Two-Step Regulation of a Meristematic Cell Population Acting in Shoot Branching in Arabidopsis

Shoot branching requires the establishment of new meristems harboring stem cells; this phenomenon raises questions about the precise regulation of meristematic fate. In seed plants, these new meristems initiate in leaf axils to enable lateral shoot branching. Using live-cell imaging of leaf axil cells, we show that the initiation of axillary meristems requires a meristematic cell population continuously expressing the meristem marker SHOOT MERISTEMLESS (STM). The maintenance of STM expression depends on the leaf axil auxin minimum. Ectopic expression of STM is insufficient to activate axillary buds formation from plants that have lost leaf axil STM expressing cells. This suggests that some cells undergo irreversible commitment to a developmental fate. In more mature leaves, REVOLUTA (REV) directly up-regulates STM expression in leaf axil meristematic cells, but not in differentiated cells, to establish axillary meristems. Cell type-specific binding of REV to the STM region correlates with epigenetic modifications. Our data favor a threshold model for axillary meristem initiation, in which low levels of STM maintain meristematic competence and high levels of STM lead to meristem initiation.


A Meristematic Cell Population in Leaf Axils
Previous in situ hybridization results showed STM expression in all stage leaf axils from examination of fixed samples, but it remains unclear if a continuous STM-expressing cell population exists during development [10,11,16]. The STM-expressing cells are closer to the meristem side of the boundary in young leaf primordia, and are closer to the leaf side of the boundary in older leaves. Thus, it has been proposed that the initial STM-expressing cells may create a separation while their neighboring cells re-differentiate as AM progenitor cells [10]. To better resolve the origin of STM expressing cells, we used live-cell imaging to determine if a continuous STM-expressing cell population exists in the leaf axil. To this end, we imaged axils of living leaf primordia that we isolated from the shoot apex and maintained in culture. As shown previously, cultured leaf primordia (P 6 and older) efficiently initiated AMs in the absence of exogenous phytohormone [9], which is distinct from de novo organogenesis [29].
By live-imaging the expression of a functional pSTM::STM-Venus reporter in P 6 and older leaves (Fig 1A), we found that cells with continuous STM-Venus expression are AM progenitors. A recent study has shown that this reporter line can fully complement stm mutation, and the enhanced boundary expression reflects the endogenous STM expression [30]. For the P 8/9 leaf primordium, which has the fewest STM-expressing cells of all the stages (see below), we observed the STM-Venus signal only in a continuous cell mass close to the incision line ( Fig  1B). The number of cells with STM-Venus signal initially decreased after 24 h of culture ( Fig  1C), but partially recovered after 48 h of culture ( Fig 1D). Starting from 72 h of culture, following a series of rapid cell divisions (Fig 1E), these cells organized into a meristem with new leaf primordia (Fig 1F and 1G). Occasionally, a few cells without initial STM-Venus signal at the first time point, but next to one or more STM-positive cells, showed detectable signal. These cells could have initially had low-level STM-signals below our detection threshold. Alternatively, their STM expression could be due to STM proteins trafficked from neighboring STMexpressing cells [31].
In early stage leaf axils, STM-Venus also persisted in the boundary region. In tissue sections, the number of STM-expressing cells gradually decreased during leaf primordia maturation from P 3 ( Fig 1H) to P 9 ( Fig 1I). Later, in P 10 and older leaves, the number of STM-expressing cells and level of STM expression increased significantly (Fig 1J), which is consistent with the liveimaging results. Quantitative measurement of leaf axil STM-Venus fluorescence intensity confirmed this STM expression dynamic pattern during leaf maturation. In particular, P 8 or P 9 have the lowest intensity in Ler, which significantly increase in the next developmental stage (Fig 1K-1N and S1A-S1C Fig). There is a small variation between individual plants with either P 8 or P 9 having the lowest STM expression, which is in line with a previous morphological analysis (10).
In addition, we also used reporter lines to follow expression of the shoot meristem central zone stem cell marker CLAVATA3 (CLV3), the shoot meristem organizing center marker WUSCHEL (WUS), and the pericycle-like cell marker J0121, which marks progenitor cells for regeneration [3,32,33]. We did not detect CLV3 or WUS expression in young Ler leaf axils until the twelfth-youngest primordium (P 12 , S1D-S1F Fig) and J0121 was not expressed at all in the leaf axil during axillary bud formation (S1G Fig), suggesting that their corresponding cell identities were not maintained.

STM-expressing Cells Are Required for AM Initiation
We next tested whether AM initiation required the STM-expressing cells. It has been reported that mild stm alleles have more active branching and show an 'abort-retry' mode of growth [34][35][36]. However, because SAM termination (in stm mutants) promotes outgrowth of axillary buds, files of cells at stage P 3 (H), later decreased to 4 files at P 8 (I), and then expanded to more cells at P 11 prior to AM initiation (J). The arrow in (J) highlights the bulged meristem. Note the expression of STM was substantially higher in the boundary than in the SAM, as shown before [30]. (K-M) Continuous transverse sections through a vegetative shoot apex, showing expression of STM-Venus (green) in leaf axils (arrowheads). Sections are ordered from most apical (K) to most basal (M); approximate distance (in micrometers) from the summit of the SAM to section is given in the bottom left-hand corner of each image. Note a significant increase in STM-Venus signal between P 9 and P 10 (L and M). (N) Relative fluorescence intensity of STM-Venus at leaf axils from P 1 to P 12 . 3 replications were done by analysis the transverse sections like in (K-M). Arrows highlight the lowest value at P 8/9 . Bars = 50 μm. branch growth may not reflect axillary bud formation. To show if STM functions in AM initiation, we analyzed the pattern of axillary bud formation in plants carrying the weak stm-bum1 allele, which can still form leaves from a partially functional SAM [37]. We found a dramatic reduction in the number of axillary buds in stm-bum1 plants, with 60% (242 out of 405) of leaves lacking axillary buds, which is distinct from Col-0 wild-type plants (Fig 2A). This reduction in axillary buds is more dramatic for rosette leaves (91% leaves lack buds). In contrast to the wild type (Figs 1I and 3A-3C), leaf axil cells in stm-bum1 plants are enlarged (Fig 3D-3F), suggesting that the leaf axil cells have undergone differentiation. On the other hand, stm-bum1 plants have reduced apical dominance, resulting in the reported enhanced branching phenotype [34][35][36].
To test if AM initiation requires STM-expressing cells, we applied laser ablation. When we ablated the cells adjacent to the STM-expressing cells, AMs initiated normally from the ablated leaf axil region (Fig 2B-2D, 16 out of 19), showing that ablation per se does not abolish AM initiation. However, after ablation of most cells within the STM-expressing cell mass in both the epidermis and internal cell layers, AMs could not initiate (Fig 2E-2G, 10 out of 10). Note that this result is in contrast to observations in shoot and root apical meristems [38][39][40][41], where neighboring cell fate can switch after ablation. Furthermore, we observed that AMs did not initiate from cultured leaves if we removed the proximal portion of the petiole containing the STM-expressing cells (S2A-S2C Fig). Taken together, our data strongly suggest that AM initiation requires the STM-expressing cells as AM progenitor cells.

Maintenance of STM Expression Requires the Leaf Axil Auxin Minimum
We next asked what regulates the maintenance of STM expression in the leaf axil. We have recently shown that the AM progenitor cells also maintained a low auxin level [9,27], suggesting that maintenance of STM expression may require the leaf axil auxin minimum. To test this hypothesis, we analyzed STM expression in pCUC2>>iaaM and pLAS::iaaM-en plants, which ectopically accumulate auxin in leaf axils and are deficient in AM initiation [9,27]. We could not detect STM expression in leaf axils of pCUC2>>iaaM plants (Fig 3G-3J). In addition, leaf axil cells in pCUC2>>iaaM plants are enlarged (Fig 3K-3O), suggesting cell differentiation. Similarly, pLAS::iaaM-en plants also have substantially reduced or undetectable leaf axil STM expression and have enlarged leaf axil cells (Fig 3P-3S).
To test if STM expression alone is sufficient for AM initiation, we introduced p35S:: STM-GR into pCUC2>>iaaM plants. In p35S::STM-GR plants [42], dexamethasone (Dex) can induce the nuclear translocation of a STM-glucocorticoid-receptor (GR) fusion protein. We aimed to test if leaf axil cells that have lost STM expression can respond to ectopic STM activity. Firstly, we detected a dramatically increase of STM expression by reverse transcription quantitative PCR (RT-qPCR) in leaf-removed leaf axil-enriched shoot apex tissues ( Fig 3T). In mature leaf axils, we found that, following Dex induction, no axillary bud could form (Fig 3U,  3V and 3Y), highlighting the importance of the low level STM expression for subsequent AM initiation. Similarly, when we introduced p35S::STM-GR into stm-bum1 plants with compromised AM initiation, we found that Dex treatment did not induce axillary buds from mature leaf axils (Fig 3Y). Taken together, these results indicate that the recently identified leaf axil auxin minimum is required to maintain low level STM expression, which is then required for later axillary buds formation.

AM Initiation Requires REV-dependent Up-Regulation of STM Expression
In contrast to pCUC2>>iaaM and pLAS::iaaM-en plants, we found that STM expression was maintained in the rev-6 mutant (Fig 4A-4E), which also lacks axillary buds [43]. In contrast to   the wild type ( Fig 1K-1M), the expression of STM does not increase in the rev-6 mutant ( Fig  4A and 4B), as it does in wild-type leaf axils, during leaf maturation (compare Figs 1B-1G and 4C-4E). Subsequently, the STM-expressing cells did not undergo active cell division to form a meristem with well-organized structure (Fig 4C-4E). The change of leaf axil STM expression in rev-6 implies that up-regulation of STM expression in P 10 and older leaves requires REV, but maintenance of STM expression does not require REV. Also in contrast to stm-bum1 and pCUC2>>iaaM, our genetic analysis indicates that over-expressing STM can suppress the AM initiation defect of rev-6 mutants (Fig 3W-3Y). Therefore, we conclude that STM expression must not only be maintained in meristematic cells, but also subsequently up-regulated for AM initiation.
To test if REV up-regulates STM expression in a cell-autonomous manner, we imaged REV distribution by using a functional pREV::REV-Venus reporter line [32]. REV-Venus is broadly expressed in the adaxial side of P 8 and younger leaves (Fig 4F-4I), but it is restricted to the center of leaf axils, especially the epidermis (L1) layer, in P 9 and older leaves (Fig 4J-4L). Furthermore, REV has stronger expression in P 9 and older leaf axils than in younger leaf axils. The leaf axil enrichment of REV is consistent with up-regulated STM expression in P 10 and older leaves, suggesting that REV up-regulates STM expression in a cell-autonomous manner.
Overexpressing alleles of REV and related HD-ZIPIII genes can induce ectopic AMs in the abaxial side leaf axils [14,44,45]. By using one such mutant, phavulota-1d (phv-1d), we observed ectopic STM expression in abaxial leaf axils prior to axillary bud initiation (S3B- S3E  Fig). By using transgenic lines overexpressing microRNA-insensitive REV and PHABULOSA (PHB), another related HD-ZIPIII gene, we detected up-regulation of STM expression in leafremoved shoot apex tissues, which are enriched with leaf axils (S3J Fig). Notably, we also detected ectopic auxin minima in abaxial leaf axils by using the auxin concentration sensor DII-Venus [46], whose strong abaxial axil signal indicates low auxin concentrations (S3F- S3I  Fig). Taken together, REV and related HD-ZIPIII proteins can promote STM expression, which, together with auxin minima, promote ectopic AM initiation.

REV Directly Up-Regulates STM Expression
To test if REV directly up-regulates STM expression, we generated functional Dex-inducible pREV::REV-GR-HA rev-6 lines (S4A- S4C Fig). We measured the effect of REV activation on the expression of STM by RT-qPCR. REV activation resulted in rapid elevation of STM mRNA levels within 2 h of treatment, with or without the protein synthesis inhibitor cycloheximide (CHX) (Fig 5A and S4D Fig), strongly suggesting that induction of STM does not require de novo protein synthesis and that STM is likely a direct target of REV. REV activation also triggered in vivo accumulation of STM-Venus, as shown by live-cell imaging (S4E-S4H Fig). Consistent with this, our recent large-scale yeast one-hybrid assay identified REV and related HD-ZIPIII proteins as binding to the STM promoter region [21].
We next performed chromatin immunoprecipitation (ChIP) assays to examine whether REV directly binds to the STM promoter in vivo. We scanned the STM genomic sequence for ATGAT, the conserved binding site for REV [47], and designed primers near identified motifs and other regions (Fig 5B). In both shoot apex tissues enriched with leaf axils and inflorescence tissues, we found that REV-GR-HA strongly associated with the regions containing multiple ATGAT motifs, but only after Dex treatment, by using antibodies against GR or HA (Fig 5C and  S4I Fig). In addition, REV-GR-HA weakly associated with seven other upstream ATGAT motifcontaining regions. A transient transfection assay in protoplasts further confirmed that REV bound to multiple ATGAT motif-containing STM genomic regions, especially the ones close to the start codon, and up-regulated STM expression (Fig 5E). These newly transformed pSTM:: LUC constructs would lack epigenetic modifications that might interfere with REV binding.
Cell Type-Specific REV Binding to the STM Region REV is widely expressed in young leaves, including in the adaxial domain and vascular tissues [43], but only up-regulates STM expression in boundary tissues-enriched samples (Fig 6A), as previously shown [11]. Furthermore, a recent ChIP-seq analysis did not identify STM as a REV-binding target in whole seedlings [47]. By using the same antibodies and protocol, we found that REV associated with the STM genomic region only in vegetative shoot apex and inflorescence tissues, but not in leaf axil region-removed mature leaves (compare Fig 5C and  5D, only data from vegetative shoot apex tissues was shown).
In animals, lineage-specific epigenetic modification of transcription factor genes leads to the fixation of stem cell fate [48]. Furthermore, the STM locus was epigenetically silenced in mature leaves containing only differentiated cells [49][50][51]. In mature leaves without STMexpressing leaf axil cells, the chromatin modification H3K27me3, which is associated with transcriptional repression, is highly enriched at the STM locus (Fig 6B and 6D). By contrast, H3K4me2 and/or H3K4me3, which are associated with transcriptional activation, are enriched at the STM locus in inflorescence tissues enriched with organ axils (Fig 6C and 6D). This histone modification pattern implies that epigenetic factors may regulate REV binding to the STM locus. In both animals and plants, the Polycomb Repressive Complex 2 (PRC2) establishes the H3K27me3 mark, which provides a docking site for PRC1 to establish a repressive chromatin configuration [52]. Mutants affecting PRC1 and PRC2 have elevated STM expression [49,50]. To test if the ectopic activation of STM expression requires REV, we introduced rev-6 into PRC mutants. We found that rev-6 mutation partly suppressed ectopic STM up-regulation (Fig 6E). ACT2 coding regions on cDNAs generated from mRNA isolated from Col-0 wild-type inflorescence, stem, mature leaf (without the leaf axil region), root, and petal, respectively. ACT2 was used as a loading control. (B and C) Results of ChIP-qPCR performed on IP with antibodies against H3K27me3 (B) and H3K4me2/3 (C) on chromatin samples extracted from Col-0 wild-type inflorescences and mature leaves. The a-e regions (indicated as in Fig 5B) were assayed. Error bars indicate SD. More controls are shown in (D). (D) ChIP enrichment test by PCR with an anti-H3K27me3 antibody and an anti-H3K4me2/3 antibody using Col-0 wildtype inflorescences and mature leaves, together with total DNA input (input) and no-antibody (mock) controls. An ACT2 promoter region was used as a negative control. (E) Up-regulation of STM expression in mutants affecting PRC1 and PRC2 requires REV. RT-qPCR analysis of STM in whole seedlings of Col-0 wild type and mutants affecting PRC1 and PRC2 w/ or w/o rev-6. The vertical axis indicates relative mRNA amount compared with the amount in wild-type plants. Error bars indicate SD.

A Meristematic Cell Population in the Leaf Axil
Plant cells, especially isolated cells, have amazing developmental plasticity, yet intact plant development follows defined patterning. Within meristems, clonal analysis and root regeneration studies suggest that meristem cells usually lack predictable destinies and that positional control is most important for plant cell fate determination [38][39][40][41]. However, distinct cell lineages emerge at later developmental stages [2,3]. In this study, we show that AM initiation is accompanied by the maintenance of a meristematic cell population, and differentiation of surrounding cells. We traced this cell population in P 6 and older leaves, and confirmed that STMpositive cells at the leaf axil are progenitors of axillary buds. Imaging results indicate that cells usually cannot acquire STM expression de novo (Fig 1B-1G), at least in P 6 and older leaves, indicating the existence of a cell lineage. Further laser ablation results show that these STMpositive cells are necessary for formation of axillary buds, whereas neighboring STM-negative cells are differentiated (Fig 2B-2G).
This leaf axil meristematic cell population relies on positional cues. Our recent studies have shown that an auxin minimum, which is associated with the leaf axil position, is required for AM initiation [9,27]. In the current work, we further demonstrate that the maintenance of the meristematic cells depends on low auxin (Fig 3G-3J and 3P-3S), which is likely determined by positional information. The observation of abaxial auxin minima and STM expression in phv-1d, which forms axillary buds at the abaxial side, also support the importance of positional cues for the maintenance of meristematic cells (S3B-S3I Fig).
Cell fate determination occurs gradually with cell cycle progression in animals [48]. Previous studies focused on cells within shoot meristems and root meristems, and found that cells from different meristematic domains can switch cell fate [38][39][40][41]. These results do not necessarily indicate that cell fate determination does not occur after additional rounds of cell cycle progression. In fact, root meristem regeneration does not occur if additional tissue beyond the meristematic zone has been removed [53]. If one assumes that cell fate determination takes place after more cell cycles in plants (than in animals), it would be conceivable that: i) cells within or close to meristems remain meristematic and can reverse cell fate, and ii) certain nondividing or slow-dividing cell types in differentiated organs may maintain a meristematic status while their neighboring cells become fully differentiated and can no longer reverse to a meristematic status. Because boundary cells are non-dividing or slow-dividing cells [54], leaf axil cells can maintain an undifferentiated status while their neighboring differentiated cells cannot.

Cell Fate Determination in Differentiated Cells
Previous studies have shown that overexpression of STM (or the related gene KNAT1) alone [55][56][57], or in combination with ectopic WUS [42,58], induces ectopic meristems. The effect of ectopic STM is highly dependent on tissue stage. As shown in a previous works [57], leaf primordia older than P 10 are not competent to ectopic STM activity (S5A Fig). For younger leaf primordia, ectopic meristems initiated only from leaf axils and the adaxial side of the proximal portion of leaf blades, especially in the sinus region between the blade and the petiole (55 out of 72 P 7 to P 9 , i.e. 72%, S5B- S5G Fig). Thus, STM alone is not sufficient to induce meristems from most cells, but is sufficient in presumably undifferentiated cells.
Similarly, we found that ectopic STM activity was insufficient to rescue axillary bud formation defects in mature leaf axils of pCUC2>>iaaM, pLAS::iaaM-en, or stm-bum1 plants, which have lost low level STM expression in leaf axil cells. By contrast, ectopic STM activity was sufficient in rev-6 maintaining low level STM expressing cells (Fig 3W and 3X). Therefore, STM expression and proper cell fate are both required for AM initiation. In tomato, recent work showed that ectopic meristems may form at the base of leaflets, where KNOX genes express, and this requires the AM initiation pathway [59]. This finding again supports the model that cell competency is required for shoot meristem formation. Epigenetic regulation is involved in the maintenance of meristematic cell competency, and STM expression serves as a marker for cell competency.

A Threshold Model for AM Initiation
Our data support the detached model for AM initiation, in which meristematic cells are detached from the SAM. When leaf primordia (P 1 ) separate from the SAM, boundary cells keep STM expression as SAM cells (Fig 1H). From P 1 to P 5 stages, the number of STM-expressing cells continues to decrease (Fig 1H-1N). Although the exact clonal relationship of early STM-expressing cells remains unknown, our data suggest that many STM-expressing cells differentiate but some may maintain STM expression. In P 6 and older leaf primordia, we used live-cell imaging to track the STM-expressing cell population (Fig 1B-1G). Notably, all cells in the enlarged STM-expressing domain are progeny of cells with previous STM expression. Our data also explain the ectopic axillary formation of phv-1d mutants. Ectopic axillary buds form in the abaxial side away from the SAM, providing key support to the de novo model [14]. PHV is highly similar to REV, and can bind to the STM promoter region in yeast. It is conceivable that ectopic PHV expression in phv-1d would result in ectopic STM expression (S3E Fig), resulting in ectopic meristematic cells in the abaxial leaf axil that initiate ectopic axillary buds.
Furthermore, our results support a 'threshold model' in which maintenance of low levels of STM expression is required but not sufficient for AM initiation, and subsequent elevated expression of STM would induce AM initiation (Fig 7). Leaf axil cells show low auxin-dependent low levels of STM expression starting at leaf primordium initiation. The early low level STM expression is required for later AM formation (Fig 3). In addition, cells lost STM expression are no longer sensitive to ectopic STM activities at a later stage. Before AM initiation, STM is up-regulated in the center of the leaf axil, triggered by REV activation, which in turn requires LAS activity [11]. We also show that this up-regulation is a local event (Fig 4I-4L), and it depends on prior, maintained STM expression (S4E-S4H Fig). We further show that REV binding to the STM promoter is tissue-specific (Fig 5), and that epigenetic regulation may underlie this cell type specificity (Fig 6), suggesting that the binding requires permissive chromatin statues. Our data favor the idea that the up-regulation of STM is causal for AM initiation, rather than a consequence of a newly formed AM, because the expression of WUS and CLV3 are still missing at the stage of initial STM up-regulation.
Leaf culture followed a previously described protocol [9]. Briefly, seedlings were grown in MS medium under short-day conditions for 15 d after seed stratification. Leaves between P 5 and P 11 were then detached from seedlings, laid flat on MS medium supplemented with 0.5 mg/L folic acid and 100 mg/L inositol, and grown for up to 30 d under the same conditions.
The pREV::REV-GR-HA construct was made by replacing the endogenous stop codon of TAC clone JAtY80N08 covering the REV genomic region with a GR-HA sequence using recombineering [61]. The construct was then transformed into rev-6 plants. Over 20 transgenic lines were obtained and lines with stringently inducible rescue phenotypes were used.

Confocal Microscopy, Optical Microscopy, and Scanning Electron Microscopy
Confocal microscopy images were taken with a Nikon A1 confocal microscope. Samples were either live-imaged or fixed and sectioned as previously described [9]. Excitation and detection wavelengths for GFP, Venus, and DsRed were as previously described [9,32]. To detect FM4-64 and PI staining, a 514 nm laser line was used for excitation and a 561 nm long-pass filter was used for detection. The modified pseudo-Schiff-PI (mPS-PI) staining was performed as described and a 488 nm laser line was used for excitation and emission was collected at 520-720 nm [62]. DAPI staining was excited at 405 nm and detected in the 425-475 nm. Autofluorescence was excited at 488 nm or 514 nm and detected in the 660-700 nm range.
Optical photographs were taken with a Nikon SMZ1000 stereoscopic microscope or an Olympus BX60 microscope equipped with a Nikon DS-Ri1 camera. Scanning electron microscopy was performed using a Hitachi S-3000N variable pressure scanning electron microscope after standard tissue preparation [9].

Laser Ablation
Laser ablations were performed on a Nikon A1 confocal microscope equipped with an Andor MicroPoint laser system consisting of a pulsed 440 nm nitrogen laser. We adjusted a variable neutral density filter to attenuate the output laser to limit damage to targeted cells, as assessed by confocal imaging. The observation of cell collapse was used to confirm successful ablation (Fig 2C and 2F).

RT-PCR and RT-qPCR
Total RNA was extracted from leaves, shoot apex tissues, or inflorescences (~6 d after bolting) of 12 plants using the AxyPrep Multisource RNA Miniprep kit (Corning). For shoot apex tissues enriched for leaf axils, leaves were manually removed from 25 d plants grown under short day conditions. For induced meristems, total RNA was extracted from leaf sinus tissues using the RNAqueous-4PCR kit (Life Technologies). First-strand cDNA synthesis was performed with 2 μg total RNA using TransScript One-Step gDNA Removal and cDNA synthesis Super-Mix (TransGen), or with 300 ng total RNA using SuperScript III reverse transcriptase (Life Technologies), and 22-mer oligo dT primers according to the manufacturer's instructions. RT-PCR analysis was performed in a 20 μL reaction using Taq DNA polymerase (TianGen) and gene-specific primers (S1 Table). Reverse transcription quantitative PCR (RT-qPCR) was performed on a Bio-Rad CFX96 real-time PCR detection system with the KAPA SYBR FAST qPCR kit (KAPA Biosystems). Relative expression by RT-qPCR was normalized to TUB6 (At5g12250). Gene-specific primers (S1 Table) were used to amplify and detect each gene. Error bars of RT-qPCR experiments in Figures are derived from three independent biological experiments, each run in triplicate.

Chromatin Immunoprecipitation
Chromatin immunoprecipitation (ChIP) experiments were performed according to published protocols [28,49]. Leaf axil-enriched shoot apex tissues (under short day conditions) or inflorescences (under long day conditions) of approximately 4-week-old Col-0 wild-type or pREV:: REV-GR-HA rev-6 plants were used. Plant material (800 mg) was harvested and fixed with 1% (v/v) formaldehyde under vacuum for 10 min. Nuclei were isolated and lysed, and chromatin was sheared to an average size of 1000 bp by sonication. The sonicated chromatin served as input or positive control. Immunoprecipitations were performed with a polyclonal antibody against GR (Affinity Bioreagents, PA1-516), a monoclonal antibody against HA (Beyotime, AH158), a polyclonal antibody against H3K27me3 (Millipore, 07-449), or a polyclonal antibody against H3K4me2/3 (Abcam, ab8580). The precipitated DNA was isolated, purified, and used as a template for PCR. RT-PCR was performed as described above (S1 Table). The data are presented as degree of enrichment of STM genomic fragments. The amount of precipitated DNA used in each assay was determined empirically such that an equal amount of ACT2 (At3g18780) was amplified. Two independent sets of biological samples were used.

Protoplast Transient Expression Assay
To produce the effector constructs, full-length REV was amplified from Arabidopsis cDNA and inserted into the pBI221 vector to generate pBI221-AP1. To generate STM promoter-driven LUC reporter genes, STM promoter regions were amplified from Arabidopsis genomic DNA. PCR fragments were inserted into the corresponding sites of the YY96 vector to produce pSTM::LUC constructs (Fig 5E, and S1 Table for primers).
Isolation of Arabidopsis protoplasts and PEG-mediated transfection were performed as described previously [28]. The reporter construct, effector plasmid, and a p35S::GUS construct (internal control) were co-transformed into protoplasts. After transformation, the protoplasts were incubated at 23°C for 12-15 h. The protoplasts were pelleted and resuspended in 100 μL of 1 × CCLR buffer (Promega). For the GUS enzymatic assay, 5 μL of the extract was incubated with 50 μL of 4-methylumbelliferyl-β-d-glucuronide assay buffer (50 mM sodium phosphate pH 7.0, 1 mM β-d-glucuronide, 10 mM EDTA, 10 mM β-mercaptoethanol, 0.1% sarkosyl, 0.1% Triton X-100) at 37°C for 15 min, and the reaction was stopped by adding 945 μL of 0.2 M Na 2 CO 3 . For luciferase activity assays, 5 μL of the extract was mixed with 50 μL of luciferase assay substrate (Promega), and the activity was detected with a Modulus Luminometer/Fluometer with a luminescence kit. The reporter gene expression levels were expressed as relative LUC/GUS ratios. Error bars in Fig 5E are derived from three independent biological experiments, each run in triplicate. showing how fluorescence intensity are measured. According to the 3D-version images, appropriate regions of STM expression at the leaf axil (red circles) are selected on sum projected images in which all the pixels are added. Note to avoid regions with positive STM-expression but not belonging to the leaf axils. Background intensity is determined by selecting a region (blue circles) next to the STM-expression region and multiplying the mean fluorescence of background readings by the area of STM-expression region. For each leaf axil, the corrected total cell fluorescence (CTCF) was then calculated by subtracting background fluorescence density from integrated density. To make data comparable between samples, relative value was transferred into from the above absolute value by setting value of P 9 at 1. Close-up of rosette leaf axils in Col-0 wild-type, rev-6, and pREV::REV-GR-HA rev-6 after mock or Dex treatment. After germination, Dex was daily applied to all leaf axils. Note the presence or absence (arrows) of an axillary bud. (B) Schematic representation of axillary bud formation in leaf axils of Col-0 wild-type plants, rev-6 plants, and pREV::REV-GR-HA rev-6 plants after mock or Dex treatment. The thick black horizontal line represents the border between the youngest rosette leaf and the oldest cauline leaf. Each column represents a single plant and each square within a column represents an individual leaf axil. The bottom row represents the oldest rosette leaf axils, with progressively younger leaves above. Green indicates the presence of an axillary bud, yellow indicates the absence of an axillary bud, and red indicates the presence of a single leaf in place of an axillary bud in any particular leaf axil. (C) Nuclear accumulation of the REV-GR-HA fusion protein after mock or Dex treatments. Protein gel blot detection of the REV-GR-HA fusion protein using crude nuclear extracts isolated from Col-0 wild-type and rev-6 plants, and pREV::REV-GR-HA plants after mock or Dex treatment. Samples were harvested 1 d after treatment. (D) RT-qPCR analysis of STM expression in pREV::REV-GR-HA rev-6 vegetative shoot apex tissues enriched with leaf axils after mock and Dex treatment. The vertical axis indicates relative mRNA amount after Dex treatment compared with the amount after mock treatment. Error bars indicate SD. (E-H) In vivo activation of STM expression by REV in pREV:: REV-GR-HA rev-6 plants. Reconstructed view of the L1 layer of a leaf axil (as shown in Fig 1B) with STM-Venus (green) expression and FM4-64 stain (red) showing the location and lineage of AM progenitor cells, with (E) being the first time point before Dex induction and elapsed time in (F-H). Selected progenitor cells are color-coded, and the same color has been used for each progenitor cell and its descendants. Arrowheads in (E-H) highlight the cut edge. (I) Enrichment of STM promoter fragment (as indicated in Fig 5B) in Dex induced pREV:: REV-GR-HA rev-6 plants. ChIP was carried out with anti-HA or anti-GR antibody, together with total DNA input (input) and no-antibody (mock) controls. STM promoter fragment 1 (see Fig 5B) was analyzed using inflorescence tissues. An APETALA1 (AP1) promoter region was used as a positive control for REV binding [47], and an ACT2 promoter region was used as a negative control. Bars