Bacterial outer membrane channel for divalent metal ion acquisition

Thomas H. Hohlea, William L. Franckb, Gary Staceyb, and Mark R. OBriana,1

a Department of Biochemistry, State University of New York, Buffalo, NY 14214; and bNational Center for Soybean Biotechnology, Division of Plant Sciences, University of Missouri, Columbia, MO 65211

Edited by Eva Kondorosi, Institute for Plant Genomics, Human Biotechnology and Bioenergy, Szeged, Hungary, and approved August 9, 2011 (received for review June 22, 2011)

metalloregulation

| regulation

embranes control the trafcking of solutes into cells and organelles, and thus play an essential role in maintaining homeostasis. Mitochondria, chloroplasts, and Gram-negative bacteria have both an outer and inner membrane in which solutes must traverse. Inner membrane permeability is highly selective because of transporters with high specicity for ligands. Conversely, the outer membrane is generally assumed to be much less discriminating toward small hydrophilic solutes because of channels with no or broad substrate specicity that permit diffusion from the environment into the intermembrane space (16). These channels are transmembrane proteins composed of amphipathic -strands that form a -barrel pore that allows solute translocation across the outer membrane. Some channels, called porins, discriminate by size and behave as a molecular sieve with a limit of about 600 Da. Other channels similar to porins bind solutes with low afnity, and thus are more selective but have relativity broad substrate specicities (7, 8). Growth phenotypes of bacterial mutants defective in genes encoding low-selectivity channels have not been demonstrated in most cases. Thus, it is difcult to conrm that a given channel prevents nutrient limitation to the cell. Metals are required for many cellular processes and must be acquired from the environment. Nutritional metals, such as manganese, copper, cobalt, and zinc are available as the divalent cation in aerobic environments, and are thus soluble. Whereas cytoplasmic (inner) membrane transporters of free metal ions are well characterized in bacteria (912), translocation across the outer membrane and into the periplasm has not been described. In principle, outer membrane pores with no or low selectivity should readily accommodate the diffusion of these small, soluble nutrients that are needed only in low quantities (4, 5). However, bacteria occupy niches in which the metal is often scarce, and cells can readily take up metals available in the low nanomolar range (13). Thus, we questioned whether simple diffusion across the outer membrane down such a shallow gradient via a nonselective pore is sufcient to satisfy the nutritional needs of the cell. The Gram-negative bacterium Bradyrhizobium japonicum is a model organism for studying metal metabolism and regulation
www.pnas.org/cgi/doi/10.1073/pnas.1110137108

Results Mn2+ acquisition, we used whole-genome microarray analysis to search for those that are coregulated with mntH, the gene encoding the inner-membrane Mn2+ transporter MntH. The mntH gene is repressed by Mur in the presence of manganese and derepressed under manganese limitation (13, 15). Thus, using wild-type cells grown in manganese-replete medium as the reference, we screened for genes that were up-regulated both in the wild-type grown under manganese limitation and in a mur strain grown in manganese-replete medium. Only 12 genes fullled both criteria (Fig. S1). Among them were mntH and irr, the latter of which was shown previously to be controlled by manganese and Mur (15, 23). The blr0095 gene, annotated as a hypothetical protein, showed the greatest response to manganese and Mur in the microarray analysis. This gene was renamed mnoP (Mn-responsive outer
Identication of a Candidate Outer Membrane Protein-Encoding Gene That Is Coregulated with mntH. To search for new genes involved in

Author contributions: T.H.H., G.S., and M.R.O. designed research; T.H.H. and W.L.F. performed research; T.H.H., W.L.F., G.S., and M.R.O. analyzed data; and T.H.H. and M.R.O. wrote the paper. The authors declare no conict of interest. This article is a PNAS Direct Submission. Data deposition: The sequence reported in this paper has been deposited in the Gene Expression Omnibus database (accession no. GSE30932).
1

To whom correspondence should be addressed. E-mail: mrobrian@buffalo.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1110137108/-/DCSupplemental.

PNAS Early Edition | 1 of 6

MICROBIOLOGY

The prevailing model of bacterial membrane function predicts that the outer membrane is permeable to most small solutes because of pores with limited selectivity based primarily on size. Here, we identied mnoP in the Gram-negative bacterium Bradyrhizobium japonicum as a gene coregulated with the inner membrane Mn2+ transporter gene mntH. MnoP is an outer membrane protein expressed specically under manganese limitation. MnoP acts as a channel to facilitate the tranlocation of Mn2+, but not Co2+ or Cu2+, into reconstituted proteoliposomes. An mnoP mutant is defective in high-afnity Mn2+ transport into cells and has a severe growth phenotype under manganese limitation. We suggest that the outer membrane is a barrier to divalent metal ions that requires a selective channel to meet the nutritional needs of the cell.

in the -Proteobacteria, a large and diverse taxonomic group that includes pathogens, symbionts, and bacteria that occupy many other ecological niches (14). B. japonicum maintains a constant intracellular manganese level over a wide range of manganese in the growth medium because of a regulated divalent manganese (Mn2+) transport activity (13). MntH is a high-afnity innermembrane Mn2+ transporter that is specically expressed under manganese limitation, allowing the uptake of extracellular Mn2+ in the nanomolar range (13). The mntH gene is transcriptionally regulated by the Mur repressor (previously referred to as Fur in this bacterium). Association of Mn2+ with Mur confers DNA binding activity on the protein, resulting in occupancy of the mntH promoter by Mur and transcriptional repression (13, 15). The responsiveness of the Fur-like regulator Mur to Mn2+ in B. japonicum and other -Proteobacteria rather than to Fe2+ (13, 1519) may reect an adaptation by this group to a more manganese-dependent metabolism. In support of this, an mntH mutant of B. japonicum has a severe growth phenotype (13), whereas other bacterial mntH mutants do not unless they are exposed to oxidative stress (20). Manganese also controls iron homeostasis in B. japonicum by affecting stability of the transcriptional regulator Irr (21, 22). In this study, we identied a regulated and selective outer membrane channel required for Mn2+ acquisition and growth under metal limitation in B. japonicum.

membrane protein) for reasons described below. The mnoP gene is not part of an operon because it has its own transcription start site (see below), and the gene downstream of it is in the opposite orientation. MnoP is predicted to contain extensive -sheet secondary structure, the signature of transmembrane outer membrane channels, and thus it was studied further. To conrm the microarray data, we examined the control of mnoP using quantitative real-time PCR (qPCR) and compared the expression prole to that of mntH (Fig. 1 A and B). The mnoP gene was strongly induced under low manganese growth conditions, and repression observed in the presence of manganese in the parent strain was relieved in a mur mutant. Thus, mnoP and mntH are regulated similarly. Moreover, mnoP mRNA was not affected by the iron status in the growth medium (Fig. 1A), showing that high expression under manganese limitation was not the result of a general stress response. In addition, some iron-responsive genes are aberrantly expressed in a B. japonicum mur mutant (24, 25), but mnoP was not.
Mur Binds the mnoP Promoter in Vivo and in Vitro in a ManganeseResponsive Manner. The mntH gene is controlled by direct binding

within the mntH (13) and irr (23) promoters (Fig. 2C). Collectively, the data show that mnoP and mntH are coregulated by direct binding of Mur to similar cis-acting elements on the respective promoters.
MnoP Is an Outer Membrane Protein and Predicted to Form a -Barrel Channel. Bioinformatic analysis (26) of MnoP shows it to be

of Mur to its promoter in the presence of manganese (13). Mur occupancy of the mnoP promoter was examined by cross-linking/ immunoprecipitation (IP). DNA that coprecipitated with antiMur antibodies was analyzed by qPCR using primers that amplify the promoter region. The promoter regions of mnoP (Fig. 1C) and mnoP (Fig. 1D) were occupied by Mur in the parent strain only when manganese was present in the growth media. Mur occupancy was independent of the iron status. As an additional control, cross-linking/IP was also carried out in a mur mutant (Fig. 1 C and D). Hence, Mur occupancy of the mnoP promoter corresponds with repression in the presence of manganese, as was observed previously for mntH. To identify the Mur binding site on the mnoP promoter, we determined the transcriptional start site using 59 RACE and then analyzed the region upstream of it (Fig. 2). DNase I footprinting analysis showed that recombinant Mur bound the mnoP promoter only in the presence of Mn2+, with maximum occupancy in the 47 to 13 region relative to the transcription start site (Fig. 2 A and B). Examination of the protected region of DNA on the mnoP promoter revealed that the sequence contains three imperfect direct hexameric repeats (Fig. 2B), similar to those observed

homologous to two families of outer membrane transmembrane -barrel proteins. The N-terminal region is similar to the surface antigen-2 superfamily, and the C-terminal portion is predicted to belong to the OprB superfamily. Pseudomonas aeruginosa OprB is a carbohydrate channel (8), but other members of the OprB superfamily have not been characterized. MnoP was modeled using the HHpred server (27) and predicted to contain 14 -sheets organized as a -barrel protein (Fig. 3A). The -barrel comprises 580 amino acids of the protein, and the remaining Cterminal 70 amino acids were unmodeled. Genes encoding MntH and MnoP tagged with c-Myc and HA, respectively, at the C-termini and under the control of their endogenous promoters were constructed and introduced into B. japonicum cells. Inner- and outer membrane fractions were prepared from cells grown in manganese-replete or manganeselimited conditions, and analyzed by Western blotting using antibodies directed against the respective tag (Fig. 3B). MnoP was found in the outer membrane fraction, in agreement with its predicted structure. MntH was conrmed to be an inner-membrane protein. Moreover, both MntH and MnoP were detected only in samples from cells grown under manganese limitation, which correlates with induction of the mRNAs that encode them. MntH appeared as a doublet of 52 and 55 kDa on SDS/PAGE, in reasonable agreement with the predicted size of 50 kDa. The larger (minor) band may be an unprocessed peptide or a modied variant. MnoP migrated at about 52 kDa, which differed substantially from the calculated mass of 69 kDa. This aberrant migration is common for outer membrane -barrel proteins (e.g., ref. 28).
MnoP Protein Facilitates Mn2+ Entry in Reconstituted Liposomes. The

outer membrane location of MnoP and its putative structure predict that it is a channel protein. To test this theory, recombinant MnoP protein was overexpressed, puried, and incorporated into liposomes, forming proteoliposomes (see Materials and Methods). Transport activity was monitored using a swelling assay (29) in

Fig. 1. Coregulation of mnoP and mntH in a manganese- and Mur-responsive manner. (A and B) Analysis of mnoP and mntH mRNA by qPCR in wild-type or mur mutant cells grown in media supplemented with or without 50 M MnCl2 and 20 M FeCl3. The data are expressed as the relative starting quantity (SQ) of mnoP or mntH mRNA normalized to the housekeeping gene gapA, and presented as the average of triplicate samples the SD. (C and D) In vivo occupancy of the mnoP and mntH promoters by Mur in cells grown as described above by cross-linking followed by IP using anti-Mur antibodies. The precipitated DNA was analyzed by qPCR using primers that amplify promoter DNA from mnoP or mntH. The data are expressed as the relative starting quantity of immunoprecipitated DNA normalized to a mock pull-down without the primary antibody.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1110137108

Hohle et al.

Fig. 2. Identication of the Mur binding site on the mnoP promoter in vitro by DNase I footprinting. (A) Protection of DNA from DNase I digestion by Mur was carried out in the presence or absence of MnCl2 using 0, 2, 5, 10, 25, or 50 nM Mur. The DNA was radio-labeled at the 59 end of the nontemplate strand with respect to the mnoP gene, and thus the 39 end is at the top of the gel. The arrows denote the sites of the three hexamer repeats. (B) The sequence of the protected region of the mnoP promoter are shown as lowercase letters. The bent arrow represents the transcription start site. The direct repeats are shown by the straight arrows. The underlined sequence shows the putative 35 region of the promoter. (C) A consensus binding site for B. japonicum Mur identied using the Mur-binding sites of mntH, mnoP, and irr as determined by the MEME algorithm. The Mur binding regions on the mntH and irr promoters were determined previously, as described in the text.

which uptake of a substrate causes the liposome to swell to maintain osmotic balance. Swelling results in a decrease in the optical density of the liposome suspension. In the absence of MnoP, the liposomes were impermeable to any of the metal salts tested, as seen by the lack of swelling (Fig. 4A). However, pro-

teoliposomes containing MnoP swelled in the presence of MnCl2 or MnSO4 (Fig. 4B). Divalent cobalt and copper (Co2+ and Cu2+) are stable in aerobic environments, and therefore CoSO4 and CuSO4 were also tested in the liposome assay (Fig. 4B). Swelling was not observed with either metal salt. These observations show that MnoP confers permeability to Mn2+ and is a selective channel.
MnoP Is Essential for High-Afnity Mn2+ Transport Activity in Cells.

Fig. 3. Localization of MnoP to the outer membrane. (A) Predicted model of MnoP -barrel structure. (B) Western blots analysis of c-myctagged MnoP or HA-tagged MntH in the outer membrane (OM) or inner membrane (IM) fractions of cells grown in the presence (+) or absence () of 50 M MnCl2. Thirty micrograms of protein was loaded onto each lane.

Previously, we reported that B. japonicum has a high-afnity Mn2+ transport activity that is induced under manganese limitation (13). To determine whether MnoP is required for this activity, we measured uptake using 5 nM 54Mn2+ as the initial concentration in an mnoP mutant grown in high- or low-manganese media, and compared them to the wild-type and mntH strains (Fig. 5). Mn2+ transport activity was very low in all strains grown in high-manganese media (Fig. 5A), conditions where mntH and mnoP are not expressed (Figs. 1 A and B, and 3B). Mn2+ transport was induced in the parent strain grown under manganese limitation, but this activity was severely diminished in the mnoP mutant, as well as in the mntH strain (Fig. 5B). Thus, MnoP is required for high-afnity Mn2+ uptake. To quantify the contribution of MnoP to cellular Mn2+ transport, we measured the initial velocity of Mn2+ uptake as a function of the Mn2+ concentration in the parent and mutant strains, and determined the apparent Km values using a Lineweaver-Burk plot analysis (Fig. 5C). The apparent Km for Mn2+ uptake by the parent strain was 87 nM (Fig. 5C). In the mnoP mutant, the apparent Km was 1,656 nM, 19-fold higher than for the parent strain. Thus, MnoP contributes substantially to highafnity Mn2+ uptake by cells. The Lineweaver-Burk plot for the mntH strain yielded a positive abscissa value (Fig. 5C); thus, the apparent Km value could not be determined, presumably because of the near absolute requirement of MntH for Mn2+ uptake.
PNAS Early Edition | 3 of 6

Hohle et al.

MICROBIOLOGY

Fig. 4. MnoP-mediated transport of divalent metals in proteoliposomes. Liposome swelling assays were initiated by the addition of MnCl2 (), MnSO4 (B), CuSO4 (), CoSO4 (;), or metal-free water () to a suspension containing 1 mg/mL reconstituted liposome, and either (A) 0 or (B) 1 g/mL recombinant MnoP. The suspension was monitored at an optical density of 400 nm (OD400) for 90 s.

MnoP Is Required for Growth Under Manganese Deciency. We wanted to address whether the defect in Mn2+ transport in the mnoP mutant limits cell growth under manganese limitation. To do this, we examined the growth of the parent, mntH, and mnoP strains on solid media containing 25 M EDDHA (ethylenediamine-di-o-hydroxyphenylacetic acid) as a metal chelator (Fig. 6). Heme was used as the iron source because heme iron is not chelated by EDDHA. The parent strain grew without manganese supplementation (Fig. 6), indicating that the cells were able to scavenge it from the medium. However, neither the mnoP nor mntH mutants grew under these conditions, showing that MnoP and MntH are necessary for growth under metal limitation. The growth phenotype of the mnoP strain was partially restored with supplementation with 5 pmol MnCl2 to a paper lter in the center of the plate, as observed by the zone growth around the lter (Fig. 6). No growth was observed when the lters were spotted with divalent salts of zinc, copper, cobalt, or nickel instead of manganese. The mntH strain had a more severe phenotype, requiring 50 pmol MnCl2 to partially restore growth (Fig. 6). These data show that MnoP is required for cell growth under manganese limitation.

Fig. 5. Mn2+ uptake by B. japonicum parent and mutants strains. Cells of the parent strain (), the mntH strain (), or the mnoP strain () were grown in (A) high- or (B) low-manganese media. At time 0, 5 nM 54Mn was added to the assay medium and aliquots were subsequently taken at various time points and counted. (C) Lineweaver Burk plot. 54Mn2+ uptake was measured for the parent strain (), mntH (), or mnoP () for 5 min at concentrations of Mn2+ ranging from 75 to 250 nM.

manganese and derepression when the metal is scarce. Second, MnoP allows the translocation of Mn2+ into proteoliposomes, but Co2+ or Cu2+ are not transported. The atomic radii of Mn2+,

Discussion The prevailing model of outer membrane structure and function predicts that small soluble substrates, such as divalent metal cations, should traverse it via channels with no or low selectivity (4, 5). Nevertheless, the present study shows that manganese acquisition requires the regulated expression of a selective outer membrane channel to translocate the metal from the environment into the periplasm. Moreover, the absence of this channel renders the outer membrane a diffusion barrier that prevents the cell from meeting its nutritional manganese requirement. This conclusion is based upon several observations. First, mnoP expression is not constitutive but is specically expressed under manganese limitation, and is coordinated with the inner membrane transporter gene mntH. Both gene promoters contain a Mur-binding site that permits occupancy in the presence of
4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1110137108

Fig. 6. Dependence on manganese supplementation for growth of the B. japonicum parent strain and mntH and mnoP mutants: 106 cells were plated on solid low-metal media supplemented with 1 M heme as an iron source and 25 M EDDHA as a metal chelator. Next, 0, 5, or 50 pmol MnCl2 was spotted onto a 6-mm lter disk located in the center of the plate. Growth around the lter disk was assessed visually. Samples were performed in triplicate, and one of each is shown in the gure.

Hohle et al.

Co2+, and Cu2+ are very similar to each other, as are the respective hydrated radii (30). Thus, the selectivity of MnoP is very unlikely to be based on size or valency alone. Third, MnoP was required for transport of Mn2+ into cells in the low nanomolar range of metal and increased the apparent afnity of the metal for cells by almost 20-fold. Finally, MnoP was necessary to support bacterial growth under manganese limitation, and underscores the physiological relevance of channel. In a few bacteria, Ni2+ (31) and perhaps Zn2+ (32) can be taken up across he outer membrane as metal chelates involving an energy-dependent receptor and auxiliary proteins, similar to what is observed for ferric (Fe3+) iron. In Helicobacter pylori, the chelated nickel-uptake system functions specically at low pH (31), but the mechanism of free Ni2+ uptake across the outer membrane is unknown. The outer membrane of P. aeruginosa is more selective than that of Escherichia coli, and outer membrane channels with some selectivity have been described (3, 4). The OprB carbohydrate channel is induced by the presence of substrate in the millimolar range and transports a wide range of structurally diverse carbon sources, such as glycerol, glucose, and trisaccharides (8). In contrast, mnoP derepression requires substrate limitation, and MnoP can discriminate between similar ions. The phosphate channel OprP is produced in response to phosphate limitation and transports numerous anions, but the induction threshold is 150 M substrate (33), which is much higher than the manganese concentration needed for mnoP expression. Although nutritional metals are needed in small quantities compared with carbohydrates or phosphate, they are often available in low amounts, and the present work indicates that simple diffusion across the outer membrane cannot support cellular needs. From this indication, it is likely that acquisition of other soluble metal nutrients requires selective channels in bacteria. Similarly, mitochondria and chloroplasts are metal-rich organelles and contain numerous -barrel outer membrane proteins, most of whose functions are not known (1). Similarity searches did not nd obvious MnoP homologs, but further studies into understanding the basis of selectivity may yield new criteria on which to base a more rened search. Materials and Methods
Strains and Media. B. japonicum strain USDA110 is the parent strain used in this study. Strains GEM4 (34), mntH (13), and mnoP (present study) are mutant derivatives of the parent strain containing a DNA cassette encoding spectinomycin and streptomycin replacing the mur, mntH, and mnoP genes, respectively. B. japonicum strains were routinely grown at 29 C in glycerolsalts yeast (GSY) medium as previously described (35). For low manganese and iron conditions, modied GSY was used, containing 0.5 g$L1 yeast extract instead of 1 g$L1, with no exogenous manganese or iron added. The actual concentrations of manganese and iron in unsupplemented media are 0.2 M and 0.3 M, respectively, as determined by atomic absorption using a Perkin-Elmer model 1100B atomic absortion spectrometer. High-metal media was supplemented with either 50 M MnCl2, 20 M FeCl3, or both. Microarray Analysis. Two comparisons were carried out: wild-type cells grown in manganese-limited media vs. manganese-replete conditions, and the mur mutant vs. the wild-type grown under manganese-replete conditions. Three biological replicates of each strain and growth condition were analyzed. RNA was isolated from B. japonicum cells using a hot phenol-extraction method and cDNA synthesized as previously described (36). Microarray analysis was carried out using oligonucleotides printed on Corning epoxycoated slides as described previously (37). The slides were read at the Microarray and Genomics Core Facility at the Roswell Park Cancer Institute (Buffalo, NY). The microarray data have been deposited in the Gene Expression Omnibus (GEO), and are accessible through GEO Series accession number GSE30932. Gene names and annotations are from the published genome database (38) (http://genome.kazusa.or.jp/rhizobase/). Analysis of RNA by qPCR. Expression levels of selected genes were determined by qPCR with iQ SYBR green supermix (Bio-Rad) using iCycler thermal cycler (Bio-Rad). RNA was isolated from B. japonicum cells using a hot phenol-ex-

traction method as previously described (36). cDNA was synthesized from 5 g total RNA using iScript cDNA synthesis kit (Bio-Rad). qPCR reactions were carried out as previously described (13). Data are expressed as average of three triplicates and the SD is represented by the error bars. Quantitative in Vivo Cross-Linking and IP. For quantitative in vivo cross-linking and IP, 200-mL cultures of parent strain USDA110 or mur strain GEM4 were grown under low- or high-manganese and low- or high-iron conditions to midlog phase (OD540 0.40.6). Cross-linking and IP were performed as described previously (39). Immunoprecipitated DNA of selected promoter regions was quantied using qPCR as described above, using 1-L coimmunoprecipitated DNA template. The data were normalized to mock pull-down sample in which primary antibody was excluded from the reaction. DNase I Footprinting Analysis. DNase I footprinting analysis examining mnoP promoter DNA protected by Mur in the presence or absence of 100 M MnCl2 was carried out as described previously (13). The reactions were carried out in the presence 0 to 50 nM Mur. Cellular Location of MnoP and MntH. MnoP-c-myctagged protein was expressed using a very low-copy plasmid (pVK102) and transcribed from the mnoP promoter in the parent strain. A gene encoding MnoP-HAtagged protein run off of the mntH promoter region was integrated into the genome of the parent strain. Next, 100 mL of cells harboring these constructs grown to midlog phase were harvested by centrifugation and washed twice in PBS (PBS). Cells were resuspended in PBS containing 2 mg/mL lysozyme. Cell suspensions were run through the French press twice before being subjected to ultracentrifugation at 185,000 g (Ti70.1 rotor, Beckman) for 90 min. The supernatant containing the soluble fraction was discarded. The pellet was resuspended in PBS containing 2% sarkosyl by rotating on a slight angle for 2 h at room temperature. The suspension was subjected to ultracentrifugation as described above, and the supernatant containing the inner-membrane fraction was saved at 20 C. The pellet was resuspended in PBS containing 0.5% SB3-14 (Sigma) detergent by rotating on a slight angle at room temperature overnight. The ultracentrifugation step was repeated, and the supernatant containing the outer membrane fraction was saved at 20 C. Proteins were quantied using the Bradford reagent (Bio-Rad) and 30 g were run on a 10% SDS/PAGE. Western blots were used to visualize MnoP-c-myc and MntH-HA using -c-Myc and -HA antibodies (Sigma), respectively, at 1:1,000 titer, incubating for 1 h. Bands were visualized using horseradish peroxidase-conjugated goat anti-rabbit IgG and developed using the Western Lightning ECL reagent (PerkinElmer). Protein Overexpression and Purication. The mnoP gene was amplied by PCR and cloned into pETite C-His vector (Lucigen) containing a C-terminal 6xHis tag. The vector with insert was transformed into chemical competent BL21 (DE3) E. coli cells. Cells were inoculated from an overnight culture grown in Luria-Bertani media containing 20 g/mL kanamycin into 1 L of fresh 2YT medium containing 2 g/mL kanamycin. Overexpression was induced in cells at the midlog phase by the addition of 0.5 mM isopropyl-1-thio--D-galactopyranoside (IPTG) at 37 C for 4 h with shaking. Cells were pelleted by centrifugation, washed in phosphate-binding buffer (50 mM NaH2PO4 and 300 mM NaCl, pH 8.0), and subsequently resuspended in 15 mL phosphate-binding buffer containing 8 M urea. Cells were disrupted by passage through a French pressure cell at 1,200 psi and claried by centrifugation at 13,000 g for 10 min. Two milliliters of Ni-NTA slurry (Qiagen) was washed in phosphate binding buffer containing 8 M urea before being added to the cleared lysate and rocked for 60 min at room temperature (22 C). The Ni-NTA slurry-protein mixture was poured into a column and washed four times with phosphate wash buffer (50 mM NaPO4, 300 mM NaCl, 8 M Urea, and 20 mM imidazole, pH 8.0) and once with phosphate wash buffer containing 10% glycerol. His-MnoP was eluted using phosphate elution buffer (50 mM NaPO4, 300 mM NaCl, 8 M Urea, and 250 mM Imidazole, pH 8.0). The puried protein was run through an FPLC buffer exchange column so that the nal buffer consisted of 0.1 mM Tris-HCl, pH 8.0, and 8 M Urea. Liposome Swelling Assay. To make liposomes, 100 mg of 1,2-dilauroyl-snglycero-3-phosphocholine (Avanti Polar Lipids) were dissolved in 1 mM chloroform in a 50-mL glass round-bottom ask. The ask was ushed with nitrogen gas until the chloroform was evaporated and a lipid lm remained. The lipids were hydrated in 1 mL 0.1 mM Tris-HCl, pH 8.0, vortexed until lipids were in solution, and subjected to ve freeze-thaw cycles using liquid nitrogen and a 37 C water bath. Next, 400 g recombinant unfolded MnoP was diluted about 12-fold in the lipid suspension to promote folding. The protein-liposome suspension was then extruded 16 times through two 200-

Hohle et al.

PNAS Early Edition | 5 of 6

MICROBIOLOGY

nM pore size polycarbonate lters (Whatman). The protein-liposome suspension was transferred to a 6-mL glass culture tube and allowed to sit at room temperature for 2 h before the swelling assay. The liposome without protein was prepared the same way as described, but instead of protein, 0.1 mM Tris-HCl, 8 M urea, pH 8.0 was added to the suspension. Liposome swelling assays (29) were performed to monitor MnoP specic transport of various substrates. The assay was initiated by the addition of 1 mM substrate or metal-free water to a solution of 0.1 mM Tris-HCl, 6.7 mM urea, 1 mg$mL1 reconstituted liposome, and either 0 or 1 g$mL1 MnoP. The suspension was monitored at an optical density of 400 nm (OD400) every 10 s for 90 s. Proteins successfully folding into liposomes were estimated by pelleting 50 L proteoliposome solution at 14,000 g for 5 min, followed by washing and resuspending the proteoliposome in 0.1 mM Tris-HCl, pH 8.0, and the protein concentration determined using the Bradford method (Bio-Rad). Mn Uptake Assay. The ability of the B. japonicum parent strain and mutants to transport 54Mn2+ was determined, as previously described (13). For de54

termination of initial velocities, 54Mn2+ uptake was measured for 5 min within the linear range for uptake. The data were analyzed by LineweaverBurk plots and the apparent KM and Vmax values reported were derived from these analyses. Growth Disk Assays. For growth disk assays, 106 cells of the parent, mntH, or mnoP strain were quickly vortexed into soft-agar medium containing 1 M heme as an iron source and 25 M EDDHA as a metal chelator at 50 C and then poured onto plates. After solidication of the agar medium, Whatman paper disks (6-mm diameter) were placed at the center of the plate and spotted with 5 L of 10 M or 1 M MnCl2 or metal-free water as a negative control. Growth around the disk was assessed visually. ACKNOWLEDGMENTS. This work was supported by National Institutes of Health Grant R01 GM067966 (to M.R.O.), National Science Foundation Grant IOS-1025752 (to G.S.), and National Institutes of Health Training Grant T32 AI707614 (to T.H.H.).

1. Blter B, Soll J (2001) Ion channels in the outer membranes of chloroplasts and mitochondria: Open doors or regulated gates? EMBO J 20:935e940. 2. Duy D, Soll J, Philippar K (2007) Solute channels of the outer membrane: From bacteria to chloroplasts. Biol Chem 388:879e889. 3. Hancock RE, Brinkman FS (2002) Function of pseudomonas porins in uptake and efux. Annu Rev Microbiol 56:17e38. 4. Nikaido H (2003) Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593e656. 5. Silhavy TJ, Kahne D, Walker S (2010) The bacterial cell envelope. Cold Spring Harb Perspect Biol 2:a000414. 6. Schirmer T (1998) General and specic porins from bacterial outer membranes. J Struct Biol 121:101e109. 7. Huang H, Hancock RE (1993) Genetic denition of the substrate selectivity of outer membrane porin protein OprD of Pseudomonas aeruginosa. J Bacteriol 175: 7793e7800. 8. Wylie JL, Worobec EA (1995) The OprB porin plays a central role in carbohydrate uptake in Pseudomonas aeruginosa. J Bacteriol 177:3021e3026. 9. Patzer SI, Hantke K (1998) The ZnuABC high-afnity zinc uptake system and its regulator Zur in Escherichia coli. Mol Microbiol 28:1199e1210. 10. Tottey S, Rich PR, Rondet SAM, Robinson NJ (2001) Two Menkes-type atpases supply copper for photosynthesis in Synechocystis PCC 6803. J Biol Chem 276:19999e20004. 11. Kehres DG, Zaharik ML, Finlay BB, Maguire ME (2000) The NRAMP proteins of Salmonella typhimurium and Escherichia coli are selective manganese transporters involved in the response to reactive oxygen. Mol Microbiol 36:1085e1100. 12. Degen O, Eitinger T (2002) Substrate specicity of nickel/cobalt permeases: Insights from mutants altered in transmembrane domains I and II. J Bacteriol 184:3569e3577. 13. Hohle TH, OBrian MR (2009) The mntH gene encodes the major Mn2+ transporter in Bradyrhizobium japonicum and is regulated by manganese via the Fur protein. Mol Microbiol 72:399e409. 14. OBrian MR, Fabiano E (2010) Mechanisms and regulation of iron homeostasis in the Rhizobia. Iron Uptake and Homeostasis in Microorganisms, eds Cornelis P, Andrews SC (Caister Academic Press, Norfolk), pp 37e63. 15. Hohle TH, OBrian MR (2010) Transcriptional control of the Bradyrhizobium japonicum irr gene requires repression by Fur and antirepression by Irr. J Biol Chem 285: 26074e26080. 16. Chao TC, Becker A, Buhrmester J, Phler A, Weidner S (2004) The Sinorhizobium meliloti fur gene regulates, with dependence on Mn(II), transcription of the sitABCD operon, encoding a metal-type transporter. J Bacteriol 186:3609e3620. 17. Daz-Mireles E, et al. (2005) The manganese-responsive repressor Mur of Rhizobium leguminosarum is a member of the Fur-superfamily that recognizes an unusual operator sequence. Microbiology 151:4071e4078. 18. Platero R, de Lorenzo V, Garat B, Fabiano E (2007) Sinorhizobium meliloti fur-like (Mur) protein binds a fur box-like sequence present in the mntA promoter in a manganese-responsive manner. Appl Environ Microbiol 73:4832e4838. 19. Anderson ES, et al. (2009) The manganese transporter MntH is a critical virulence determinant for Brucella abortus 2308 in experimentally infected mice. Infect Immun 77:3466e3474.

20. Anjem A, Varghese S, Imlay JA (2009) Manganese import is a key element of the OxyR response to hydrogen peroxide in Escherichia coli. Mol Microbiol 72:844e858. 21. Puri S, Hohle TH, OBrian MR (2010) Control of bacterial iron homeostasis by manganese. Proc Natl Acad Sci USA 107:10691e10695. 22. Qi Z, OBrian MR (2002) Interaction between the bacterial iron response regulator and ferrochelatase mediates genetic control of heme biosynthesis. Mol Cell 9: 155e162. 23. Friedman YE, OBrian MR (2003) A novel DNA-binding site for the ferric uptake regulator (Fur) protein from Bradyrhizobium japonicum. J Biol Chem 278: 38395e38401. 24. Friedman YE, OBrian MR (2004) The ferric uptake regulator (Fur) protein from Bradyrhizobium japonicum is an iron-responsive transcriptional repressor in vitro. J Biol Chem 279:32100e32105. 25. Yang J, Sangwan I, OBrian MR (2006) The Bradyrhizobium japonicum Fur protein is an iron-responsive regulator in vivo. Mol Genet Genomics 276:555e564. 26. Geer LY, Domrachev M, Lipman DJ, Bryant SH (2002) CDART: Protein homology by domain architecture. Genome Res 12:1619e1623. 27. Sding J, Biegert A, Lupas AN (2005) The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res 33(Web Server issue): W244eW248. 28. Cavard D (2002) Assembly of colicin A in the outer membrane of producing Escherichia coli cells requires both phospholipase A and one porin, but phospholipase A is sufcient for secretion. J Bacteriol 184:3723e3733. 29. Nikaido H, Rosenberg EY (1983) Porin channels in Escherichia coli: Studies with liposomes reconstituted from puried proteins. J Bacteriol 153:241e252. 30. Burgess J (1978) Metal Ions in Solution (Ellis Horwood Ltd, Sussex), p 481. 31. Schauer K, Gouget B, Carrire M, Labigne A, de Reuse H (2007) Novel nickel transport mechanism across the bacterial outer membrane energized by the TonB/ExbB/ExbD machinery. Mol Microbiol 63:1054e1068. 32. Stork M, et al. (2010) An outer membrane receptor of Neisseria meningitidis involved in zinc acquisition with vaccine potential. PLoS Pathog 6:e1000969. 33. Sukhan A, Hancock REW (1996) The role of specic lysine residues in the passage of anions through the Pseudomonas aeruginosa porin OprP. J Biol Chem 271: 21239e21242. 34. Hamza I, Hassett R, OBrian MR (1999) Identication of a functional fur gene in Bradyrhizobium japonicum. J Bacteriol 181:5843e5846. 35. Frustaci JM, Sangwan I, OBrian MR (1991) Aerobic growth and respiration of a aminolevulinic acid synthase (hemA) mutant of Bradyrhizobium japonicum. J Bacteriol 173:1145e1150. 36. Yang J, et al. (2006) Bradyrhizobium japonicum senses iron through the status of haem to regulate iron homeostasis and metabolism. Mol Microbiol 60:427e437. 37. Chang WS, et al. (2007) An oligonucleotide microarray resource for transcriptional proling of Bradyrhizobium japonicum. Mol Plant Microbe Interact 20:1298e1307. 38. Kaneko T, et al. (2002) Complete genomic sequence of nitrogen-xing symbiotic bacterium Bradyrhizobium japonicum USDA110. DNA Res 9:189e197. 39. Small SK, Puri S, Sangwan I, OBrian MR (2009) Positive control of ferric siderophore receptor gene expression by the Irr protein in Bradyrhizobium japonicum. J Bacteriol 191:1361e1368.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1110137108

Hohle et al.

Supporting Information
Hohle et al. 10.1073/pnas.1110137108

Fig. S1. Identication of genes responsive to manganese and Mur. Genes repressed by manganese and derepressed in a mur mutant were identied by whole-genome microarray analysis. (A) Graphical representation and overlap of genes that are up-regulated in the parent strain in response to manganese limitation (290) and genes that are up-regulated in a mur background compared with the parent strain grown in high-manganese media (207). (B) Twelve genes found in both microarray analyses are listed.

Hohle et al. www.pnas.org/cgi/content/short/1110137108

1 of 1