inhibitors of, 407 Zoospores, Ulva, signaling mechanisms of, 262–265,
Vir proteins, in crown gall disease, 292, 301–302 421–422

Chemical Communication Among Bacteria

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Chemical Communication Among Bacteria

Diunggah oleh

Hak Cipta:

Format Tersedia

Chemical

Copyright © 2008 ASM Press

Library of Congress Cataloging-in-Publication Data

Chemical communication among bacteria / edited by Stephen C.Winans and

All Rights Reserved

Cover: A Staphylococcus aureus-Pseudomonas aeruginosa co-culture biofilm. Aggregates of S.

I. CELL-CELL SIGNALING DURING

6. Heterocyst Development and Pattern Formation

II. CELL-CELL SIGNALING IN MUTUALISTIC

16. Acylated Homoserine Lactone Signaling in Marine

III. PRODUCTION, DETECTION, AND QUENCHING

IV. EUKARYOTIC QUORUM SENSING

Thomas Bovbjerg Rasmussen

Susanne B. von Bodman

lthough a few groups of bacteria have long been known to communicate

normal function of CtpB is to fine-tune the membrane-embedded metalloprotease B. We

effects often indicate that there is signaling Population-density-dependent control of

also directly activates genes involved in produc-

in molecular detail. RapA, RapB, and RapE REGULATION OF ComA BY

Transcription of the rapG phrG and rapH

tumefaciens cells. Cell-cell signaling also regu- CELL-CELL SIGNALING AND

Portetelle, S. Porwollik, A. M. Prescott, E. sion of the Bacillus subtilis pleiotropic regulatory

FIGURE 1 Model for pheromone-inducible conjugative transfer of pCF10. Recipient cell

attractive. The C terminus may provide a exogenous pheromone production likely is

response to pheromone without undergoing Transcription of prgQ in

FIGURE 4 Structural consequences of inhibitor and pheromone binding on PrgX

pheromone inhibitor, iPD1, excreted by Streptococ- J. W. Harmala, H. Hirt, and G. M. Dunny.

rapidly in the absence of a carbon source and

that any of the lipid attractants regulate ECM

PCC 7120, Anabaena variabilis ATCC 29413,

Sensing Nitrogen Limitation

activity (105). Minigenes and heterologous transporters and periplasmic oligopeptide-

at intercalary positions between heterocysts, ACKNOWLEDGMENTS

strain PCC 7120:cloning,expression,and inactiva- differentiation in the cyanobacterium Anabaena sp.

FIGURE 1 Two well-characterized -butyrolactones are A-factor of S. griseus (A), which

FIGURE 2 Three -butyrolactone-regulated pathways.The best-characterized pathway controls antibiotic

3. Ainsa, J. A., N. J. Ryding, N. Hartley, K. C. 12. Bongiorni, C., S. Ishikawa, S. Stephenson, N.

evaluating a new biological perspective of mutations to enzymes required for synthesis of

mutants. Interestingly, cysE strains showed Another explanation we believe worth

otes and multicellular organisms. In humans, between NO concentrations and regulation of

functionally common bacterial virulence factors. Vivanco. 2004.Pseudomonas aeruginosa-plant root

104. Zhang, X. S., R. Garcia-Contreras, and 105. Zumft, W. G. 2004. Denitrification by

presence of both signals, 3-O-C12-HSL and activate expression of the quorum-controlled

directly as they possess a putative RpoS- secretion functions and exopolysaccharide

TABLE 1 Characteristics of acyl-HSL/LuxR-type receptor interaction

LuxR-type Acyl-HSL required Binding

REFERENCES 14. Devine, J. H., G. S. Shadel, and T. O. Baldwin.

phosphatases depending on their autoinducer- Phosphorylated LuxO is a transcriptional

TcpPH membrane complex, resulting in a pos- regulates quorum-sensing-controlled pheno-

162 ■ GEISINGER AND NOVICK

11. SIGNAL INTEGRATION AND VIRULENCE GENE REGULATION IN S.AUREUS ■ 163

TABLE 2 Known accessory gene regulation and transcription units in S. aureus

TWO-COMPONENT SYSTEMS (Fig. 1).These variations have resulted in at least

autoinduction precedes P3 induction. It is pleiotropic transcription factor Rot (39) that

be successful, QS allows concerted massive (18).RsmA is a homologue of the E.coli regula-

CONCLUDING REMARKS REFERENCES

FIGURE 1 Stewart’s wilt. (A) Typical parallel streaking symptoms due to

wceG2 wceO wzx2

FIGURE 3 Model of the converging EsaI/EsaR quorum-sensing and Rcs environmental

A.L.Carlier and S.B.von Bodman,unpublished notype is readily reversed by exogenous AHL.

transcriptional regulators belonging to the modifications determine a wide range of Nod

FIGURE 1 Arrangement of plasmid transfer and replication genes in different plasmids

traR promoter in response to CinI-made 3- ficient to totally block 3-OH-C14:1-HSL pro-

Recipient-induced plasmid transfer is The mechanisms of induction of the traR