Members of the GCN5 Histone Acetyltransferase Complex Regulate PLETHORA-Mediated Root Stem Cell Niche Maintenance and Transit Amplifying Cell Proliferation in Arabidopsis  

Abstract

The PLETHORA (PLT) stem cell transcription factors form a developmentally instructive protein gradient in Arabidopsis thaliana roots. Histone acetylation is known to facilitate gene transcription and plays an important role in developmental processes. Here, we show that histone acetyltransferase GCN5 (for general control nonderepressible 5) attenuates the PLT gradient. Based on genetic evidence, we establish that GCN5 is essential for root stem cell niche maintenance and acts in the PLT pathway. The GCN5-associated factor ADA2b (for alteration/deficiency in activation 2b) is also positioned in the PLT pathway and regulates PLT expression, similar to GCN5. Both GCN5 and ADA2b mediate proliferation of the transit amplifying cells, but ADA2b does not affect stem cell niche maintenance. Overexpression of PLT2 rescues the stem cell niche defect of gcn5 mutants, indicating that GCN5 regulation of PLT expression is essential for maintenance of the root stem cell niche. We conclude that histone acetylation complexes play an important role in shaping a developmentally instructive gradient in the root.

The Arabidopsis GCN5 histone acetyltransferase complex attenuates the recently discovered gradient expression of the PLETHORA genes and thereby regulates root stem cell niche maintenance and proliferation. Therefore, chromatin modifications play an important role in stem cell maintenance and in shaping a developmentally instructive gradient in the root.

INTRODUCTION

Root growth is maintained by the organizing center or quiescent center (QC) and its surrounding stem cells (van den Berg et al., 1995). The stem cell niche in the root is specified by two parallel transcription factor pathways: the PLETHORA (PLT) pathway (Aida et al. 2004; Blilou et al. 2005) and the SHORT-ROOT (SHR)/SCARECROW (SCR)/RETINOBLASTOMA RELATED (RBR) pathway (Helariutta et al., 2000; Nakajima et al., 2001; Sabatini et al., 2003; Wildwater et al., 2005). PLT proteins act as dose-dependent regulators of root development (Galinha et al., 2007). High levels of PLT2 maintain the stem cells, intermediate levels facilitate transit amplifying cell divisions that define the meristem region, and low levels allow progression of differentiation. PLT gene transcription is auxin responsive, requires auxin response transcription factors (ARFs) and forms gradients that are thought to be a readout of an underlying auxin gradient (Aida et al., 2004; Galinha et al., 2007; Grieneisen et al., 2007), but additional factors that regulate the PLT gradient have hitherto not been identified.

Chromatin modifications play an important role in gene transcription. Both the nucleosomal histone core and the histone tails can be posttranslationally modified by acetylation, methylation, phosphorylation, ubiquitination, and many other modifications. Histone acetylation is thought to enhance the accessibility of the DNA and facilitate transcription (reviewed in Lee and Workman, 2007; Shahbazian and Grunstein, 2007).

There are four histone acetyltransferase (HAT) families and GCN5 (for general control nonderepressible 5), a member of the Gcn5 N-acetyltransferase (GNAT) family, is the best-studied HAT in yeast and mammals (reviewed in Baker and Grant, 2007; Lee and Workman, 2007; Nagy and Tora, 2007; Shahbazian and Grunstein, 2007). Arabidopsis thaliana also has four HAT families, and the Arabidopsis GCN5 homolog of the GNAT family has been best characterized (Pandey et al., 2002; Chen and Tian, 2007). Similar to yeast and mammals, Arabidopsis GCN5 contains a bromodomain and has been shown to acetylate histone 3 in vitro (Stockinger et al., 2001; Mao et al., 2006; Earley et al., 2007), and global histone 3 acetylation is reduced in gcn5 mutants (Bertrand et al., 2003). More specifically, histone 3 lysine 14 (H3K14) and H3K27 acetylation marks are reduced at defined loci in gcn5 mutants (Benhamed et al., 2006).

Yeast GCN5 is found in the SAGA (for Spt-Ada-Gcn5-Acetyl transferase) and ADA complexes together with ADA2, and similar complexes exist in human and Drosophila melanogaster (Nagy and Tora, 2007). Also, in Arabidopsis, GCN5 interacts in vitro with the two homologs ADA2a and ADA2b (Stockinger et al., 2001; Mao et al., 2006). Arabidopsis ADA2b enhances the HAT activity of GCN5 (Mao et al., 2006), which is also found in yeast. A subset of genes is regulated by the Arabidopsis GCN5 complex, since expression of ∼5% of the genes is changed in gcn5 and ada2b mutants (Vlachonasios et al., 2003; Benhamed et al., 2008).

The Arabidopsis GCN5 complex has been implicated in developmental processes. Mutation of GCN5 and ADA2b causes pleiotropic defects in the shoot and root (Bertrand et al., 2003; Vlachonasios et al., 2003; Benhamed et al., 2006). The flower meristem defects of gcn5 mutants are probably caused by ectopic expression of homeodomain protein WUSCHEL (WUS) (Bertrand et al., 2003). It is unclear whether WUS is a direct target of GCN5. Mutation of GCN5 restores the expression of WUS in topless (tpl) mutants and rescues their shoot formation (Long et al., 2006). The TPL transcriptional corepressor acts as a repressor of root identity in the shoot during embryogenesis. Recently, TPL was found to support the repressive activity of Indole-3-Acetic Acid 12 (IAA12)/BODENLOS, which counteracts Auxin Response Factor 5 (ARF5)/MONOPTEROS required for embryonic root development (Szemenyei et al., 2008). In addition to its role in development, the GCN5 complex is also involved in transcriptional responses to environmental stress, for example, in cold-regulated gene expression (Stockinger et al., 2001; Vlachonasios et al., 2003; Mao et al., 2006) and light-regulated gene expression (Benhamed et al., 2006). Unlike ADA2b, ADA2a is not involved in the response to abiotic stress and development (Hark et al., 2009).

Here, we show that GCN5 and ADA2b specifically increase PLT1 and PLT2 expression levels and modulate stem cell maintenance, meristem zonation, and cell expansion. Overexpression of PLT2 rescues the stem cell niche defect of gcn5 mutants, indicating that GCN5 affects root stem cell niche maintenance mainly through regulation of PLT expression.

RESULTS

GCN5 Regulates Stem Cell Niche Maintenance and PLT Expression Levels

To investigate the role of GCN5 in root development, we obtained T-DNA insertions in the 9th intron (SALK_048427) and the 1st intron (SALK_150784) of the GCN5 gene, also called histone acetyltransferase of the GNAT family 1 (HAG1). These alleles were described previously as hag1-5 and hag1-6, respectively (Long et al., 2006). The hag1-5 allele lacks the bromodomain, while the hag1-6 allele disrupts both the acetyltransferase domain and the bromodomain. Disruption of GCN5 causes pleiotropic defects in the shoot as described before (Bertrand et al., 2003; Vlachonasios et al., 2003; Benhamed et al., 2006) and shorter roots (Figures 1A and 1B  

Figure 1.

GCN5 Is Required for Stem Cell Niche Maintenance and PLT1 and PLT2 Expression.

(A) Phenotype of the wild-type and hag1-6 mutant seedlings 7 DAG.

(B) to (D) Root length (B), meristem size (C), and size of differentiated epidermis cells 7 DAG (D) of hag1-6 and ada2b-3 seedlings compared with those of the wild type. Error bars indicate 95% confidence interval, n = 30 to 50.

(E) to (H) QC25 expression (turquoise) and starch granules stained by lugol in the wild type (E) and hag1-6 ([F] to [H]) 7 DAG. Arrow indicates columella stem cells.

(I)  GCN5_pro:GCN5:GFP 8 DAG.

(J) Expression of PLT2_pro:PLT2:YFP in hag1-6 embryos.

(K) and (L) Expression of PLT1_pro:PLT1:YFP in the wild type (K) and hag1-6  (L) 4 DAG.

(M) and (N) Expression of PLT2_pro:CFP in the wild type (M) and hag1-6  (N) 7 DAG.

(O) and (P) Expression of PLT2_pro:PLT2:GFP in the wild type (O) and hag1-6  (P) 4 DAG.

Bars = 20 μm in (E) (scale in [F] to [H] as in [E]) and 50 μm in (I) and (J) (scale in [K] to [P] as in [I]).

; see Supplemental Figures 1A and 1C online) (Vlachonasios et al., 2003). hag1-5 seedlings show a less severe phenotype than hag1-6 mutants (see Supplemental Figures 1A to 1C online), and hag1-5 mutant plants are able to produce viable seeds. hag1-6 mutant seedlings possess a more severe phenotype, with shorter roots and smaller meristem sizes compared with those of the wild type and the hag1-5 mutant (Figures 1A to 1C; see Supplemental Figures 1A and 1B online). The root meristem of hag1-6 mutants is not maintained and differentiates at 14 d after germination (DAG) (Figure 1C). The size of differentiated epidermis cells is significantly smaller than those of the wild type (Figure 1D; t test, P = 6.5 × 10⁻⁶). hag1-6 mutant plants are infertile and were maintained as a heterozygous population. GCN5 is ubiquitously expressed in all cells of the root using a GCN5 translational fusion (Figure 1I), which confirms digital in situ hybridization data (Birnbaum et al., 2003).

In the wild type, a layer of columella stem cells is present between the QC (marked by promoter trap QC25) and differentiated columella cells containing starch granules (Figure 1E). By contrast, hag1-6 mutant roots show starch granules next to the QC (20%, n = 30 roots) (Figures 1G and 1H; see Supplemental Table 1 online) or in the QC (30%, n = 30 roots) (Figures 1F and 1G; see Supplemental Table 1 online). The number of differentiated columella layers in hag1-6 mutants is significantly smaller than in the wild type (Figures 1E and 1F; t test, P = 2.0 × 10⁻³). In addition, QC25 is occasionally absent in all (13.3%, n = 30 roots) (see Supplemental Table 1 online) or some QC cells (30%, n = 30 roots) (Figure 1H; see Supplemental Table 1 online). Thus, a defect in QC specification correlates with stem cell malfunctioning and root meristem differentiation.

Neither expression of SCR (see Supplemental Figures 1K and 1L online) nor localization of SHR (see Supplemental Figures 1N and 1O online) is affected in hag1-6 mutants compared with the wild type. By contrast, in situ hybridization with a PLT1 probe revealed that expression of PLT1 is reduced (see Supplemental Figures 1D and 1E online). Protein levels of a PLT1 translational fusion (Galinha et al., 2007) are consistently downregulated in hag1-6 mutants (Figures 1K and 1L; see Supplemental Figure 1P online). Similarly, transcriptional and translational fusions of PLT2 show severely reduced expression in hag1-6 mutants (Figures 1M to 1P; see Supplemental Figure 1Q online). In hag1-6 embryos, the expression of the translational fusion of PLT2 is not affected (Figure 1J), indicating that GCN5 affects PLT2 expression postembryonically.

PLT proteins regulate the PIN proteins (putative auxin efflux carriers), which tune the position of the auxin maximum (Blilou et al., 2005; Galinha et al., 2007). Therefore, auxin distribution was analyzed using the DR5:β-glucuronidase (GUS) construct (Ulmasov et al., 1997). The position of the auxin maximum is not changed in hag1-6 mutants (see Supplemental Figures 1F and 1G online). Strikingly, the expression of cyclinB1;1:GUS (marker for the G2/M phase of the cell cycle) (Colon-Carmona et al., 1999) is reduced or even absent in hag1-6 mutants when cells are still dividing (see Supplemental Figures 1H to 1J and 1M online), similar to plt1-4 plt2-2 double mutants (Aida et al., 2004). CycB1;1 and CycB1;3 transcripts are also lower in hag1-6 mutant roots (Tukey's honestly significant difference test, P < 0.05), whereas CycB1;4 is unaffected (see Supplemental Figures 5D to 5F online), indicating that several but not all CyclinB1 transcripts are regulated by GCN5. Collectively, our data suggest that GCN5 regulates stem cell maintenance and proliferation of transit amplifying cells through the modulation of PLT levels.

ADA2b Regulates PLT Expression Levels and Proliferation of Transit Amplifying Cells

The GCN5 coactivator complexes in yeast and mammals contain the ADA2 protein (reviewed in Nagy and Tora, 2007). Similarly, Arabidopsis ADA2a and ADA2b interact with GCN5 in vitro (Stockinger et al., 2001; Mao et al., 2006). To investigate the role of ADA2b in stem cell niche maintenance, we obtained T-DNA insertion lines with insertions in the 5th intron (ada2b-1) (Vlachonasios et al., 2003) and in the 9th exon of ADA2b (SALK_019407), hereafter called ada2b-3. The ada2b mutants show pleiotropic defects in the shoot and have shorter roots than the wild type, as described previously (Figures 1B and 2A) (Sieberer et al., 2003; Vlachonasios et al., 2003). ada2b-3 mutant roots are shorter than those of hag1-6 mutants (Figure 1B) due to the combination of a smaller mature cell size (Figure 1D; t test, P = 1.9 × 10⁻⁸) and a smaller meristem zone that remains constant in size (Figure 1C). A translational ADA2b fusion is constitutively expressed in all cells of the root (Figure 2F  

Figure 2.

ADA2b Is Required for PLT1 and PLT2 Expression.

(A) Phenotype of the wild type and ada2b-3 mutant 7 DAG.

(B) and (C) QC25 expression (turquoise) and starch granules in the wild type (B) and ada2b-3  (C) 7 DAG. Arrow indicates columella stem cells.

(D) and (E) Whole-mount in situ hybridization with the PLT1 probe in the wild type (D) and ada2b-3 seedlings (E) 3 DAG.

(F)  ADA2b_pro:ADA2b:GFP 6 DAG.

(G) Expression of PLT2_pro:PLT2:YFP in ada2b-3 embryos.

(H) and (I) Expression of PLT1_pro:PLT1:YFP in the wild type (H) and ada2b-3  (I) 4 DAG.

(J) and (K) Expression of PLT2_pro:CFP in the wild type (J) and ada2b-3  (K) 7 DAG.

(L) and (M) Expression of PLT2_pro:PLT2:GFP in the wild type (L) and ada2b-3  (M) 4 DAG.

Bars = 20 μm in (B) and (D) (scale in [C] as in [B]; scale in [E] as in [D]) and 50 μm in (F) and (G) (scale in [H] to [M] as in [F]).

), confirming digital in situ hybridization data (Birnbaum et al., 2003).

A layer of columella stem cells is present in ada2b-3 mutants between the QC and differentiated columella cells (Figures 2B and 2C). Furthermore, ada2b-3 roots continue to grow for at least 21 d without loss of meristematic activity (100%, n = 14). This indicates that, in contrast with GCN5, the stem cell niche is not affected by mutation of ADA2b. However, the number of differentiated columella layers is significantly reduced in ada2b-3 mutants (Figures 2B and 2C; t test, P = 4.4 × 10⁻⁶), like in hag1-6 mutants.

Expression of the SCR transcript (see Supplemental Figures 2A and 2B online) and localization of SHR protein (see Supplemental Figures 2C and 2D online) is not changed in ada2b-3 mutants, but transcript and protein levels of PLT1 and PLT2 are reduced (Figures 2D, 2E, and 2H to 2M; see Supplemental Figures 2I and 2J online). ADA2b affects PLT2 expression postembryonically, since PLT2 is expressed in ada2b-3 embryos (Figure 2G). Expression of cyclinB1;1:GUS is completely absent in ada2b-3 mutants (see Supplemental Figures 2G and 2H online) when cells still divide. The location of the auxin maximum as monitored by DR5:GUS is not affected in ada2b-3 mutants (see Supplemental Figures 2E and 2F online), although auxin levels in the columella are reduced. Together, these data show that ADA2b is required for proper PLT1 and PLT2 expression, similar to GCN5. In addition, ADA2b affects proliferation of the transit amplifying cells and therefore meristem size.

GCN5 Acts in the PLT Pathway

We investigated the genetic interaction of the GCN5 mutant hag1-6 with shr, scr, and plt mutants to assess whether GCN5 acts solely through the PLT pathway or whether it might influence the SHR/SCR/RBR pathway posttranslationally. hag1-6 shr-2 double mutant seedlings possess shorter roots and a smaller meristem at 7 DAG than either of the single mutants (Figures 3D to 3F  

Figure 3.

GCN5 Acts in the PLT Pathway.

(A) to (C) Root tip of hag1-6  (A), plt1-4 plt2-2  (B), and hag1-6 plt1-4 plt2-2  (C) mutants 4 DAG. (Note: the scale of the hag1-6 meristem is different from that of the others.)

(D) to (F) Root tip of hag1-6  (D), shr-2  (E), and hag1-6 shr-2  (F) 7 DAG.

(G) and (H) Root tip of RBR RNAi  (G) and hag1-6 RBR RNAi  (H) plants 6 DAG.

Bars = 50 μm (scale in [C] to [H] as in [B] and [D]).

; see Supplemental Figures 3B and 3E online). This indicates that GCN5 acts in a pathway parallel to SHR. RBR acts genetically downstream of the SHR target, SCR, and reduction of RBR leads to additional columella stem cell layers (Figure 3G). However, reduction of RBR expression is not able to rescue meristem size or root length of hag1-6 mutants (see Supplemental Figure 3C online), and additional columella stem cell layers are only transiently observed (Figure 3H; see Supplemental Figure 3F online). Together, these data indicate that GCN5 acts independently of the SHR/SCR/RBR pathway.

To assess whether GCN5 acts in the PLT pathway, hag1-6 plants were crossed to plt1 plt2 double mutant plants (Aida et al., 2004). The root length and meristem size of hag1-6 plt1-4 plt2-2 triple mutants is similar to that of plt1-4 plt2-2 double mutants (Figures 3A to 3C; see Supplemental Figures 3A and 3D online), confirming that GCN5 acts in the PLT pathway.

ADA2b Acts in the PLT Pathway

Combinations with null mutants (Fukaki et al., 1998; Aida et al., 2004) were made to assess whether ADA2b acts through the PLT pathway or the SHR/SCR/RBR pathway. Compared with the single mutants, shorter roots and smaller meristem size are found in both ada2b-3 scr-4 (Figures 4G to 4I  

Figure 4.

ADA2b Acts in the PLT Pathway.

(A) to (C) Root tip of ada2b-3  (A), plt1-4 plt2-2  (B), and ada2b-3 plt1-4 plt2-2  (C) 4 DAG.

(D) to (F) Root tip of ada2b-3  (D), shr-2  (E), and ada2b-3 shr-2  (F) 5 DAG.

(G) to (I) Root tip of ada2b-3  (G), scr-4  (H), and ada2b-3 scr-4  (I) 7 DAG.

(J) to (L) Root tip of RBR RNAi  (J), ada2b-3  (K), and ada2b-3 RBR RNAi  (L) 6 DAG.

Bars = 50 μm (scale in [B] to [F] as in [A], scale in [H] and [I] as in [G], and scale in [K] and [L] as in [J]).

; see Supplemental Figures 4C and 4F online) and ada2b-3 shr-2 double mutant seedlings (Figures 4D to 4F; see Supplemental Figures 4B and 4E online). Additional columella stem cell layers are only transiently found upon RBR reduction in ada2b-3 mutants (Figures 4J to 4L; see Supplemental Figure 4I online), and neither meristem size nor root length of ada2b-3 mutants is rescued by RBR reduction (see Supplemental Figures 4G and 4H online). These additive interactions suggest that ADA2b acts independently of the SHR/SCR/RBR pathway, similar to GCN5.

Root size is further reduced in ada2b-3 plt1-4 plt2-2 triple mutants compared with plt1-4 plt2-2 double mutants (see Supplemental Figure 4A online). However, meristem size of ada2b-3 plt1-4 plt2-2 triple mutants and plt1-4 plt2-2 double mutants is similar (Figures 4A to 4C; see Supplemental Figure 4D online). These data indicate that ADA2b acts through PLT1 and PLT2 to determine the activity of the transit amplifying cells but that it acts through other factors in the elongation or differentiation zone.

The gcn5 Stem Cell Niche Defect Can Be Rescued by Overexpression of PLT2

To determine whether overexpression of PLT genes can bypass the stem cell niche defects in hag1-6 mutants, a 35S_pro:PLT2:GR construct was introduced (Galinha et al., 2007). Shoot development is arrested when dexamethasone is applied to 35S_pro:PLT2:GR seedlings upon germination (see Supplemental Figure 5A online). However, proximal meristem size increases substantially after a short induction (Figures 5A and 5B  

Figure 5.

Overexpression of PLT2 Rescues the hag1-6 Mutant.

(A) to (E) Overexpression of PLT2 without ([A] and [C]) and 3 d after 2 μM dexamethasone (DEX) treatment ([B], [D], and [E]) of 35S_pro:PLT2:GR ([A] and [B]) and hag1-6 35S_pro:PLT2:GR ([C] to [E]) plants 8 DAG.

(F) and (G) Quantification of the number of cortical cells in the meristem (F) and elongating epidermal cells (G), 8 DAG, induced by 35S_pro:PLT2:GR in the wild type and hag1-6, 3 d after 2 μM dexamethasone treatment at 5 DAG. Error bars indicate 95% confidence interval, n = 25 to 32.

(H) Phenotype of 35S_pro:PLT2:GR and hag1-6 35S_pro:PLT2:GR without or 3 d after 2 μM dexamethasone (+DEX) treatment 8 DAG.

Bar = 50 μm (scale in [B] to [E] as in [A]).

; Galinha et al., 2007) without affecting shoot development severely (Figure 5H). In the hag1-6 background, the size of the proximal meristem is significantly increased after a short induction with dexamethasone (Figures 5C to 5F; see Supplemental Table 2 online; t test, P = 1.7 × 10⁻⁴). Expression of cycB1;1 and cycB1;3 is regained in hag1-6 mutant roots by PLT2 overexpression (see Supplemental Figures 5C to 5E online). In addition, the size of the elongation zone and the number of epidermis cells in the elongation zone of hag1-6 mutants are significantly increased by the induced overexpression of PLT2 (Figure 5G; see Supplemental Table 3 online; t test, P = 1.1 × 10⁻⁵). The reduced size of hag1-6 differentiated epidermis cells (Figure 1D) is not rescued by overexpression of PLT2. This is not surprising because induced PLT2 overexpression by itself causes a reduction in size of differentiated epidermis cells in the wild type (see Supplemental Figure 5B and Supplemental Table 4 online). At 13 DAG, 47% (n = 15) of the rescued hag1-6 seedlings possess an enlarged meristem, indicating full rescue of stem cell activity, whereas 53% of the seedlings are differentiated around that time, similar to the hag1-6 mutant (Figure 1C). Together, these data show that overexpression of PLT2 can rescue the hag1-6 defects in the root stem cell niche and the elongation zone, which confirms that GCN5 acts in the PLT pathway.

DISCUSSION

Here, we provide evidence that GCN5 regulates stem cell niche maintenance and proliferation of the transit amplifying cells through the PLT pathway. GCN5 and ADA2b have specific roles in root development despite their pleiotropic phenotypes. We established that GCN5 and ADA2b are positioned in the PLT pathway, based on genetic interactions with null mutants (Figures 3 and 4; see Supplemental Figures 3 and 4 online) and on their specific regulation of the postembryonic level of PLT1 and PLT2 expression, while the expression of the key root meristem regulators, SHR and SCR, remains unaltered (Figures 1 and 2; see Supplemental Figures 1 and 2 online).

PLT protein dosage determines distinct cellular responses: (1) high levels specify the stem cell niche; (2) intermediate levels maintain cell division in transit amplifying meristem cells; and (3) low levels allow progression of cell expansion and terminal differentiation (Galinha et al., 2007). Overexpression of PLT2 rescues the stem cell niche defect in gcn5 mutants, confirming that GCN5 regulates stem cell niche maintenance through its effect on PLT levels (Figure 5). gcn5 mutants differentiate later than plt1 plt2 double mutants (Aida et al., 2004), which is likely caused by residual PLT1 and PLT2 activity in the gcn5 mutant background. Expression of PLT2 is affected postembryonically, suggesting that the GCN5 complex acts as a postembryonic booster of PLT expression, rather than in the establishment of PLT expression domains. Both GCN5 and ADA2b mediate proliferation of the transit amplifying cells, possibly in part by regulating cycB1;1 and cycB1;3 expression (see Supplemental Figures 1, 2, and 5 online). High expression of cycB1;1 has been shown to enhance shoot and root growth rates (Doerner et al., 1996; Li et al., 2005). Surprisingly, the stem cell niche is maintained in ada2b mutants, even though PLT1 and PLT2 expression levels are reduced, as in the gcn5 mutant, where stem cells are lost (Figures 1 and 2; see Supplemental Figures 1 and 2 online). This puzzling result suggests the presence of a compensatory mechanism for lower PLT levels in the stem cell niche that is dependent on GCN5 but not ADA2b. Regardless of the mechanism involved, our mutant analysis indicates that GCN5 and ADA2b do not act in the same complex and on similar targets in stem cells and stem cell daughters (Figures 3 and 4; see Supplemental Figures 3 and 4 online). The idea of complex-specific target genes is consistent with published microarray experiments (Vlachonasios et al., 2003).

Our observation that GCN5 and ADA2b regulate PLT1 and PLT2 expression levels (Figures 1 and 2; see Supplemental Figures 1 and 2 online) fits with the observation that histone acetylation mediated by GCN5-containing complexes quantitatively facilitates transcription in mammals (reviewed in Baker and Grant, 2007; Nagy and Tora, 2007; Shahbazian and Grunstein, 2007). Recruitment of these complexes to target promoters is mediated by transcription factors, also in Arabidopsis (Stockinger et al., 2001; Mao et al., 2006). However, whether GCN5 and ADA2b act directly at the PLT1 and PLT2 promoter is uncertain. Recent genome-wide data suggest that PLT genes may not be direct GCN5 targets (Benhamed et al., 2008).

At least one alternative scenario can be envisioned. GCN5 might regulate PLT levels by modulating the inhibition of ARFs, which regulate auxin-mediated gene expression (Long et al., 2006; Szemenyei et al., 2008). Intriguingly, the suppression in tpl gcn5 double mutants of the embryonic ectopic root meristem in the tpl mutant alone (Long et al., 2006) can be explained when considering that ectopic PLT activity activates root identity in shoots (Aida et al., 2004; Galinha et al., 2007) and that GCN5 promotes PLT activity (Figure 1; see Supplemental Figure 1 online). In analogy to connections between GCN5 and ARF activity in the Arabidopsis embryo (Long et al., 2006; Szemenyei et al., 2008), GCN5 activity might counteract IAA-mediated repression of ARF activity to upregulate PLT levels postembryonically. However, ARFs have not yet emerged as direct targets of GCN5 regulation (Benhamed et al., 2008).

In mammalian embryonic stem cells, two transcription factors (octamerbinding transcription factor 4 (OCT4) and SRY-box 2 (SOX2)) are essential and sufficient to induce the stem cell state and impose the unique chromatin state present in stem cells by regulating chromatin factors (reviewed in Niwa, 2007a, 2007b; Jaenisch and Young, 2008; Hochedlinger and Plath, 2009). In addition, stem cell transcription factors regulate chromatin factors, which then feed back on their own activity (Loh et al., 2007). In analogy, the PLT stem cell transcription factors are necessary and sufficient to induce root stem cell niches (Aida et al., 2004; Galinha et al., 2007). It will be interesting to find out whether an analogous feedback loop exists in which PLT proteins directly or indirectly induce GCN5 expression and GCN5 influences the chromatin state of PLT promoters or PLT regulators to maintain pluripotency in Arabidopsis stem cells. Interestingly, ectopic expression of the mammalian cell cycle regulatory transcription factor c-Myc increases the frequency of induced pluripotent stem cells (reviewed in Jaenisch and Young, 2008). c-Myc is required for the global maintenance of active chromatin in mammalian progenitor cells by regulating GCN5 expression directly (Knoepfler et al., 2006). In analogy to the mammalian scenario, Arabidopsis GCN5 might contribute to an open chromatin state that reinforces the stem cell state in parallel with PLT activity.

METHODS

Plant Materials and Growth Conditions

Arabidopsis thaliana ecotypes Columbia (Col) and Wassilewskija (Ws) were used. hag1-5 (SALK_048427), hag1-6 (SALK_150784), and ada2b-3 (SALK_019407) (Col) were obtained from the Signal Insertion Mutant Library (http://signal.salk.edu) and were confirmed by PCR-based genotyping. ada2b-1 (Ws) was kindly provided by Vlachonasios et al. (2003). hag1-6 and ada2b-3 mutant plants were unable to produce functional flower meristems; therefore, they were maintained as a heterozygous population.

The origins and ecotypes of other markers and mutants are as follows: QC25 (Ws) (Sabatini et al., 2003); SHR_pro:GFP:SHR (Col) (Nakajima et al., 2001); PLT1 _pro:PLT1:YFP, PLT2 _pro:CFP, PLT2 _pro:PLT2:YFP, and 35S_pro:PLT2:GR (Col) (Galinha et al., 2007); DR5:GUS (Col) (Ulmasov et al., 1997); cyclinB1;1:GUS (Col) (Colon-Carmona et al., 1999); scr-4 (Ws) (null allele) and shr-2 (Col) (null allele) (Fukaki et al., 1998); RBR RNAi (Col) (Wildwater et al., 2005); and plt1-4 plt2-2 (Ws) (null allele combination) (Aida et al., 2004).

Plants were grown as described before (Sabatini et al., 1999).

Transgenic Plants

The GCN5 translational fusion (GCN5_pro:GCN5:GFP) was constructed by PCR amplifying 3.3 kb of the promoter from Col genomic DNA using the following primers (gateway recombination sites are in lowercase letters, and gene-specific regions are indicated in uppercase letters): 5′-ggggacaactttgtatagaaaagttgttCCGTTCAATTTAAAGAAATCCAACAA-3′ and 5′-ggggactgcttttttgtacaaacttgCGAAGCAGTATAGTGAAGGTGATTGA-3′, and the GCN5 genomic coding sequence was amplified using primers 5′-ggggacaagtttgtacaaaaaagcaggctATGGACTCTCACTCTTCCCACCTC-3′ and 5′-ggggaccactttgtacaagaaagctgggttTTGAGATTTAGCACCAGATTGGAGA-3′. These fragments were placed in a pGreenII vector (Hellens et al., 2000) containing the norflurazon resistance gene (Heidstra et al., 2004) together with a GFP coding sequence (fused to the 3′-end of the genomic fragment).

In the same way, the ADA2b translational fusion (ADA2b _pro:ADA2b:GFP) was constructed by PCR amplifying 1.8 kb of the promoter using primers 5′-ggggacaactttgtatagaaaagttgTTCTCCTGAAACGTTCATTGACAT-3′ and 5′-ggggactgcttttttgtacaaacttgTGATGCCCAAGTAGAAAATTGGATT-3′, and the ADA2b genomic coding sequence was amplified using primers 5′-ggggacaagtttgtacaaaaaagcaggctATGGGTCGCTCTCGAGGGAAC-3′ and 5′-ggggaccactttgtacaagaaagctgggttAAGTTGAGCAATACCCTTCTTCACA-3′ from Col genomic DNA (gateway recombination sites are in lowercase letters, and gene-specific regions are indicated in uppercase letters).

For the construction of SCR _pro:GFP, a 2.4-kb HindIII fragment of the SCR promoter was generated by PCR using primers pSCRF (5′-TCTCTATGAAAAGTGGAAATTTACCTGGAA-3′) and pSCRR (5′-GGAGATTGAAGGGTTGTTGGTCGTG-3′) and cloned into pGII226-GFP_ERT.

To construct the pPLT2 _pro:PLT2:GFP translational fusion, a 5876-bp fragment of the PLT2 promoter, the PLT2 genomic coding sequence and GFP were cloned into pGII0227. Transgenic plants were generated using the floral dip method (Clough and Bent, 1998).

Microscopy and in Situ Hybridization

Light microscopy and fluorescence microscopy were performed as described (Willemsen et al., 1998). For confocal microscopy, dissected embryos were mounted in 7% glucose. Starch granules and β-glucuronidase activity were visualized as before (Willemsen et al., 1998). Root length, meristem size, and differentiated epidermis cell size were determined as described (Willemsen et al., 1998). Fluorescence levels were quantified as before (Galinha et al., 2007). Whole-mount in situ hybridization was performed manually (Hejatko et al., 2006) using the PLT1 probe as described (Aida et al., 2004).

Real-Time Quantitative RT-PCR

Roots were collected 8 DAG without treatment or 3 d after DEX treatment at 5 DAG. Total RNA was extracted with on-column DNase treatment (RNeasy; Qiagen), treated again with DNase in solution (RQ1 RNase-Free DNase; Promega), and used for reverse transcription (SuperScript III First-Strand Synthesis Supermix for qRT-PCR kit; Invitrogen). The cDNA equivalent of 10 ng of total RNA was used on a LightCycler 480 II real-time PCR system (Roche) with LightCycler 480 SYBR Green 1 Master reagents (Roche). In all experiments, three biological replicates of each sample and two technical (PCR) replicates were performed. To normalize the qPCR data, four reference genes (At4g34270, At2g28390, At2g32170, and At4g26410) (selected from Czechowski et al., 2005) were used, and the stability of the reference genes was tested using geNorm software (Vandesompele et al., 2002).

Accession Numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: GCN5 (AT3G54610), ADA2b (AT4G16420), PLT1 (AT3G20840), PLT2 (AT1G51190), SCR (AT3G54220), SHR (AT4G37650), RBR (AT3G12280), CycB1;1 (AT4G37490), CycB1;3 (AT3G11520), and CycB1;4 (AT2G26760).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure 1. GCN5 Does Not Affect SCR Expression and SHR Localization, but Regulates PLT1, PLT2, and cycB1;1 Expression.

• Supplemental Figure 2. ADA2b Regulates PLT1, PLT2, and cycB1;1 Expression, but Not SCR Expression and SHR Localization.

• Supplemental Figure 3. GCN5 Acts in the PLT Pathway.

• Supplemental Figure 4. ADA2b Acts in the PLT Pathway.

• Supplemental Figure 5. Overexpression of PLT2 in the hag1-6 Mutant.

• Supplemental Table 1. GCN5 Is Required for Stem Cell Niche Maintenance.

• Supplemental Table 2. Overexpression of PLT2 Rescues Meristem Size of the hag1-6 Mutant.

• Supplemental Table 3. Overexpression of PLT2 Rescues the Size of the Elongation Zone in the hag1-6 Mutant.

• Supplemental Table 4. Overexpression of PLT2 Does Not Influence Differentiated Epidermis Cell Size of the hag1-6 Mutant.

Acknowledgments

We thank K. Vlachonasios for providing us with ada2b-1 seeds and R. Leito, F. Kindt, and I. Rieu for technical assistance.

Author notes

1

Current address: Faculty of Biology, Schaenzlestrasse 1, University of Freiburg, D-79104 Freiburg, Germany.

²

Address correspondence to .

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantcell.org) is: Ben Scheres ().

Online version contains Web-only data.

www.plantcell.org/cgi/doi/10.1105/tpc.108.065300