Anda di halaman 1dari 14

The Plant Cell, Vol. 18, 1667–1680, July 2006, www.plantcell.

org ª 2006 American Society of Plant Biologists

The BAM1/BAM2 Receptor-Like Kinases Are Important


Regulators of Arabidopsis Early Anther Development W

Carey L.H. Hord,a,b,1 Changbin Chen,a,1 Brody J. DeYoung,c,2 Steven E. Clark,c and Hong Maa,b,3
a Department of Biology and the Huck Institutes for the Life Sciences, Pennsylvania State University, University
Park, Pennsylvania 16802
b Integrative Biosciences Graduate Degree Program, Pennsylvania State University, University Park, Pennsylvania 16802
c Department of Molecular, Cellular, and Developmental Biology, University of Michigan, Ann Arbor, Michigan 48109

Anther development involves the formation of several adjacent cell types required for normal male fertility. Only a few genes
are known to be involved in early anther development, particularly in the establishment of these different cell layers.
Arabidopsis thaliana BAM1 (for BARELY ANY MERISTEM) and BAM2 encode CLAVATA1-related Leu-rich repeat receptor-
like kinases that appear to have redundant or overlapping functions. We characterized anther development in the bam1
bam2 flowers and found that bam1 bam2 anthers appear to be abnormal at a very early stage and lack the endothecium,
middle, and tapetum layers. Analyses using molecular markers and cytological techniques of bam1 bam2 anthers revealed
that cells interior to the epidermis acquire some characteristics of pollen mother cells (PMCs), suggesting defects in cell
fate specification. The pollen mother-like cells degenerate before the completion of meiosis, suggesting that these cells are
defective. In addition, the BAM1 and BAM2 expression pattern supports both an early role in promoting somatic cell fates
and a subsequent function in the PMCs. Therefore, analysis of BAM1 and BAM2 revealed a cell–cell communication
process important for early anther development, including aspects of cell division and differentiation. This finding may have
implications for the evolution of multiple signaling pathways in specifying cell types for microsporogenesis.

INTRODUCTION on cell lineage (Szymkowiak and Sussex, 1996). For instance, it


was shown in both leaf and root that cells are able to respond to a
Understanding the mechanisms that regulate cell division and new position and differentiate accordingly when they are dis-
differentiation is an essential goal for developmental biologists. placed from one layer into an adjacent layer, irrespective of their
In plants, cellular differentiation occurs as an interplay between lineages (Stewart and Dermen, 1975; van den Berg et al., 1995).
a cell’s lineage and its relative position within a structure Positional signals can be mediated by either apoplastic mech-
(Szymkowiak and Sussex, 1996; Berger et al., 1998; Scheres anisms, via receptors, or symplastic mechanisms, through plas-
and Benfey, 1999). The relative importance of these determining modesmata (Hake and Freeling, 1986; Fletcher et al., 1999; Rojo
factors appears to vary depending on the specific developmental et al., 2002; Kwak et al., 2005). Apoplastic cell–cell communica-
process. For example, analysis using chimeras between Sola- tion is often achieved via receptor–ligand interactions (Bergmann,
num luteum and tomato (Solanum lycopersicum) leaf cell layers 2004). For example, the leucine-rich repeat receptor-like kinase
indicated that trichome development was determined by the (LRR-RLK) SCRAMBLED is required for normal epidermal cell
genotype of the differentiating L1 cells (autonomous develop- type patterning in the Arabidopsis thaliana root (Kwak et al.,
ment), whereas leaf shape was controlled by the genotype 2005). In addition, the ERECTA (ER) and ERECTA-LIKE1 (ERL1)
of the neighboring cell layer (nonautonomous development) and ERL2 genes encode LRR-RLKs that control position-
(Jorgenson and Crane, 1927; Szymkowiak and Sussex, 1996). dependent guard cell differentiation (Shpak et al., 2005). Hence,
For several aspects of plant development, cell fate specifica- positional cues via receptor-mediated intercellular signaling from
tion has been shown to rely more on positional information than adjacent cells help control cell division and differentiation (Hake
and Freeling, 1986; Berger et al., 1998; Kwak et al., 2005).
In the male reproductive organ of flowering plants, the anther,
1 These authors contributed equally to this work. the differentiation of sporogenous and parietal cell types is essen-
2 Current address: Department of Biology, Indiana University, Bloo- tial for the propagation of the species. In Arabidopsis, develop-
mington, IN 47405.
3 To whom correspondence should be addressed. E-mail hxm16@ ment of the male gametophytes, the pollen grains, occurs within
psu.edu; fax 814-863-1357. the four lobes of the anthers (Goldberg et al., 1993; Sanders et al.,
The authors responsible for distribution of materials integral to the 1999). Within each lobe, cells divide and differentiate to form
findings presented in this article in accordance with the policy described distinct somatic cell layers surrounding the developing repro-
in the Instructions for Authors (www.plantcell.org) are: Steven E. Clark ductive cells, or the pollen mother cells (PMCs). The proper
(clarks@umich.edu) and Hong Ma (hxm16@psu.edu).
W
Online version contains Web-only data.
formation and development of these somatic cell layers is critical
Article, publication date, and citation information can be found at for the development and eventual release of pollen grains
www.plantcell.org/cgi/doi/10.1105/tpc.105.036871. (Mariani et al., 1990, 1991; Denis et al., 1993; Ross et al., 1996;
1668 The Plant Cell

Sanders et al., 1999; Zhao et al., 2002; Ma, 2005; Hord and Ma, division, they fail to complete cytokinesis and degrade com-
2006). pletely by anther stage 7 (Ma, 2005; Hord and Ma, 2006). EMS1/
Anther development in Arabidopsis has been divided into spe- EXS, SERK1, and SERK2 encode LRR-RLKs, indicating an es-
cific stages according to morphological characteristics (Sanders sential role for a position-dependent intercellular signaling event
et al., 1999). In the emergent anther primordium (stage 1), there (Canales et al., 2002; Zhao et al., 2002; Albrecht et al., 2005;
are three distinct cell layers derived from the floral meristem: Colcombet et al., 2005). TPD1 encodes a small putatively se-
from outer to inner they are L1, L2, and L3 (Sanders et al., 1999). creted protein that may act in the same pathway as EMS1/EXS
The four corners of the anther primordia develop into the four and SERK1/2 (Yang et al., 2003, 2005; Albrecht et al., 2005;
lobes during further anther development, as described previ- Colcombet et al., 2005). In addition to its role in anther develop-
ously (Goldberg et al., 1993; Sanders et al., 1999). The L1 layer ment, SERK1 was initially shown to promote somatic embryo-
develops into the epidermis, and the L2-derived cell layers de- genesis (Schmidt et al., 1997; Hecht et al., 2001).
velop via a series of periclinal (parallel to the adjacent outer LRR-RLKs constitute one of the largest gene families in
surface and creating additional cell layers) and anticlinal (per- Arabidopsis; however, relatively few of these genes have known
pendicular to the outer surface and increasing cell number in a functions (Shiu and Bleecker, 2001, 2003). For instance, CLAV-
layer) cell divisions. At stage 2, L2 cells called archesporial cells ATA1 (CLV1) is a LRR-RLK that acts with CLV3 to limit the size of
divide periclinally to form primary parietal (outer) and primary the shoot apical and floral meristems and may promote the tran-
sporogenous (inner) cells. Then at stage 3, periclinal division of sition of central zone cells to peripheral zone cells (Clark et al.,
the primary parietal cells forms the inner and outer secondary 1996; Fletcher et al., 1999; Gallois et al., 2002; Rojo et al., 2002;
parietal cells. Subsequently, in each of the four anther lobes, cells Lenhard and Laux, 2003). In addition, BRASSINOLIDE INSEN-
of one of the secondary parietal layers divide again at stage 4, SITIVE1 (BRI1) and BRI1-ASSOCIATED RECEPTOR KINASE
resulting in three somatic layers at stage 5: endothecium, middle (BAK1/SERK3) are involved in the brassinosteroid signaling path-
layer, and tapetum, which immediately surrounds the L2-derived way (He et al., 2000; Wang et al., 2001; Li et al., 2002; Nam and Li,
PMCs. The tapetum is important for providing the developing 2002).
pollen with nutrients and materials (Mariani et al., 1990, 1991; BAM1 (for BARELY ANY MERISTEM) and BAM2 encode LRR-
Denis et al., 1993). The L3 layer gives rise to the vascular and RLKs (DeYoung et al., 2006) that share high levels of amino acid
connective tissues of the anther. sequence identity and form a four-gene monophyletic clade
To date, only a few genes are known to be involved in early along with CLV1 and BAM3 (Shiu and Bleecker, 2001; DeYoung
anther cell division and cell differentiation events, including et al., 2006). Single mutants in these genes do not exhibit any
SPOROCYTELESS/NOZZLE (SPL/NZZ), EXCESS MICROSPO- obvious morphological defects, indicating that they have redun-
ROCYTES1/EXTRA SPOROGENOUS CELLS (EMS1/EXS), dant functions (DeYoung et al., 2006). By contrast, bam1 bam2
SOMATIC EMBRYOGENESIS1 (SERK1), SERK2, and TAPETUM double mutants display multiple developmental defects, includ-
DETERMINANT1 (TPD1) (Schiefthaler et al., 1999; Yang et al., ing a reduction of meristem size, altered leaf shape, size, and
1999; Canales et al., 2002; Zhao et al., 2002; Yang et al., 2003, venation, male sterility, and reduced female fertility (DeYoung
2005; Ito et al., 2004; Albrecht et al., 2005; Colcombet et al., et al., 2006). Here, we show that the bam1 and bam2 mutations
2005; Ma, 2005; Hord and Ma, 2006). SPL/NZZ was shown to affect normal cell division and differentiation during early anther
promote microsporogenesis under the control of AGAMOUS in development. The bam1 bam2 double mutant does not produce
whorl three floral organs (Ito et al., 2004). In the spl and nzz the somatic cell layers that are derived from archesporial cells;
mutants, the L2-derived cells do not develop properly and are instead, they only form cells that resemble PMCs, which then de-
unable to form PMCs (Schiefthaler et al., 1999; Yang et al., 1999). grade before completing meiosis. The BAM1 and BAM2 genes
The detailed descriptions of the spl and nzz mutants differ are expressed in the area of archesporial cells as early as stage 2
somewhat. The spl mutant was reported as having primary and also preferentially in sporogenous cells and PMCs at later
sporogenous and secondary parietal cells (Yang et al., 1999), stages. The very early BAM1/2 expression pattern and the early
whereas the description that the nzz mutant forms an undiffer- morphological defects suggest that these genes promote cell
entiated mass of archesporial cells (Schiefthaler et al., 1999) division and differentiation, including the specification of the pa-
suggests that SPL/NZZ may act at the stage when the arche- rietal cells that give rise to the endothecium, middle layer, and
sporial cells divide to form primary sporogenous cells and pri- tapetum. BAM1/2 expression in sporogenous cells and PMCs
mary parietal cells. SPL/NZZ has been cloned and encodes a and the degeneration of PMCs in the bam1 bam2 double mutant
putative transcription factor (Schiefthaler et al., 1999; Yang et al., suggest that these genes also play a role in the development and/
1999). or function of PMCs.
The ems1/exs, serk1 serk2, and tpd1 mutations affect cellular
differentiation by altering cell fate at a later stage than spl/nzz RESULTS
(Schiefthaler et al., 1999; Yang et al., 1999; Canales et al., 2002;
Zhao et al., 2002; Yang et al., 2003; Albrecht et al., 2005; Anther Development in the bam1 bam2 Mutant Is Abnormal
Colcombet et al., 2005). Instead of forming the four normal L2-
derived cell types, these mutants form endothecium, middle bam1 bam2 mutant plants are male sterile, and the mutant
layer cells, and a greater than normal number of PMCs. The anthers fail to produce pollen (DeYoung et al., 2006). To better
mutant anther lobes completely lack the nutritive tapetum layer. understand the overall morphological differences between the
Although the excess PMCs enter meiosis and undergo nuclear mutant and wild-type anthers, scanning electron microscopy
BAM1/BAM2 and Anther Development 1669

images of dissected buds were examined. At approximately 1 bam1 bam2 anther (27.9 6 3.5; n ¼ 11) was slightly smaller than
flower stage 8 (Smyth et al., 1990), the size and shape of wild- the number in the wild type (Landsberg erecta [Ler]) (31.9 6 5.6;
type and mutant anthers appeared similar (Figures 1A and 1D). n ¼ 44). This reduction is much less dramatic than the size
Subsequently, mutant anthers had lobes that appeared less full reduction of the bam1 bam2 inflorescence meristem (DeYoung
(Figures 1B and 1E) at approximately late stage 9, although the et al., 2006). At stage 2, both wild-type and mutant anthers ap-
stage of the mutant flowers was difficult to determine at this res- peared to have a similar structure and overall cell patterning, but
olution because of the abnormal size and morphology of other the cells in the mutant remained slightly larger than those of the
floral organs. Close to stage 10, the mutant anthers had a wild type (Figures 2B and 2E). At anther stage 3, the wild-type
shriveled appearance, suggesting that the locules had collapsed anther has well-defined lobes and the archesporial cells have
(Figures 1C and 1F). Anther development appeared to be defec- undergone a periclinal cell division, forming the primary sporog-
tive at or before flower stage 9, during which several key devel- enous and primary parietal cells, which lie approximately parallel
opmental processes occurred, including the establishment of the to the epidermis at the outermost part of each lobe (Figure 2C).
five cell layers and meiosis. The anther morphology of bam1-1 Mutant anthers at stage 3 appeared to have fewer and larger
bam2-1 (data not shown) is similar to that of bam1-3 bam2-3. cells in each lobe (Figure 2F). Although evidence of a periclinal
cell division was observed occasionally, in general there were no
bam1 bam2 Is Defective in Formation of Anther Cell Layers clearly defined primary parietal and primary sporogenous cells.
At stage 4, wild-type sporogenous cells are visible at the center
To better understand the defect in anther development, we of each lobe (Figure 2G). At this stage, the primary parietal cells
prepared and analyzed transverse sections of wild-type and have divided periclinally to form the outer secondary parietal and
bam1-3 bam2-3 anthers (Figure 2). At stage 1 of anther devel- the inner secondary parietal cells. At stage 4 in the mutant, the
opment, cells from all three cell layers, L1, L2, and L3, appeared width and thickness of the anther began to appear substantially
slightly larger in the bam1 bam2 anthers than in the wild type greater than the wild-type dimensions (Figure 2I). Although at
(Figures 2A and 2D). At the same time, the average number (6SD) times there appeared to have been cell divisions with orienta-
of subepidermal cells (L2 and L3) in a cross section of the stage tions close to periclinal and anticlinal, the mutant anthers still did
not form the cell layers that are characteristic of normal anthers,
and the cellular pattern appeared to be disorganized.
In the wild-type anther at stage 5, the PMCs are formed at the
center of each lobe and are surrounded sequentially, from inner
to outer, by the tapetum, middle layer, and endothecium, with the
epidermis encasing the entire anther (Figure 2H). With the ex-
ception of the epidermis, the mutant anthers never formed these
organized sporophytic cell layers (Figure 2J). Instead, their
enlarged cells appeared slightly fewer in number (see below)
and less organized. In addition, the nuclei of these cells appeared
unusually large, a characteristic of meiotic cells. Meiosis in the
wild type commences during anther stage 6, and the PMCs be-
come isolated from each other and from the tapetum as a thick
callose wall is formed around them (Figure 2K). Some cells within
the bam1 bam2 anther lobes had callose around them, whereas
some other cells exhibited signs of degradation (Figure 2L).
Eventually, most or all of the L2-derived cells in the mutant de-
graded, causing the lobes to collapse and the anther to appear
shriveled (Figure 1F).
The cells in the bam1 bam2 mutant anthers appeared larger
and fewer in number than their wild-type counterparts. To obtain
Figure 1. Scanning Electron Microscopy Images of Dissected Buds. more quantitative information, the number (Figure 3A) and di-
Several floral organs have been removed to allow for better visualization mensions of the L2-derived cells in each lobe were analyzed
of the anther defects. using cross sections. At stage 5 in the wild type, the average
(A) to (C) Wild-type (Ler) floral buds. number 6 SD of PMCs per lobe was 3.8 6 0.9, and the total
(A) A floral bud at stage 8 (anther stage 4). number of L2-derived cells per lobe was 36.9 6 3.8 (Figure 3A).
(B) A floral bud at stage 9 (anther stage is between 5 and 7). By contrast, the number of L2-derived cells in the double mutant
(C) A floral bud at stage10 (anther stage is between 7 and 8).
was 15.8 6 1.8. Although this number is obviously smaller than
(D) to (F) bam 1-3 bam 2-3 double mutant (Ler) floral buds.
the total number of L2-derived cells in the wild-type, it is also
(D) A floral bud near stage 8. The anthers appear similar to those of the
wild type (A).
much larger than the average number of PMCs per lobe in the
(E) A mutant floral bud near stage 9 showing anthers that are less full than wild type, by greater than fourfold. The average length (16.0 6
the wild type (B). 5.4 mm; parallel to the epidermis) and height (14.1 6 5.0 mm;
(F) A floral bud near stage 10, with anthers that have apparently col- perpendicular to the epidermis) of the mutant cells were signif-
lapsed locules. icantly greater than those of wild-type PMCs (5.8 6 2.5 mm long,
1670 The Plant Cell

Figure 2. Semithin Sections of Anthers.


(A) to (C), (G), (H), and (K) Wild type (Ler).
(D) to (F), (I), (J), and (L) bam 1-3 bam 2-3 (Ler).
(A) A wild-type anther at stage 1.
(B) A wild-type anther at stage 2.
(C) A wild-type anther at stage 3.
(D) A mutant anther at stage 1.
(E) A mutant anther at stage 2.
(F) A mutant anther at stage 3 does not have normally organized L2-derived cell layers.
(G) A wild-type anther at stage 4.
(H) A wild-type anther at stage 5 showing five distinct cell layers in each lobe.
(I) A mutant anther at stage 4 has the epidermis, but the distinct developing cell layers are missing.
(J) A mutant anther at stage 5, which appears to completely lack the normal subepidermal cell layers. The cells in each lobe appear enlarged.
(K) A wild-type anther at stage 6 showing thick callose surrounding the meiotic sporogenous cells.
(L) A mutant anther at stage 6; the L2-derived cells begin to degrade.
ar; archesporial cells; c, callose; en, endothecium; ep, epidermis; isp, inner secondary parietal cells; mi, middle layer; osp, outer secondary parietal
cells; pmc, pollen mother cells; pml, pollen mother-like cells; pp, primary parietal cells; ps, primary sporogenous cells; sl, sporogenous-like cells; t,
tapetal cells. Bar ¼ 20 mm; all panels are of the same magnification.

4.7 6 2.1 mm high). Interestingly, although the length of the division defect, we counted the number of cell walls per lobe,
mutant juxtaepidermal cells at stage 5 did not differ significantly interior to the epidermis and exterior to the PMCs, that appeared
from those of the corresponding wild-type cells (16.0 6 5.4 mm to have arisen via a periclinal cell division for stages 4 to 5 (Figure
for the mutant and 16.1 6 4.2 mm for the wild type), the height of 3B). The wild type at stage 4 had an average of 3.5 6 0.8
the cells was significantly greater in the mutant (14.1 6 5.0 mm for periclinal cell division events per lobe, which was not statistically
the mutant and 6.9 6 1.6 mm for the wild type). Together, these different from the lobes transitioning from stage 4 to 5, which had
observations indicate that cell division and/or cell expansion 5.9 6 2.2. However, at stage 5, the wild type had an average of
were altered in the bam1 bam2 mutant. In particular, it seems that 10.8 6 1.9 periclinal cell division events per lobe, which was
cells in the double mutant expanded without some of the cell significantly higher than both stage 4 and stage 4 to 5. In the
divisions that normally produced the secondary parietal cells or bam1 bam2 mutant, the numbers of periclinal cell divisions were
subsequent cell layers in the wild type. similarly small at stages 4 (1.6 6 0.9) and 5 (1.3 6 0.7) and
The cell layers in the anther are normally formed by periclinal severely reduced compared with those seen at stages 4 and 5 in
cell divisions of the subepidermal cells. To better quantify the cell the wild type. Sections of bam1-1 bam2-1 at stages 5 and 6 (data
BAM1/BAM2 and Anther Development 1671

the PMCs, which are normally found interior to the tapetum. To


test whether the L2-derived mutant cells have additional prop-
erties of PMCs, we examined the expression of known cell
identity markers. In wild-type anthers, the meiotic genes
ATRAD51 and SDS are strongly expressed in PMCs at stage 6
in the center of each lobe (Figures 4A and 4B, respectively)
(Azumi et al., 2002; Li et al., 2004). In stage 6 bam1 bam2 anthers,
ATRAD51 and SDS appeared to be expressed (Figures 4D and
4E, respectively) in the L2-derived cells that had not yet de-
graded, including the cells that occupied a position immediately
interior of the epidermis. This finding suggests that the L2-
derived PMLs in the bam1 bam2 anther share molecular prop-
erties with normal PMCs. To further verify the lack of tapetal cells
in the mutant anther, we examined the expression of DYT1,
which is very strongly expressed in the wild-type tapetal cells at
late anther stage 5 and stage 6 (Figure 4C) (W. Zhang, Y. Sun, and
H. Ma, unpublished results). No DYT1 expression was detected
in the bam1 bam2 mutant anthers at stages 5 or 6 (Figure 4F; data
not shown), indicating the absence of tapetal cells.

PMLs in bam1 bam2 Anthers Can Enter but Fail to


Complete Meiosis

The ems1 mutant lacks the tapetum layer and produces excess
PMCs, which proceed to meiosis II (Zhao et al., 2002). The
expression of meiotic genes in bam1 bam2 PMLs suggests that
they may also undergo meiosis. To test this hypothesis, we
Figure 3. Comparison of the Number of Cells and Cell Division Events analyzed 49,6-diamidino-2-phenylindole (DAPI)–stained chro-
between Wild Type and bam1 bam2. mosome spreads. More than 200 cells from mutant anthers
were examined. Most cells appeared to be entering or undergo-
(A) Average numbers of cells per cross section in each lobe.
ing meiosis, on the basis of their chromosome morphology.
(B) Average numbers of cell division events per cross section in each
Among these, at least one-quarter of them were at the leptotene
lobe that were parallel to the epidermis, between the epidermis and the
PMCs. or zygotene stage of prophase I. In addition, ;10% of the cells
MS, PMCs; SE, subepidermal L2-derived cells. Error bars indicate SD. had pachytene-like chromosomes, at midprophase I. Therefore,
a substantial fraction of the PMLs had entered meiosis. Amaz-
not shown) appeared to be very similar to those of bam1-3 ingly, two of the examined cells were at metaphase I and one was
bam2-3. Similar anther sections of bam1-3 bam2-3 plants car- at anaphase I (data not shown). The chromosomes in the meta-
rying a ProER-BAM1-FLAG construct (DeYoung et al., 2006) were phase I cells were highly condensed and aligned but appeared
normal (data not shown). to be partially degraded (data not shown). In addition, in the
In summary, bam1 bam2 mutant anthers did not form the anaphase I cell, the chromosomes segregated unequally be-
normal somatic cells layers in the L2-derived position. The cells tween the two poles of the meiocyte. Furthermore, chromosomal
that formed in their place were larger, fewer in number, and fragments were frequently observed at different stages of mei-
appeared to be randomly organized; these cells degraded near osis (Figure 4J; data not shown). To verify that the cells able to
anther stage 6. Periclinal cell division events in the mutant were enter meiosis were not confined to the center of the lobes, we
reduced compared with the wild type. Although the cells formed performed DAPI staining of sectioned anthers. Whereas in the
by the bam1 bam2 anthers were significantly larger than the wild type, the four sporophytic cell layers surrounding the meiotic
normal sporophytic cells, they appeared similar to the PMCs cells were clearly observable (Figure 4G), in the bam1 bam2 an-
formed in wild-type anthers. We conclude that the bam1 bam2 thers only the epidermis was seen encasing and juxtaposed to a
anther formed pollen mother-like cells (PMLs) in place of the mass of randomly organized PMLs, many of which were evi-
three normal somatic cell types. These PMLs were clearly abnor- dently undergoing meiosis (Figure 4H).
mal, because they degenerated before producing microspores In summary, the L2-derived cells in bam1 bam2 anthers
(see below for more characterization). possessed attributes of PMCs and were partially able to enter
meiosis, indicating that the L2-derived cell fates are altered in the
L2-Derived Cells in bam1 bam2 Anthers Express mutant anthers. On the other hand, the PMLs exhibited defects
PMC Markers at meiotic stages much earlier than the PMCs in the ems1/exs,
serk1 serk2, and tpd1 mutants, with most cells unable to com-
The L2-derived cells in bam1 bam2 mutant anthers have size, plete prophase I before degeneration. Therefore, BAM1 and
shape, and distribution that are similar to the characteristics of BAM2 are also important for normal PMC development.
1672 The Plant Cell

Expression of BAM1 and BAM2 Supports Their Function in


Anther Development

To obtain clues about the mechanisms of BAM1 and BAM2


action, we analyzed their expression during early anther devel-
opment (Figure 5). Overall, BAM1 and BAM2 had very similar
expression patterns. Both were expressed in the archesporial
cells at anther stage 2 (Figures 5A and 5H) and in the primary
parietal and primary sporogenous cells near the lateral edges of
stage 3 anthers (Figures 5B and 5I). Interestingly, at stage 4, both
genes appeared to be expressed predominantly in the sporog-
enous cells and might have a low level of expression in the L2-
derived secondary parietal cells (Figures 5C and 5J). At stage 5,
BAM1 and BAM2 were highly expressed in the PMCs, with a very
low level of expression in the tapetum (Figures 5D and 5K).
During anther stage 6, both BAM1 and BAM2 were very strongly
expressed in the PMCs and tapetum and might be weakly
expressed in the middle layer (Figures 5E and 5L). Similarly, at
stage 7, strong BAM1 expression was seen in the tapetum and a
lower level was seen in the tetrads (Figure 5F). By anther stage 9,
there was no detectable BAM1 and BAM2 expression (data not
shown). Also, there was no detectable expression in bam1 bam2
double mutant tissues (data not shown). In summary, BAM1 and
BAM2 appear to be expressed at stage 2 in the archesporial cells
and at stage 3 in the primary sporogenous and primary parietal
cells; subsequently, they are preferentially expressed in the spo-
rogenous cells at anther stage 4, after which their expression
becomes restricted to the tapetum and PMCs. These expression
patterns support an early function that promotes the formation of
the primary parietal cells, which are the progenitors of the L2-
Figure 4. Molecular and Cytological Analyses of Wild-Type and bam1 derived somatic cell layers, and a later function in support of
bam2 Anthers at Stage 6. PMC development.
(A) to (C), (G), and (I) Wild type.
(D) to (F), (H), and (J) bam1 bam2.
(A) and (D) ATRAD51 was expressed in wild-type PMCs (A) at a high Possible Relationships of BAM1/2 with SPL and EMS1
level and in bam1 bam2 (D) interior to the epidermis (arrows).
As BAM1/2 appear to have a role in promoting the differentiation
(B) and (E) SDS was expressed in the wild type (B) at the center of the
of the parietal cell type, in situ hybridization experiments were
lobe at the position of the PMCs and in bam1 bam2 (E) in the cells interior
and juxtaposed to the epidermis (arrows). performed with two known anther genes: SPL/NZZ and EMS1/
(C) and (F) DYT1 expression. In the wild type (C), DYT1 signal was EXS (Figure 6). From anther stages 2 and 4, there was no obvious
detected in the tapetal layer, but no detectable DYT1 signal was seen in difference in SPL/NZZ expression between the wild type (Figures
bam1 bam2 (F). 6A and 6B) and bam1 bam2 (Figures 6D and 6E). At stage 5,
(G) and (H) Sections of wild-type and bam1 bam2 anthers stained with however, when SPL/NZZ expression was restricted to the PMCs
DAPI. in the wild type (Figure 6C), SPL/NZZ expression in the bam1
(G) A wild-type lobe with PMCs undergoing meiosis in the center sur- bam2 anther had expanded to all or most of the L2-derived cells
rounded by the tapetal layer, middle layer, endothecium, and epidermis. (Figure 6F), consistent with these cells being similar to PMCs.
(H) A bam1 bam2 lobe showing a greater number of PMCs only sur-
Conversely, BAM1 expression in the spl mutant appeared normal
rounded by the epidermis and connective tissue.
at stage 2 (cf. Figures 5A and 6G), but in the L2-derived cells, it
(I) and (J) DAPI-stained chromosome spreads of wild-type and bam1
bam2 meiotic cells showing pachytene-like chromosomes. Fragments of continued to be restricted to the juxtaepidermal cells at later
chromosome can be seen in the bam1 bam2 meiotic cells (J) (arrow) but stages (Figures 6H and 6I). Also, BAM1 expression in the spl
not in the wild type (I). mutant appeared to extend into the epidermal cells. This finding
en, endothecium; ep, epidermis; mi, middle layer; pmc, pollen mother is consistent with the failure of the spl mutant to form sporog-
cells; pml, pollen mother-like cells; t, tapetum. Bar ¼ 20 mm in (A) and (G) enous cells and PMCs and suggests that the L2-derived cells
and 10 mm in (I). (A) to (F), (G) and (H), and (I) and (J) are of the same may remain undifferentiated or archesporial-like. Therefore,
magnification. BAM1/2 expression and SPL/NZZ expression at anther stage 2
seem independent of each other; on the other hand, by stage 5,
BAM1/2 appear to play a role in restricting SPL/NZZ expression
to the PMCs, and SPL/NZZ seems to promote PMC-preferential
expression of BAM1.
BAM1/BAM2 and Anther Development 1673

Figure 5. BAM1 and BAM2 in Situ Hybridization of Wild-Type Ler Anthers.

(A) to (F) BAM1 expression.


(G) Sense control probe.
(H) to (L) BAM2 expression.
(A) and (H) Stage 2 anthers, showing that BAM1 and BAM2 are expressed in lateral L2-derived archesporial cells (arrows).
(B) and (I) Stage 3 anthers; BAM1 and BAM2 are expressed predominantly in the primary sporogenous cells.
(C) and (J) Stage 4 anthers; BAM1 and BAM2 expression is seen predominantly in the sporogenous cells.
(D) and (K) Stage 5 anthers; both genes were highly expressed in PMCs and may have a very low level of expression in the tapetal layer.
(E) and (L) Stage 6 anthers showing that BAM1 and BAM2 are very strongly expressed in the tapetum and PMCs and faintly expressed in the middle layer.
(F) Stage 7 anther showing strong BAM1 expression in the tapetum and somewhat weaker expression in the tetrads.
(G) Sense probe showing no cross-hybridization.
ms, microspores in tetrads; pmc, pollen mother cells; t, tapetal cells. Bar ¼ 20 mm; all panels are of the same magnification.

For EMS1/EXS expression at early stages, there was no clear stages 4 and 5, BAM1 expression was seen in the PMCs of ems1
difference in the expression pattern between the wild type and (Figures 6Q and 6R), consistent with the normal BAM1 expres-
bam1 bam2 (Figures 6J and 6M; data not shown). At early wild- sion in PMCs at these stages. Therefore, BAM1/2 may be in-
type stage 5, EMS1/EXS expression was strong in the PMCs and volved in regulating the reduction of EMS1 expression in the
moderate in the tapetum (Figure 6K). At late stage 5/early stage 6, PMCs, but EMS1 does not seem to affect the BAM1 expression
EMS1/EXS expression was almost exclusively in the tapetum and pattern.
very weak, if present, in the PMCs (Figure 6L). The bam1 bam2
stage 5 anthers had strong and expanded EMS1/EXS expression DISCUSSION
in the PMLs (Figure 6N), filling the lobes. Furthermore, strong
EMS1/EXS expression continued in the L2-derived cells at stage BAM1 and BAM2 Are Important for Normal Cell Division
6 (Figure 6O), suggesting that the PMLs were different from and Differentiation in Early Anther Development
normal PMCs, in agreement with other studies described above.
BAM1 expression in the ems1 mutant was similar to that of We have demonstrated here that BAM1 and BAM2 together are
the wild type through stage 4 (cf. Figures 5A and 6P). During important for normal early anther development. Our phenotypic
Figure 6. In Situ Hybridizations of BAM1/2, SPL/NZZ, and EMS1/EXS.
(A) to (C) SPL expression in wild-type anthers.
(D) to (F) SPL expression in bam1 bam2 anthers.
(G) to (I) BAM1 expression in spl anthers.
(J) to (L) EMS1 expression in wild-type anthers.
(M) to (O) EMS1 expression in bam1 bam2 anthers.
(P) to (R) BAM1 expression in ems1 anthers.
(A) and (D) SPL expression in wild-type and bam1 bam2 anthers, respectively, was seen in most of the L2-derived cells at stage 2 (arrows).
(B) and (E) SPL expression in wild-type and bam1 bam2 anthers, respectively, was strongest at the center of each lobe and somewhat fainter in the
remaining L2-derived cells at stage 4.
(C) In the wild type at stage 5, strong SPL expression was seen in the PMCs.
(F) In bam1 bam2 at stage 5, strong SPL expression was seen throughout the lobe in the PMLs.
(G) BAM1 expression in the spl anther at stage 2 was seen in the archesporial cells (arrows).
(H) BAM1 expression in the spl anther at stage 4 was confined mainly to the cells immediately adjacent to the epidermis.
(I) BAM1 expression in the spl anther at stage 5 was seen in the epidermis and juxtaepidermal cells at the lateral edges of the lobes (arrow).
(J) EMS1 expression in the wild type at stage 3 was seen in the L2-derived cells and in parts of the epidermis (arrows); this may be broader than what
was seen in the bam1 bam2 anther at this stage (M).
(K) EMS1 expression in the wild type at early stage 5 was predominantly in the PMCs and weakly in the tapetum.
(L) EMS1 expression in the wild type at late stage 5/early stage 6 was almost exclusively in the tapetum.
(M) EMS1 expression in the bam1 bam2 anther at stage 3 was in the L2-derived cells (arrows).
(N) and (O) EMS1 expression in the bam1 bam2 anther at early stage 5 and early stage 6, respectively, was seen throughout the anther lobes in PMLs.
(P) BAM1 expression in the ems1 anther at stage 2 was seen in the archesporial cells (arrows).
(Q) BAM1 expression in the ems1 anther at stage 4 was seen predominantly in the center of the lobe and weakly in the other L2-derived cells.
(R) Strong BAM1 expression was seen in the ems1 anther at stage 5 in the PMCs.
pmc, pollen mother cell; pml, pollen mother-like cell; t, tapetum. Bar ¼ 20 mm; all panels are of the same magnification.
BAM1/BAM2 and Anther Development 1675

studies indicate that bam1 bam2 double mutant anthers have


only a slightly reduced number of cells at very early stages, in
contrast with the dramatic size reduction of bam1 bam2 mer-
istems (DeYoung et al., 2006). Therefore, the bam1 bam2 anther
defects do not seem to be a direct consequence of a greatly
reduced anther primordium that resulted from the small mutant
meristem. Nevertheless, analysis of subsequent anther stages
clearly indicates a dramatic reduction of L2-derived cell num-
bers, with the bam1 bam2 anther producing fewer than half of the
wild-type number of subepidermal cells per lobe, suggesting that
cell division is decreased significantly in the mutant. In addition,
our analyses showed that the bam1 bam2 mutant failed to
specify the normal identity of L2-derived somatic cell layers,
completely lacking the endothecium, middle layer, and tapetum.
Moreover, morphological, cytological, and molecular studies
indicate that the L2-derived cells in the mutant had properties of
PMCs. Furthermore, BAM1 and BAM2 were expressed in ar-
chesporial cells at stage 2, and in the bam1 bam2 double mutant,
abnormal cell patterning was visible by stage 3. As BAM1 and
BAM2 encode LRR-RLKs, we propose that they meditate a de- Figure 7. Model for the Control of Early Anther Cell Differentiation by
BAM1 and BAM2.
velopmental signal that promotes the differentiation of arche-
sporial cells at stage 2 (Figure 7A). Our results strongly support (A) BAM1/2 function in early anther development. In the wild-type anther
the idea that BAM1 and BAM2 contribute to the differentiation of by stage 3, archesporial cells have divided to form primary parietal (pp)
the parietal and perhaps sporogenous cell types at stage 3. It is and primary sporogenous (ps) cells. In the absence of BAM1 and BAM2
possible that BAM1 and BAM2 regulate the asymmetric cell function, the asymmetric division of the archesporial cells might be
abnormal, resulting in the lack of primary parietal cells and the formation
division of the archesporial cells into the primary sporogenous
of primary sporogenous-like cells. At anther stage 5, the wild-type anther
and primary parietal cells.
has three clearly defined concentric cell layers surrounding the pollen
In the wild-type anther, the number of somatic cells increases mother cells; in the bam1 bam2 anther, however, only PMLs are observed
considerably during stages 4 and 5, whereas the sporogenous in the area interior of the epidermis. Therefore, BAM1 and BAM2 define a
cells expand in size but do not change dramatically in number cell–cell signaling pathway critical for the differentiation of the primary
during this period. In the bam1 bam2 anther, the decrease in L2- parietal and primary sporogenous cells. Abbreviations for cell types are
derived cell number is accompanied by the change of somatic as in Figure 2.
cells to PMLs. In the wild type, PMCs are larger than the L2- (B) Left, interaction of CLV1, CLV3, and WUS in controlling the stem cell
derived somatic cells; therefore, it is possible that the reduction pool in meristems. Right, proposed interaction of BAM1/2, SPL, and a
in number and expansion in size of the mutant L2-derived cells CLE-type gene for a putative ligand for BAM1/2, regulating cell fates in
the anther. Arrows indicate positive genetic interactions; lines with a bar
are a reflection of these cells being similar to PMCs. A similar
at the end denote negative interactions; open arrows represent a positive
phenomenon was seen in the ems1/exs, tpd1, and serk1 serk2
effect on the cellular process.
mutants, which possess a total number of PMCs that is less than
the combined number of tapetum and PMCs seen in the wild type
but is greater than the number of PMCs seen in the wild type later in anther development. In addition, although the L2-derived
(Canales et al., 2002; Zhao et al., 2002; Yang et al., 2003; cells in the mutant resemble PMCs, they are not normal. Some of
Albrecht et al., 2005; Colcombet et al., 2005). It is possible that the PMLs in the bam1 bam2 mutant were able to initiate meiosis,
when progenitor cells are directed to form PMCs (or PMLs), they but none was able to complete meiosis I, and all mutant PMLs
divide less frequently and develop into larger cells. An alternative eventually degenerated. Sporogenesis requires the coordinated
explanation is that in these mutants a negative regulation of PMC development of multiple adjacent cell types that probably in-
proliferation is lacking as a result of the absence of the adjacent volves cell–cell communication; therefore, it is plausible that
somatic cells, allowing the sporogenous cells to proliferate ab- BAM1/2 might be needed to receive signaling directed toward
normally. Therefore, the functions of BAM1 and BAM2 in regu- the PMCs. It is possible that the bam1 bam2 PMLs did not fully
lating cell division and differentiation may be coupled, and the and/or properly differentiate; the strong and expanded EMS1/
normal functions of these genes are important for the formation EXS expression in the bam1 bam2 anther, which persisted
of correct cell types in the anther. through early stage 6, suggests that the PMLs might have
some somatic cell properties. Therefore, BAM1 and BAM2 may
BAM1 and BAM2 May Play a Role in Sporogenous Cells promote the normal differentiation of PMCs. In addition, BAM1
and BAM2 may be required in the meiocytes for the completion
In addition to the BAM1/BAM2 function in very early anther devel- of meiosis. Another possibility is that one or more of the somatic
opment, strong BAM1 and BAM2 expression was seen from an- cell layers may be required for meiosis. For example, the tapetum
ther stages 4 to 6, preferentially in sporogenous cells and PMCs. may provide materials and nutrients needed for normal meiosis,
This finding suggests possible additional roles for these genes or BAM1 and BAM2 may mediate a response to a developmental
1676 The Plant Cell

signal(s) normally released from neighboring cells that promotes plants called the CLEs (Cock and McCormick, 2001; Carles and
the meiotic process. LRR-RLKs constitute a very large gene Fletcher, 2003); however, clv3 mutants are fertile (Clark et al.,
family, but only a small number of them have been character- 1996), suggesting that either CLV3 is not a ligand for BAM1/2 or
ized functionally. If BAM1 and BAM2 indeed promote PMC that the CLV3 gene is functionally redundant with another gene.
differentiation and/or function, this would be a new function for Nevertheless, the biochemical similarity of BAM1/2 and CLV1
LRR-RLKs. It is possible that other LRR-RLKs are involved in suggests that the ligand for BAM1/2 may be a member of the CLE
this complex process, perhaps by interacting with BAM1 and family (Cock and McCormick, 2001).
BAM2. Further genetic and molecular studies are needed to Our results and previous studies strongly support the hypoth-
uncover and better elucidate LRR-RLK functions in anther de- esis that both SPL/NZZ and BAM1/BAM2 act very early in anther
velopment. development to promote the formation of several cell types. The
spl/nzz mutants do not produce any sporogenous cells, whereas
Model for BAM1/2 Function in Anther Development the bam1 bam2 mutant generates an abnormally large number of
PMLs, which had several properties of PMCs. Therefore, SPL/
BAM1 and BAM2 regulate several aspects of development, NZZ and BAM1/2 seem to act in opposing ways in regulating the
including meristem size and leaf development (DeYoung et al., number of sporogenous(-like) cells. Furthermore, our results on
2006). We show here that they are important for the formation of the expression of these genes suggest that, although the ex-
somatic cell types in early anther development. Detailed analysis pression patterns of these genes at stage 2 are not affected by
of the bam1 bam2 anthers suggests that the mutant defect in each other, BAM1/2 seem to limit the domain of SPL/NZZ
somatic cell formation is not likely to be a direct consequence of expression and SPL/NZZ promotes BAM1 expression in the
the reduced meristem. Therefore, we believe the BAM1/2 func- central sporogenous cells. Therefore, BAM1/2 and SPL/NZZ, as
tion in anther development is different from that in regulating well as the ligand for BAM1/2, may form a regulatory loop (Figure
meristem size. BAM1 and BAM2 are closely related in sequence 7B), somewhat similar to the interactions between CLV1, WUS,
to CLV1 (DeYoung et al., 2006) (Figure 8), which acts in the and CLV3 (Figure 7B) (Carles and Fletcher, 2003). It is likely that
meristem (Clark et al., 1996). CLV1 may achieve the correct wild-type BAM1/2 negatively regulate SPL/NZZ expression in-
balance of the central zone and periphery zone in the meristem directly, possibly via the specification of somatic L2-derived cell
by limiting cell proliferation in the central zone cells and/or types, which have reduced SPL/NZZ expression. In addition, the
promoting the transition of central zone cells to peripheral zone expanded EMS1/EXS expression in the stage 5 bam1 bam2 an-
cells (a form of differentiation) (Clark et al., 1996, 1997; Gallois ther may also reflect a change in cell type, because EMS1/EXS is
et al., 2002) (Figure 7B). We found that BAM1 and BAM2 expressed in the early PMCs. However, normal EMS1/EXS
negatively regulate the number of sporogenous cells at the cen- expression is reduced in meiocytes and strong in the tapetum;
ter of the anther lobes, seemingly by promoting the differentiation therefore, the strong EMS1/EXS expression in the bam1 bam2
of the peripheral somatic cells and/or possibly by reducing the PMLs at the time of meiosis might be an indication of a defect in
division of sporogenous cells. Therefore, the BAM1/BAM2 func- meiocyte development in bam1 bam2, supporting a role of the
tion in the early anther may be to promote differentiation (and BAM1/2 genes in meiocytes.
limit proliferation) in a manner analogous to the role of CLV1 in the
meristem (Figure 7B). Just as CLV1 acts to restrict the prolifer- BAM1 and BAM2 Function and the Evolution of Sporophytic
ation of the cells in which it is expressed and may promote the Cell Types
differentiation of adjacent cells, BAM1/2 expression and the
phenotype conferred by bam1 bam2 suggest that BAM1 and In the ems1/exs, serk1 serk2, tpd1, and bam1 bam2 mutants, cell
BAM2 may restrict sporogenous cell proliferation while promot- differentiation is altered to the end that one or more normal cell
ing the differentiation of the adjacent parietal cells. layers do not form, but PMCs or PMLs form in their place. These
Further support for the functional similarity between BAM1/2 results support the idea that a default pathway leads to the
and CLV1 came from transgenic experiments. It was shown that formation of PMCs and that the development of the other cell
high levels of BAM1 or BAM2 were able to partially rescue the layers requires differentiation mediated by additional signaling
phenotype conferred by clv1-11, and expression of CLV1 using pathways. It is known that nonflowering plants have relatively
the ER promoter, which is active in the anther, could completely simple structure in the microsporangium. For example, in
rescue the phenotype conferred by bam1 bam2 (DeYoung et al., leptosporangiate ferns, a series of precise cell divisions results
2006). These results suggest that these highly similar proteins in the formation of the sporangium, which has an outer wall (one
have retained some conserved biochemical activities and, when cell layer thick), and a two-cell-layered tapetum, which sur-
expressed outside of their normal expression domains, are able rounds and provides nutrients to the developing spore mother
to interact with components of related signaling pathways cells (Raven et al., 1999). Spores are released when the sporo-
(DeYoung et al., 2006). Thus, although CLV1 normally functions phytic sac opens, without the highly decorated wall found on
to limit meristem stem cell population size and/or to promote the pollen grains. It appears that as vascular plants evolved, partic-
transition of central zone cells to peripheral zone cells, when ularly angiosperms, they acquired additional developmental path-
expressed in the developing anther it can function in the place of ways that allowed for greater complexity in the structure of the
BAM1/2 to promote the differentiation of parietal cells and to limit male sporangium, which is the anther lobe in angiosperms. Our
sporogenous cells. CLV3 is thought to encode the ligand for results support the hypothesis that through evolution, cell divi-
CLV1 and is a member of a conserved gene family in flowering sions that would have resulted in the formation of sporogenous
BAM1/BAM2 and Anther Development 1677

Figure 8. Neighbor-Joining Tree of Selected Arabidopsis, Rice, and Poplar LRR-RLK Amino Acid Sequences.
Gene identifier numbers starting with At indicate genes from Arabidopsis thaliana; names of genes with functional information are given after the
identifier numbers. Gene identifier numbers starting with Os indicate genes from rice (Oryza sativa); MSP1 is also shown as Os MSP1. Pt indicates genes
from poplar (Populus trichocarpa), with temporary names given according to sequence similarity to the closest Arabidopsis genes. Bootstrap values are
shown near the relevant nodes.

cells in an ancestor were restricted or otherwise altered to tors. Distinct homologs of each member of the BAM/CLV1 clade
produce other cell types, among which the earliest and most have been identified in poplar (Populus trichocarpa) and rice
important appears to be the tapetum. (Oryza sativa) (Figure 8), suggesting that the function of BAM1
The specification of primary parietal cell fate by the BAM1 and and BAM2 might be conserved in other flowering plants and
BAM2 LRR-RLKs represents a novel signaling pathway that has different from the roles of CLV1 and BAM3. It is known that the
implications for the evolution of the sporophytic cell types in- homolog of EMS1 in rice, MSP1, has a very similar function to
volved in sporogenesis. The development of PMLs seems to be a that of EMS1; the rice msp1 mutant also forms excess PMCs in
default pathway, and the development of the somatic cell types the anther and concomitantly lacks the tapetum (Nonomura et al.,
is achieved through the specification of a progenitor cell type 2003). Further investigation of the rice homologs of BAM1 and
(parietal cells) as a result of the acquisition of new signaling path- BAM2 is needed to test whether they also have conserved
ways, perhaps like that defined by the BAM1 and BAM2 recep- functions.
1678 The Plant Cell

METHODS scope with a SPOT II RT camera or a Nikon E800 microscope coupled


with a Nikon D50 SLR digital camera. All images were then edited using
Plant Material and Growth Conditions Photoshop 7.0 (Adobe Systems).
A fragment of the BAM1 cDNA from 821 to 2250 bp after the ATG
Two mutant alleles were found for each of the BAM1 (At5g65700) and initiation codon was amplified using the primers 59-ATCTAGATCTTCTC-
BAM2 (At3g49670) genes and were described in detail previously CGGTCCATTAACTTGG-39 and 59-ACTCGAGTCTGTGTCTTATCCTTCC-
(DeYoung et al., 2006). Briefly, the bam1-1 and bam1-3 alleles were TAAG-39 and cloned into the T/A site of the pCRII-TOPO vector (Invitrogen)
generated in the Columbia ecotype and backcrossed into the Ler eco- (DeYoung et al., 2006). The resulting construct (pMC 3021) was linearized
type. The bam1-1 allele contains a dSpm insertion and is from the SLAT with NotI or SpeI and transcribed using SP6 or T7 to generate the antisense
collection. bam1-3 is a SALK T-DNA line. The bam2-1 and bam2-3 alleles and sense probes, respectively. The BAM2 probes were similarly synthe-
are in the Ler ecotype, and each contains a Ds insertion. bam2-1 is from a sized using in vitro transcription from a linearized fragment of the BAM2
launching pad line, and bam2-3 is from the Cold Spring Harbor TRAPPER cDNA, which was amplified with the gene-specific primers 59-ATCTAGA-
collection. The two pairs of double mutants used in this study were bam1-1 CATTTACAGGGACAATAACTCAA-39 and 59-ACTCGAGCTGTGTCT-
bam2-1 and bam1-3 bam2-3. Ler plants were used as the wild-type control. GATTCTACCTAGC-39 and cloned into pCRII-TOPO, resulting in pMC
Arabidopsis thaliana seeds were planted directly, or transplanted after 3022. The plasmid was linearized with BglII and transcribed using SP6 for
germinating on Murashige and Skoog plates, on potting mixture and were the antisense probe. The SDS, EMS1, and ATRAD51 probes were synthe-
grown with a 16-h-light/8-h-dark cycle at 18 to 238C. sized as described previously (Azumi et al., 2002; Zhao et al., 2002; Li et al.,
2004). The DYT1 probe was synthesized using a cDNA clone (W. Zhang, Y.
Characterization of the Mutant Phenotype Sun, and H. Ma, unpublished data). The SPL/NZZ probe was synthesized
as described previously (Sieber et al., 2004).
Scanning electron microscopic analysis was performed on flowers from
various stages as described previously (Dievart et al., 2003). Some organs
were removed to expose the inner floral organs. Flower buds and inflo- Construction of the Phylogenetic Tree
rescences were prepared for sectioning using a method described pre- To generate the phylogenetic tree shown in Figure 8, BAM1, BAM2,
viously (Owen and Makaroff, 1995) with some minor modifications (Zhao BAM3, CLV1, ER, EMS1, and BRI1 protein sequences from Arabidopsis
et al., 2002). Semithin (0.5 mm) sections were made using either a were used to perform a BLASTp search of The Institute for Genomic
Reichert-Jung Ultracut E ultramicrotome or an Ultracut UCT ultramicro- Research rice (Oryza sativa) pseudomolecules:protein sequences data-
tome (both Leica Microsystems) and were stained with 0.1% toluidine base with no filter. The four members of the BAM/CLV1 clade were also
blue in 0.1% Na2B4O7 for 30 s. Images were photographed using an used to search the Populus trichocarpa genomic sequences at the Plant-
Olympus BX51 microscope and a SPOT II RT Slider digital camera with Tribes website (http://www.floralgenome.org/cgi-bin/tribedb/tribe.cgi).
SPOT software version 3.5.8 for Windows (Diagnostic Instruments). Redundant sequences were removed. Full-length sequences were im-
Images were edited using Photoshop 7.0 (Adobe Systems). Measure- ported into ClustalX (Plate-Forme de Bio-Informatique), and a multiple
ment of cell length and height was done using GIMP version 2.2.4 sequence alignment was performed (see Supplemental Figure 1 online).
software (http://www.gimp.org). For both the wild type and the mutant, Phylogenetic analysis was conducted using MEGA version 3.0 (Kumar
only the cells juxtaposed to the epidermis of stage 5 anther transverse et al., 2004) (http://www.megasoftware.net/index.html) and the neighbor-
sections were measured. Average cell number and standard deviations joining algorithm as well as a bootstrap analysis with 1000 replicates to
were calculated using Microsoft Excel. test the significance of the nodes. Default parameters were used, includ-
DAPI staining and chromosome spreads were performed as described ing random seed initiation, the amino:Poisson correction model, and
previously (Ross et al., 1996). DAPI staining of tissue sections was done uniform rates among sites, and gaps were deleted.
with young inflorescences that were fixed by methanol:acetone (4:1)
fixative for 45 min on ice. They were then embedded in wax and sectioned
at 10 mm thick. Dewaxed slides were directly stained with DAPI. Images Accession Numbers
were taken using a Nikon E800 microscope and a Hamamatsu C4742 Arabidopsis Genome Initiative locus identifiers for the genes mentioned in
digital camera with Image Pro Plus software version 4.5.1.27 for Windows this article are At5g65700 (BAM1) and At3g49670 (BAM2).
(Media Cybernetics).

Supplemental Data
In Situ Hybridization Experiments
The following material is available in the online version of this article.
Nonradioactive RNA in situ hybridization was performed essentially as
Supplemental Figure 1. Multiple Sequence Alignment.
described (Xu et al., 2002). Young inflorescences of wild type and bam1
bam2 double mutants were fixed in FAA (3.5% formaldehyde, 50%
alcohol, and 5% acetic acid) fixative for at least 2 h at room temperature.
The tissue was then dehydrated and embedded in Fisher paraffin. Ten- ACKNOWLEDGMENTS
micrometer-thick sections were made using a Shandon Finesse paraffin
microtome (Thermo Electron) and were mounted onto slides so that each We thank Bridget Leyland for help in collecting our quantitative data and
slide had a similar number of wild-type and bam1 bam2 sections. All with plant care and Khushboo Nangia, Sarah Suchy, and Gavilange
slides were dewaxed with Histo-Clear (National Diagnostics), treated with Nestor for help in plant care and sectioning. We also thank Steven Hord
protein kinase for 30 min, and then dehydrated and baked at 428C for at for assistance in collecting data, editing, and preparing images. We
least 2 h. The dried slides were used immediately or stored at –808C for up thank Kay Schneitz and Patrick Sieber for the SPL/NZZ template, Wei
to 6 months. RNA probes labeled with digoxigenin were used for the Zhang for the DYT1 template, and Wei Hu for help in phylogenetic anal-
hybridization. After hybridization, anti-digoxigenin alkaline phosphatase ysis. We also thank Julian Schroeder and Sacco de Vries for commu-
and NBT/BCIP (nitroblue tetrazolium chloride/5-bromo-4-chloro-3-indolyl nicating results before publication. This work was supported by grants
phosphate, toluidine salt; Roche Diagnostics) were used to detect any from the Department of Energy (DE-FG02-02ER15332) and the National
hybridization signal. Images were taken using a BX51 Olympus micro- Institutes of Health (R01 GM-63871-01) to H.M., by National Institutes of
BAM1/BAM2 and Anther Development 1679

Health Grant R01 GM-62962-05 to S.E.C., and by funds from the Goldberg, R.B., Beals, T.P., and Sanders, P.M. (1993). Anther devel-
Department of Biology and the Huck Institutes of the Life Sciences opment: Basic principles and practical applications. Plant Cell 5,
at Pennsylvania State University. This research used plant materials 1217–1229.
generated with support from National Science Foundation Grant Hake, S., and Freeling, M. (1986). Analysis of genetic mosaics shows
0215923. C.L.H.H. was partially supported by the Integrative Bioscience that the extra epidermal cell divisions in Knotted mutant maize plants
Graduate Degree Program at the Pennsylvania State University. B.J.D. are induced by adjacent mesophyll cells. Nature 320, 621–623.
was partially supported by the Cellular Biotechnology training program. He, Z., Wang, Z.Y., Li, J., Zhu, Q., Lamb, C., Ronald, P., and Chory, J.
(2000). Perception of brassinosteroids by the extracellular domain of
the receptor kinase BRI1. Science 288, 2360–2363.
Received August 4, 2005; revised April 26, 2006; accepted May 11, 2006; Hecht, V., Vielle-Calzada, J.P., Hartog, M.V., Schmidt, E.D., Boutilier,
published June 2, 2006. K., Grossniklaus, U., and de Vries, S.C. (2001). The Arabidopsis
SOMATIC EMBRYOGENESIS RECEPTOR KINASE 1 gene is ex-
pressed in developing ovules and embryos and enhances embryo-
REFERENCES genic competence in culture. Plant Physiol. 127, 803–816.
Hord, C.L.H., and Ma, H. (2006). Genetic control of anther cell division
Albrecht, C., Russinova, E., Hecht, V., Baaijens, E., and de Vries, S. and differentiation In Cell Division Control in Plants, D.P.S. Verma and
(2005). The Arabidopsis thaliana SOMATIC EMBRYOGENESIS Z. Hong, eds (Heidelberg, Germany: Springer-Verlag), in press.
RECEPTOR-LIKE KINASES1 and 2 control male sporogenesis. Plant Ito, T., Wellmer, F., Yu, H., Das, P., Ito, N., Alves-Ferreira, M.,
Cell 17, 3337–3349. Riechmann, J.L., and Meyerowitz, E.M. (2004). The homeotic pro-
Azumi, Y., Liu, D., Zhao, D., Li, W., Wang, G., Hu, Y., and Ma, H. tein AGAMOUS controls microsporogenesis by regulation of SPOR-
(2002). Homolog interaction during meiotic prophase I in Arabidopsis OCYTELESS. Nature 430, 356–360.
requires the SOLO DANCERS gene encoding a novel cyclin-like protein. Jorgenson, C.A., and Crane, M.B. (1927). Formation and morphology
EMBO J. 21, 3081–3095. of Solanum chimeras. J. Genet. 18, 247–273.
Berger, F., Haseloff, J., Schiefelbein, J., and Dolan, L. (1998). Kumar, S., Tamura, K., and Nei, M. (2004). MEGA3: Integrated
Positional information in root epidermis is defined during embryogen- software for molecular evolutionary genetics analysis and sequence
esis and acts in domains with strict boundaries. Curr. Biol. 8, 421–430. alignment. Brief. Bioinform. 5, 150–163.
Bergmann, D.C. (2004). Integrating signals in stomatal development. Kwak, S.H., Shen, R., and Schiefelbein, J. (2005). Positional signaling
Curr. Opin. Plant Biol. 7, 26–32. mediated by a receptor-like kinase in Arabidopsis. Science 307,
Canales, C., Bhatt, A.M., Scott, R., and Dickinson, H. (2002). EXS, a 1111–1113.
putative LRR receptor kinase, regulates male germline cell number Lenhard, M., and Laux, T. (2003). Stem cell homeostasis in the Arabidop-
and tapetal identity and promotes seed development in Arabidopsis. sis shoot meristem is regulated by intercellular movement of CLAV-
Curr. Biol. 12, 1718–1727. ATA3 and its sequestration by CLAVATA1. Development 130, 3163–3173.
Carles, C.C., and Fletcher, J.C. (2003). Shoot apical meristem main- Li, J., Wen, J., Lease, K.A., Doke, J.T., Tax, F.E., and Walker, J.C.
tenance: The art of a dynamic balance. Trends Plant Sci. 8, 394–401. (2002). BAK1, an Arabidopsis LRR receptor-like protein kinase, inter-
Clark, S.E., Jacobsen, S.E., Levin, J.Z., and Meyerowitz, E.M. (1996). acts with BRI1 and modulates brassinosteroid signaling. Cell 110,
The CLAVATA and SHOOT MERISTEMLESS loci competitively reg- 213–222.
ulate meristem activity in Arabidopsis. Development 122, 1567–1575. Li, W., Chen, C., Markmann-Mulisch, U., Timofejeva, L., Schmelzer,
Clark, S.E., Williams, R.W., and Meyerowitz, E.M. (1997). The CLAV- E., Ma, H., and Reiss, B. (2004). The Arabidopsis AtRAD51 gene is
ATA1 gene encodes a putative receptor kinase that controls shoot dispensable for vegetative development but required for meiosis.
and floral meristem size in Arabidopsis. Cell 89, 575–585. Proc. Natl. Acad. Sci. USA 101, 10596–10601.
Cock, J.M., and McCormick, S. (2001). A large family of genes that Ma, H. (2005). Molecular genetic analyses of microsporogenesis and
share homology with CLAVATA3. Plant Physiol. 126, 939–942. microgametogenesis in flowering plants. Annu. Rev. Plant Biol. 56,
Colcombet, J., Boisson-Dernier, A., Ros-Palau, R., Vera, C.E., and 393–434.
Schroeder, J.I. (2005). Arabidopsis SOMATIC EMBRYOGENESIS Mariani, C., De Beuckeleer, M., Truettner, J., Leemans, J., and
RECEPTOR KINASES1 and 2 are essential for tapetum development Goldberg, R.B. (1990). Induction of male sterility in plants by a
and microspore maturation. Plant Cell 17, 3350–3361. chimaeric ribonuclease gene. Nature 347, 737–741.
Denis, M., Delourme, R., Gourret, J.P., Mariani, C., and Renard, M. Mariani, C., Goldberg, R.B., and Leemans, J. (1991). Engineered male
(1993). Expression of engineered nuclear male sterility in Brassica sterility in plants. Symp. Soc. Exp. Biol. 45, 271–279.
napus. Plant Physiol. 101, 1295–1304. Nam, K.H., and Li, J. (2002). BRI1/BAK1, a receptor kinase pair
DeYoung, B.J., Bickle, K.L., Schrage, K.J., Muskett, P., Patel, K., mediating brassinosteroid signaling. Cell 110, 203–212.
and Clark, S.E. (2006). The CLAVATA1-related BAM1, BAM2 and Nonomura, K., Miyoshi, K., Eiguchi, M., Suzuki, T., Miyao, A.,
BAM3 receptor kinase-like proteins are required for meristem function Hirochika, H., and Kurata, N. (2003). The MSP1 gene is necessary to
in Arabidopsis. Plant J. 45, 1–16. restrict the number of cells entering into male and female sporogenesis
Dievart, A., Dalal, M., Tax, F.E., Lacey, A.D., Huttly, A., Li, J., and and to initiate anther wall formation in rice. Plant Cell 15, 1728–1739.
Clark, S.E. (2003). CLAVATA1 dominant-negative alleles reveal func- Owen, H.A., and Makaroff, C.A. (1995). Ultrastructure of microsporo-
tional overlap between multiple receptor kinases that regulate meri- genesis and microgametogenesis in Arabidopsis thaliana (L.) Heynh.
stem and organ development. Plant Cell 15, 1198–1211. ecotype Wassilewskija (Brassicaciae). Protoplasma 185, 7–21.
Fletcher, J.C., Brand, U., Running, M.P., Simon, R., and Meyerowitz, Raven, P.H., Evert, R.F., and Eichhorn, S.E. (1999). Biology of Plants,
E.M. (1999). Signaling of cell fate decisions by CLAVATA3 in Arabi- 6th ed. (New York: Worth Publishers), pp. 453–454.
dopsis shoot meristems. Science 283, 1911–1914. Rojo, E., Sharma, V.K., Kovaleva, V., Raikhel, N.V., and Fletcher, J.C.
Gallois, J.L., Woodward, C., Reddy, G.V., and Sablowski, R. (2002). (2002). CLV3 is localized to the extracellular space, where it activates
Combined SHOOT MERISTEMLESS and WUSCHEL trigger ectopic the Arabidopsis CLAVATA stem cell signaling pathway. Plant Cell 14,
organogenesis in Arabidopsis. Development 129, 3207–3217. 969–977.
1680 The Plant Cell

Ross, K.J., Fransz, P., and Jones, G.H. (1996). A light microscope atlas Stewart, R.N., and Dermen, H. (1975). Flexibility in ontogeny as shown
of meiosis in Arabidopsis thaliana. Chromosome Res. 4, 507–516. by the contribution of the shoot apical layers to leaves of periclinal
Sanders, P.M., Bui, A.Q., Weterings, K., McIntire, K.N., Hsu, Y., Lee, chimeras. Am. J. Bot. 62, 935–947.
P.Y., Truong, M.T., Beals, T.P., and Goldberg, R.B. (1999). Anther Szymkowiak, E.J., and Sussex, I.M. (1996). What chimeras can tell us
development defects in Arabidopsis thaliana male-sterile mutants. about plant development. Annu. Rev. Plant Physiol. Plant Mol. Biol.
Sex. Plant Reprod. 11, 297–322. 47, 351–367.
Scheres, B., and Benfey, P.N. (1999). Asymmetric cell division in van den Berg, C., Willemsen, V., Hage, W., Weisbeek, P., and
plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 505–537. Scheres, B. (1995). Cell fate in the Arabidopsis root meristem deter-
Schiefthaler, U., Balasubramanian, S., Sieber, P., Chevalier, D., mined by directional signalling. Nature 378, 62–65.
Wisman, E., and Schneitz, K. (1999). Molecular analysis of NOZZLE, Wang, Z.Y., Seto, H., Fujioka, S., Yoshida, S., and Chory, J. (2001).
a gene involved in pattern formation and early sporogenesis during BRI1 is a critical component of a plasma-membrane receptor for plant
sex organ development in Arabidopsis thaliana. Proc. Natl. Acad. Sci. steroids. Nature 410, 380–383.
USA 96, 11664–11669. Xu, Y.Y., Chong, K., Xu, Z.H., and Tan, K.H. (2002). Practical tech-
Schmidt, E.D., Guzzo, F., Toonen, M.A., and de Vries, S.C. (1997). A niques of in situ hybridization with RNA probe. Chin. Bull. Bot. 19,
leucine-rich repeat containing receptor-like kinase marks somatic 234–238.
plant cells competent to form embryos. Development 124, 2049– Yang, S.L., Jiang, L.X., Puah, C.S., Xie, L.F., Zhang, X.Q., Chen, L.Q.,
2062. Yang, W.C., and Ye, D. (2005). Overexpression of TAPETUM DE-
Shiu, S.H., and Bleecker, A.B. (2001). Receptor-like kinases from TERMINANT1 alters the cell fates in the Arabidopsis carpel and
Arabidopsis form a monophyletic gene family related to animal recep- tapetum via genetic interaction with EXCESS MICROSPOROCYTES1/
tor kinases. Proc. Natl. Acad. Sci. USA 98, 10763–10768. EXTRA SPOROGENOUS CELLS. Plant Physiol. 139, 186–191.
Shiu, S.H., and Bleecker, A.B. (2003). Expansion of the receptor-like Yang, S.L., Xie, L.F., Mao, H.Z., Puah, C.S., Yang, W.C., Jiang, L.,
kinase/Pelle gene family and receptor-like proteins in Arabidopsis. Sundaresan, V., and Ye, D. (2003). TAPETUM DETERMINANT1 is
Plant Physiol. 132, 530–543. required for cell specialization in the Arabidopsis anther. Plant Cell 15,
Shpak, E.D., McAbee, J.M., Pillitteri, L.J., and Torii, K.U. (2005). 2792–2804.
Stomatal patterning and differentiation by synergistic interactions of Yang, W.C., Ye, D., Xu, J., and Sundaresan, V. (1999). The SPORO-
receptor kinases. Science 309, 290–293. CYTELESS gene of Arabidopsis is required for initiation of sporogen-
Sieber, P., Petrascheck, M., Barberis, A., and Schneitz, K. (2004). esis and encodes a novel nuclear protein. Genes Dev. 13, 2108–2117.
Organ polarity in Arabidopsis. NOZZLE physically interacts with mem- Zhao, D.Z., Wang, G.F., Speal, B., and Ma, H. (2002). The EXCESS
bers of the YABBY family. Plant Physiol. 135, 2172–2185. MICROSPOROCYTES1 gene encodes a putative leucine-rich repeat
Smyth, D.R., Bowman, J.L., and Meyerowitz, E.M. (1990). Early flower receptor protein kinase that controls somatic and reproductive cell
development in Arabidopsis. Plant Cell 2, 755–767. fates in the Arabidopsis anther. Genes Dev. 16, 2021–2031.

Anda mungkin juga menyukai