Altered number and size of pavement cells in the leaves with modified TCP activity

The class II TCP genes are heterochronic regulators of leaf development and their function has been implicated in the repression of cell division [22] and the promotion of differentiation [17]. In order to examine the specific role of the five redundant, miR319-targeted, class II TCP genes–TCP2, 3, 4, 10 & 24—in cell division/ differentiation of Arabidopsis leaves, we compared the mature first leaf pair and their epidermal cells of wild type (Col-0) with those in the loss-of-function mutants tcp2;4;10 [27] and jaw-D, where all the five TCP transcripts are down-regulated due to the over-expression of miR319a [15]. The tcp2;4;10 and jaw-D leaves measured >1.5 times larger than the Col-0 leaves (Fig 1A), as was reported previously [9,17]. The number of the abaxial epidermal cells in the mutant leaves increased ~2.5-fold (Fig 1B), suggesting that the increased mutant leaf area is due to an excess of cell proliferation. To determine if class II TCP activity is sufficient to repress proliferation, we induced a miR319-resistant, dexamethasone (DEX)-inducible form of TCP4 protein (mTCP4-GR) expressed under ProTCP4 and Pro35S [27]. DEX-grown, homozygous ProTCP4:mTCP4:GR or Pro35S:mTCP4:GR plants in the Col-0 background (hereafter referred to as Col-0;GR or 35S;GR, respectively) showed smaller leaves with reduced cell number (Fig 1C and 1D and S1A–S1D Fig). A more pronounced relative reduction in the final leaf size (by ~70%; Fig 1C and 1E) and cell number (by >80%; Fig 1D and 1F) were observed in the DEX-grown ProTCP4:mTCP4:GR plants in the jaw-D background (hereafter referred to as jaw-D;GR) compared to the mock treatment, demonstrating that TCP4 is a strong cell proliferation inhibitor. Similar negative effect of TCP4 on leaf size and cell number was also observed in three independent transgenic lines where three different gain-of-function forms of TCP4 were expressed constitutively under a leaf primordia-specific promoter (ProBLS:rTCP4:GFP) [17], a ubiquitous promoter (Pro35S:TCP4:3F6H) [39] or the endogenous promoter (ProTCP4:TCP4:VP16) [23] (S1F–S1I Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Perturbation in number and size of pavement cells in leaves with altered activity of miR319-targeted TCP genes.

(A) and (B) Mature first leaves (A, left), average area of their lamina (A, right), and number (B) of their abaxial pavement cells. Sample size (N), 10–12. Scale bar, 2 mm. (C) and (D) Images of mature first leaves (C, left) from plants grown in the absence (M, Mock) or presence (DEX) of 12 μM dexamethasone and average area of their lamina (C, right), and number (D) of their abaxial pavement cells. N, 12–15. Scale bar, 2 mm. Col-0;GR, jaw-D;GR, 35S;GR and 35S;sGR indicate Col-0;ProTCP4:mTCP4:GR; jaw-D;ProTCP4:mTCP4:GR, Col-0;Pro35S:mTCP4:GR and Col-0;Pro35S:TCP4:GR, respectively. (E) and (F) Mature leaf area (E) and number of pavement cells (F) from DEX-treated Col-0;ProTCP4:mTCP4:GR and jaw-D;ProTCP4:mTCP4:GR plants relative to respective mock-treated values. Note that the relative value for jaw-D;ProTCP4:mTCP4:GR is smaller than that of Col-0;ProTCP4:mTCP4:GR since the denominator (mock value) for the former is more than that for the latter. (G) Average pavement cell area of leaves shown in (A). (H) and (I) Outline of the abaxial pavement cells (H) of leaves shown in (C) and their average area (I). Scale bar, 100 μm. For (G) and (I), total 125–150 cells per leaf at different regions were measured and averages from 5–8 leaves are shown. (J) and (K) Frequency distribution of pavement cell size in mature first leaves based on their size, of indicated genotypes (J), and of jaw-D;ProTCP4:mTCP4:GR plants grown without (Mock) or with (DEX) 12 μM dexamethasone (K). (L) Mature rosette leaves from 32-day old jaw-D;ProTCP4:mTCP4:GR plants grown in mock control (Mock), in continuous 12 μM dexamethasone (DEX), in dexamethasone for first 10 days and then transferred to mock (D→M), and in mock for first 10 days and then transferred to dexamethasone (M→D). The numbers indicate leaf positions. Scale bar, 1 cm. Error bars, * and ns indicate SD, p <0.05 and not significant, respectively. Unpaired Student’s t-test was used to determine significant differences.

https://doi.org/10.1371/journal.pgen.1007988.g001

Concomitant with the increased number, the area of mature pavement cells reduced to almost half in jaw-D and tcp2;4;10 leaves (Fig 1G and S2A Fig). Induction of TCP4 activity, however, had little effect on the final cell area in the Col-0;GR and 35S;GR leaves (Fig 1H and 1I and S1E Fig) and merely restored the cell size defect found in the jaw-D;GR leaves to the wild-type level (Fig 1I and S2B Fig). Pavement cell area also remained unaltered in the three other dominant lines of TCP4, ProBLS:rTCP4:GFP, Pro35S:TCP4:3F6H and ProTCP4:TCP4:VP16 (S1J and S1K Fig). These results suggest that the class II TCP proteins are required for pavement cell maturation during leaf development but are not sufficient to promote pavement cell expansion per se, even though they do so in the hypocotyl [27]. Contrary to the miR319-resistant inducible form of TCP4, constitutive expression of a miR319-susceptible form of TCP4 protein in the homozygous Pro35S:TCP4:GR plants (hereafter referred to as 35S;sGR) showed no effects on leaf size, cell number and pavement cell area (Fig 1C, 1D and 1I), demonstrating that the effects of mTCP4:GR induction described above resulted specifically from miR319 activity on the class II TCP transcripts and not from artifacts due to GR fusion or to the site of insertion [27].

The requirement of TCP genes for developmental cell maturation was also apparent in the reduced differentiation status of the pavement cells in the TCP loss-of-function leaves. The abaxial surface of mature Col-0 leaf is composed mostly of large epidermal cells with jigsaw shape (Fig 1H, S2A and S2C Fig), a characteristic marker for differentiation [17,40]. By contrast, mature tcp2;4;10 and jaw-D leaves were composed mostly of smaller cells (Fig 1J, S2A and S2C Fig), a differentiation defect that was totally rescued by TCP4 induction (Fig 1K). However, induction of TCP4 in the Col-0 background did not alter the cell size distribution (S2D Fig).

Kinematic analysis of TCP-mediated leaf growth

The results described above suggest that the class II TCP proteins restrict cell number and promote cell maturation during leaf development. To examine their role in cell proliferation/ maturation at early growth phase, we compared the kinematics of growth of the first pair of jaw-D;GR leaves under TCP4 inductive and non-inductive conditions. Col-0 leaf blade expanded slowly at an early growth stage, from 6–10 days after stratification (DAS), during which the average area of the abaxial pavement cells remained small (~500 μm² or less) and their number increased exponentially (Fig 2A–2C, S3A and S3B Fig). At 10 DAS, the blade acquired ~10% of its final area and ~75% of its final pavement cell number. Later, the blade expanded at an exponential rate up to 16–18 DAS, before reaching a plateau at ~22 DAS (Fig 2A and S3A Fig). The pavement cells also expanded faster during this period, and continued to do so beyond 22 DAS (Fig 2C and S3B Fig). The number of pavement cells reached its maximum at the mid-log phase of blade expansion and no further proliferation was observed beyond 12 DAS (Fig 2B), at which stage the blade acquired approximately half of its final area. These kinematics results are in general agreement with the earlier reports on Arabidopsis leaf published by other laboratories [5,41].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Kinematic growth analysis of leaves with altered TCP4 activity.

(A) to (F) Kinematic growth parameters of the first leaf pair of jaw-D;ProTCP4:mTCP4:GR plants grown in the absence (Mock) or presence (DEX) of 12 μM dexamethasone are compared to the wild type (Col-0) control. (A) Average leaf area. Magnified data at 6 days after stratification (DAS) for Col-0 (open bar) and mock-grown jaw-D;GR (filled bar) leaves are shown in inset. (B) Average number of abaxial pavement cells per leaf (note that the Y-axis scale is in log₂ scale). (C) Average area of abaxial pavement cells. (D) and (E) Frequency distribution of smaller (D, <1500 μm²) and larger (E, >3000 μm²) pavement cells are shown. For each time point, total 30–40 cells per leaf at base, mid or tip [5,41] were measured and averages from 5–7 leaves are shown. (F) Stomatal index expressed as percentage of total epidermal cells. N, 9–12 leaves. Error bars indicate SD and * indicate p <0.05.

https://doi.org/10.1371/journal.pgen.1007988.g002

The overall trend of the growth kinematics of jaw-D leaf blade was by and large similar to Col-0, with a slow (6–10 DAS), exponential (10–18 DAS) and saturating (18–22 DAS) phase. However, the jaw-D blades started expanding at a rate faster than that of Col-0 from the earliest stage measured (6 DAS; inset in Fig 2A), and consequently its area remained higher throughout the growth. The increased jaw-D blade area was due to excess number, and despite reduced size, of the pavement cells. The jaw-D cells accumulated at a faster rate than Col-0, but the proliferation ceased at almost similar growth stage (~12 DAS) in both the genotypes, resulting in ~2.5 times more cells in the mature jaw-D leaf compared to Col-0 (Figs 1B and 2B). As in Col-0, the size of the jaw-D pavement cells remained small in the slow phase (6–10 DAS) and then increased exponentially until 16 DAS, before saturating rather abruptly (Fig 2C). However, the average area of jaw-D pavement cells remained 1.6 ± 0.16 fold smaller than that in Col-0 throughout the growth phase of the leaf (S1 Table).

All the cellular defects of the jaw-D leaf were rescued by continuous induction of TCP4 throughout the growth phases (Fig 2A–2C and S3A Fig). The DEX-grown jaw-D;GR blade expanded at the same rate as Col-0 until 12 DAS and then at a slower rate, before ceasing to expand at 18 DAS to a final area that was smaller than the Col-0 blade (Fig 2A and S3A Fig). Upon TCP4 induction, the pavement cells accumulated in the jaw-D;GR blade for almost the same duration (till ~12 DAS) as in Col-0, but at a slower rate, resulting in fewer cells at maturity compared to Col-0 (Figs 2B and 1D). While both blade area and cell number were smaller than the Col-0 values in the jaw-D;GR leaves grown under DEX, the average size of the pavement cells was rescued up to the Col-0 value at all stages of leaf growth (Figs 2C and 1I).

The growth kinematics described above shows that class II TCP activity is essential for reducing the total number, and for increasing the average size, of the pavements cells during leaf development. A growing leaf blade, however, is composed of cells of heterogeneous size and proliferation status; the smaller, dividing cells are located at the proximal end and the larger, differentiating cells towards more distal side [4]. To examine if TCP4 activity is specific to a stage of cell maturation, we determined the proportions of smaller (arbitrarily defined as <1500 μm² in area) and larger (>3000 μm²) pavement cells in expanding Col-0 lamina (S2C Fig) and compared with jaw-D. During early growth (6–8 DAS), nearly all Col-0 pavement cells measured <1500 μm², and their proportion sharply declined to ~30% at 22 DAS as the blade expanded (Fig 2D and S3C Fig). The larger cells started appearing at 10 DAS and steadily increased to ~40% at 22 DAS (Fig 2E). The fraction of the smaller cells in the jaw-D;GR blade without TCP4 induction, however, declined at a slower rate and stabilized at >60% at 16 DAS (Fig 2D and S3D Fig). Consequently, the jaw-D;GR larger cells increased in proportion rather slowly and reached the maximum value of little over 10% at 16 DAS (Fig 2E). TCP4 induction in the jaw-D;GR blade restored the proportions of smaller and larger cells close to the Col-0 levels at all growth stages (Fig 2D and 2E and S3E Fig).

Taken together, these kinematic results suggest that a major function of the class II TCP proteins is to convert the smaller pavement cells to larger cells, possibly by promoting their proliferation to differentiation competence. The differentiation-inducing activity of the TCP proteins is also apparent in the altered stomatal index (the proportion of epidermal cells assuming stomatal lineage), a measure of leaf maturation [41], in the mutant leaves. 20.5 ± 1.2% Col-0 epidermal cells assumed stomatal identity at 6 DAS, a value that steadily increased during leaf maturation and reached to a maximum value of 37.4 ± 2.6% at 22 DAS (Fig 2F). However, only 13.8 ± 2.7% jaw-D;GR cells were converted to stomatal lineage at 6 DAS, and this value increased to a maximum of 26.7 ± 2.8% at 22 DAS. Induction of TCP4 activity in the jaw-D;GR leaves nearly restored the defect in stomatal index to wild-type level.

Temporal control of leaf growth by TCP4

All jaw-D;GR rosette leaves were smaller with smoother margin when grown with continuous presence of DEX (Fig 1L), indicating that TCP4 induction is active throughout the vegetative growth. However, when jaw-D;GR seedlings were grown under inductive condition for first 10 days and then transferred to non-inductive medium, the first few leaves resembled those grown under continuous DEX, but the leaves that emerged later showed progressively more jaw-D-like phenotype. Conversely, when the seedlings were grown under non-inductive condition for first 10 days and then transferred to inductive condition, the first few leaves resembled jaw-D leaves but the later leaves were smaller with smoother margin (Fig 1L). This suggests that DEX application can be used to induce TCP4 activity at any point of time during plant development.

To determine the developmental timing of TCP4 function during leaf maturation, we induced TCP4 activity in the jaw-D;GR leaves after growing them in the non-inductive medium for various number of days after stratification (4 to 15 DAS; Fig 3 and S4A Fig) and compared their blade area, cell number and cell size at maturity (29 DAS). Without TCP4 induction, the jaw-D;GR lamina grew to 32.6 ± 2.5 mm² that contained 17450 ± 1336 pavement cells of average area 1870 ± 170 μm² (Fig 3). On the other hand, the lamina under continuous TCP4 induction grew to 12.0 ± 1.7 mm² with much fewer cells (4130 ± 579) of the average area 2920 ± 140 μm². The blade area progressively increased from 12.0 ± 1.7 mm² to the non-inductive value of 32.6 ± 2.5 mm² when TCP4 was induced at 4, 5 or 6 DAS, even though their final cell numbers were much fewer than the non-inductive value. When TCP4 was induced at 6 DAS, its blade area at maturity was close to the jaw-D value (30.4 ± 2.3 mm²), even though its cell number was almost half of that in jaw-D leaf (Fig 3A–3C). This is because the defect in the pavement cell size was rescued in these leaves to the wild-type value (3130 ± 170 μm²) (Fig 3D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Temporal induction of cell differentiation by TCP4.

TCP4 activity was induced at indicated DAS in the jaw-D;ProTCP4:mTCP4:GR seedlings by shifting them from mock to dexamethasone (12 μM)-containing medium and the mature first leaf at 29 DAS was analyzed. (A) Images of mature first leaves after shifting them from mock to DEX at indicated DAS. Leaves grown in continuous absence (Mock) or presence (DEX) of dexamethasone are shown as controls. (B) to (D) Average leaf area (B), and average number (C) and size (D) of abaxial pavement cells of leaves that are shown in (A). N, 12–15. For (D), total 120–140 cells per leaf were measured and averages from 5–7 leaves are shown. Mock and DEX (gray filled circles) indicate values for plants continuously treated with either mock solvent or dexamethasone, respectively. Broken lines are drawn through the values of the mock-treated controls. Error bars indicate SD. Scale bar, 5 mm.

https://doi.org/10.1371/journal.pgen.1007988.g003

When TCP4 was induced at 7, 8, 9 or 10 DAS, the final jaw-D;GR blade area progressively increased to a maximum of 46 mm², nearly 1.5-fold bigger than the jaw-D blade (Fig 3A and 3B and S5A and S5B Fig). When TCP4 was induced at 10 DAS, the mature first leaf had pavement cell number (15420 ± 1220) that was close to the jaw-D value, but the cells (2940 ± 240 μm²) were larger than the jaw-D cells (1870 ± 170 μm²), ultimately yielding a mature leaf that was bigger than even the jaw-D leaf. It may be noted here that the jaw-D leaves were at the peak of its exponential phase of proliferation at 10 DAS, when almost all pavement cells were smaller in size (<1500 mm²; Fig 2B and 2D). When TCP4 was induced at 11, 12, 13, 14 and 15 DAS, the mature blade area progressively reduced to the jaw-D value (Fig 3A and 3B). At 15 DAS of TCP4 induction, the blade area, cell number and cell area were comparable to the respective jaw-D values, demonstrating that TCP4 induction at 15 DAS or later failed to rescue the size defect of the jaw-D pavement cells. At 15 DAS, the jaw-D cells ceased to proliferate and the proportion of the smaller cells are at its minimum level (Fig 2B and 2D).

When TCP4 was induced at various developmental stages of Col-0;GR and 35S;GR leaves, similar effect on final blade area was observed (S5C–S5E Fig). Leaves of both genotypes grew to ~26 mm² under non-inductive condition and to 9.2 ± 2.0 mm² (Col-0;GR) and 12.7 ± 1.0 mm² (35S;GR) upon continuous TCP4 induction (Fig 1C and S5C–S5E Fig). When TCP4 was induced at 4, 6, 8, 9 and 10 DAS in these plants, the final blade area progressively increased to the un-induced value but, unlike in jaw-D;GR, never surpassed it. This is consistent with the observation that TCP4 failed to promote pavement cell expansion beyond the wild type level (Figs 1I and 3D). On the contrary to the Col-0;GR and 35S;GR leaves, when DEX-induction was performed in the expanding 35S;sGR blades, the final blade area remained more or less unaffected at all stages of TCP4 induction (S5C and S5F Fig), implying that the collective effects of TCP4 induction on jaw-D;GR, Col-0;GR and 35S;GR leaf growth dynamics described above resulted from miR319 activity on the TCP4 transcript.

These results, together with those shown in Fig 2, strongly suggest that TCP4 promotes maturation in proliferating pavement cells, perhaps by committing them to differentiation, within the developmental window of the proliferative phase.

TCP4 irreversibly changes the cell fate from proliferation to differentiation

Developmental cell maturation is a unidirectional process where the fully expanded pavement cells do not revert back to proliferative fate under normal circumstances. The transcripts of the miR319-regulated class II TCP genes that induce leaf differentiation are abundant in young lamina and their level steadily declines as the lamina matures [42]. Therefore, it is possible that these TCP genes activate the proliferation→differentiation transition in the pavement cells in an irreversible manner. To test this, we grew jaw-D;GR seedlings under TCP4 inductive condition for 0.5, 1, 2, 3, 4, 5 or 6 DAS, shifted them to non-inductive condition until 29 DAS (S4B Fig) and monitored leaf area and cell number of their mature first leaves (Fig 4A–4D). Arabidopsis seeds germinated (emergence of radicle) at 1 DAS and the first leaf pair initiated at 2 DAS (Fig 4A). When the plants were grown for 0.5 or 1 DAS in the presence of DEX, there was no effect of TCP4 induction on the final leaf area (Fig 4B and 4C), possibly because the leaf primordia were barely formed at 1 DAS. When TCP4 was induced for 2 DAS, the leaf margin defect was totally rescued (Fig 4B) and leaf area reduced to almost half (Fig 4C). TCP4 induction for 4 DAS yielded mature leaf as small as what was observed with continuous DEX induction (Fig 4B and 4C), suggesting that miR319-mediated inhibition of class II TCP transcripts for the first 2 days post leaf initiation (that is, 4 DAS), when the average length of leaf primordium was 274 ± 35 μm (Figs 4A and 5B), is crucial for normal leaf growth.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. TCP4 irreversibly commits the proliferating cells of leaf primordia to differentiation.

(A) Scanning electron micrographs of jaw-D;ProTCP4:mTCP4:GR seedlings after indicated DAS. A germinated seed (1 DAS) and the meristematic region of seedlings are shown. Scale bars, 500 μm (1, 6 DAS) and 100 μm (2, 3, 4 DAS). (B) Mature first leaf (29 DAS) of jaw-D;ProTCP4:mTCP4:GR seedlings grown in the presence of 12 μM dexamethasone for the indicated DAS and then shifted to medium without dexamethasone. Leaves grown in continuous absence (Mock) or presence (DEX) of dexamethasone are shown as controls. Scale bar, 5 mm. (C) Average size of leaves shown in (B). N, 12–15 leaves. (D) Number of pavement cells in leaves shown in (B). (E) GUS reporter analysis of first leaf from 8-day old jaw-D;ProTCP4:mTCP4:GR X ProCyclinD3;2:GUS seedlings grown in the presence of dexamethasone for the indicated DAS and then shifted to medium that lacks dexamethasone. Mock indicates seedlings grown without dexamethasone. (F) Average area of mature first leaves (N, 12–15) of jaw-D;ProTCP4:mTCP4:GR seedlings treated with 12 μM dexamethasone for indicated durations at 4 DAS. (G) Average area (N = 10–15) of mature first leaves of jaw-D;ProTCP4:mTCP4:GR seedlings treated with 12 μM dexamethasone for 24 hours at indicated DAS. Mock indicates no treatment; DEX, continuous dexamethasone treatment. (H) Average area of mature first leaves from jaw-D;ProTCP4:mTCP4:GR seedlings treated with 2 mM hydroxyurea (HU) and/or 12 μM dexamethasone (DEX) for 36 hours at 3 DAS. Mock indicates no treatment. N, 8–12 leaves. (I) and (J) Mature first leaf (I) from seedlings grown in the presence of dexamethasone for the indicated DAS and then shifted to medium that lacks dexamethasone, and their average size (J). N, 10–15 leaves. (K) Nuclear ploidy analysis of first leaf pair from 8-day old seedlings grown in the absence (Col-0, M) or presence (DEX) of 12 μM dexamethasone. Averages from biological triplicates are shown. Error bars indicate SD. * indicates p <0.05. Unpaired Student’s t-test was used.

https://doi.org/10.1371/journal.pgen.1007988.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Spatio-temporal expression analysis of miR319-TCP4 module.

(A) and (B) GUS reporter analysis in young leaves of 19-day old seedlings (A) and in the first leaf pair in 4-day old seedlings (B). White arrows point to the leaf base. (C) GUS activity in the fifth leaf at two developmental stages. (D) ProTCP4:TCP4:GUS activity in 10-day old mock grown Col-0 and jaw-D;GR seedlings. (E) and (F) GUS activity of YUC5 promoter in the first leaf (E) and in 5-day old seedlings grown in the absence (Mock) or presence (DEX) of 12 μM dexamethasone (F). The boxed regions in the top panel of (F) are magnified below to highlight GUS activity in the first leaf. Numbers indicate leaf length in mm (A) (B) (C) and (E) or leaf position (D).

https://doi.org/10.1371/journal.pgen.1007988.g005

Consistent with the reduction of blade area, TCP4 induction for 4 DAS also reduced the pavement cell number to 4550 ± 910, similar to the value obtained with continuous TCP4 induction (Fig 4D). Lower cell proliferation activity was also reflected in the reduction of the cell cycle marker CyclinD3;2 [43] in young leaves. Up to 1 DAS of TCP4 induction did not reduce the GUS activity to a noticeable extent in 8-day old jaw-D;ProTCP4:mTCP4:GR X ProCyclinD3;2:GUS seedlings (Fig 4E). However, the GUS activity much reduced when TCP4 was induced for 2 or 4 DAS, and totally disappeared at 6 DAS of induction.

In order to determine the minimum duration of TCP4 activity that is sufficient to induce differentiation commitment in proliferating cells, we provided pulse of TCP4 function of various durations from 0 to 48 hours in 4-day old jaw-D;GR seedlings and compared their mature leaf area at 29 DAS. Leaves without TCP4 induction grew to a maximum area of 36 ± 3.8 mm² (Fig 4F). A 3-hour pulse of TCP4 induction reduced the area to a noticeable extent, and the area continued to reduce with longer pulses, finally stabilizing to a minimum value (23.2 ± 2.0 mm²) with a 24-hour pulse. This suggests that TCP4 activity for the duration of 24 hours is sufficient to impart commitment to differentiation in a dividing cell. Consequently, a 24-hour DEX pulse at 4, 6, 8, 9, 10 and 12 DAS (S6 Fig) altered the mature leaf area of mock-grown jaw-D;GR seedlings to nearly the same extent (Fig 4G) as in continuous DEX-induction starting from the respective DAS (Fig 3B). When we treated 3-day old jaw-D;GR seedlings with hydroxyurea, a widely used cell cycle inhibitor that induces reversible G1/S arrest [44,45], only for a duration of 36 hours, the mature first leaf grew as big as the untreated control (Fig 4H), implying that reversible cell arrest for a short while does not adversely affect the overall cell proliferation in a growing leaf. However, similar treatment with DEX reduced the final leaf size to a great extent (Fig 4H), suggesting that TCP4 initiates irreversible cell cycle arrest in a proliferating leaf.

When TCP4 activity was induced in the 35S:GR seedlings for 0.5, 1, 2, 3, 4, 5 or 6 DAS, the final leaf area progressively reduced with increased duration of induction (Fig 4I, 4J and S7 Fig), as observed for jaw-D;GR leaves (Fig 4B and 4C). However, when similar induction experiment was performed on 35S;sGR seedlings, the final leaf area remained unaltered, suggesting that miR319 activity at early stage of leaf growth contributes to blade expansion.

Increased cell maturation upon TCP4 induction was also reflected in the elevated level of nuclear ploidy, a typical cell differentiation marker [46], in cells of DEX-induced leaves. Flow cytometry analysis of the nuclei isolated from 8-day old Col-0 leaves revealed a ploidy distribution of 57.5% 2C, 39.6% 4C and 2.9% 8C cells (Fig 4K). In the jaw-D;GR leaves grown in mock condition, the proportion of 2C nuclei increased modestly to 69.5% and that of 4C (28.7%) and 8C (1.7%) nuclei decreased, suggesting a less differentiation status of this cell population. Conversely, when TCP4 function was induced in these leaves, the fraction of 2C nuclei reduced to 34.8%, with a concomitant increase in the proportion of 4C (55.7%) and 8C (9.5%) nuclei. Together these results demonstrate that TCP4 makes the proliferative cells exit division cycle and enter into endoreplication.

MiR319-mediated spatial regulation of class II TCP transcripts establishes the transition zone

To correlate the function of the miR319-TCP module with their expression dynamics in expanding leaf blade, we localized miR319c and TCP4 promoter activity and TCP4 protein at various growth stages of transgenic leaves expressing GUS reporter. In a 19-day old ProTCP4:GUS seedling, strong GUS activity was detected throughout the blade of the 9^th rosette leaf (1.2 mm long), which was progressively restricted towards the proximal end in the more mature 8^th, 7^th and 6^th leaves (Fig 5A). Since the maturation states of younger to older leaves in an Arabidopsis rosette recapitulate the ontogeny of a leaf on a specific node [17], it can be concluded that TCP4 promoter activity progressively gets restricted towards leaf base as the blade differentiates, before disappearing completely, as observed previously for TCP4 and its orthologues [8,14,47]. GUS activity in the ProCyclinD3;2:GUS leaves [43] resembled that of ProTCP4:GUS leaves (Fig 5A), suggesting that TCP4 promoter is active in the proliferative zone of the expanding leaf blade.

GUS reporter was detected in a small domain at the base of the young 9^th leaf (0.8 mm long) of PromiR319c:GUS seedlings (Fig 5A), and expression was reduced and restricted to more proximal region in the 8^th leaf (1.9 mm long); no detectable GUS activity was found in the 7^th and the 6^th leaves. These results are in agreement with the earlier reports [48], and suggests that miR319c promoter is active during a small temporal window at the earliest stage of leaf growth. Consequently, the TCP4:GUS reporter activity in young ProTCP4:TCP4:GUS leaf at the 9^th position was not detected in the corresponding basal region where miR319c promoter is active at this stage, suggesting that miR319 activity excludes TCP4 transcript [15] and its protein product (Fig 5A) from its expression domain.

To compare the relative expression domains of miR319c, TCP4 transcript and TCP4 protein in young leaf primordia where miR319 promoter is active, we studied the GUS expression pattern in 4-day old transgenic seedlings where the first leaf pair just initiated (Fig 4A). In ~300 μm long first leaf, TCP4 and CyclinD3;2 promoters were active throughout the leaf, whereas the miR319 promoter was detected in a small region at the very base (Fig 5B and S8 Fig). Activity of the TCP4:GUS fusion protein was, however, detected only in the distal half of young leaf primordia (Fig 5B and 5D and S8 Fig), and was excluded from the proximal region where miR319c promoter is active. Similar expression patterns have been reported for endogenous TCP4 transcript in wild-type [15], and for a miR319-susceptible form of ProTCP4:GUS transcript in transgenic leaf primordia [47].

In young fifth leaf primordium (~300 μm long), barely detectable TCP4:GUS activity was found towards the distal extreme (Fig 5C), implying that miR319 activity is present in a broader basal domain in this leaf. In agreement with this, strong GUS activity was detected across the basal half of a PromiR319c:GUS leaf primordium of similar length (Fig 5C). At a later growth stage, PromiR319c activity was restricted more towards the base, where it reduced the TCP4:GUS signal. Since TCP4:GUS activity retracted from the distal end in these leaves owing to the lack of ProTCP4 activity in differentiated cells, the TCP4:GUS protein was detected in a band-like domain in the middle of the leaf with diffused proximal and distal boundaries (Fig 5A, 5C and 5D), which has been proposed as the transition zone where proliferating cells acquire differentiation competence [3,4,8,24]. Ectopic expression of miR319 throughout the leaf blade completely abolished the TCP4:GUS activity in the jaw-D;GR X ProTCP4:TCP4:GUS seedlings (Fig 5D and S8 Fig), highlighting the importance of class II TCP activity within the transition zone in leaf maturation.

Cells proximal to the transition zone are expected to remain proliferative whereas the distal cells would differentiate. In agreement with this, the promoter activity of YUCCA5 (YUC5), a known differentiation marker [27,49], was initiated at the distal end in a ~300 μm long first leaf of 4-day old ProYUC5:GUS seedlings, and the expression extended towards more proximal direction as the leaf matured (Fig 5E). The GUS signal reduced when the class II TCP genes were down-regulated in the jaw-D;GR X ProYUC5:GUS leaves (Fig 5F), and increased when gain-of-function TCP4 activity was induced by DEX application, suggesting that the distal expression of YUC5 in young leaves is indeed mediated by TCP4 activity.

TCP4-induced auxin response is required for pavement cell maturation

The expression dynamics of TCP4 described above, together with the kinematic data (Figs 2–4), suggests that TCP4 activity imparts irreversible differentiation fate to the proliferating cells within the transition zone of an expanding leaf blade. Auxin is known to play a key role in tissue maturation by coordinating the cell division with differentiation [1,50]. TCP4 function has been linked to the auxin biosynthesis and response in hypocotyl cells [27]. We observed that TCP4 activation increased GUS reporter activity in both jaw-D;GR X ProYUC5:GUS and jaw-D;GR X ProDR5:GUS rosettes (Figs 5F and 6A), demonstrating that TCP4 promotes YUC5 transcription and auxin response in developing leaves. In agreement with this observation, the levels of many SAUR and HD-ZIP II transcripts, which are known targets of auxin signaling [34,51,52], were altered upon TCP4 induction (S9 Fig) when a previously reported jaw-D;GR microarray dataset was analyzed [27]. Fourteen SAUR transcripts were also elevated upon heightened auxin response or upon external application of auxin (Fig 6B) in the microarray datasets available in the GENEVESTIGATOR database [33,34,51–53]. The SAUR and HD-ZIP II genes are known to be involved in multiple developmental events including cell proliferation and differentiation [34,37,50,51], suggesting a link between TCP4 function and auxin responsive genes in cell differentiation. RT-qPCR analysis demonstrated that the transcripts of SAUR19, 20, 21, 24, 62 and 63 increased in the jaw-D;GR seedlings grown under constitutive TCP4 activation (Fig 6C). In addition, GUS reporter activity in the jaw-D;GR X ProSAUR63:GUS seedlings increased upon dexamethasone treatment (Fig 6D), suggesting that TCP4 activates the SAUR genes in planta. SAUR activation was abolished in the jaw-D;GR seedlings when TCP4 was induced for 4 h in the presence of cycloheximide, a general protein synthesis inhibitor (Fig 6E) [54], suggesting that the promotion of SAUR by TCP4 is an indirect effect, possibly mediated by elevated auxin response.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. TCP4-mediated cell differentiation requires auxin signaling.

(A) GUS reporter assay of 17-day old seedlings grown in the absence (Mock) or presence (DEX) of dexamethasone. (B) Expression analysis of 14 SAUR genes in previously reported microarray datasets of indole-3-acetic acid (IAA)-treated seedlings using the signature tool of Genevestigator database (https://www.genevestigator.com/gv/plant.jsp). Top 12 transcriptome profiles of IAA-treated microarrays are shown. (C) Relative transcript levels of SAUR genes determined by RT-qPCR analysis in 9-day old jaw-D;ProTCP4:mTCP4:GR seedlings grown in the absence (Mock) or presence (DEX) of 12 μM dexamethasone. Averages from independent biological triplicates are shown. YUC5 level has been used as a positive control. (D) GUS analysis of 4-day old seedlings grown on mock and DEX containing medium. Scale bar, 1 mm. (E) Relative transcript levels of indicated genes estimated by RT-qPCR analysis in 9-day old jaw-D;ProTCP4:mTCP4:GR seedlings treated either with 40 μM cyclohexamide (CHX) or with a combination of 40 μM cyclohexamide and 20 μM dexamethasone (CHX DEX) for 4 hrs. PP2A/TUB2 was used as internal controls. (F) to (J) Average leaf area (F), number of total pavement cells (G), average pavement cell area (H) and relative cell number (J) in mature first leaves of indicated genotypes grown in the absence (Mock) or presence (DEX) of dexamethasone are shown. 30-day old rosettes pictures of indicated genotypes are shown in (I). Cell number in the DEX-grown leaf was normalized to that of Mock-grown leaf (J). (K) Frequency distribution of pavement cell size in mature first leaves. j;G X C and j;G X s indicate jaw-D;ProTCP4:mTCP4:GR X Col-0 and jaw-D;ProTCP4:mTCP4:GR X shy2-2, respectively. M and D indicate Mock and DEX, respectively. Total 130–150 cells from 5–7 leaves were analyzed for (H) and (K). Error bars indicate SD. Scale bar, 2 mm (I). N, 12–15 leaves (F).

https://doi.org/10.1371/journal.pgen.1007988.g006

To examine whether the elevated auxin response by TCP4 is capable of inducing leaf cell maturity, we studied the effect of TCP4 induction on pavement cell area/number in short hypocotyl2-2 (shy2-2), where cells are insensitive to auxin response due to the sequestration of AUXIN RESPONSE FACTORs (ARF6 and 8) by a dominant, non-degradable form of AUX/INDOLE-3-ACETIC ACID3 (IAA3) inhibitor protein [55]. Mature first leaves of jaw-D;GR X Col-0 heterozygous seedlings grew to 29 ± 2.6 mm², consisting of 18850 ± 1670 pavement cells with average area of 1550 ± 165 μm² (Fig 6F, 6G and 6H). TCP4 induction in these leaves reduced the blade area to 19.0 ± 2.2 mm² and cell number to 5240 ± 650, while increased the average cell area to 3490 ± 180 μm². The average area of mature jaw-D;GR X shy2-2 leaves (Fig 6I) under non-inductive condition was 18.0 ± 2.3 mm², with the total cell number 14080 ± 1660 and average cell area 1350 ± 200 μm² (Fig 6F, 6G and 6H), possibly indicating a TCP-independent role for auxin response in leaf blade expansion. TCP4 induction, however, failed to reduce the jaw-D;GR X shy2-2 leaf area to any noticeable extent (Fig 6F), even though it increased the pavement cell area ~2-fold, as in jaw-D;GR X Col-0 control leaves (Fig 6H). While TCP4 induction reduced the total pavement cell number by ~75% in the jaw-D;GR X Col-0 leaves, it reduced the jaw-D;GR X shy2-2 cell number by ~50% (Fig 6G and 6J), suggesting that at least a part of the TCP4 effect on the repression of cell number is mediated by IAA3-dependent suppression of auxin signaling.

The pavement cell population of the mature jaw-D;GR X Col-0 and jaw-D;GR X shy2-2 leaves showed similar size distribution under un-induced condition (Fig 6K), with ~70% cells of smaller area (<1500 μm²) and <10% cells of larger area (>3000 μm²). Upon TCP4 induction, the fraction of the smaller cells decreased to ~25% and that of larger cells increased to ~50% in jaw-D;GR X Col-0, indicating a rescue of the cell size defect by TCP4 activity. The extent of rescue was, however, less (~40% smaller cells and ~30% larger cells) in the jaw-D;GR X shy2-2 leaves, suggesting that TCP4 partly requires auxin response to induce cell maturity.

TCP4 activates HAT2 independent of auxin response

HAT2 and its closest homologue HAT1 have been recognized as immediate auxin responsive genes and control cell proliferation in leaves and cell elongation in hypocotyls, functions that overlap with miR319-targeted TCP genes [22,27,33,34]. HAT1 and HAT2 transcripts were consistently up-regulated in multiple independent microarray datasets performed upon external auxin application (Fig 7A) [34]. HAT2 transcript was also up-regulated at both 2 and 4 h of TCP4 induction in an earlier microarray experiment (S9 Fig) [27]. RT-qPCR analysis showed that the transcript of HAT2 alone, and not of HAT1 or the clade members, was up-regulated upon constitutive TCP4 induction in the Pro35S:mTCP4:GR seedlings (Fig 7B). HAT2 was induced within 1 h of TCP4 induction (Fig 7C), and the level was maintained for at least 4 h. In addition, HAT2 level was elevated in TCP4:VP16 seedlings, a gain-of-function line of TCP4 with enhanced transcriptional activity [23], while that was significantly reduced in the loss-of-function line jaw-D (Fig 7D). Together, these results suggest that HAT2 is an early downstream target of TCP4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. TCP4 directly activates HAT2 transcription.

(A) Comparative transcript levels of Class II HD-ZIP genes in Arabidopsis seedlings in previously reported microarray datasets upon auxin treatment using the signature tool of Genevestigator database (https://www.genevestigator.com/gv/plant.jsp). Top 9 transcriptome profiles are shown. (B) and (D) Relative transcript levels of class II HD-ZIP genes in jaw-D;ProTCP4:mTCP4:GR (B) and HAT2 in indicated genotypes (D) determined by RT-qPCR in 9-day old seedlings grown without dexamethasone (Mock) or in the continuous presence 12 μM dexamethasone (DEX). (C) RT-qPCR analysis of HAT2 expression in 9-day old jaw-D;ProTCP4:mTCP4:GR seedlings treated with 12 μM dexamethasone for indicated durations. (E) RT-qPCR analysis of indicated transcripts in 9-day old jaw-D;ProTCP4:mTCP4:GR seedlings treated either with DEX alone or with DEX plus and combination of NPA and L-Kyn, pharmacological inhibitors of auxin transport and synthesis respectively, for 2 hrs. Solvent treatment (Mock) has been used a negative control. (F) HAT2 transcript levels determined by RT-qPCR in 9-day old seedlings grown without (Mock) or with (DEX) 12 μM dexamethasone. j;GR X C indicate jaw-D;ProTCP4:mTCP4:GR X Col-0 and and j;GR X s indicate jaw-D;ProTCP4:mTCP4:GR X shy2-2. Transcript levels were initially normalized to PP2A/TUB2 internal control and then compared to respective mock controls. (G) and (H) HAT2 transcript levels analyzed by RT-qPCR in 9-day old jaw-D;GR (jaw-D;ProTCP4:mTCP4:GR) (G) or Pro35S:GR (Pro35S:mTCP4:GR) (H) seedlings treated either with 40 μM cyclohexamide (CHX) or with a combination of 40 μM cyclohexamide and 20 μM dexamethasone (CHX DEX) for 4 hrs. (I) Schematic representation of the HAT2 locus. Exons are shown in gray box and translation start site is indicated with an arrow. BS1-BS3 represent the putative TCP4-binding motifs that are used for EMSA experiments in (J). R1-R4 indicate the regions that are used for ChIP-qPCR analysis shown in (K). (J) Electrophoretic mobility shift assay demonstrating the retardation of the oligonucleotides corresponding to the TCP4-binding sites in the HAT2 locus shown in (I) by recombinant MBP-fused TCP4 domain (MBP-TCP4D). 250-fold of un-labeled oligonucleotides was used as competitors. For super shift, anti-MBP antibody was used. (K) ChIP-qPCR analysis of the HAT2 chromatin regions R1-R4 (shown in I) with anti-FLAG antibody. LOX2 and TUB2 were used as positive and negative controls, respectively. For RT-qPCR analyses in (B-H) and Chip-qPCR in (K), averages of biological triplicates are shown. Error bars indicate SD. * indicates p <0.05, ns indicates not significant, Unpaired Student’s t-test was used.

https://doi.org/10.1371/journal.pgen.1007988.g007

Since class II TCP proteins induce auxin biosynthesis [27,56] and HAT2 is an early auxin responsive gene, it would appear that HAT2 activation is an indirect effect of TCP4 induction. However, we found that TCP4 induction activates HAT2, but not the SAUR genes, in the jaw-D;ProTCP4:mTCP4:GR seedlings even when both auxin biosynthesis and polar auxin transport were inhibited by administering a combination of L-kynurenine and N-1-naphthylphthalamic acid (Fig 7E) [57]. Moreover, TCP4 induction in the jaw-D;GR X shy2-2 seedlings, where auxin response was genetically depleted by expressing a degradation-resistant form of IAA3 [55], also promoted HAT2 expression (Fig 7F). These results suggest that TCP4 is capable of promoting HAT2 expression independent of auxin response, and possibly by direct transcriptional activation.

HAT2 is a direct target of TCP4

Consistent with this hypothesis, TCP4 induction increased HAT2 transcript level in jaw-D;ProTCP4:mTCP4:GR and Pro35S:mTCP4:GR seedlings even when new protein synthesis was blocked by applying cycloheximide (Fig 7G and 7H). This suggests that TCP4 directly activates HAT2 possibly by binding to its genomic locus. Analysis of HAT2 genomic sequence yielded three putative TCP4 binding elements (BS1-3); two in the promoter region and the third in the coding region (Fig 7I). Electrophoretic mobility shift assay showed that recombinant MBP (maltose binding protein)-tagged TCP4 protein [20] binds to oligonucleotides corresponding to these elements with sequence specificity (Fig 7J). To test whether the TCP4 protein is recruited to the HAT2 chromatin in planta, we performed chromatin immuno-precipitation (ChIP) assay in the Pro35S:TCP4:3F6H transgenic line using anti-FLAG antibody [39]. TCP4 recruitment was observed in the HAT2 upstream regulatory regions that corresponded to the TCP4 binding motifs (Fig 7K), but no recruitment was found in the HAT2 coding regions. Consistent with this result, formaldehyde-assisted isolation of regulatory elements (FAIRE) experiment [58] suggested that the HAT2 chromatin assumed more open conformation in the regions corresponding to the putative TCP4-binding sites and not in a distant region when TCP4 was activated in the Pro35S:mTCP4:GR seedlings by dexamethasone treatment (S10A Fig).

Even though we identified one putative TCP4 binding site in the HAT1 upstream regulatory region and another in the coding region (S10B Fig), we suspected that HAT1 is not a target of TCP4 since its transcript level remained unaltered upon TCP4 induction (Fig 7B). ChIP assay also suggested that TCP4 protein is not recruited at the HAT1 locus (S10C Fig). Based on these results, and on the observation of a general overlap in the expression levels of TCP4 and HAT2 at various developmental stages (S11 Fig), we conclude that HAT2, and not HAT1, is a direct target of TCP4.

CIN-TCPs require HD-ZIP II transcription factors for limiting cell number and leaf size

Both TCP4 and HAT2 restrict cell number in leaves and promote cell expansion in hypocotyl [27,34]. Since TCP4 directly activates HAT2 (Fig 7), it is likely that TCP4 requires HAT2 activity in planta to promote developmental cell maturation. To test this, we crossed the jaw-D;GR line to the hat1;hat2 double mutant, since there is a high degree of functional redundancy between HAT1 and HAT2 and their single mutants cause little phenotypic alterations [59]. However, the dominant jaw-D phenotype was suppressed in the jaw-D;GR X hat1;hat2 F₁ individuals possibly due to transgene silencing, a phenomenon that commonly occurs with the viral 35S promoter [60,61]. To overcome this, we established the Col-0;GR;hat1;hat2 line and analyzed its leaf parameters upon TCP4 induction. The Col-0;GR leaves, which resemble Col-0 under non-inductive condition [27], were reduced to ~38% in area upon TCP4 induction, primarily due to a corresponding reduction in the number of pavement cells (Fig 8A–8C) that were mildly but significantly larger than the un-induced cells (Fig 8D). Though the mature hat1;hat2 leaves resembled wild type leaves in shape and size (Fig 8A), they were made up of an increased number of pavement cells that were smaller in area (Fig 8B and 8D), an indication of differentiation defect frequently observed in the mutants of CIN-TCP genes (Fig 1) [17]. When TCP4 was activated in the absence of HAT1 and HAT2 gene function in the hat1;2;GR line, leaf blade area reduced only to ~70% with a corresponding reduction in the pavement cell number. Further, TCP4 activation failed to sustain the increase in the pavement cell size in the absence of HAT function (Fig 8D). Taken together, these results suggest that TCP4 requires HAT2, and perhaps HAT1 as well, to limit cell proliferation and leaf size.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. TCP4 requires the HD-ZIP II transcription factors for leaf maturation.

(A) to (D) Mature first leaves (A, left) of plants grown without (Mock) or with (DEX) 12 μM dexamethasone, their average area (A, right), total number of pavement cells (B), cell number upon TCP4 induction relative to mock-induction (C) and average abaxial pavement cell area (D). Sample number, 10–13 leaves. Col-0;GR and hat1;2;GR indicate Col-0;ProTCP4:mTCP4:GR and hat1;2;ProTCP4:mTCP4:GR, respectively. For (A), *** indicates p <0.0001 (Unpaired Student’s t-test was performed). (E) to (I) Mature first leaves from 29-day old seedlings of indicated genotypes (E, left), their average area (E, right), total pavement cell number (F), cell number relative to Col-0 (G), outline of abaxial pavement cells (H) and their average area (I). N, 8–12 leaves. jaw-D;GR and HAT2 indicate jaw-D;ProTCP4:mTCP4:GR and Pro35S:HAT2, respectively. For (D) and (I) total 125–150 cells per leaf were measured and averages from 3–7 leaves are shown. Error bars indicate SD. * indicates p <0.05. Unpaired Student’s t-test was used. (J) A schematic diagram proposed to describe the mechanism of TCP4-induced commitment to differentiation of proliferating cells in a wild type leaf. The proliferative basal zone (green) remains devoid of TCP4 activity due to the action of miR319 and expresses cyclins and GRF [22,65]. The transition zone (red) above the base expresses miR319-regulated class II TCP proteins that induce cell commitment within this region via auxin response as well as HAT2 activation. TCP4 activity declines in the differentiated zone towards the tip (white), which expresses the direct targets of TCP4 such as YUC5 and miR396 [22,27]. Scale bars, 2 mm (A), 5 mm (E) and 100 μm (H).

https://doi.org/10.1371/journal.pgen.1007988.g008

It has been shown that overexpression of HAT2 inhibits the leaf size and promotes the cell elongation in hypocotyl similar to that of TCP4 activity in leaf and hypocotyl, respectively [27,34,59]. Pro35S:HAT2 line shows only a mild effect in reducing leaf size and cell number in the wild-type background (60). However, this line reduced the size of the jaw-D leaf by ~30% and its pavement cell number by ~50% (Fig 8E–8G). Further, Pro35S:HAT2 rescued the small-size phenotype of the jaw-D pavement cells nearly to the wild type level (Fig 8H and 8I).

Since HAT1 and HAT2 are functionally redundant partners and their overexpression yields comparable changes in leaf and hypocotyl [33–35,37], we tested whether HAT1 overexpression can also rescue jaw-D phenotype. Similar to TCP4 dominant effect (Fig 8A), overexpression of HAT1 in the Pro35S:JAIBA:GFP line (JAB:GFP) [37] reduced leaf size to a third of the wild type value with ~60% decrease in pavement cell number (S12A–S12C Fig), demonstrating that HAT1 is a strong cell proliferation inhibitor in developing leaves. Moreover, HAT1 overexpression rescued the effect of class II TCP down-regulation on the leaf size and pavement cell number of the jaw-D;GR leaves nearly to the Col-0 values (S12A–S12D Fig). Intriguingly, overexpression of HAT1 resulted in a significant decrease in pavement cell area (S12E and S12F Fig), a phenotype that contrasts TCP4 overexpression. This negative effect of HAT1 on pavement cell size remained unaltered even in the jaw-D;GR background, possibly indicating an effect of HAT1 independent of TCP4.

Taken together, these genetic interaction studies show that TCP4 induces pavement cell maturation in part via direct activation of HAT2, and not HAT1. However, HAT1 (along with HAT2) is likely to be an indirect target of TCP4 as it is an auxin response gene (Fig 7A) [34] and TCP4 is known to promote auxin level through YUC5 activation (Fig 6A) [27]. Consequently, overexpression of each of these clade-members (HAT1 and HAT2) rescued jaw-D phenotype (Fig 8E–8I and S12 Fig).

Plant materials and growth conditions

Col-0 was used as the wild type control in all experiments. The mutant lines tcp2;4;10, jaw-D, shy2-2 and hat1;2 were reported previously [9,15,33,55]. The Col-0;Pro35S:mTCP4:GR #2 line was screened and established in the T₂ generation. The cloning methodology and primers information was previously reported [27]. The homozygous transgenic lines Col-0;ProTCP4:mTCP4:GR, jaw-D;ProTCP4:mTCP4:GR, Col-0;Pro35S:mTCP4:GR, Col-0;Pro35S:TCP4:GR, ProYUC5:GUS, ProDR5:GUS, ProCyclinD3;2:GUS, PromiR319c:GUS, ProSAUR63:GUS, Pro35S:JAB:GFP, Pro35S:HAT2, Pro35S:TCP4:3F6H, ProBLS:rTCP4:GFP and ProTCP4:TCP4:VP16 used in this study are all established lines reported previously [17,23,27,37,39,43,48,59,67,68]. The F₁ genetic crosses of jaw-D;ProTCP4:mTCP4:GR X ProYUC5:GUS, Col-0 X ProTCP4:GUS, jaw-D;ProTCP4:mTCP4:GR X ProTCP4:GUS, Col-0 X ProTCP4:TCP4:GUS, jaw-D;ProTCP4:mTCP4:GR X ProTCP4:TCP4:GUS, jaw-D;ProTCP4:mTCP4:GR X ProDR5:GUS, jaw-D;ProTCP4:mTCP4:GR X ProSAUR63:GUS, Col-0 X Pro35S:HAT2 and jaw-D;ProTCP4:mTCP4:GR X Pro35S:HAT2 were used in this study. Col-0;ProTCP4:mTCP4:GR line was established in the hat1;2 background by genetic cross in F₄ generation. The homozygous jaw-D;ProTCP4:mTCP4:GR;Pro35S:JAB:GFP line was established by genetic cross in F₃ generation. All experiments were performed under long day conditions (16 hours light/8 hours dark at 22°C). 12 μM DEX was used for TCP4 induction in all the experiments unless specified otherwise.

Plasmid constructs and transgenic lines

The 2.16 kb upstream region of TCP4 including 5’ untranslated region was amplified using the forward primer 5’-AATTGACCCTTTTCTATCATGC-3’ and the reverse primer 5’-TGGTAGAGCATATTCGTCGAGA-3’, pfu DNA polymerase and Col-0 genomic DNA and cloned into pGEMT-Easy vector by TA cloning (Promega; pGEMT-Easy-ProTCP4) [27] and then moved into pCAMBIA 1304 using NcoI and SacI digestion to generate ProTCP4:GUS construct.

The TCP4 cDNA fragment of pBSKS-TCP4:GR [27] was digested with the SalI and BamHI and moved into pCAMBIA-1391Xa to generate TCP4:GUS cassette. The NotI released 2.16 kb ProTCP4 fragment of pGEMT-Easy-ProTCP4, described above, was end filled with Klenow DNA polymerase and cloned upstream of the TCP4:GUS cassette by blunt end ligation to generate ProTCP4:TCP4:GUS construct. The transgenic lines were generated in Col-0 background by Agrobacterium tumefaciens-mediated floral dip method [69].

Measurement of leaf size, epidermal cell size and cell number

For the measurement of leaf size the crinkled leaves were flattened and photographed using a Canon PowerShot S110 camera. Leaf area measurement was made in the photographs using Image J software (rsbweb.nih.gov/ij/). The first leaf from 29-day old plants were cleared with 70% alcohol and processed in choral hydrate:water:glycerol (8:2:1) solution and lactic acid for one to three weeks. In another method, leaves were collected in 70% ethanol followed by shifting them to ethanol:glacial acetic acid (7:1) and kept overnight (12–14 h) at room temperature. After clearing, the leaf tissues were treated with 1 M KOH for maximum 30 min in constant rotating followed by two times washing with double-distilled water. Image of abaxial epidermal cells were taken using differential interference contrast microscope (Olympus, USA) and analyzed with Image J software. For cell size analysis, 400–750 cells area was measured from different fields of 4–7 leaves and averaged to obtain the final cell size. The total number of cells per leaf was calculated by using the average cell size and leaf area.

Kinematic analysis

The kinematic analysis of Col-0 and mock and DEX-grown jaw-D;GR leaf development was performed as described earlier [5,9,41,70]. Initially seedlings were grown on Murashige Skoog (MS) medium up to 10 days and then the plants were transplanted to soil. For DEX induction, the MS medium was supplemented with 12 μM DEX or ethanol control. For continuous DEX induction, seedlings were transplanted to DEX (12 μM)- or ethanol (0.04%)-treated soil and sprayed with DEX or ethanol on alternate days until experiment was completed. One of the first leaf pair from at least 9–12 plants was used for each time point of kinematic study. The leaf area, cell number and cell size were analyzed as described above. The stomatal index was analyzed as described earlier [5,41].

Gene expression and GUS reporter analyses

Total RNA samples were isolated by Trizol (Sigma, USA) method after treatment with ethanol solvent (Mock) or 12 μM dexamethasone. To check the relative expression of respective genes, quantitative RT-PCR was performed using SYBER Green qPCR kit (SensiFAST SYBR Lo-ROX Kit, Bioline, USA) according to the manufacturer’s instructions. Results were analyzed using ABI Prism 7900HT SDS software (Applied Biosystems, USA) and ΔΔCT values were determined after normalization with internal control. Intensity ratios were calculated using the formula 2^-ΔΔCT. Detailed experimental protocols for dexamethasone induction and primer sequences for RT-qPCR were described earlier [27,70]. GUS assay, EMSA and FAIRE experiments were performed according to the earlier protocols [9,27]. The list of primers used in this study is provided in S2 Table.

Flow cytometry

1^st pair of leaves from 8-day old seedlings of indicated genotypes were collected after removing hypocotyl and cotyledons. Ploidy levels were measured using the protocol provided earlier [71,72]. In brief, leaves were placed in ice-cold Galbraith’s buffer (45 mM MgCl₂, 30 mM sodium citrate, 20 mM 4-morpholinepropane sulfonate, and 1% Triton X-100, pH 7.0) and finely chopped using sharp razor blade followed by homogenizing in the buffer. The homogenate was filtered through Miracloth (22–25 μm) and treated with 10 μl/ml RNaseA (1mg/ml stock) for 20–30 min followed by nuclear DNA staining with 50 μl/ml of propidium iodide (1 mg/ml stock) and kept in dark at 4°C for 30 min. Samples were run through BD FACS verse (USA) flow cytometer and analyzed using BD FACS unit software. It was considered that the left-most peak corresponds to 2C [71,72]

Chromatin immunoprecipitation (ChIP) assay

ChIP protocol was followed as mentioned earlier [39]. In brief, approximately 1.5–2 gm of fresh 10-day old Pro35S:TCP4:3F6H seedlings were harvested and crushed into fine powder in liquid nitrogen. The powder was homogenized in the nuclei extraction buffer 1 (0.4 M sucrose, 10 mM Tris-HCl pH 8.0, 10 mM MgCl₂, 5 mM β-mercaptoethanol, 0.1 mM PMSF, 1 mM Na₃VO₄, 1 mM NaF, and Complete protease inhibitor cocktail tablets [Roche]) and proceeded for cross-linking. 37% formaldehyde was used for crosslinking by gentle rotating in 4°C for 10 min and 2 M glycine was added to stop the reaction. After quenching the cross-linked samples were filtered through two layers of Miracloth and centrifuged for 10 min at 10 k rpm at 4°C. The pellets containing the nuclei were resuspended and washed twice with extraction buffer 2 (0.25 M sucrose, 10 mM Tris-HCl pH 8.0, 10 mM MgCl₂, 1% (w/v) Triton X-100, 5 mM β-mercaptoethanol, 0.1 mM PMSF, 1 mM Na₃VO₄, 1 mM NaF, and Complete protease inhibitor cocktail tablets) and transferred to ice-cold 1.5mL Eppendorf tubes followed by centrifugation at maximum rotor speed for 10 min at 4°C to isolate the nuclei. The isolated nuclei were lysed using Nuclei lysis buffer (50 mM Tris-HCl pH 8.0, 10 mM EDTA, 1% SDS, 1 mM PMSF, and Complete protease inhibitor cocktail tablets) and proceeded for sonication to shear the chromatin (~ 500 bp to 1 kb size fragments) using Bioruptor (Condition; 30 sec ON 45 sec OFF, 45 cycles). The sheared chromatin were diluted in ChIP dilution buffer (16.7 mM Tris-HCl pH 8.0, 167 mM NaCl, 1.1% Triton X-100, 1.2 mM EDTA, 0.1 mM PMSF, 1 mM Na₃VO₄, 1 mM NaF, and Complete protease inhibitor cocktail tablets) to make the volume up to 1.5 mL and centrifuged at maximum speed for 5 min at 4°C. From the collected chromatin solution, 30 μL (2%) was kept as input for the assay and rest of the solution was divided into two parts from which one was immunoprecipited using anti-FLAG antibody (Sigma; F1804; 1 mg/mL) and the other using Anti-IgG antibody (Sigma; A9044). Magnetic Dynabeads Protein G (Life technologies) were used for Immunoprecipitation for which first the beads were pretreated with the anti-FLAG antibody (Sigma; F1804; 1 mg/mL) or anti-IgG antibody for 20 min by rotating at room temperature and diluted in ChIP dilution buffer after washing twice with 1X PBST. The antibody-bound beads were incubated with chromatin solution (26 μL Dynabeads and 3 μL antibody per 600 μL of chromatin solution) for 20 min in ice followed by 2 h rotating at 4°C. After immunoprecipitation, the immunocomplex was washed using low salt buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 2 mM EDTA) and high salt buffer (20 mM Tris-HCl pH 8.0, 500 mM NaCl, 0.1% SDS, 1% Triton X-100, 2 mM EDTA) and TE buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA). To elute the immunocomplex from the beads, 50 μL of nuclei lysis buffer was added and incubated for 20 min at 65°C and the elution was repeated twice. Eluted immunocomplex along with the input samples were reverse cross-linked by adding 6 μL of 5 M NaCl and incubating at 65°C overnight. Proteinase K and RNase A was treated to digest all proteins and RNA, respectively, and finally DNA were cleaned up by phenol:chloroform:isoamyl alcohol (15:24:1) followed by precipitation using sodium acetate and 100% ethanol. The final extracted DNA was dissolved in 50 μL of nuclease-free water and proceeded for qPCR using 0.8 μL DNA as template per reaction.

To quantify the relative enrichment on the respective gene element, qPCR was performed using 0.8 μL of immunoprecipited DNA and SYBER Green qPCR kit (SensiFAST SYBR Lo-ROX Kit, Bioline, USA) in a 10 μL reaction volume according to manufacturer’s protocol. Relative enrichment was calculated using the formula 0.02 X 2^{(Ct input- Ct IP)} X 100 and normalized to IgG control. The final fold enrichment was plotted in relative to Col-0 value.