Threshold-dependent transcriptional discrimination underlies stem cell homeostasis

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.

Keywords: CLAVATA3; WUSCHEL; cis-regulatory module; gene regulation; shoot apical meristem.

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.

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.)

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.

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.)

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. 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.

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.)