Threshold-dependent transcriptional discrimination underlies stem cell homeostasis

Significance

Various mechanisms have been proposed to explain dose-dependent transcriptional regulation mediated by morphogen gradients in animal development. In plant development, on the other hand, transcriptional mechanisms that underlie dose-dependent modulation of gene expression have not been discovered, despite the well-documented importance of positional signals in cell-fate specification. Here we show that the stem cell-promoting transcription factor WUSCHEL (WUS) regulates transcription in a concentration-dependent manner, activating transcription at a lower level and repressing transcription at a higher level, thus leading to the transcriptional control of its own negative regulator. Our work also shows that WUS binds a group of tightly clustered cis elements, each with different affinities; this binding suggests a buffering mechanism in maintaining a stable CLAVATA3 (CLV3) expression.

Abstract

Transcriptional mechanisms that underlie the dose-dependent regulation of gene expression in animal development have been studied extensively. However, the mechanisms of dose-dependent transcriptional regulation in plant development have not been understood. In Arabidopsis shoot apical meristems, WUSCHEL (WUS), a stem cell-promoting transcription factor, accumulates at a higher level in the rib meristem and at a lower level in the central zone where it activates its own negative regulator, CLAVATA3 (CLV3). How WUS regulates CLV3 levels has not been understood. Here we show that WUS binds a group of cis-elements, cis- regulatory module, in the CLV3-regulatory region, with different affinities and conformations, consisting of monomers at lower concentration and as dimers at a higher level. By deleting cis elements, manipulating the WUS-binding affinity and the homodimerization threshold of cis elements, and manipulating WUS levels, we show that the same cis elements mediate both the activation and repression of CLV3 at lower and higher WUS levels, respectively. The concentration-dependent transcriptional discrimination provides a mechanistic framework to explain the regulation of CLV3 levels that is critical for stem cell homeostasis.

Various mechanisms have been proposed to explain dose-dependent transcriptional regulation mediated by morphogen gradients in animal development (1). Transcriptional mechanisms that underlie dose-dependent modulation of gene expression in plant development have not been discovered, but cell-fate specification is known to rely heavily on positional cues (2). Shoot apical meristems (SAMs) contain a set of pluripotent stem cells in the central zone (CZ); the stem cell descendants that are displaced into the adjacent peripheral zone (PZ) differentiate as lateral organs (3), and the stem cell descendants that are displaced into the rib meristem (RM), located beneath the CZ, differentiate and become part of the stem (3). In Arabidopsis SAMs, WUSCHEL (WUS), a homeodomain transcription factor that is expressed in the RM, has been shown to provide cues for stem cell maintenance (4, 5). WUS synthesized in few cells of the RM has been shown to migrate into adjacent cells where it accumulates at a lower level than in the RM (6). WUS has been shown to activate CLAVATA3 (CLV3) expression (7). CLV3 is a secreted peptide that activates a receptor kinase pathway to restrict WUS transcription (8–11). Earlier studies have shown that a precise regulation of CLV3 level is critical to control WUS transcription (12, 13). However, the mechanisms underlying transcriptional regulation of CLV3 have remained elusive.

WUS has been shown to bind the regulatory regions of type A ARABIDOPSIS RESPONSE REGULATORS (ARRs), negative regulators of cytokinin signaling, and repress their transcription (14). The function of at least two type A ARRs, ARR7 and ARR15, have been shown to be critical for activating CLV3 expression (15). Because the type A ARRs lack the DNA-binding domain, they might play an indirect role in CLV3 activation. EMSAs have shown that WUS binds the regulatory sequences in CLV3, and transient transfection studies in Arabidopsis leaf protoplasts have shown that these regulatory sequences activate CLV3 transcription (6). However, further studies are required to understand the in vivo significance of WUS binding in transcriptional regulation of CLV3. It also is not known how CLV3 is transcribed in only a few cells of the CZ despite the presence of the WUS protein in a much broader spatial domain (6). WUS has been shown to interact with HAIRY MERISTEM (HAM) family proteins HAM1 and HAM2, which are expressed in the RM and lateral edges of the PZ (16, 17). Perhaps HAM proteins may provide a spatial context for expression of CLV3 in the CZ.

Several studies have shown that WUS also represses the transcription of many genes (14, 18, 19), including several differentiation-promoting transcription factors, thus preventing premature differentiation of stem cell descendants (19). The transient transcriptional assays in Arabidopsis leaf protoplasts have shown that WUS can function as both an activator and a repressor (20). WUS has been shown to bind the transcriptional corepressor proteins TOPLESS (TPL) and -TOPLESS RELATED (TPR) (21, 22). The TPL and TPR proteins also have been shown to interact with HISTONE DEACETYLASE19 (HDA19) to form a transcriptional repression complex (23). These studies suggest that WUS may exist in different protein complexes that allow it to function both as a repressor and as an activator; however, the association of WUS with the components of a transcriptional activation complex has not been established. Furthermore, WUS has been shown to bind cis elements containing TAAT core sequences in the regulatory regions of activated target genes CLV3 (6) and AGAMOUS (24) as well as in the regulatory regions of repressed target genes (14, 18, 19). However, the mechanisms through which WUS can discriminate between transcriptional activation and repression by using similar cis elements are not understood. Here we used in vivo assays and biochemical analysis to investigate WUS-mediated regulation of CLV3 transcription. Our analysis shows that WUS functions as an activator at lower concentrations and as a repressor at higher concentrations. Our finding that WUS binds the same cis elements as a monomer at lower concentrations and as a dimer at higher concentrations suggests that WUS-mediated transcriptional discrimination may involve changes in the binding behavior of WUS to DNA. Moreover, we find that WUS binds a collection of closely spaced cis elements with different affinities that may buffer variations in WUS levels to maintain CLV3 transcription. Finally, we discuss the significance of concentration-dependent transcriptional regulation in the context of stem cell homeostasis and the mechanisms of transcriptional regulation.

Results

A WUS-Binding cis-Regulatory Module Regulates CLV3 Expression and Stem Cell Homeostasis.

Our efforts to understand the regulation of the transcriptional control of CLV3 led to the identification of six WUS-binding cis elements. Five of these elements, 950, 970, 997, 1007, and 1060 (the numbers represent the position of cis elements measured in base pairs from the start codon) (Fig. 1A and Fig. S1 A–D) were clustered within 110 bp in the 3′ region, collectively referred to as the “cis-regulatory module” (CRM). To understand the in vivo significance of the CRM, we generated a deletion by removing the CRM in the CLV3 regulatory region (hereafter referred to as “pCLV3”) driving the Histone2B-modified YFP (H2B-mYFP) reporter (25). The mutations in the WUS-binding cis element −1080 located upstream of the start codon identified in an earlier study (6) did not influence pCLV3 expression (Fig. S1E). The deletion of the CRM led to low levels of expression of the reporter in the L1 and the L2 cell layers and the RM, showing that the WUS-binding CRM is required for activating CLV3 (Fig. 1B). The mutant pCLV3-ΔCRM promoter lacking the CRM failed to complement fully the overproliferated clv3-2 mutant SAMs and the overproliferated floral meristems (FMs), as revealed by the presence of a higher number of carpels (Fig. 1 C and D and Fig. S2). These results show that the WUS-binding CRM is required for maintaining the CLV3 level needed for SAM and FM maintenance.

Fig. 1.

A CRM regulates CLV3 expression. (A and B) 3D-reconstructed top and side views of inflorescence meristems showing wild-type pCLV3::H2B-mYFP (A) and deletion of the CRM (nucleotides 943–1067), also referred to as “pCLV3-ΔCRM” (B). H2B-mYFP expression is shown in yellow; FM4-64 labeling is shown in red. The white arrows indicate different cell layers. (Scale bars: 10 μm.) (C and D) The phenotypic complementation analysis with the CLV3 genomic region containing the mutated WUS-binding cis elements. The inflorescence meristem height (C) and the number of carpels (D) in clv3-2 plants transformed with the wild-type pCLV3 and various CLV3 mutant promoters: pCLV3-TM (970 and 997 double mutant); pCLV3-SM (950, 970, 997, 1007, and 1060 quintuple mutant); pCLV3-ΔCRM (nucleotides 943–1067 deletion); and the two higher-affinity mutants pCLV3-970-M1 and pCLV3-970-M4 expressing the CLV3 genomic region. In all CLV3 promoters, the upstream −1080 cis element is mutated. The error bars represent SE. Different letters indicate statistical differences between cis lines (P < 0.001) as determined by Tukey’s Honest Significant Difference (HSD) tests.

Fig. S1.

Characterization of the WUS-binding cis elements in CLV3. (A) Sequence comparison of the WUS-binding cis elements of pCLV3 (−1080, 1060, 1007, 997, 970, and 950); TAAT cores are underlined. (B) Sequences of mutated TAAT cores in WUS-binding cis elements of pCLV3 are underlined and highlighted in gray. (C) EMSA showing WUS binding to the wild-type and mutated cis elements 950, 970, 997, 1060, and 1007. (D) A comparison of WUS binding to the oligos containing a 2-bp mutation or a 4-bp mutation in the TAAT core of the 1007 cis element. Black arrowheads indicate the monomer WUS–DNA complex. (E) 3D top views of SAMs showing the pCLV3(−1080M)::H2B-mYFP, the 300-bp deletion pCLV3(Δ-1200 to -900)::H2B-mYFP, and the double mutant (−1080 and 997) pCLV3(997M)::H2B-mYFP. Neither the single (−1080) nor the double (−1080 and 997) mutant altered CLV3 expression. H2B-mYFP (yellow) is overlaid on FM4-64 (red). (Scale bar: 10 μm.)

Fig. S2.

Complementation analysis with pCLV3 mutant promoters. (A–H) Images of siliques of wild-type (A), clv3-2 (B), and clv3-2 plants transformed with the CLV3 genomic rescue construct expressed from the mutant pCLV3(WT)::gCLV3 mutant (C), mutant pCLV3(TM)::gCLV3 mutant (D), mutant pCLV3(SM)::gCLV3 (E), mutant pCLV3(ΔCRM)::gCLV3 (F), mutant pCLV3(970-M1)::gCLV3 (G), and mutant pCLV3(970-M4)::gCLV3 (H). (Scale bars: 2 mm.) (I–O) Top views of 3D-reconstructed meristems of clv3-2 (I) and) clv3-2 mutants carrying CLV3 genomic constructs expressed from pCLV3(WT)::gCLV3 9J0 (J), mutant pCLV3(TM)::gCLV3 (K), mutant pCLV3(SM)::gCLV3 (L), mutant pCLV3(ΔCRM)::gCLV3 (M), mutant pCLV3(970-M1)::gCLV3 (N), and mutant pCLV3(970-M4)::gCLV3 (O). (Scale bars: 80 μm in I; 20 μm in J–O.) (P) Graphical sketch showing the spatial landmarks (flower primordia) used for measuring the inflorescence meristem height.

To understand the significance of the five cis elements in the CRM in CLV3 regulation (Fig. 2A and Fig. S3), we introduced mutations that abolish WUS binding (Fig. 2 B–F and Fig. S1 A–D) and that also contained a point mutation in the upstream −1080 cis element (n = 15) (Fig. S1E) into the pCLV3::H2B-mYFP reporter (Fig. 2 G and H). The mutations in the highest-affinity cis element 970 (the double mutant pCLV3-DM) led to a decrease in reporter fluorescence in the L1 and the L2 layers and an increase in the L4 layer (Fig. 2 C, I, and J and Fig. S4 A–C) (n = 12). The mutation in intermediate-affinity cis element 997 (Fig. 2D) did not significantly alter CLV3 expression (Fig. S1E). However, mutations in the 970 and the 997 elements (the triple mutant pCLV3-TM) resulted in a decrease in expression levels in the L1 and the L2 layers along with a significant increase in expression in layers 4–6 (Fig. 2 K and L and Fig. S4 A–E) (n = 16). The pCLV3-TM also revealed a higher number of cells that expressed the mutant reporter, some of which showed reporter expression at extremely low levels even in the L1 and L2 layers (Fig. 2 K and L). An increase in the number of cells expressing pCLV3-TM and higher levels of the mutant reporter in deeper layers suggest that the higher-affinity cis elements 970 and 997 may inhibit lower-affinity cis elements from activating CLV3. Additional mutations in the lower-affinity cis elements 950 (Fig. 2B) and 1060 (Fig. 2F) were introduced to create a quintuple mutant promoter, pCLV3-QM, which led to a severe decrease in expression in the L1 layer and higher expression in the L4 layer (Fig. 2 M and N and Fig. S4 A–C). An overall reduction in the number of pCLV3-expressing cells was observed in all cell layers (Fig. 2 M and N and Fig. S4 E–F) (n = 12). A decrease in the number of cells in pCLV3-QM suggests that the lower-affinity cis elements 950 and 1060 can activate CLV3 transcription in a broader spatial domain. The mutation in the final lower-affinity cis element 1007 (Fig. 2E) was introduced to generate a sextuple mutant, pCLV3-SM. The pCLV3-SM revealed a severe reduction in pCLV3 expression (Fig. 2 O and P) that was comparable to the reporter expression observed upon deletion of the entire CRM (Fig. 1B). Taken together these results show that WUS uses the same cis elements to activate and repress CLV3 expression. Furthermore, these results show that higher-affinity cis elements may repress transcription from the lower-affinity cis elements to prevent misexpression of CLV3 in deeper cell layers and thus suggest interactions among cis elements.

Fig. 2.

The same cis elements mediate activation and repression of CLV3 expression. (A) Schematic of the CLV3 gene showing the location of the 3′ CRM. The DNA sequence from +933 to +1080 is shown in black, and the TAAT core containing WUS-binding elements on the CRM are labeled in red. (B–F) EMSAs (Left) and WUS–DNA saturation curves (Right) were performed using different concentrations of the WUS (amino acids 1–134) DNA-binding domain bound to radiolabeled oligonucleotides of 950 (B), 970 (C), 997 (D), 1007 (E), and 1060 (F) cis elements. Black arrowheads show the WUS–DNA complex. (G–P) All reporters carry mutations in −1080, the 5′ cis element. Schematic representations of the reporter constructs are annotated with wild-type (cyan) and mutant (red) cis elements on their respective inflorescence (G, I, K, M, and O) and vegetative (H, J, L, N, and P) meristems. Side views show inflorescence and vegetative SAMs of wild-type pCLV3 (G and H); the 970 cis-element double mutant (pCLV3-DM) (I and J); the 970 plus 997 cis-element triple mutant (pCLV3-TM) (K and L); the 950, 970, 997, and 1060 cis-element quintuple mutant (pCLV3-QM) (M and N); and the 950, 970, 997, 1007, and 1060 cis-element sextuple mutant (pCLV3-SM) (O and P). In G, I, K, M, and O H2B-mYFP expression is shown in yellow and FM4-64 labeling is shown in red. In H, J, L, N, and P the H2B-mYFP expression (yellow) is overlaid onto the bright-field images. In G–P the white arrows show different cell layers. (Scale bars: 10 μm.)

Fig. S3.

Estimation of the binding affinities of WUS to the five cis elements in the CRM and to the higher-affinity (970-M4) mutant cis element. (A) EMSA using different concentrations of WUS (amino acids 1–134) bound to radiolabeled oligonucleotides of five cis elements in the CRM and the 970-M4. The three replicates for each cis element were used for estimating K_d values. Concentration range of WUS (amino acids 1–134) is stated in micromolars above each gel. Black arrowheads indicate the WUS–DNA complex. (B–G) WUS–DNA saturation curves for cis elements 970 (B), 997 (C), 1007 (D), 950 (E), 1060 (F), and 970-M4 (G). Quantification details are provided in Experimental Procedures in the main text.

Fig. S4.

The effect of cis-element deletions on CLV3 expression. (A) Sample images of different cis-element mutants of pCLV3::H2B-mYFP used for quantification of fluorescence from nuclear-bound regions. Four centrally located cells (within the red lines) were considered. (B) Quantification of the number of cells with detectable expression found in different cell layers of cis-element mutations. (C) The average fluorescence was quantified from centrally located cells of five independent SAMs. Expression from pCLV3(−1080M)::H2B-mYFP was used as the wild-type reference for the pCLV3(DM)::H2B-mYFP mutant (970 and −1080), the pCLV3(TM)::H2B-mYFP mutant (970, 997, and −1080), the pCLV3(QM)::H2B-mYFP mutant (950, 970, 997, 1060, and −1080), and pCLV3(970-M4)::H2B-mYFP. The error bars represent the SE of each sample set. A single asterisk denotes statistical significance (P < 0.05) as determined by two-tailed Student’s t test between pCLV3(−1080M)::H2B-mYFP and pCLV3(TM)::H2B-mYFP. (D and E) RNA in situ hybridization patterns of pCLV3(WT)::H2B-mYFP (D) and pCLV3(TM)::H2B-mYFP (E) using mGFP5 as the anti-sense probe. (F) Side view of seedling SAMs with eGFP-WUS expressed from the pWUS showing cells in L1 with low (yellow lines) and high (white lines) fluorescence. (Scale bars: 10 μm.)

WUS Binds to cis Elements in a Concentration-Dependent Manner.

The misexpression of the cis-element mutant promoters of CLV3 observed in the previous section suggests that WUS may repress CLV3 transcription in the RM, where it accumulates at a higher level, and activate in the L1 and the L2 cell layers, where it accumulates at a lower level (6). To understand the biochemical basis by which WUS can mediate both transcriptional activation and repression by binding the same cis elements, we tested the binding behavior of cis elements at increasing concentrations of the WUS protein. Five cis elements formed a low molecular weight complex at lower WUS levels and a higher molecular weight complex with increasing WUS concentrations (Fig. 3 A–E and Fig. S5 P–S). We tested whether the concentration-dependent switch is caused by the dimerization of WUS, which was found to be critical for WUS function. WUS contains two homodimerization domains, homodimerization domain1 (HOD1) and homodimerization domain2 (HOD2). HOD1 consists of a single amino acid residue located within the loop that connects the second and third α-helices of the homeodomain. HOD2 consists of a 74-aa region (amino acids 134–208) in the central part of the protein (26). A truncated form of WUS (amino acids 1–134) lacking HOD2 bound CLV3 cis elements with affinity comparable to that of the full-length protein. However, the truncated form required about four times higher WUS levels to produce a higher molecular weight complex, revealing that HOD1 alone is sufficient to dimerize at higher WUS levels (Fig. 3 F–H). A WUS mutant protein lacking both HOD1 and HOD2, WUS (amino acids 1–134:G77E), bound cis elements with lower affinity but failed to produce a higher molecular weight complex (Fig. 3I). The HOD1 mutant version with the intact HOD2, WUS (amino acids 1–208:G77E), was unable to bind DNA (Fig. 3J), unlike with the HOD1 mutant lacking the HOD2, WUS (amino acids 1–134:G77E). The severe loss of binding of the HOD1 single mutant with an intact HOD2 may be caused by the predominance of intermolecular interactions between WUS–WUS homodimers that might have outcompeted the DNA–WUS interactions. Alternately, the intact HOD2 might have inhibited the mutated HOD1 from binding DNA through a drastic conformational change in the homeodomain or by enhancing steric hindrance. In summary, these results show that the two HOD domains are required for dimerization of WUS at higher concentrations and that HOD1 also is necessary for DNA binding.

Fig. 3.

DNA promotes homodimerization of WUS. (A–E) EMSAs showing the binding of five cis elements of the CLV3 CRM, 950 (A), 970 (B), 997 (C), 1007 (D), and 1060 (E), to increasing concentrations of full-length WUS (amino acids1–292) which contains both HOD1 and HOD2. (F–J) EMSA showing the binding of the 970 cis element to increasing concentrations [0, 1× (0.5 ng/μL), 2×, 4×, and 8×] of full-length WUS (amino acids 1–292) (F), truncated WUS (amino acids 1–208) (G), truncated WUS (amino acids 1–134) lacking the HOD2 (H), truncated WUS (amino acids 1–134) containing the HOD1 (G77E) mutation (I), and truncated WUS (amino acids 1–208) containing the HOD1 (G77E) mutation (J). SEC experiments were performed using 0.3 µM, 3 µM, and 15 µM of full-length purified recombinant WUS protein. (K) Dot blot analyses of SEC-collected fractions containing WUS protein complexes were visualized by anti-WUS antibodies. (L) The WUS dimer/monomer ratio of WUS (amino acids 1–292) protein concentration and K_d was estimated from the saturation-fitting hyperbolic curve using GraphPad Prism 5 software. (M) Immuno dot blot analyses of SEC-collected fractions of the HOD1 (G77E) and HOD2 (Δ amino acids134–208) double mutant using anti-WUS antibody. Positions of WUS monomer, dimer, and multimer complexes are shown. Elution positions of molecular mass standards (BSA: 66 kDa; carbonic anhydrase: 29 kDa; and cytochrome C: 12 kDa) are marked. The position of the void volume (Vo) (∼100 KDa) is marked.

Fig. S5.

The DNA promotes homodimerization of WUS. (A–D) EMSAs showing recombinant full-length WUS bound to radiolabeled oligonucleotides that contained individual cis elements (950, 970, 997, and 1060) found in CLV3. The binding behavior at increasing WUS concentrations is shown: (A) 0; (B) 1× (0.5 ng/μL); (C) 4× (2 ng/μL); (D) 16× (8 ng/μL). Black and gray arrowheads indicate positions in the gel that show lower and higher molecular weight complexes, respectively. (E–P) Dot blot analyses of SEC-collected fractions containing WUS protein complexes visualized by anti-WUS antibodies. SEC experiments were performed using 0.3 µM (E–H), 3 µM (I–L), and 15 µM (M–P) of full-length purified recombinant WUS (amino acids 1–292) protein. (Q) The fractions corresponding to the WUS dimer and monomer were pooled to measure the protein concentration. The table summarizes the number of fractions pooled, protein concentrations of dimers and monomers, and the dimer/monomer ratios. Procedural details can be found in Experimental Procedures in the main text. (R) The WUS dimer/monomer ratio was presented as a function of total WUS (amino acids 1–292) protein concentration using GraphPad Prism 5 software, and the K_d was estimated from the saturation-fitting hyperbolic curve. (S) Comparison of SEC experiments using 0.3 µM, 3 µM, and 15 µM of bacterially expressed purified full-length WUS (amino acids 1–292). Shown are immuno dot blot analyses of SEC-collected fractions using anti-WUS antibody. The positions of WUS monomer, dimer, and multimer complexes are shown. Elution positions of the molecular-mass standards are marked: BSA, 66 kDa; carbonic anhydrase, 29 kDa. The position of the void volume (Vo) ∼100 kDa, is marked.

Decreasing the Homodimerization Threshold of cis Elements Leads to CLV3 Repression.

The results from the analysis of cis element deletion suggested that WUS may activate CLV3 transcription at lower concentrations and repress CLV3 transcription at higher levels. The biochemical analysis revealed that the same cis elements could bind WUS as monomers at lower concentrations and as dimers at higher levels. The EMSA experiments also show that WUS formed dimers at lower WUS concentrations in interactions with the higher-affinity cis elements than in interactions with the lower-affinity cis elements (Fig. 3 A–E and Fig. S5 A–D). Thus the affinity and the dimerization thresholds are inversely correlated. These experiments also suggest that the WUS-binding sequence contains information that determines WUS-binding affinity and homodimerization thresholds. Taken together, these observations suggested that WUS may repress CLV3 transcription by forming homodimers at higher concentrations. Such a model predicts that decreasing the homodimerization threshold of cis elements should lead to repression of CLV3 in the CZ, where WUS is present at a lower level.

To test further the importance of the WUS-binding sequence in promoting WUS homodimerization, we analyzed the self-association behavior of WUS in solution (without DNA) by using size-exclusion chromatography (SEC) at WUS concentrations ranging from 0.3 to 15 µM. The SEC analysis revealed that even a 50-fold increase in WUS concentration was unable to convert the monomeric pool into dimeric or higher-order complexes (Fig. 3 K–M and Fig. S5 E–S). However, in the presence of DNA, a mere two- to fourfold increase in WUS concentration was sufficient to switch the monomeric WUS–DNA complex completely into a dimeric WUS–DNA complex (Fig. 3 A–E and Fig. S5 A–D). Taken together, these results show that the WUS-binding sequence promotes homodimerization or multimerization of WUS (EMSA experiments may not have distinguished dimeric complexes from multimeric complexes) and that the homodimerization threshold is inversely correlated to cis-element affinity.

Based on the biochemical analysis presented in the previous section, we hypothesized that a much higher-affinity cis element might dimerize WUS at lower concentrations in the CZ and repress CLV3. As part of our efforts to understand DNA features that determine affinities and dimerization thresholds, we systematically mutated each base pair within and outside the TAAT core of the high-affinity cis element 970, which contains two tandem TAAT cores. Nucleotides both within and outside the TAAT core either increased or decreased affinity to WUS, suggesting that a higher-order structure may influence binding affinities (Fig. S6). Earlier studies have shown that the homeodomain proteins can bind sequences that deviate slightly from the canonical TAAT core-containing sequences (27, 28). To take advantage of this ability of homeodomain proteins to bind noncanonical TAAT core sequences, we considered two of these mutant cis elements, 970-M4 and 970-M1, for further analysis with respect to WUS binding. Estimation of the binding affinity revealed that 970-M4 bound WUS at an affinity approximately three times higher (K_d = 0.05830 μM) (Fig. S3G) than that of the wild-type 970 cis element (Fig. S3B). The 970-M1 mutant also bound WUS at a concentration range comparable to that of 970-M4 (Fig. 4 A and B). The WUS protein that contained only HOD1 was able to dimerize with 970-M1 and 970-M4 (Fig. 4 A and B) at much lower levels than with the wild-type 970 (Fig. 2C). Moreover, the full-length WUS protein dimerized with 970-M4 at lower levels than with the wild-type 970 cis element (Fig. 4 C and D). Taken together, these results show that the two higher-affinity cis elements also exhibited lower dimerization thresholds.

Fig. S6.

Nucleotides within and outside the TAAT core modulate WUS binding. (A) Sequences showing single base substitutions of the 970 cis element. (B) EMSA comparing WUS (amino acids 1–134) binding to wild-type and mutated 970 cis elements shown in A. Note the higher-affinity cis elements including 970-M1 and 970-M4, which were used for further in vivo analysis. The black arrowhead indicates the WUS monomer.

Fig. 4.

Increasing the cis-element affinity lowers the dimerization threshold, leading to CLV3 repression. (A and B) EMSAs showing binding of truncated WUS (amino acids 1–134) lacking the HOD2 at increasing concentrations to mutant versions of the 970 cis elements 970-M1 (A) and 970-M4 (B). The sequence is described in Fig. S6A. The numbers above the autoradiograms indicate the WUS concentration in nanograms per microliter. Compare with the wild-type 970 cis element in Fig. 2C. Note dimerization at WUS levels in 970-M4 and 970-M1. (C and D) EMSAs showing the binding of wild-type 970 (C) and mutated 970 cis element (970-M4) (D) to increasing concentrations [0, 1× (0.5 ng/μL), 2×, 4×, 8×, and 16×] of the full-length WUS (amino acids 1–292). Black arrowheads indicate monomers, and gray arrowheads indicate dimers. (E–J) Side views of wild-type (E–G) and clv3-2 (H–J) inflorescence meristems showing H2B-mYFP expression in mutated pCLV3-(−1080M) (E and H), mutated pCLV3-970-M1 (F and I), and mutated pCLV3-970-M4 (G and J). (K and L) Side views of inflorescence meristems showing pWUS:eGFP-WUS expression in wild-type (K) and clv3-2 (L) plants. H2B-mYFP (yellow in E–J) and eGFP-WUS (green in K and L) are superimposed on FM4-64–stained (red) inflorescence meristems. In K and L the white arrows show different cell layers. (Scale bars: 10 μm in E–G and K; 15 μm in H–J and L.)

In our efforts to understand the role of dimerization in CLV3 regulation, we analyzed higher-affinity mutant reporters and CLV3 genomic constructs. Introduction of the 970-M4 and 970-M1 mutations into the pCLV3::H2B-mYFP reporter resulted in a dramatic decrease in the reporter expression except in deeper cell layers, where it was expressed at lower levels (Fig. 4 E–G and Fig. S4 B and C) (n = 12) that were comparable to the expression detected upon deletion of the CRM (Fig. 1B). Consistent with the lower levels of the reporter, complementation assays using the 970-M4 and 970-M1 mutant promoters expressing the CLV3 genomic region failed to complement fully the SAM and the FM defects of clv3-2 mutants (Fig. 1 C and D and Fig. S2). To test further whether the repression is WUS dependent, we introduced the mutant reporters into clv3-2 mutants that accumulate WUS at extremely low levels in the L1 layer despite its higher synthesis in deeper layers (Fig. 4 K and L) (29). Both 970-M4 and 970-M1 were reactivated in the L1 layer of clv3-2 SAMs (Fig. 4 I and J). The recovery of expression of the 970-M4 and 970-M1 reporters suggests that the repression caused by higher-affinity binding could be alleviated by decreasing the WUS concentration. Taken together these results show that increasing the cis-element affinity decreases the dimerization threshold and is sufficient to repress CLV3.

CLV3 Expression Is Sensitive to WUS Dosage.

The cis element manipulation analysis reveals that WUS activates CLV3 transcription at lower concentrations and represses CLV3 transcription at higher concentrations. Earlier studies have shown that ectopic overexpression of WUS leads to activation of CLV3, a result that is not consistent with higher WUS levels leading to CLV3 repression (30, 31). However, these studies only analyzed WUS mRNA patterns and did not analyze WUS protein distribution patterns following ectopic activation. The results presented in the companion article in this issue (26) show that ectopic overexpression of WUS leads to protein destabilization. eGFP-WUS expressed from a CZ-specific promoter (pCLV3::LhG4;6XOP::eGFP-WUS) accumulated at much lower level (n = 20) (Fig. 5B) than eGFP-WUS expressed from the pWUS (Fig. 5A). These SAMs also revealed higher levels of pCLV3 (n = 5) (Fig. 5 D and E and Fig. S7 A and B), in agreement with earlier studies (30, 31). The ubiquitous overexpression of WUS using a dexamethasone (Dex)-inducible system (35S::eGFP-WUS-GR) also destabilized WUS protein (26). Specifically, within 6 h of Dex application, the protein became unstable in cells located in the CZ. The region of lower protein accumulation expanded radially into adjacent PZ cells within 12 h of Dex treatment and reached outer edge of the PZ within 24 h of Dex treatment (26). We also monitored the pCLV3 response to the induction of 35S::eGFP-WUS-GR and found an increase in pCLV3 levels within 12 h of Dex treatment and subsequent radial expansion of the reporter expression into adjacent PZ cells within 24 h of Dex treatment (Fig. 5 G–I). These results reveal that CLV3 activation correlates with a reduction in WUS levels.

Fig. 5.

The CLV3 promoter is sensitive to WUS dosage. (A–C) 3D top views of inflorescence meristems expressing pWUS::eGFP-WUS (A), pCLV3::LhG4;6XOP::eGFP-WUS (B), and pCLV3::LhG4;6XOP::NLS-eGFP-WUS (C). eGFP-WUS is shown in green, and FM4-64 is shown in red. (D–F) Side views of vegetative SAMs expressing wild-type-pCLV3 (H2B-mYFP) overlaid on bright-field images in wild-type (D), pCLV3::LhG4;6XOP::eGFP-WUS (E), and pCLV3::LhG4;6XOP::NLS-eGFP-WUS (F). H2B-mYFP is shown in yellow. (G–I) 3D top views of 35S::WUS-GR inflorescence meristems showing pCLV3::H2B-mYFP mock treated (G) or treated with 10 μM Dex for 12 h (H) or 24 h (I). (A-I) (Scale bars: 10 μm in A, B, and D–I; 30 μm C.) (J–L) 3D top views of 35S::GR-LhG4; 6XOP::amiR-WUS inflorescence meristems mock treated (J) or treated with Dex for 4 d (K) or 8 d (L). M, N, and O are side views of J, K, and L, respectively. (P) Quantification of the WUS transcript levels in Dex- and mock-treated seedlings expressing amiR-WUS. Error bars represent SD. (Q–S) Side views of 7-d-old wild-type (Q), wus-1^−/− (R), and wus-1^+/− (S) vegetative SAMs showing the wild-type-pCLV3 (H2B-mYFP). In M–O and Q–S the white arrows show different cell layers. (Scale bars: 20 μm J–Q and S; 50 μm in R.)

Fig. S7.

pCLV3 is sensitive to WUS dosage. (A–C) 3D reconstructions of inflorescence SAMs showing pCLV3::H2B-mYFP in wild type (A), pCLV3::LhG4;6XOP::eGFP-WUS (B), and pCLV3::LhG4;6XOP::NLS-eGFP-WUS (C). H2B-mYFP is shown as yellow. (Scale bars: 10 μm in A and B; 15 μm in C.) (D and E) Average fluorescence intensity from centrally located cells (D) and number of cells with detectable expression (E) found in different cell layers of pCLV3(−1080M)::H2b-mYFP (mutant −1080 was used as wild type) and wus-1 heterozygous background. The error bars represent the SE of each sample set.

A close spatiotemporal correlation between the dilution of WUS levels and the increase in pCLV3 expression suggested that lower WUS levels may activate CLV3 transcription. We tested this hypothesis by transiently depleting WUS using a Dex-inducible two-component system to activate an artificial microRNA (amiRNA) directed against WUS (Fig. 5P). Within 4 d of Dex treatment, a dramatic increase in pCLV3 along with radial expansion of the reporter was observed, and continued treatment with Dex for 8 d led to lower pCLV3 expression (Fig. 5 J–O) (n = 12). The wus-1 heterozygous mutants also expressed pCLV3 at higher level, whereas homozygous mutants failed to express pCLV3 at a detectable level (Fig. 5 Q–S and Fig. S7 D and E). The radial expansion of the pCLV3 upon partial depletion of WUS could be caused by the overpopulation of the monomeric form of WUS, which has been shown to diffuse farther in inflorescence meristems (26, 29), showing that the regulation of transcription and protein distribution are coupled. Taken together, these experiments show that the decrease in WUS levels leads to the activation of CLV3 expression until WUS levels fall below a certain threshold.

Because a decrease in WUS level led to CLV3 activation, we next tested whether an increase in WUS level can repress CLV3. Expression of WUS carrying a strong nuclear localization signal, NLS-eGFP-WUS, by using the pCLV3::LHG4 driver revealed an intense fluorescence in the nuclei of a few cells, showing that an enhanced nuclear accumulation improves WUS protein stability (Fig. 5C) (n = 20). These SAMs revealed a severe decrease in pCLV3 expression (n = 5) (Fig. 5F and Fig. S7C), showing that higher WUS levels repress CLV3. We cannot rule out the possibility that the repression could be indirect; however, the results from the manipulation of cis elements and WUS levels in tandem show that CLV3 transcription is maintained over a window of WUS levels bound by the activation and the repression thresholds.

The CLV3 promoter lacking the CRM, pCLV3(ΔCRM), partially complemented the enlarged SAM (Fig. 1C) and FM (Fig. 1D) phenotypes of clv3-2 mutants. An earlier study has shown that at 16% of wild-type levels the CLV3 mutant promoter was able to complement partially clv3-2 mutant phenotypes (31), indicating that the meristem maintenance requires only a small amount of CLV3. A relatively better floral meristem rescue (based on carpel number) was observed with the pCLV3-SM mutant than with the higher-affinity mutants pCLV3(970-M1) and pCLV3(970-M4) and with the mutant promoter lacking the entire CRM, suggesting that pCLV3-SM may contain relatively higher CLV3 levels (Fig. 1D). EMSA shows that the nucleotides outside the TAAT core also can influence WUS binding (Fig. S6) and that WUS also can bind a partially mutated cis element (Fig. S1D). These observations suggest that at higher levels WUS may still bind the mutant cis elements or elements elsewhere within the CRM and can explain the higher levels of CLV3 in the pCLV3-SM mutant than seen with the deletion of the CRM. Therefore, a collective WUS binding to the CRM may be important in regulating CLV3 transcription. In clv3-2 mutants the expression of the mutant reporter pCLV3(TM) was decreased in the inner layers, where WUS accumulates at higher levels, and was increased in the L1 layer, where WUS accumulates at lower levels (Fig. 6), suggesting that the remaining cis elements respond to WUS levels. A similar resetting of the pCLV3(SM) in clv3-2 mutants suggests that the remaining regulatory elements located within and outside the CRM may also respond to WUS levels. The resetting of the mutant promoters in clv3-2 also suggests that the dosage sensitivity of CLV3 depends partly on the CRM.

Fig. 6.

The mutant CLV3 promoters are repressed in the inner layers and are activated in the outer layers of clv3-2 mutants. (A–C) Side views of clv3-2 inflorescence meristems expressing wild-type pCLV3 (H2B-mYFP) (A), cis-element mutant pCLV3-TM (H2B-mYFP) (B), and cis-element mutant pCLV3-SM (H2B-mYFP) (C). H2B-mYFP (yellow) is superimposed on FM4-64-stained (red) SAMs. The white arrows indicatate different cell layers. (Scale bars: 15 μm.)

Discussion

Our work shows that WUS mediates the activation and repression of CLV3 by using the same set of cis elements within the CRM. WUS levels vary not only among cells located in different cell layers but also between adjacent cells of a given cell layer (Fig. S4F). The observation that WUS binds each cis element within the CRM with different affinities may provide a mechanism to buffer these variations in maintaining CLV3 transcription that, in turn, has been shown to regulate WUS transcription in the RM (7, 10). The higher levels of WUS repressing CLV3 and lower levels activating CLV3, as proposed here, may not be able to produce a stable system if a strict linear relationship exists between levels of WUS transcription and the amount of WUS protein. However, we have observed lower levels of WUS in the L1 layer of clv3-2 mutants despite higher synthesis in the RM, suggesting additional tiers of WUS regulation. Earlier studies have shown that CLV3 promoter is highly active in the L1 layer of clv mutants (10), as is consistent with lower levels of WUS activating CLV3 transcription. The lower levels of WUS protein accumulation in the L1 layer of clv3-2 mutants could be caused by a direct role of CLV3-mediated signaling in posttranslational regulation of WUS. Alternately it might be an indirect consequence of higher protein synthesis resulting from enhanced transcription, because our work shows that overexpression of WUS leads to protein instability (26). In either scenario, higher levels of WUS-repressing CLV3 will destabilize WUS, leading to CLV3 activation, which in turn stabilizes the WUS protein and leads to CLV3 repression, thus forming a feedback loop that connects concentration-dependent transcriptional regulation of CLV3 to the regulation of WUS protein levels. The cis element manipulation analysis also suggests interactions among cis elements. Perhaps higher WUS levels might induce cooperative interactions among cis elements, as shown in the case of BICOID binding to the HUNCHBACK CRM (32); these cooperative interactions could lead to enhanced dimerization and transcriptional repression. Biochemical analysis of the collective WUS-binding behavior to the CRM, along with the in vivo analysis of the importance of spacing between cis elements and relative orientations of cis elements, may provide further insights into the nature of interaction among cis elements.

Several alternate models can explain WUS-mediated concentration-dependent transcriptional regulation. The WUS-mediated repression of CLV3 may involve the quenching of an independent activator of CLV3 by the WUS homodimers. Such a mechanism would lead to higher levels of CLV3 upon WUS removal or deletion of WUS-binding cis elements. We have observed CLV3 up-regulation upon partial depletion of WUS, which can explain the quenching mechanism. However, wus null mutants do not express CLV3, which is not consistent with the quenching mechanism. The quenching mechanism also fails to explain the decrease in CLV3 expression observed upon mutating WUS-binding cis elements unless the independent activator also binds the same cis elements. Alternately, spatially localized coactivators in the CZ or corepressors in the RM could mediate the localized activation of CLV3. Biochemical studies have shown that the HAM proteins expressed in both the RM and the PZ interact with WUS (16, 17). The genetic analysis revealed that HAM proteins require WUS function in regulating SAM maintenance (17). At present we do not understand the mechanisms by which HAM proteins promote SAM maintenance; however, the CLV3 repression observed in the L1 and the L2 layers of higher-affinity cis element mutants presented here shows that WUS can repress CLV3 in a HAM-independent fashion. Perhaps HAM proteins may provide an additional layer of regulation in potentiating the WUS-mediated repression in inner layers, either by increasing the binding affinity of WUS to cis elements or by promoting homodimerization of the WUS protein. Moreover, the switch in CLV3 expression observed in cis-element mutation analysis cannot support the CZ-localized coactivator model.

We suggest that WUS might function with an activation complex and a repression complex present in both the CZ and the RM. At lower concentrations, WUS could bind as a monomer or as a heterodimer that could recruit activation machinery. At higher concentrations, WUS homodimers may fail to engage an activation complex, may favor recruitment of ubiquitously expressed corepressor complex, or may prevent the CRM located in the 3′ region from engaging with the proximal promoter. Future work on both transcriptional coregulators and understanding the importance of the location of the CRM may lead to further insights into the concentration-dependent modulation of gene expression.

Experimental Procedures

Plant Growth, Genotypes, and Microscopy.

Arabidopsis plant-growth procedures for live imaging and phenotypic analysis were followed as described in earlier studies (6, 13, 33). All transgenic lines were generated in the Landsberg erecta background. wus-7, wus-1, and clv3-2 have been described previously (4, 8, 34). Preparation for live imaging, Dex treatment, the optics, the microscopy platform (Zeiss 510 LSM), image acquisition, and image reconstruction have been described in earlier studies (6, 13, 33). Vegetative SAM sections (see following section) were imaged using a Leica SP5 system fitted with an argon laser that was activated to 20%. The images were taken using a 40× objective and a 2.5× digital zoom. eGFP-tagged constructs were imaged with excitation at 488 nm, and emission was collected between 500 and 550 nm. mYFP-tagged constructs were imaged by using either 488 nm or 514 nm excitation, and emission was collected between 550 and 600 nm. The bright-field images were captured by simultaneous scanning.

Fluorescence Quantification.

Image stacks (.lif files) from the Leica SP5 microscopy were imported into ImageJ software using the LOCI Bio-Formats plugin. Image stacks then were separated into individual xy .tiff images using the Image → Stacks → Stack to Images function. Fluorescent channels from these images were isolated using the Image → Color → Split Channel function. The region of interest (ROI) manager was loaded using the Analyze → Tools → ROI Manager function, and circular regions of interest were drawn around the nuclei expressing H2B-mYFP or eGFP:WUS. The measure tool in the ROI manager was used to determine the average fluorescence within each region of bounded nuclei, and these values were assigned to their respective cells as numbered by ROI selection. Cells were assigned to SAM layers as L1, L2, L3, and so forth. These values were averaged across each layer, and this process was repeated in five plants per construct to obtain the quantification values used to compare fluorescence intensity values across layers.

EMSA.

His-WUS fusion full-length protein WUS1–292, truncated WUS proteins WUS1–208 and WUS1–134, the HOD1 single mutant WUS1–208:G77E, and the HOD1/HOD2 double mutant WUS1–134:G77E were generated using the oligonucleotides listed in Table S1. The amplified DNA fragments were cloned into pET28a plasmid (Novagen) for recombinant expression in Escherichia coli Bl21 cells as described previously (6, 19). His-WUS fusion proteins were purified as described earlier (6, 19). All oligonucleotides used for EMSAs are listed in Table S1. Oligonucleotide radiolabeling, DNA–protein binding reactions, and electrophoresis were performed as described earlier (6). Purified full-length His-WUS and truncated and mutated protein versions were diluted in 20 μM Hepes-KOH at pH 7.8, 100 μM KCl, and 1 μM EDTA at 1× concentration (0.5 ng/μL) for binding assays. For K_d calculation, WUS binding was quantified by densitometry from three independent saturation curves using Quantity One 1-D Analysis Software (Bio-Rad), and the fraction of the oligos bound by WUS at each protein concentration was calculated. Graphical representation and nonlinear fit for a saturation-binding curve and K_d calculation were performed using GraphPad Prism 5.0 software. The data are presented as the mean ± SE of three independent experiments.

Table S1.

Primers used in this study

Construct		Primer name	Sequence
pCLV3::H2b-mYFP
H2b-mYFP	H2b-mYFP-BamH-Fw		CAGGATCCATGGCGAAGGCAGATAAGAAACCAGCGG
	H2b-mYFP-BamH-Rev		GTGGATCCTTATTTGTATAGTTCATCCATGCCATGTG
pCLV3::H2b-mYFP mutant promoter
pCLV3-1080M	pCLV3-1080M-Fw		CATATGATCCATTCAATTTATGTTTTTTC
	pCLV3-1080M-Rev		AGCCTTGCCGGCGCCGTATCGAGGG
pCLV3-1200, -900	pCLV3-1200, -900-Fw		AAGCATATAACTGTTTCCAGATTAAAC
	pCLV3-1200, -900-Rev		CAGATTCCGTTTTGCTTCGTTACTC
pCLV3 950M	pCLV3 950M –Fw		CGTACCCCCAAATTTTCCCAACGGTACATTGC
	pCLV3 950M-Rev		TTTTCAATTGTCAATGCAAATACCCCATGG
pCLV3 970M	pCLV3 970M –Fw		GGTATTTGCATTGACAATTGAAAACGTAC
	pCLV3 970M –Rev		CCATGGATGTGATAGTCACAATTAAAC
pCLV3 997M	pCLV3 997M –Fw		GTGACTATCACATCCATTAATTATTTGC
	pCLV3 997M –Rev		AATGGAACATACAATAATAAAAATGATGATG
pCLV3 1007M	pCLV3 1007M –Fw		GATTCGATGATGTGGTGGGAAGG
	pCLV3 1007M –Rev		ATCATCATCATTTTTGGGGTTGTATGTT
pCLV3 1060M	pCLV3 1060M –Fw		GTCGGTTCCCCTTATCCTTCCCACCACATCATC
	pCLV3 1060M –Rev		TTTGGGGCAGTGACAGGCAGTGTCAGTG
pCLV3 943,1067	pCLV3 1067 -Fw		CCGACTTTGGGGCAGTGACAG
	pCLV3 943 -Rev		CCCAACGGTACATTGCTTTGG
EMSA probes
950	950-Fw		AACGTACTAATAAATTTTCCCAACGGTA
	950-Rev		TACCGTTGGGAAAATTTATTAGTACGTT
950M	950M-Fw		AACGTACCCCCAAATTTTCCCAACGGTA
	950M-Rev		TACCGTTGGGAAAATTTGGGGGTACGTT
970	970-Fw		CACATCCATTAATTATTTGCATTGACAATTG
	970-Rev		CAATTGTCAATGCAAATAATTAATGGATGTG
970M	970M-Fw		CACATCCATCCCCTATTTGCATTGACAATTG
	970M-Rev		CAATTGTCAATGCAAATAGGGGATGGATGTG
997	997-Fw		TTATTGTATGTTTAATTGTGACTAT
	997-Rev		ATAGTCACAATTAAACATACAATAA
997M	997M-Fw		TTATTGTATGTTCCATTGTGACTAT
	997M-Rev		ATAGTCACAATGGAACATACAATAA
1007	1007-Fw		ACATACAATAATAAAAATGATGATGAT
	1007-Rev		ATCATCATCATTTTTATTATTGTATGT
1007M	1007M-Fw		ACATACAAGGGGAAAAATGATGATGAT
	1007M-Rev		ATCATCATCATTTTTCCCCTTGTATGT
1060	1060-Fw		GTCGGTTTAATTTATCCTTCCCA
	1060-Rev		TGGGAAGGATAAATTAAACCGAC
1060M	1060M-Fw		GTCGGTTCCCCTTATCCTTCCCA
	1060M-Rev		TGGGAAGGATAAGGGGAACCGAC
970-M1	970-M1-Fw		CACATCCAGTAATTATTTGCATTGACAATTG
	970-M1-Rev		CAATTGTCAATGCAAATAATTACTGGATGTG
970-M2	970-M2-Fw		CACATCCATGAATTATTTGCATTGACAATTG
	970-M2-Rev		CAATTGTCAATGCAAATAATTCATGGATGTG
970-M3	970-M3-Fw		CACATCCATTGATTATTTGCATTGACAATTG
	970-M3-Rev		CAATTGTCAATGCAAATAATCAATGGATGTG
970-M4	970-M4-Fw		CACATCCATTAGTTATTTGCATTGACAATTG
	970-M4-Rev		CAATTGTCAATGCAAATAACTAATGGATGTG
970-M5	970-M5-Fw		CACATCCATTAAGTATTTGCATTGACAATTG
	970-M5-Rev		CAATTGTCAATGCAAATACTTAATGGATGTG
970-M6	970-M6-Fw		CACATCCATTAATGATTTGCATTGACAATTG
	970-M6-Rev		CAATTGTCAATGCAAATCATTAATGGATGTG
970-M7	970-M7-Fw		CACATCCATTAATTGTTTGCATTGACAATTG
	970-M7-Rev		CAATTGTCAATGCAAACAATTAATGGATGTG
970-M8	970-M8-Fw		CACATCCATTAATTAGTTGCATTGACAATTG
	970-M8-Rev		CAATTGTCAATGCAACTAATTAATGGATGTG
970-M9	970-M9-Fw		CACATCCATTAATTATGTGCATTGACAATTG
	970-M9-Rev		CAATTGTCAATGCACATAATTAATGGATGTG
Cloning of WUS deletion and mutation in pET-28a for protein expression in E. coli
WUS-NDE	Forward		CATATGATGGAGCCGCCACAGCATCA
WUS-FL-XHO	Reverse		CTCGAGCTAGTTCAGACGTAGCTCA
WUS-208-XHO		Reverse	CTCGAGTCCACCTACGTTGTTGTAATTCATAG
WUS-134-XHO		Reverse	CTCGAGCGTGATGATGGTGAAGTAGAGGATG
WUS-7M		WUS-7M-Fw	GAGACAGTTCGAAAAGATTGAGG
		WUS-7M-Rev	AGCCTTGCAGTGATCTTC
qRT-PCR primers
WUS		WUS-Rev	GAGCTTTAATCCCGAGCGACACCGG
		WUS-Fw	GAAGACGGGGGAATGGGATGAGATT
UBQ10		UBQ10-Fw	GATCTTTGCCGGAAAACAATTGGAGGA
		UBQ10-Rev	CGACTTGTCATTAGAAAGAAAGAGATACA
WUS		WUS-Rev	GAG CTT TAA TCC CGA GCG ACA CCG G
WUS		WUS-Fw	ATC ATG CAA GCT CAG GTA CTG AAT GT
amiRWUS cloning
amiRWUS		WUS-I miR-s	gaTATTAATCACTAGCGAAGCGTtctctcttttgtattcc
precursor		WUS-II miR-a	gaACGCTTCGCTAGTGATTAATAtcaaagagaatcaatga
		WUS-III miR	gaACACTTCGCTAGTCATTAATTtcacaggtcgtgatatg
		WUS-IV miR	gaAATTAATGACTAGCGAAGTGTtctacatatatattcct
amiRWUS		WUS-oligo A (GW)	CACCTGCAAGGCGATTAAGTTGGGTAAC
pENTR		WUS-oligo B	GCGGATAACAATTTCACACAGGAAACAG
pENTR eGFP-WUS and pENTR NLS-eGFP-WUS for 6xOP cloning
		NLS-eGFP-Fw-GW	CACCATGGAGCAGAAGCTGATCTCCGAGGAGGAC
		eGFP-WUS-Fw-GW	CACCATGGTGAGCAAGGGCGAGGAGCTGTTCACC

SEC and WUS Detection and Quantification.

Full-length WUS (WUS1–292) and the HOD1 and HOD2 double mutant [WUS (1–134:G77E)] were expressed in E. coli BL21 cells, purified from soluble lysates using the his-tag protocol (Ni-NTA His⋅Bind Resins; Novagen), and subsequently dialyzed against 20 mM Hepes (pH 7.8) and 100 mM KCl buffer. Resin Bio-Gel P-100 (Bio-Rad) was hydrated in 20 mM Hepes (pH 7.8) and 100 mM KCl buffer and was packed on a 0.7 × 20 cm chromatography column (Bio-Rad) following the manufacturer’s instructions. SEC was performed with the dialyzed full-length (WUS1–292) and the HOD1 and HOD2 double-mutant [WUS (1-134:G77E)] forms suspended in 20 mM Hepes (pH 7.8) and 100 mM KCl buffer at room temperature, with a flow rate of 0.05 mL/min. Elutions of 250-mL fractions were collected, and the presence of WUS protein was analyzed in each elution using the Bio-Dot system (Bio-Rad) using anti-WUS antibody (4). To calculate the WUS dimer K_d, SEC was performed using 0.3-, 3-, and 15-µM dialyzed full-length WUS (WUS1–292). Concentrations of WUS monomer and WUS dimer were measured upon pooling the eluted fractions representing the two WUS species using the BCA Protein Assay Reagent and following the manufacturer’s instructions for enhanced sensitivity (Thermo Scientific Pierce). The ratio of WUS dimer/WUS monomer was determined. Graphical representation and K_d calculation were performed using GraphPad Prism 5.0 software. The data are presented as the mean ± SE of four independent experiments.

Plasmid Constructs, Generation of Transgenic Lines, and Phenotypic Complementation Analysis.

The deletions of pCLV3 were derived from the CLV3 promoter (SI Experimental Procedures) as described in an earlier study (31, 35). Details on the constructs used for expression and phenotypic complementation analysis of the mutant CLV3 promoter and the misexpression of eGFP-WUS, NLS-eGFP-WUS, and Dex-inducible amiRNA for WUS silencing are given in SI Experimental Procedures.

For quantitative RT-PCR (qRT-PCR), yeast two-hybrid analysis, RNA in situ analysis, and phenotypic complementation analysis of clv3-2 with the pCLV3 mutant promoters, see SI Experimental Procedures.

SI Experimental Procedures

Plant Seedling Sectioning for Confocal Microscopy and RNA in Situ Hybridization.

Seedlings were germinated on MS plates for 7 d, embedded in 5% agarose, hand cut with a fine razor, mounted on glass slides, and imaged as described in Experimental Procedures, Plant Growth, Genotypes, and Microscopy in the main text Wherever appropriate, seedlings were germinated on Murashiga and Skoog (MS) plates containing 10 μM Dex. Tissue preparation, sectioning of plants for RNA in situ analysis, probe synthesis, hybridization, and detection were performed as described earlier (www.its.caltech.edu/∼plantlab/protocols/insitu.html). The mGFP5 probe used in this study has been described in an earlier study (13).

Plasmid Constructs and Generation of Transgenic Lines: CLV3 Promoter.

The reporter construct pCLV3::H2B-mYFP was generated by introducing the BamH1-containing coding sequence of H2B-mYFP described in an earlier studies (31, 35). The CLV3 promoter contained the region 1,500 bp upstream and 1,200 bp downstream in the pGREEN binary vector. Mutations and deletions of pCLV3 were performed by inverse PCR using 5′ phosphorylated oligonucleotides (Table S1). A minimum of eight plants from independent transgenic lines of each promoter mutation construct were obtained and analyzed for H2B-mYFP fluorescence in both vegetative and inflorescence SAMs. The actual number of independent transgenic plants screened for each mutant promoter is stated in parentheses at appropriate places in the text. To account for the variability across different insertion lines, fluorescence intensity was quantified from five SAMs derived from independent transgenic lines.

Complementation Analysis.

The SAM size quantification was carried out on plants that were ∼5 wk old. A minimum of nine SAMs from six independent transgenic lines for each mutant promoter were used for quantifying SAM height. The adaxial junction of the fifth primordia was used as reference to measure the SAM height (see the schematic in Fig. S2P for details). Height was measured from the base of the fifth primordia to the SAM tip. For multiple comparisons, statistical analysis was performed by one-way ANOVA followed by Tukey’s HSD test using the soft R Project (v.3.1.2) package.

Plasmid Construct and Generation of Transgenic Lines: Manipulation of Wus Levels.

To generate the pCLV3::LhG4, 6XOP::eGFP-WUS and pCLV3::LhG4, 6XOP::NLS-eGFP-WUS lines, the coding sequences of eGFP-WUS and NLS-eGFP-WUS were cloned into the pENTR vector (Invitrogen). The recombination reaction was carried out with the 6XOP pzp222 binary vector to create 6XOP::eGFP-WUS and 6XOP::NLS-eGFP-WUS, respectively. The independent transgenic lines carrying these constructs were crossed to the driver lines-pCLV3::LhG4 described in an earlier study (12).

To generate a Dex-inducible amiRNA for silencing WUS, the 21-nt amiRNA sequence against WUS was designed based on guidelines outlined at wmd3.weigelworld.org/cgi-bin/webapp.cgi. The sequence showing no or the fewest possible off-targets was chosen for further construction of the amiRNA. For the amplification of the amiRNA precursor, several rounds of PCR were carried out using the primers listed in Table S1. The final amplification product of the overlapping PCR, coding for the amiR-WUS precursor, was first cloned into the pENTR vector. The LR reaction with pENTR amiR-WUS was set up with the 6xOP-pzp222 binary vector to create 6xOP::amiR-WUS. The independent lines carrying this construct were crossed to the 35S::GR-LhG4 driver line to generate 35S::GR-LhG4;6xOP::amiR-WUS. The number of independent transgenic lines and combinations tested in each case are stated in parentheses at appropriate places in the text.

qRT-PCR.

For qRT–PCR, RNA was isolated from SAMs expressing amiR-WUS that had been treated with Dex or mock treated for 4 or 8 d using the RNeasy kit (Qiagen). cDNA was reverse-transcribed using ThermoScript RT (Invitrogen). qRT-PCR reactions were performed using the sensiMix SYBR kit (Bioline) on a Bio-Rad iQ5 Cycler. Analyses were performed for duplicate samples, and quantification was standardized to ubiquitin (UBQ10) mRNA levels. Primers used in this study are listed in Table S1.