ABSTRACT
The apical portion of the Arabidopsis globular stage embryo gives rise to the cotyledons and the shoot apical meristem (SAM). The SHOOT MERISTEMLESS (STM) gene is required for SAM formation during embryogenesis and for SAM function throughout the lifetime of the plant. To more precisely define the development of molecular pattern in the apical portion of the embryo, and the role of the STM gene in the development of this pattern, we have examined AINTEGUMENTA (ANT), UNUSUAL FLORAL ORGANS (UFO) and CLAVATA1 (CLV1) expression in wild-type and stm mutant embryos. The transcripts of these genes mark subdomains within the apical portion of the embryo. Our results indicate that: (1) the molecular organization characteristic of the vegetative SAM is not present in the globular embryo but instead develops gradually during embryogenesis; (2) radial pattern exists in the apical portion of the embryo prior to and independent of STM with STM expression itself responding to radial information; (3) the embryonic SAM consists of central and peripheral subdomains that express different combinations of molecular markers and differ in their ultimate fates; and (4) STM activity is required for UFO expression, STM is required for maintenance but not onset of CLV1 expression and the pattern of ANT expression is independent of STM.
INTRODUCTION
The shoot apical meristem (SAM) of angiosperm plants gives rise to new leaves and stem in a predictable and regular pattern. The SAM arises during embryogenesis. The presumptive cotyledons flank the presumptive SAM in the apical half of the globular embryo. By the heart stage of embryogenesis, the presumptive SAM consists of three layers of cells that are the precursors to the L1, L2 and L3 layers (Barton and Poethig, 1993). As the cotyledons grow out and bend over the top of the embryo (the ‘walking-stick stage’) these cell layers adopt division patterns characteristic of the SAM. These divisions result in the formation of a small mound of cells with the tunica/corpus organization typical of the angiosperm SAM.
The Arabidopsis SHOOT MERISTEMLESS (STM) gene is required for the initiation and maintenance of the SAM. Embryos homozygous for strong loss-of-function mutations in the STM gene form cotyledons and other embryonic structures but never establish a population of self-renewing stem cells (Barton and Poethig, 1993). In stm mutant embryos, the three precursor layers to the SAM fail to undergo the cell divisions that generate the tunica/corpus organization of the SAM. Thus, the stm phenotype is detectable anatomically at the walking stick stage when no mound of cells recognizable as the embryonic SAM is formed. Plants homozygous for weak stm alleles form SAMs but these terminate prematurely (Clark et al., 1996; Endrizzi et al., 1996; Felix et al., 1996).
The STM gene encodes a Knotted-like homeodomain containing protein and is therefore likely to act as a transcriptional regulator in promoting SAM development (Long et al., 1996). Within the embryo, the STM gene is expressed in the presumptive SAM beginning at the late globular stage and expression continues in the developing SAM throughout embryogenesis. Postembryonically, STM is expressed throughout the vegetative, inflorescence and floral meristems and is excluded from presumptive organ primordia.
The close correlation between lack of STM expression and organ primordium formation suggests that STM promotes SAM development by negatively regulating leaf fate and/or positively regulating ‘meristem’ cell fate within the apex. Here, we explore the development of apical pattern in wild-type and stm mutant embryos by analyzing the expression of the AINTEGUMENTA (ANT), UNUSUAL FLORAL ORGANS (UFO), and CLAVATA1 (CLV1) genes. In addition to establishing which aspects of apical embryonic pattern STM is required for, and which it is not required for, this analysis shows that the apical pattern characteristic of the SAM develops gradually during embryogenesis.
The ANT, UFO and CLV1 genes were chosen because their transcripts are limited to subdomains of apical meristems. The UFO and CLV1 subdomains fall within the domain of STM expression (Clark et al., 1997; Lee et al., 1997), and the ANT gene shows a complementary expression pattern to that of STM (Elliott et al., 1996). The UFO gene is required for the correct placement of floral organs and for the correct specification of floral organ fate (Levin and Meyerowitz, 1995; Wilkinson and Haughn, 1995). If UFO has a role outside the inflorescence, it must be redundantly encoded by another gene since ufo mutants exhibit normal embryonic and vegetative development. The CLV1 gene is required to limit SAM size (Leyser and Furner, 1992; Clark et al., 1993) and is expressed in the L3 and possibly the L2 layers of the SAM (Clark et al., 1997). The ANT gene is expressed in presumptive organs throughout the plant, including the cotyledons (Elliott et al., 1996). In the flower, ANT plays a role in the outgrowth of sepals, stamens and petals. However, since the mutant phenotype of ant homozygotes is limited to the flower (Elliott et al., 1996; Klucher et al., 1996), if ANT has a role in embryonic or vegetative development, this role must be redundantly performed by another gene.
MATERIALS AND METHODS
Growth conditions and genetics
Soil-grown plants were grown in Metromix 200 (Grace Sierra Co.) under continuous light (100 microeinsteins/m2/second) in Conviron growth chambers at 24°C.
stm-1 is a strong mutant allele and has been described previously (Barton and Poethig, 1993; Long et al., 1996). The stm-11 allele is due to a nonsense mutation at codon 196. In all characters tested, i.e. extent of cotyledon fusion and rate of escape, stm-11 is a stronger mutant allele than stm-1 and thus more closely approximates a null allele.
In situ hybridization
In situ hybridization was performed as described by Long et al. (1996). A detailed protocol is posted at http://www.wisc.edu/genetics/CATG/barton/index.html.
Double labeling: To detect both STM and UFO transcripts in the same section, antisense STM transcripts were labeled with digoxigenin-11-UTP and antisense UFO transcripts were labeled with fluorescein-12-UTP. The two probes were hybridized in approximately equimolar amounts. STM transcripts were detected first using anti-digoxigenin antibody (diluted 1:1250) to which alkaline phosphatase had been conjugated (Boehringer Mannheim). Fast Red was used as the substrate. The slides were then mounted in water, photographed and placed in 1× SSC for 2.5 hours at 65°C to deactivate the alkaline phosphatase. Slides were placed in 1× NTE for 10 minutes and blocked in 1% BSA in 100 mM Tris 7.5, 150 mM NaCl, 0.3% Triton X-100 for 45 minutes. UFO transcripts were detected using anti-fluorescein antibodies (Boehringer Mannheim) conjugated to alkaline phosphatase (diluted 1:6000 and incubated at room temperature for 2 hours). Western Blue was used as the substrate, slides were mounted in GelMount (Fisher Scientific) and rephotographed.
Probes: Antisense STM transcripts were made from meriHB1, a 1.1 kb cDNA clone as described by Long et al. (1996). Antisense UFO transcripts were made from pDW22.1 as in Lee et al. (1997). For ANT, a 516 base pair fragment (corresponding to nucleotides 613 to 1129 of the ANT coding region (Klucher et al., 1996)) was amplified from Landsberg erecta genomic DNA using PFU polymerase (Stratagene) and the following primers: ANT1 – 5′ATATATCTCG-AGGGTGGAGGATTTCTTTGGGACCC3′ and ANT2 – 5′ATATATTCTAGAAACGCCTCGGTATTGAGAAGTTCG-3′. This fragment does not contain the AP2-like repeats found in ANT to prevent crosshybridization to other AP2 like genes. For CLV1, two probes were used which gave identical results. The first probe contained a 1206 base pair fragment (corresponding to nucleotides 48 to 1254 of CLV1; Clark et al., 1997) using the primers CLV1a – 5′AGACTCGAGTCTCACTGAGAGGACAC3′ and CLV1b – 5′AGAATCTAGAGTTAGAGAGAATTAAC3′. The second probe contained a 326 base pair fragment (corresponding to nucleotides 2787 to 3113) using the primers CLV1c – 5′ACACCTCGAGTTCGGAGTGGTTTTGTTGG3′ and CLV1d – 5′ATCTTCTAGACATATTTATCTTGCTTGGG3′.
Stages of embryogenesis
Globular stage embryos are spherical with more than 16 cells. Transition stage embryos are no longer spherical but the developing cotyledons do not yet extend above the embryonic apex. In heart stage embryos the developing cotyledons extend above the apex but make up less than one half of the length of the embryo. In torpedo stage embryos, the cotyledons make up about half of the embryo length. Walking-stick embryos are those in which the cotyledons bend over the apex.
RESULTS
The pattern of STM expression differs in central and peripheral regions of the embryo
The STM gene is expressed in the SAM progenitor cells beginning at the late globular stage and continuing throughout embryogenesis (Long et al., 1996). Its expression is confined to the presumptive SAM and excluded from the presumptive cotyledons. However, both the histological analysis of SAM formation and the analysis of STM expression were confined primarily to the longitudinal, frontal plane of the embryo. (See Fig. 1 for definition of section planes.) We have extended the analysis of STM expression in the developing embryo to the sagittal and transverse planes. All of the analysis presented is based on the observation of serial sections of complete embryos. STM mRNA is found in a continuous band between the cotyledons of the heart-shaped embryo (Fig. 2C-F). This band of expression begins in the globular embryo where it is found first in the peripheral regions and subsequently enlarges to include the central regions. Expression is first detected in one or two cells in the globular stage embryo; these cells are always slightly displaced from the center of the embryo (Fig. 2A). Subsequently, the domain of STM expression enlarges such that expression is detected on both sides of the globular/transition stage embryo (see sagittal section in Fig. 2B). At this stage, the peripheral regions of the presumptive SAM express detectable levels of STM mRNA, but the central region does not. By the early heart stage of embryogenesis, expression has spread and is found in a continuous band between the presumptive cotyledons (Fig. 2C,D). We will refer to the peripheral regions of STM expression as the embryonic peripheral region (EPR) and the central region as the embryonic central region (ECR). The observation that expression of STM in the peripheral regions of the embryonic SAM can be separated from that in the central regions of the embryonic SAM shows that differences in radial positional information are present within the presumptive SAM at this early stage and that STM expression is sensitive to these differences. Sections behind and in front of that shown in Fig. 2B contain the presumptive cotyledons and lack STM expression. In the hundreds of globular embryos we have examined, we have never observed STM expression in the entirety of the apical half of the globular embryo. Thus, cotyledons and leaves differ in the type of cells from which they derive; all of the cells of true leaves originate from cells that express STM, but the bulk of the cotyledons originate from cells that have never expressed STM.
The band of STM expression persists through the heart stage but as the embryo advances through the torpedo and the walking-stick stages of embryogenesis, expression in the EPR decreases relative to the expression in the ECR (Fig. 2F,H,J). By the time the embryo is fully developed, expression is reduced or undetectable in the EPR but remains strong in the ECR (Fig. 2K,L). The depth of STM expression in the EPR and the ECR also differ. While expression in the ECR is confined to the region above the o′ boundary, expression in the EPR in the heart and torpedo stage embryos extends below the o′ boundary into the basal half of the embryo (Fig. 2F,H). (The o′ boundary is formed at the eight cell stage of embryogenesis and divides the embryo proper into upper and lower halves.)
In addition to having different STM expression patterns, the EPR and the ECR differ in their ultimate fates. Cells in the ECR participate in making the meristematic dome that forms between the cotyledons in the wild-type embryo, while cells in the EPR participate in forming the separations between the cotyledons. This can be seen in the wild-type seedling in Fig. 3. The two cotyledons wrap around the apex and meet at their bases in the periphery of the medial sagittal plane. Since plant cells do not migrate, the cells in the valleys where the cotyledons meet must derive from the EPR. In agreement with this, the bases of the cotyledons meet along a region that is three to four cells in diameter in the wild-type – approximately the same number of cell files found in the EPR at the torpedo stage (Fig. 2H).
In stm seedlings, the meristematic dome fails to form from the ECR and the cotyledon petioles are often fused (Fig. 3; Barton and Poethig, 1993; Clark et al., 1996; Endrizzi et al., 1996). The interdigitation of cells in the fused regions indicates that the fusion is due to outgrowth of the region normally responsible for separation rather than to post-genital fusion. If STM acts cell autonomously, STM function in the ECR is required for the meristematic dome to form while STM function in the EPR is required to prevent outgrowth at the junction of the cotyledon bases. Thus one simple interpretation of STM function, that it is required to promote cell division in the SAM, holds only for the ECR. The opposite is true in the EPR where STM is required to prevent cell division.
Effect of STM on apical embryonic pattern
Since the STM gene encodes a predicted homeodomain containing protein, it likely functions by regulating, either positively or negatively, the transcription of other genes. We have determined when and in what pattern three genes expressed in vegetative apices, the UFO, CLV1 and ANT genes, are expressed in the embryo during apical development. We have also determined whether their expression requires STM.
UFO expression in wild-type and stm embryos
The UFO gene is expressed in vegetative, inflorescence and floral meristems where it is found at low levels in the center of the SAM and at higher levels towards the periphery of the SAM (Lee et al., 1997). We have examined the expression of UFO in the embryo and find that it is limited to a subset of STM- expressing cells and is dynamic during embryogenesis. UFO transcripts are first found in a small group of cells at the presumptive apex at the early heart stage (Fig. 4A,B). UFO transcripts are limited to a subset of cells in the STM expression domain; they do not accumulate in the EPR (Fig. 4B). UFO expression at this early stage is limited to the presumptive L2 and L3 layers and does not extend into the L1 (or outer) layer of the embryo. UFO expression becomes more intense in the late heart/ early torpedo stage embryo where it is expressed between the cotyledons and may extend into the presumptive L1 (Fig. 4C,D). Toward the torpedo stage of embryogenesis, UFO expression resolves into a cup-shaped domain at the base of the ECR (Fig. 4E-H). In the fully developed embryo, a faint cup-shaped region of UFO expression persists around the SAM (Fig. 4G). In addition, two brightly staining regions of UFO expression are detectable at the outer boundaries of the SAM in sagittal section (Fig. 4F).
To determine the extent of overlap between the domains of expression of UFO and STM, labelling for both genes was done on the same section. In both heart and torpedo stage embryos, the domain of UFO expression falls entirely within the domain of high STM expression we call the ECR (Fig. 5). In later stages, we were not able to determine the degree of overlap of UFO and STM expression due to lower levels of UFO mRNA in older embryos.
The temporal and spatial pattern of UFO expression make it a candidate for positive regulation by STM. To test this, we assayed UFO expression in self-progeny embryos from parents of genotype stm-1/+ or stm-11/+. 29/122 mid-heart stage embryos from an stm/+ parent failed to express UFO (Fig. 4I). This is consistent with the 25% stm embryos expected in these siliques. (Siliques were cut in half, fixed and embedded. The embryos counted were from a total of 12 siliques and three separate in situ hybridization experiments. All sections of each embryo scored were examined). In wild-type controls (six separate in situ hybridization experiments in which greater than 200 embryos were examined), we never observed embryos at this stage that failed to express UFO. At later stages, when stm embryos could be detected morphologically, we find that they also fail to express UFO. We conclude that STM is required for the expression of UFO and that the STM gene product is active by the early heart stage of embryo development. Note that at the early heart stage of embryogenesis, the morphology of the stm embryo is indistinguishable from that of wild-type, so we expect that the cells normally expressing UFO in the wild-type embryo are present in the stm mutant embryo.
CLV1 expression in wild-type and stm embryos
The CLV1 gene, like the UFO gene, is expressed in vegetative, inflorescence and floral meristems where it is found in the central region in the L3 and possibly the L2 layers (Clark et al., 1997). We have examined its expression in the embryo and find that it is expressed in the L2 and L3 layers of the developing SAM beginning at the heart stage of embryogenesis (Fig. 6A,B). This is later than both UFO and STM are first detected. Early heart stage embryos, such as the one in Fig. 6A, exhibited light and variable staining. By late heart stage, expression was detected in all wild-type embryos examined (n=150; data collected from 3 separate in situ hybridization experiments). CLV1 expression, as seen in transverse section, is limited to a subset of cells in the ECR (Compare Fig. 6C with D). As embryogenesis proceeds, it remains expressed in the ECR (Fig. 6E,F,G,H).
In late heart to torpedo stage embryos from self-fertilized stm-11/+ plants (i.e. embryos between stages shown in Fig. 6B and E), 95 exhibited wild-type CLV1 expression, 19 showed severely reduced expression and expression was absent in 8. (The same result was seen in three separate in situ hybridization experiments. The data were collected from the experiment that gave the strongest signal.) Together the reduced expression and absent expression categories of embryos make up 22% of total embryos. Later in embryonic development, embryos that were morphologically distinguishable as stm mutants exhibited very low or no CLV1 expression (Fig. 6I,J,K). The embryos shown in Fig. 6J,K were chosen to illustrate the domain of CLV1 expression in stm mutants but are atypical in the high level of expression detected. We conclude that the onset of CLV1 expression during embryogenesis does not absolutely require STM and that STM is required for CLV1 expression to be maintained at wild-type levels. Note that whereas the effect of the stm mutation on UFO transcript accumulation is observed before any detectable morphological defect is seen in stm mutant embryos, CLV1 expression is affected around the time stm mutant embryos deviate morphologically from wild-type embryos.
In addition to expression in the developing SAM, two small (one or two cells) and transient patches of CLV1 expression were detected in some heart stage embryos (not shown). These patches were outside the procambium and just below the cotyledons in the medial longitudinal plane.
ANT expression in wild-type and stm embryos
The ANT gene is expressed in the developing cotyledons in the embryo and excluded from the presumptive SAM, thus showing a near opposite pattern of expression to that of STM (Elliott et al., 1996). ANT is first detected in the 32 cell stage globular embryo where it is found in a few cells in the apical portion of the embryo. We find that ANT expression forms a ring in the apical portion of the embryo by the late globular to transition stage of embryogenesis (Fig. 7A,B). Thus, the pattern of STM and ANT expression in the embryo is not exactly complementary since the domains overlap in the EPR. The ring-shaped pattern of ANT expression in the late globular to transition stage embryo is consistent with the expression of this gene in presumptive organs since the cotyledons wrap around and encircle the apical portion of the embryo as shown in Fig. 3. ANT expression continues in the heart stage embryo (Fig. 7C,D). As the cotyledons grow out, the region of ANT expression becomes progressively limited to a plane that separates the cotyledons into near equal ad- and abaxial regions (Elliot et al., 1996 and our observations).
ANT is excluded from the ECR. If STM is required to negatively regulate ANT in the ECR, we would expect to see the region of ANT expression expandinto the ECR in stm mutant embryos. However, in 133/133 self-progeny embryos from parents of genotype stm-11/+, we detected no changes in the expression pattern of ANT and conclude that negative regulation of ANT in the ECR is not dependent on STM activity. In addition, since STM and ANT transcripts coexist in the EPR, the presence of STM transcript alone is not sufficient to negatively regulate ANT expression.
DISCUSSION
Establishing molecular domains in the apical portion of the early embryo
Our results show that pattern develops stepwise in the apical portion of the Arabidopsis embryo (Fig. 8). The ANT gene is expressed in a peripheral ring and is excluded from the center of the late globular/early transition stage embryo. The pattern of ANT expression, including its exclusion from the central region of the embryo, does not require STM. STM expression, first detected slightly later than ANT expression, is expressed in a stripe across the apical half of the transition/early heart stage embryo and divides the apical portion of the embryo into two bilaterally symmetrical halves. UFO transcripts appear in the presumptive SAM at the early heart stage of development and require STM for their accumulation. By the torpedo stage of embryogenesis, UFO expression resolves into a cup-shaped domain within the SAM. CLV1 transcripts appear first at the heart stage in a small number of cells in the center of the SAM and do not require STM for their accumulation.
Our experiments suggest a model for embryonic apical development in which presumptive cotyledon fate is specified in a ring-shaped domain in the late globular embryo. This is followed by the development of bilateral symmetry, one aspect of which is the specification of the presumptive SAM. As the presumptive SAM develops, it becomes progressively divided into additional domains.
Four observations indicate that radial information (i.e. signals that distinguish central regions of the embryo from peripheral regions) is present in the apical portion of the globular embryo and does not require STM function. First, the ‘ring’ pattern of ANT expression, including the exclusion of its transcript from the ECR, is not dependent on wild-type STM function; it is found prior to the first detectable STM expression and is not altered in stm mutant embryos. Second, the STM gene itself is expressed in a temporal pattern in the radial dimension: STM is present first in the EPR and only subsequently in the ECR. Third, in the absence of STM function, the EPR behaves differently from the ECR — cells in the ECR appear to differentiate without dividing while cells in the outer region divide extensively, ultimately developing as cotyledon petiole. Fourth, the expression of CLV1 in the centralmost region of the developing SAM does not require STM.
In addition to exhibiting different cell fates along the radial axis, Arabidopsis embryos develop with bilateral symmetry. The STM expression pattern is an early molecular indication of bilateral symmetry in the embryo. There are several cases in which a lack of bilateral symmetry is correlated with a failure to form a SAM. For instance, cup-shaped cotyledon mutants (Aida et al., 1997) and topless mutants (Long and Barton, unpublished) in Arabidopsis frequently lack bilateral symmetry. These mutants fail to form a SAM, and cotyledon tissue develops in the apical region of the embryo in a radially symmetrical pattern. The absence of normal bilateral symmetry may result in an inability to activate the STM gene, thus explaining the failure to develop a SAM.
The STM-dependent accumulation of UFO in the transition/early heart stage embryo indicates that STM functions by this stage. It also suggests that UFO is a direct target of STM activation. Since the expression of UFO and STM do not correspond in a one-for-one manner, additional factors must act in combination with STM to direct UFO expression. UFO is initially expressed in a block of cells in the ECR but, as the ECR grows, becomes limited to a cup-shaped region just inside the limits of the ECR. One possibility is that proximity to the EPR, in combination with STM, is required for UFO expression.
The function of UFO in the embryonic and vegetative SAMs, if any, is unknown. Several functions of a gene expressed at the boundary between primordium and meristem can be imagined. Boundaries of gene expression in other systems can act as sources of morphogens (Cohen, 1996). In addition, the boundary between meristem and leaf is likely to be a site at which cell division is carefully controlled. Support for the latter idea comes form the observation that the Antirrhinum FIMBRIATA gene product (FIM is the snapdragon ortholog to UFO) interacts with proteins implicated in cell cycle control (Ingram et al., 1997).
It has been hypothesized that the apical portion of the globular embryo represents the first incarnation of the SAM (Kaplan and Cooke, 1997). This hypothesis is based chiefly on the chain of logic that since cotyledons and leaves are very similar to each other and since leaves are derived from a SAM, cotyledons must also be derived from a SAM. Several findings presented here and elsewhere (Long et al., 1996) are not consistent with this interpretation. First, cotyledons and true leaves differ in their origin; the majority of the cells from which the cotyledons are derived have never expressed STM, while true leaves derive entirely from a lineage of STM-expressing cells. Thus, the embryonic cells that give rise to the bulk of the cotyledons never possess the stem cell-like character of giving rise to both cells like themselves and cells destined to differentiate. Second, several characteristics of the vegetative SAM, both molecular and morphological, are not present in the globular stage embryo. These include the tunica/corpus organization of the SAM and genes expressed in the SAM such as UFO, STM and CLV1. Instead, these characteristics develop gradually during embryogenesis and this program of development requires STM for at least some aspects.
The role of STM in embryonic SAM function
Different authors have emphasized various, in some cases overlapping, roles for the STM gene in SAM function. For instance, Barton and Poethig (1993) concluded that STM was required for the initiation of the SAM during embryogenesis because in its absence, no population of stem cells was formed and cells within the presumptive SAM failed to divide and organize into a tunica/corpus organization. The organization of cells into tunica and corpus is a hallmark of angiosperm SAMs and was the only marker of SAM fate available at the time. Meyerowitz (1997) has focused on the requirement of STM for cell division in the SAM, since in its absence fewer cell divisions take place. Chuck and Hake (1996) have stressed the requirement for STM in preventing differentiation of cells within the SAM. This is similar to the interpretation by Endrizzi et al. (1996), that STM is required for the maintenance of an undifferentiated population of cells within the SAM.
Our observations suggest that, even in the relatively simple developing embryo, STM functions in a complex manner that can be described adequately only by combining the above concepts. We propose a model in which STM acts differently in the embryonic peripheral region (EPR) and in the embryonic central region (ECR). Our model posits that in the EPR, STM is required to inhibit organ outgrowth and subsequent differentiation but not organ specification, whereas in the ECR, STM is required to inhibit differentiation and to initiate a SAM-specific program of development.
The model is based on the observation that the EPR and the ECR differ in several ways. They differ in the fates of their cells both in the presence and the absence of STM, the time of onset and the persistence of STM expression, and in the distribution of the UFO, ANT and CLV1 transcripts. Let us first consider the EPR. Cells in the EPR are destined to form the separations between the cotyledons. To do so, cells in this region must be restricted in their outgrowth relative to the adjacent regions. In the absence of STM function, cells in the EPR form a region of fused cotyledon petiole instead of the normal separation between cotyledons. The morphology of the fused petiole region in stm mutants, with its extensively interdigitated cells, is most consistent with the fusion forming by inappropriate division and outgrowth of cells of the EPR. Thus, STM is required to inhibit the outgrowth of these cells but not to prevent their specification as part of the cotyledons since ANT expression, a marker for organ primordia, in the EPR overlaps with STM. Since cells in the EPR ultimately differentiate, STM is required to delay their differentiation rather than prevent it indefinitely. Consistent with this, the expression of STM in the EPR is transient – decreasing during embryogenesis and ultimately disappearing completely in the germinated seedling (data not shown).
This model requires that the instructions as to what portion of the cotyledon the EPR cells will form, change over time. Thus, in the complete absence of STM activity, cells in the EPR grow out prematurely, and develop as cotyledon petiole. In the wild-type, cells in the EPR do not differentiate until STM levels are below some critical level, presumably past the heart stage of development. At this point, the instructions for cotyledon development have changed so that only a more limited basal region of the cotyledon petiole develops. (Note: this is similar to the maturation schedule hypothesis for maize leaf development (Muehlbauer et al., 1997).)
In the ECR, STM is required to prevent differentiation. In the absence of STM function, cells in the ECR differentiate instead of forming a small mound of cells as they do in the wild-type. This mound of cells includes the ‘stem cells’ that are responsible for the growth and development of the shoot system. As opposed to cells in the EPR, these cells persistently express STM. We propose that, in addition to preventing differentiation in the ECR, STM is required for the implementation of a SAM-specific program of development in the ECR through the activation of a SAM-specific battery of genes. UFO is an example of a SAM-specific gene whose expression depends on STM.
Not all gene expression in the developing embryonic SAM requires STM; the CLV1 transcript appears in stm mutant embryo apices. However, maintenance of CLV1 expression at normal levels requires STM. With regard to embryonic SAM formation, stm is epistatic to clv1 (Clark et al., 1996). This observation is consistent with two models of gene interaction. In a regulatory model, CLV1 could negatively regulate STM; in the absence of STM function, the presence or absence of its negative regulator, CLV1, is indistinguishable and the stm phenotype is observed. In an assembly model, STM acts to promote the development of a tissue, in this case an apical meristem (or some intermediate stage in apical meristem formation). The CLV1 gene then acts to control the size of this tissue, the apical meristem. In the absence of the tissue, the presence or absence of CLV1 is indistinguishable and the stm phenotype is observed. Our results are consistent with both models and in addition indicate that STM is a positive regulator of CLV1.
What might explain the difference in action of STM in the EPR versus the ECR? The relative levels of STM protein may differ in the two regions with lower levels dictating EPR activity while higher levels dictate ECR activity. Alternatively, STM may act in combination with other factors whose presence or absence differs in the EPR and the ECR.
Genes such as PINHEAD/ZWILLE (McConnell and Barton, 1995; Moussian et al., 1998) and WUSCHEL (Laux et al., 1996) which are required for the formation of a functional SAM during embryogenesis may be specifically required for ECR development. Mutants in these genes show normal cotyledon separation and premature SAM termination.
A model for dual STM functions helps to explain the two phenotypes seen in plants that over- or ectopically express other KNOTTED-like genes. In some cases, overexpression results in the formation of ectopic SAMs on the leaf surface (Sinha et al., 1993; Chuck and Hake, 1996). This is consistent with the conclusion that KNOTTED-like genes have the ability to activate a SAM-specific program of development, analogous to the role proposed for STM in the ECR. In other cases, ectopic expression of KNOTTED-like genes causes increased lobing or division of leaves (Lincoln et al., 1994; Chuck et al., 1996; Hareven et al., 1996). Lobing can be explained by KNOTTED-like genes preventing the growth and differentiation of cells specified as leaf. This is analogous to the role proposed for STM in the EPR. Just as in the embryo, the difference in phenotypes between plants that overexpress KNOTTED-like genes could result from differences in the level or the context of gene expression.
The model for dual action of STM may extend to the peripheral and central zones of postembryonic meristems. The localization of STM in the EPR and the fate of cells in this region parallel that in the PZ of the vegetative SAM. STM is expressed between developing organ primordia in the vegetative and floral meristems (Long et al., 1996) and is required to prevent organ fusion in flowers (Felix et al., 1996). Likewise, STM is expressed in the central zone of vegetative and floral meristems (Long et al., 1996) and is required to prevent premature meristem termination (Clark et al., 1996; Endrizzi et al., 1996).
We would like to thank Jane McConnell for help in making the plasmid containing the ANT sequences. We also thank Detlef Weigel and Ilha Lee for sharing data prior to publication and Heidi Barnhill for expert help with Scanning Electron Microscopy. We thank Matthew Evans, Anita Fernandez, Rebecca Joy, Karyn Lynn and Scott Woody for helpful comments on the manuscript and members of the Barton lab for helpful discussions along the way. This work was funded by the NSF. J. L. received support from the USDA/DOE/NSF Cooperatively funded Arabidopsis Training Grant. This is manuscript number 3510 from the Laboratory of Genetics.