Stem-Cell Homeostasis and Growth Dynamics Can Be Uncoupled in the Arabidopsis Shoot Apex

See allHide authors and affiliations

Science  28 Oct 2005:
Vol. 310, Issue 5748, pp. 663-667
DOI: 10.1126/science.1116261


The shoot apical meristem (SAM) is a collection of stem cells that resides at the tip of each shoot and provides the cells of the shoot. It is divided into functional regions. The central zone (CZ) at the tip of the meristem is the domain of expression of the CLAVATA3 (CLV3) gene, encoding a putative ligand for a transmembrane receptor kinase, CLAVATA1, active in cells of the rib meristem (RM), located just below the CZ. We show here that CLV3 restricts its own domain of expression (the CZ) by preventing differentiation of peripheral zone cells (PZ), which surround the CZ, into CZ cells and restricts overall SAM size by a separate, long-range effect on cell division rate.

Pattern formation in the SAM is a dynamic process that results from active orchestration of spatial and temporal patterns of gene expression and of cellular behavior, by cell-cell communication. In Arabidopsis thaliana, the SAM consists of several hundred cells divided into functional domains that are characterized by different cellular behaviors and by different patterns of gene expression (1, 2). Progeny of CZ cells enter into differentiation pathways when they enter the surrounding meristematic regions: the flanking PZ, where leaf and flower primordia are formed, and the RM beneath the CZ, where cells of the stem form. The functional domains of the SAM, CZ, PZ, and RM are established in embryonic development and maintain their relative proportions throughout postembryonic life, even though cells are continually diverted to differentiation pathways. The cells of the CZ signal to the RM by producing the product of the CLV3 gene, a small extracellular protein thought to be the ligand for the CLAVATA1 receptor kinase, expressed in some RM cells (3, 4). CLV1 acts, at least in part, by down-regulating the activity of the homeodomain protein WUSCHEL (WUS), also expressed in some RM cells (5, 6). WUS acts, in turn, to up-regulate CLV3 expression in the overlying CZ by means of an unknown diffusible signal (7). CLV3 thus regulates its own expression through a feedback network involving CLV3-expressing cells in the CZ and the cells of the RM (8). Loss-of-function clv3 (or clv1) mutants (9, 10) orgain-of-function wus mutants (7) have a greatly enlarged SAM, which has an enormous CZ and an enlarged RM as well. Nevertheless, how the mutant phenotype results from alteration of the communication between CZ and RM is not known. It has been speculated that the enlarged meristem seen in clv loss-of-function mutants results from increased rates of cell division in the meristem, but also that it results from a decreased rate of departure of cells from the meristem to form differentiated tissues (911). There are at least three possibilities for the mechanism of enlargement of the CZ in clv mutants. The cells of the CZ could divide more rapidly in the mutants, while the transition of CZ daughters into PZ pattern of gene expression proceeds unperturbed; the cells on the boundary of the CZ and PZ could remain CZ cells after division, rather than transitioning to a PZ pattern of gene expression when pushed out of the CZ; or cells of the PZ could be respecified to become CZ cells. Which of these possibilities is correct is not known, nor is the relation between control of CZ size and control of meristem size in clv mutants.

Ordinary analysis of the mutants gives clues to the network of interactions but is not adequate to define the function of individual genes in the network of interacting cells. This is because loss or gain of function of one component affects the entire network, so that the ultimate mutant phenotype, assessed long after the initial effects of the mutation, is the result of a series of events that affects the expression of many genes. To understand the function of the CLV3 gene, we have developed new methods for turning it off in a living, wild-type meristem and then for following in real time the changes in the organization of the SAM and in the rates and patterns of cell division in it. By doing this, we can see the immediate effect of loss of CLV3 function, rather than the eventual effect, and have started to answer the questions of how loss of CLV3 function leads to an enlarged CZ and to an enlarged meristem.

In order to test the immediate effects of perturbation in CLV signaling, transgenic plants capable of conditional CLV3 silencing were generated. The gene construct introduced has a transcribed region that is capable of silencing CLV3 by double-stranded RNA interference (dsRNAi); this coding region, which makes a foldback CLV3 RNA, has already been described (12). Conditional induction of the foldback RNA was achieved by use of a two-component transcription activation system with a glucocorticoid-activated artificial transcription factor, GR-LhG4, which is capable of binding to and activating genes adjacent to a multimerized LhG4 binding sequence, 6xOP (13). Plants transgenic for the GR-LhG4 gene attached to a strong constitutive promoter (the cauliflower mosaic virus 35S coat protein gene promoter), p35S::GR-LhG4-N, were crossed to plants transformed with a p6xOPΩ::CLV3dsRNAi construct, and doubly transgenic F1 plants were selected. These F1 plants were treated with 10 μM dexamethasone (Dex) from the time of germination until flowering and then were analyzed for SAM defects. Four out of 35 independent RNAi lines examined showed the clv3 mutant phenotype, massive overgrowth of the SAM (Fig. 1, A and B). Progeny of these lines inherited the ability to suppress CLV3 activity and were used in further analyses. Inducible silencing of the CLV3 gene was confirmed using reverse transcription polymerase chain reaction (RT-PCR) to show loss of CLV3 transcripts after Dex treatment (Fig. 1K).

Fig. 1.

Dex-inducible clv3 phenotypes. (A) SAM of 30-day-old wild-type plant. (B) SAM of 30-day-old plant harboring constructs capable of silencing endogenous CLV3 treated with Dex starting from germination. (C) A three-dimensional (3D) reconstructed view of L1 layer of 18-day-old mock-treated SAMs labeled with FM4-64 (red) and pCLV3::mGFP5-ER (green) marking the CZ. Primordia at different stages of development are marked. (D) A 3D reconstructed view of the L1 layer of 18-day-old Dex-treated plant harboring constructs capable of inducing CLV3 silencing, marked with FM4-64 (red) and pCLV3::mGFP5-ER (green). Note both the expansion of the CZ and increased SAM size. Primordial outgrowths are marked as P. (E) The reconstructed side view of (C), depicting pCLV3::mGFP5-ER (green) expression in different clonal layers (arrows) and the CZ and the PZ (arrow). (F) The reconstructed side view of (D) depicting expanded CZ (green) and the PZ (arrow). (G) and (H) Longitudinal sections showing the GFP RNA expression domain of pCLV3::mGFP5-ER in SAMs either mock-treated (G) or treated with Dex (H) for a period of 7 days after bolting. (I) and (J) Longitudinal sections showing the WUS RNA expression domain in SAMs mock-treated or exposed to Dex, respectively, for a period of 7 days after bolting. (K) Ethidium bromide–stained agarose gel showing RT-PCR products amplified with CLV3-specific (arrow) and LIPASE-specific (arrowhead) primers from mock-treated SAMs or those treated with Dex for 6 days. Scale bar in (C) and (D), 20 μM.

P35S::GR-LhG4-N; p6xOPΩ::CLV3dsRNAi plants were crossed to plants transgenic for a CZ reporter construct, pCLV3::mGFP5-ER, which express a green fluorescent protein (GFP) from the CLV3 promoter, localized to the endoplasmic reticulum (ER). In untreated or mock-treated plants, GFP-ER fluorescence reflects the expected CZ domain of expression, with fluorescence restricted to the central tip of the SAM, extending from the epidermal to the fourth layer of cells (Fig. 1, C and E). Dex-treated plants, for a period of 16 to 18 days starting from germination, displayed an expanded fluorescence domain (Fig. 1, D and F), concordant with the expanded domain of CLV3 expression observed by in situ hybridization in clv mutants (3). RNA in situ hybridization revealed an expanded domain of GFP RNA, which ruled out the possibility that the expanded fluorescence domain is due to GFP protein movement or to protein stability, which would cause the GFP to persist through cell divisions (Fig. 1, G and H). Plants that were Dex treated for a period of 7 days also displayed an expanded domain of WUS expression (Fig. 1, I and J), and the WUS expression domain appeared patchy, as it consisted of cells with variable levels of expression (fig. S2, A to D).

A fourth transgene was added to the lines to enable the detection of all of the plasma membranes in the SAM and, therefore, all of the cell expansions and divisions that occur during observation. This was accomplished by crossing a triply transgenic plant to a plant bearing a p35S::YFP 29-1 transgene, which expresses a plasma membrane–localized yellow fluorescent protein (YFP) in all of the cells of the SAM (14). Inflorescence meristems of plants carrying all four constructs were treated with either Dex or a control solution and then imaged at 12- or 24-hour intervals in different experiments. Within 24 hours of Dex treatment, a moderate increase in pCLV3 promoter activity was observed within the native domain, and its expression intensified by 48 hours, coupled with the radial expansion of the pCLV3::mGFP5-ER expression domain (Fig. 2, I to K). In the following 24-hour window, the expansion of pCLV3 promoter activity continued, but with a dramatic increase in SAM size, both in height and width (n = 10) (Fig. 2, L and R). Expansion of the CZ and the increase in SAM size continued for the next 2 days, the duration for which live imaging was performed (Fig. 2, M, N, S, and T). Mock-treated plants (n = 10) and Dex-treated plants lacking the CLV3dsRNAi construct (n = 8) served as controls and did not show consistent changes in SAM size, or in the CLV3 expression domain, with time (Fig. 2, A to H; fig. S1, A to D). During the course of the expansion of the CLV3 expression domain and the increase in SAM size, the process of flower primordium formation continued unperturbed as monitored by the appearance of primordial outgrowths (Fig. 2, I to N), in comparison with mock-treated SAMs (Fig. 2, A to D). Primordial initiation always occurred outside the perimeter of the expanding CLV3 expression domain (and therefore in the PZ), which demonstrated the functional expansion of the CZ (Fig. 2, L to N). The expansion of the CZ was detected earlier than the increase in overall SAM size, whereas the process of organ primordium initiation continued unchanged, which showed that meristem reorganization precedes excess meristem growth and that increased SAM size is not merely due to a reduction in the rate of PZ cells entering differentiation pathways.

Fig. 2.

Increase in SAM and the CZ is not due to a reduced rate of differentiation. (A to D) and (I to N) are the time series of mock- and Dex-treated SAMs, respectively. Total time elapsed, post treatment, is indicated on individual panels. The 3D views of the L1 layer are labeled with 35S::YFP 29-1 (plasma membrane localized YFP, red) and p CLV3:: mGFP5-ER (green). Primordia at different stages are marked and the individual primordia are color-coded to track the same primordia over time. (E to H) The 3D side views of the SAMs in (A to D), respectively. (O to T) The 3D side views of the SAMs in (I to N), respectively. (I to N) Note gradual expansion of the pCLV3::mGFP5-ER domain and increase in SAM size upon Dex treatment, while primordial morphogenesis continues unperturbed. Increase in SAM height is evident (O to T). Also note that the radial expansion of the CZ precedes the increase in SAM size (compare I to K with L to N). Arrows point to the CZ. Scale bar, 30 μM.

We next tested which of the three hypotheses for the expansion of the CZ after reduction of CLV3 activity is correct—increased rate of cell division in the CZ, decreased transition of CZ cells to PZ cells at the CZ boundary, or switching of PZ cells to a CZ pattern of gene expression—the opposite of the normal differentiation pathway. We did this by monitoring the expansion of the CZ and, meanwhile, using the plasma membrane marker to follow cell division in CZ and PZ cells (Fig. 3, A to C and G to I). Images were acquired about every 12 hours for a 48-hour period after Dex treatment and reconstructed in three dimensions to allow the same sets of cells to be tracked over time. Within 24 hours of Dex treatment, preexisting CZ cells began to express the pCLV3 marker at an elevated level, and at the same time, some preexisting PZ cells that had not divided after Dex treatment began to show pCLV3 activity (Fig. 3, G and H, arrows point to the same cells at different times). The radial expansion of the CLV3 domain was readily visible by 40 hours of treatment, with larger numbers of former PZ cells starting to express pCLV3 (Fig. 3, G and I). Cell divisions were monitored by the appearance of new cross walls and were not correlated with expansion of the CZ. Similar expansion of the CZ was also noticeable in layers below the epidermal, L1, layer (Fig. 3, J to L). Mock-treated plants showed subtle and slow changes in the pCLV3 expression levels and domain, with fluctuations in the level of fluorescence in individual CZ cells located at the periphery (fig. S1, A to D, arrows) and with a limited expansion of the pCLV3 domain (Fig. 3, A and B, arrows). However, they did not show a consistent radial expansion as seen in Dex-treated SAMs. This analysis revealed that the gradual expansion of the CZ after reduction in CLV3 activity results from respecification of PZ cells to CZ fate in a way that is uncoupled from growth, as assayed by CLV3 expression. The fluctuations of pCLV3 expression in mock-treated plants may indicate that PZ cell respecification normally occurs at the edge of the CZ and that the size of the CZ in wild type is regulated by maintenance of a dynamic balance between CZ and PZ cells.

Fig. 3.

Expansion of the CZ is due to the respecification of PZ cell identity. (A to C) and (G to I) Reconstructed views of the L1 layer of SAMs either mock-treated or treated with Dex, respectively, and followed for a period of 48 hours. Total time elapsed, post treatment, is indicated. The CZ behavior (pCLV3::mGFP5-ER, green) is followed, along with a ubiquitously expressed cell membrane marker (35S::YFP29-1, red). (A to C) Note only subtle changes to the CZ with time in mock-treated SAM (arrows point to the same cells in images acquired at successive intervals). (G to I) Dex-treated SAM showing both increased expression of pCLV3 expression within the native domain and also a radial expansion over time (arrows point to the same cells in images acquired at successive intervals). Note the expansion of CZ identity in the first 40 hours without a noticeable effect on the SAM size. (D to F) and (J to L) Side views of SAMs represented in (A to C) and (G to I), respectively, and the arrows in each panel indicate outer limits of the CZ. Scale bar, 20 μM.

To better understand the expansion of the SAM that follows CZ reorganization, we analyzed cell division patterns in the period between 60 and 108 hours after Dex treatment. During this period, maximal SAM growth and continued expansion of the CZ occurs (Fig. 4, F to J). Lineage reconstruction was performed for all of the cells of the outermost layer of the SAM, the L1 layer, of Dex-treated plants (n = 3), based on images acquired at 12-hour intervals. Mock-treated plants (n = 3) and Dex-treated plants that did not carry the RNAi construct (n = 2) served as controls. As these two control genotypes showed no significant differences, the control data were pooled. Cell division activity was expressed as the size of individual lineages [the number of progeny cells at 108 hours that descended from a single cell present at 60 hours (Fig. 4M)]. Cells of Dex-treated plants had larger lineages compared with the controls, which demonstrated an increase in cell division after compromise of CLV3 function. The lineage data were represented by superimposing each lineage on a reconstruction of the L1 at the 108-hour time point (Fig. 4, K and L). This revealed that the increase in cell division rate is observed distant from the meristematic center (Fig. 4L) and, therefore, represents a long-distance effect of events that started in the CZ.

Fig. 4.

A long-range effect of CLV3 activity on cell division rates. (A to E) and (F to J) Reconstructed views of the L1 layer of the SAMs mock-treated or treated with Dex, respectively, starting 60 hours after treatment. Total time elapsed, post treatment, is shown. The CZ behavior (pCLV3::mGFP5-ER, green) is followed with a cell division marker (35S::YFP29-1, red). (F to J) Expansion of the CZ continues as more of the PZ cells begin to acquire CZ identity (arrows point to the same cells in images acquired at successive intervals) along with an increase in SAM size. (M) The images from the same SAM were used to score for cell division events to reconstruct lineages within the 48-hour period (60 to 108 hours after treatment) and expressed as size of the individual lineages against mean number of lineages per SAM. Error bars represent standard deviation. (K) and (L) The spatial distribution of lineages represented in (M) in control and Dex-treated SAMs, respectively; CZ size is shown in green. The individual lineages are color-coded. The same color has been assigned to more than one lineage as long as they do not abut each other. White arrows in (K) point to three-celled lineages. White arrows in (L) point to five-celled lineages and the yellow arrows indicate four-celled lineages. Scale bar for (A to J), 20 μM.

This study provides evidence that CLV3 signaling in meristems mediates both cell fate specification and growth control through inhibition of cell division rate, as well as that the processes can be temporally uncoupled. The gradual radial expansion of the CZ that follows reduction in CLV3 activity indicates a process of PZ cell respecification that either spreads from the CZ with time or reflects a gradient of response to a CZ-promoting activity. If the signal does spread with time, it could do so either by diffusion or by stepwise communication between adjoining cells. The presently proposed function of CLV3 is to confine the expression of the transcription factor WUS to a limited set of cells as a result of CLV1 activation, and this is consistent with the WUS in situ hybridization results. A real-time analysis of WUS levels and the WUS-expression domain in the reorganizing meristems of CLV3-RNAi plants should yield further insights into the relation between cell-cell communication in the SAM and its organization and growth, as should manipulations of WUS activity during live imaging. It is not yet known whether the long-range effect of CLV3 in repressing cell division is mediated through its effect on WUS activity. Earlier studies have proposed that WUS is required for CZ specification and that SHOOTMERISTEMLESS (STM), a homeodomain protein, is required for division of PZ cells (15). Such a model would predict that CLV-mediated growth inhibition should impinge on STM and its regulatory cascades. It will be illuminating to test the function of STM and its regulation by the CLV-WUS network in real-time experiments such as those introduced here. However, it is also possible that the increased cell division rates that follow CLV3 removal could be an indirect consequence of the influence exerted by the expanding CZ on adjacent PZ cells. Meristems with alterations in WUS and STM activity, like the experiments related here in which CLV3 activity has been manipulated, should allow the dissection of the influence of signaling between cells on cell-fate specification, organization, and growth in SAMs, despite their being highly coupled in space and in time in wild-type plants.

Supporting Online Material

Materials and Methods

Figs. S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article