Arabidopsis MSI1 Is Required for Negative Regulation of ... - Cell Press

Molecular Plant

•

Volume 2

•

Number 4

•

Pages 675–687

•

July 2009

RESEARCH ARTICLE

Arabidopsis MSI1 Is Required for Negative Regulation of the Response to Drought Stress Cristina Alexandrea,2, Yvonne Mo¨ller-Steinbacha,2, Nicole Scho¨nrocka,b,2, Wilhelm Gruissema and Lars Henniga,c,1 a Department of Biology and Zurich-Basel Plant Science Center, ETH Zurich, CH-8092 Zurich, Switzerland b Present address: Brain and Mind Research Institute, University of Sydney, 100 Mallett St, Camperdown, NSW 2050, Australia c Department of Biology, ETH Zurich, LFW E17, CH-8092 Zurich, Switzerland

ABSTRACT Arabidopsis MSI1 has fundamental functions in plant development. MSI1 is a subunit of Polycomb group protein complexes and Chromatin assembly factor 1, and it interacts with the Retinoblastoma-related protein 1. Altered levels of MSI1 result in pleiotropic phenotypes, reflecting the complexity of MSI1 protein functions. In order to uncover additional functions of MSI1, we performed transcriptional profiling of wild-type and plants with highly reduced MSI1 levels (msi1-cs). Surprisingly, the known functions of MSI1 could only account for a minor part of the transcriptional changes in msi1-cs plants. One of the most striking unexpected observations was the up-regulation of a subset of ABA-responsive genes eliciting the response to drought and salt stress. We report that MSI1 can bind to the chromatin of the drought-inducible downstream target RD20 and suggest a new role for MSI1 in the negative regulation of the Arabidopsis drought-stress response. Key words: Abiotic/environmental stress; water relations; chromatin structure and remodeling; transcriptome analysis; development.

INTRODUCTION MSI1-like proteins exist in all eukaryotes and function as subunits of diverse protein complexes involved in various aspects of chromatin assembly and dynamics (Hennig et al., 2005). MSI1-like proteins are WD40 repeat proteins that are predicted to fold into a seven-bladed b-propeller structure. The structure of the Drosophila MSI1-like protein p55 has been solved and revealed that p55 binds to the first helix of histone H4 via a binding pocket located on the side of the b-propeller structure (Song et al., 2008). Arabidopsis has five genes encoding MSI1-like proteins (MSI1-5) (Ach et al., 1997; Kenzior and Folk, 1998; Hennig et al., 2003b), of which MSI1 has been studied in most detail. The function of the closely related MSI2 and MSI3 genes is unknown. MSI4 and MSI5 form another pair of closely related MSI1-like genes in Arabidopsis, and MSI4 was found to be allelic to FVE, a gene from the autonomous floral induction pathway (Ausin et al., 2004; Kim et al., 2004). MSI4 is required to repress the floral repressor FLC, and fve mutants are late flowering as a result of increased FLC expression (Michaels and Amasino, 2001). Similar to MSI1-like proteins in other eukaryotes, Arabidopsis MSI1 participates in various protein complexes. First, MSI1 interacts with the Retinoblastoma-related protein RBR1 (Ach et al., 1997; Jullien et al., 2008). Second, MSI1 is part of Chro-

matin Assembly Complex CAF-1 (Kaya et al., 2001; Exner et al., 2006). Third, MSI1 is a subunit of PRC2 (Polycomb repressive complex 2)-like complexes including the FIS2- and the EMF2complexes (Ko¨hler et al., 2003a; Scho¨nrock et al., 2006a) and interacts with EMF1 (Calonje et al., 2008). Because MSI1 is an essential component of the FIS2-complex, msi1 null mutants are gametophytic lethal and initiate seed development in the absence of fertilization (Ko¨hler et al., 2003a; Guitton et al., 2004; Guitton and Berger, 2005; Leroy et al., 2007). Reduced levels of MSI1 in transgenic MSI1 antisense lines is not lethal but revealed functions of MSI1 in flowering time control and trichome development (Bouveret et al., 2006; Exner et al., 2006, 2008). Strong reduction of MSI1 in transgenic MSI1 co-suppression lines (msi1-cs), which contain only about 5% of wild-type MSI1 levels, causes loss of repression

1

To whom correspondence should be addressed at address c. E-mail: lars. [email protected], fax +41 1 632 1044, tel. +41 1 632 2244.

2

These authors contributed equally to this work.

ª The Author 2009. Published by the Molecular Plant Shanghai Editorial Office in association with Oxford University Press on behalf of CSPP and IPPE, SIBS, CAS. doi: 10.1093/mp/ssp012, Advance Access publication 27 March 2009 Received 8 December 2008; accepted 20 February 2009

676

|

Alexandre et al.

d

MSI1 Represses the Response to Drought Stress

of some EMF2-target genes and strong defects in vegetative and reproductive development leading to sterility (Hennig et al., 2003b; Scho¨nrock et al., 2006a). The multiplicity of MSI1 interactions is a challenge for understanding MSI1 function in plant development. To investigate the requirement of MSI1 on a genome-wide scale, we performed transcriptional profiling of msi1-cs plants. Here, we report that despite a significant overlap of affected gene sets, RBR1, CAF-1, and PRC2-functions can only account for less than 15% of the gene expression changes in msi1-cs. Through extensive transcriptome data mining, we unexpectedly observed enrichment in the msi1-cs dataset for ABA-responsive genes, specifically salt and osmotic-stress-related genes. These gene expression changes are physiologically relevant, and therefore we propose a new role for MSI1 in the negative regulation of the drought stress response.

RESULTS Transcript Profiling Reveals Functional Overlap of MSI1 with CAF-1 and RBR One prominent consequence of the strong reduction of MSI1 levels in msi1-cs plants is abnormal late leaf development. After apparently normal early leaf development, approximately 3 weeks after germination, fully expanded rosette leaves begin to bulge and become highly asymmetric (Hennig et al., 2003b). This phenotype does not occur in fas1 or fas2, which are mutants of the other Arabidopsis CAF-1 subunits (Reinholz, 1966; Leyser and Furner, 1992; Kaya et al., 2001; Hennig et al., 2003b). Similar to the fas1 and fas2 mutants, however, msi1-cs but not segregating MSI1 overexpressing plants (msi1-oe) had highly over-branched trichomes on rosette leaves (Exner et al., 2006, 2008 and Supplemental Figure 1). To better understand the effects of reduced MSI1 levels on global gene expression, we used Affymetrix ATH1 microarrays to profile transcripts of wild-type and msi1-cs leaves at a time in development when the msi1-cs phenotype was barely detectable (21 DAG). Median cv values for replicate array sets were between 3.1 and 4.6%, demonstrating the high quality of the data. We identified 515 up-regulated and 337 downregulated probe sets (Supplemental Tables 1 and 2). Because MSI1 is part of CAF-1 (Kaya et al., 2001; Exner et al., 2006), we hypothesized that genes with altered expression in msi1cs plants are also affected by loss of CAF-1. In agreement with this hypothesis, there was a significant overlap between the genes affected in msi1-cs or in CAF-1 mutants (p = 1.0E–28) (Figure 1A). The overlap was most significant for up-regulated genes and included AtDMC1 (AT3G22880), a H3.2 replacement histone variant gene (AT1G13370), cyclin B1;1 (AT4G37490), and a DNA polymerase delta subunit gene (AT1G09815). Increased expression of H3.2 and CYCLIN B1;1 was confirmed by RT–PCR in independent leaf samples of msi1-cs plants as well as in leaves of an independent MSI1 co-suppression line (MSI1OEa3) (Figure 2).

Figure 1. Overlapping Gene Sets Are Controlled by MSI1, CAF-1, and the E2F–DP–RBR1 Pathway. (A) Venn diagram representations of overlapping gene sets with altered expression in msi1-cs and in the CAF-1 mutants fas1 and fas2 (Scho¨nrock et al., 2006b). (B) Venn diagram representations of overlapping gene sets with altered expression in msi1-cs and in E2Fa–DPa overexpressing plants (Vandepoele et al., 2005). The upper and lower panels in (A) and (B) represent up-regulated and down-regulated genes, respectively. (C) Increased expression of cell-cycle-phase-specific genes in msi1-cs plants. 9910 genes identified by Menges et al. (2003) were sorted into 14 bins, based on their maximal expression during the cell cycle. Shown are the mean 6 SE of the signal log ratios (SLR) between expression signals in wild-type and msi1-cs plants for all genes in a bin. Shaded areas represent periods of S-phase and mitosis (M), respectively.

In addition to its function in CAF-1, MSI1 interacts with RBR1 (Ach et al., 1997; Jullien et al., 2008), which, like RB in animals, is proposed to repress the E2F-DP pathway of S-phase entry in Arabidopsis as well (van den Heuve and Dyson, 2008). As expected, several genes were similarly affected in msi1-cs and in lines that ectopically express E2Fa-DPa and thus bypass the RBR1-dependent cell cycle block (Figure 1B). Similarly to the comparison to CAF-1 mutants, the overlap was most significant for up-regulated genes (p = 3E–23) and included

Alexandre et al.

d


|

677

not overlap more than expected by chance (Table 1). It is possible that PRC2-dependent effects on gene expression in msi1-cs could be masked by stronger and more widespread PRC2-independent effects. Nevertheless, many PRC2 targets appear to remain silent in msi1-cs. Persistent silencing of PRC2 targets in msi1-cs plants is most likely due to the residual MSI1 protein level and is in agreement with the much more severe phenotype of strong alleles of PRC2-subunit genes (Yoshida et al., 2001; Chanvivattana et al., 2004). It is not clear, however, why PRC2 targets do not respond uniformly to reduced MSI levels. Together, transcriptome data of msi1-cs plants revealed an interesting overlap with CAF-1 and RBR1 but not with PRC2 function. Despite the high significance (p values , 1E–20), the overlap with CAF-1 and RBR1 involved only about 10% of the genes up-regulated in msi1-cs and only 3% of the down-regulated genes. This suggests that transcriptional effects in msi1-cs plants are not mainly caused by defects in CAF-1, RBR, or PRC2 activity, but are rather caused by other MSI1 functions. Figure 2. The Expression of Replacement Histone H3.2 and CYCLIN B1;1 Genes Is Up-Regulated in msi1-cs. (A) CYCB1;1 and H3.2 average signals from microarrays. Error bars indicate the range of the measurements. Numbers above bars represent fold change versus wild-type. (B) Confirmation of CYCB1;1 and H3.2 up-regulation in independent samples of the original msi1-cs line MSI1OEc2 and of the independent msi1-cs line MSI1OEa3. RNA was extracted from 26-day-old rosette leaves for all genotypes. GAPDH was used as reference gene.

ATORC4 (AT2G01120), three genes for DNA-replication proteins (RFC-like AT1G63160, RPA2 AT2G24490, AT5G49010), PCNA2 (AT2G29570), a ribonucleoside-diphosphate reductase gene (TSO2, AT3G27060), E2FC (AT1G47870) and 13 histone genes (AT5G65360, AT5G59870, AT5G59970, AT5G22880, AT5G10400, AT5G10390, AT3G53730, AT3G53650, AT3G46320, AT1G07660, AT2G38810, AT1G09200, AT2G28740). Thus, it is likely that the RBR-dependent block of E2Fa-DPa-mediated S-phase activation is at least partially impaired in msi1-cs plants. This notion is fully supported by the significant enrichment of S-phase-specific genes among the up-regulated genes in msi1-cs (Table 1). Similarly, a significant up-regulation of S-phase-specific genes became visible when plotting average fold changes in msi1-cs for cell-cycle-specific genes (Figure 1C). Arabidopsis MSI1 is also part of PRC2-like complexes, and some PRC2 targets including AGAMOUS and AGL19 are activated in msi1-cs plants (Hennig et al., 2003b; Scho¨nrock et al., 2006a). PRC2-like complexes repress their target genes via trimethylation of histone H3 lysine 27 (H3K27me3) (Ko¨hler and Villar, 2008). Therefore, a strong enrichment for genes with H3K27me3 could be expected in the gene set with increased expression in msi1-cs. The sets of genes up-regulated in msi1-cs and those carrying H3K27me3 marks, however, did

ABA-Responsive Genes Are Activated in msi1-cs Plants In order to test whether reduction of MSI1 levels preferentially affects particular cellular functions or pathways, we analyzed the distribution of Gene Ontology (GO) terms in the sets of upand down-regulated genes. Analysis of top-level GO terms (GO-SLIM) revealed a strong enrichment of genes for nuclear proteins and for chromatin-related proteins among the upregulated genes (Supplemental Tables 3 and 4). This observation is consistent with CAF-1 defects expected in msi1-cs. Down-regulated genes showed a strong enrichment for plastid proteins. When using the entire set of GO terms instead of GO-SLIM, the terms ‘nucleosome’ (p = 6.8E–17), ‘nucleosome assembly’ (p = 1.5E–09), and ‘response to abscisic acid stimulus’ (p = 8.6E–05) were enriched among up-regulated genes, and the term ‘chloroplast’ (p = 7.3E–07) was enriched among the down-regulated genes. While the GO terms ‘nucleosome’ and ‘nucleosome assembly’ can be directly related to the role of MSI1 in CAF-1, the small p-values for the terms ‘response to abscisic acid stimulus’ and ‘chloroplast’ were unexpected. To test whether the identification of enriched GO-terms in our dataset is sensitive to selection thresholds for p-values and fold changes, we used PAGE (parametric analysis of gene set enrichment, Kim and Volsky, 2005) to search for non-randomly distributed GO terms in the entire dataset. This analysis revealed that three terms had a distribution that was highly significant for up-regulation: ‘nucleosome’ (p = 2.7E–21), ‘nucleosome assembly’ (p = 3.6E–19), and ‘response to abscisic acid stimulus’ (p = 2.4E–14). The tendency for down-regulation was most significant for the terms ‘chloroplast’ (p = 8.6E–56), ‘chloroplast thylakoid membrane’ (p = 1.1 E–46), ‘plastoglobule’ (p = 7.5E–16), and ‘chloroplast thylakoid lumen’ (p = 3.7E–15). Thus, the results of GO enrichment analysis and PAGE closely correspond, and the observed GO

678

|

Alexandre et al.

d


term enrichments are independent of arbitrary p-value or fold change thresholds. The identification of the term ‘response to abscisic acid stimulus’ with very small p-values prompted us to investigate the distribution of known hormone-responsive genes (Nemhauser

et al., 2006) in msi1-cs data in more detail. The analysis of genes up-regulated in msi1-cs revealed a highly significant enrichment of ABA-inducible genes (p = 1E–40) and significant enrichment of MeJA (methyl jasmonate) inducible genes (p = 1E–12) (Figure 3A). Considerable activation of other

Table 1. Functional Classification of Genes with Altered Expression in msi1-cs into Cell-Cycle Phase Categories (Menges et al., 2003) and H3K27me3-Containing Genes (Zhang et al., 2007). Observed

Expected

Log (enrichment)

p (enriched)

p (depleted)

Up-regulated S

77

20

1.9

8.3E–25

1.00

G2

1

0

2.1

1.00

1.00

M

7

6

0.2

0.97

1.00

G1

19

10

0.9

0.01

1.00

Down-regulated 11

13

–0.2

1.00

1.00

G2

S

0

0

n.a.

1.00

1.00

M

3

4

–0.4

1.00

1.00

G1

8

6

0.3

0.83

1.00

84

83

0.02

0.82

1.00

37

54

1.00

0.01

Up-regulated H3K27me3 Down-regulated H3K27me3

–0.5

n.a., not applicable.

Figure 3. ABA-Responsive Genes Are Up-Regulated in msi1-cs. (A) Functional classification of genes with altered expression in msi1-cs into hormone-response categories (Nemhauser et al., 2006). Shown are p-values for significant enrichment of hormone-responsive genes. (B, D) Average signals from microarrays for four selected genes found up-regulated in msi1-cs (At2g39030, MYB2, LTP3, and LTP4). Error bars indicate the range of the measurements. Numbers above bars represent fold change versus wild-type. (C, E) Confirmation of microarray results for two independent MSI1-co-suppression lines (msi1-cs line and MSI1OEa3). RNA was extracted from 26-day-old rosette leaves.

Alexandre et al.

hormone responses was not detected in msi1-cs plants. Crosstalk between ABA and MeJA signaling has been reported (Munemasa et al., 2007), and it is possible that the less pronounced enrichment of MeJA-responsive genes is a secondary effect of the activation of ABA-inducible genes in msi1-cs. Despite the strong enrichment of ABA-inducible genes among genes with increased expression in msi1-cs, analysis of gene set overlaps revealed that only about 10% of genes reported to be ABA-inducible (Nemhauser et al., 2006) were activated in msi1-cs plants. Because microarray data can generate false-positive results, we tested the expression of four known ABA-responsive genes that appeared up-regulated in the msi1-cs microarrays in independent samples. Up-regulation of all tested genes was confirmed in msi1-cs plants (Figure 3B–3E). In addition, the four genes were also up-regulated in the independent MSI1-cosuppression line MSI1OEa3. The identification of many genes related to ABA responses in msi1-cs by GO analysis, PAGE and overlap with published microarray data suggests that a reduced MSI1 level leads to biological relevant changes in the expression of genes responsive to ABA.

Transcript Levels of Genes for ABA Synthesis or Signaling Are Not Increased in msi1-cs The increased expression of ABA-responsive genes in msi1-cs could be caused by increased ABA synthesis or by constitutively activated ABA signaling. When analyzing genes for ABA synthesis, we found no indication for increased transcription in msi1-cs (Supplemental Table 5). A key enzyme of ABA biosynthesis is 9-cis-epoxycarotenoid dioxygenase (NCED), and NCED3 is strongly induced by dehydration and high salinity (Yamaguchi-Shinozaki and Shinozaki, 2006). In msi1-cs plants, NCED genes were either repressed or not changed, and NCED3, in particular, was three-fold repressed. These results suggest that there was no transcriptional activation of ABA biosynthesis in msi1-cs leaves. ABA- and stress-induced gene expression is mediated by a network of signaling pathways. Major transcriptional activators are ABA-responsive element (ABRE) binding factors (ABF), a pair of MYC and MYB transcription factors and drought-responsive element (DRE) binding factors (DREB), which are also called cold-responsive element (CRT) binding factors (CBF) (Yamaguchi-Shinozaki and Shinozaki, 2006). Despite the strong activation of ABA-responsive genes in msi1-cs, expression of most of the transcription factor genes related to ABA signaling was unchanged. To the contrary, DREB1 and DREB2, which are involved in ABA-independent responses to osmotic stress and cold, were strongly repressed. MYB2 was the only known positive regulator of ABA responses that was strongly induced, but MYB2 is itself activated by ABA and MYB2 is involved in late responses to ABA. Thus, the MYB2 induction in msi1-cs plants could well be a secondary consequence of the activation of primary ABA-inducible genes.

d


|

679

Genes Responding to Osmotic and Salt Stress Are Specifically Up-Regulated in msi1-cs ABA signaling is involved in the response to abiotic stress, and different abiotic stresses often affect similar genes. Therefore, we tested whether up-regulated genes in msi1-cs plants usually respond to many or only to specific stresses using the AtGenExpress stress reference dataset (Kilian et al., 2007) (Supplemental Figure 2). The distribution of Shannon entropy as a measure for specificity revealed that genes that were up-regulated in msi1cs usually respond to fewer abiotic stresses (lower entropy) than down-regulated genes or the complete gene set. In order to identify abiotic stresses that activate many of the genes up-regulated in msi1-cs, we performed clustering of genes up-regulated in msi1-cs according to their stress response by SOTA (Self Organizing Tree Algorithm, Herrero et al., 2001). This analysis resulted in a major cluster of 288 genes, which were specifically induced by osmotic and salt stress, but generally only weakly by cold, drought, oxidative, UV-B, genotoxic, wound, or heat stress (Figure 4A). Cluster 2 of the up-regulated genes in msi1-cs consisted of 123 genes that were generally repressed by osmotic and salt stress. Of the remaining three smaller clusters, only genes in cluster 4 had a highly characteristic expression signature because these genes showed strong activation by genotoxic stress. Next, we tested whether there was any substructure in cluster 1 by performing SOTA specifically on the genes from cluster 1 (Figure 4B). As observed before, most genes in this cluster respond to osmotic and salt stress, but some genes respond to additional stresses. Of the 288 genes in cluster 1, 63 respond strongly to ABA, but not to cold or other stresses. A further 62 genes do not respond strongly to ABA within 3 h but respond to cold. Another 33 genes respond rapidly and strongly but transiently to heat stress, and 130 genes respond predominantly to osmotic and salt stress. Thus, many genes that can be activated by osmotic and salt stress were up-regulated in msi1-cs. This result was confirmed by Principal Component Analysis (PCA) (Figure 4C). The first principle component in the stress and msi1-cs dataset separated msi1-cs and the responses to ABA, osmotic, and salt stress from the other stress and control samples. This component reflects common up-regulation of genes in these but not the other samples. The second principle component separated msi1-cs from the response to ABA, osmotic, and salt stress and reflects opposite gene regulation in msi1-cs and stress responses. Genes with such opposite regulation include cell-cycle genes, which are up-regulated in msi1-cs, but repressed in response to stress to prevent cell proliferation under critical conditions. Together, these results demonstrate that only a specific subset of stress-responsive genes, namely genes that respond to osmotic and salt stress, was activated in msi1-cs plants.

Genes that Are Up-Regulated in msi1-cs Often Contain ABA-Responsive Elements or TCP-Binding Sites Co-expressed genes often share common transcription factor binding sites or motifs in their promoters. We searched for

680

|

Alexandre et al.

d


Figure 4. Osmotic and Salt Stress-Responsive Genes Are Up-Regulated in msi1-cs. (A) Cluster analysis of gene expression after ABA treatment and during abiotic stress. Self-organizing tree clustering (SOTA) of the 515 genes up-regulated in msi1-cs. (B) SOTA of the 288 genes from cluster 1 in (A). Gene expression data after ABA treatment and after various stresses was taken from Goda et al. (2008) and Kilian et al. (2007), respectively. Each horizontal line represents the average expression profile of one cluster. The color scale at the top represents the expression level (–2 to +2). The tree structure on the left represents cluster similarities. Numbers on the right are labels for clusters and numbers of probesets in each cluster. (C) Principle component analysis of gene expression at 3 h after ABA treatment, at 24 h after abiotic stress treatment, and in msi1-cs. The applied stresses were cold, osmotic, salt, drought, genotoxic, oxidative, UV-B, wounding, and heat (for details, see Kilian et al., 2007).

over-represented motifs using the genes in the five SOTA clusters, which revealed several consensus motifs. The major cluster 1 was characterized by the presence of the ABA-responsive element (ABRE, PyACGTGGC) and ABRE-like motifs (C/G/ T)ACGTG(G/T) (A/C) that are bound by ABFs (Guiltinan et al., 1990; Choi et al., 2000) (Figure 5A). In addition, genes from cluster 1 often contained CACG-like motifs that function in drought stress and overlap with ABRE-like motifs but are

bound by NAC transcription factors ANAC019, ANAC055, and ANAC072 (Tran et al., 2004) (Figure 5B). Cluster 2 was characterized by the presence of site II motifs (TGGGCC/T) that function in cell proliferation and are bound by TCP-domain transcription factors (Kosugi et al., 1995; Kosugi and Ohashi, 2002) (Figure 5C). Several other motifs were identified, but could not be matched to data in the literature (data not shown).

Alexandre et al.

Together, the promoter motif analysis revealed likely mechanisms for activation of two large groups of genes in msi1-cs. First, activation of genes in cluster 1, which are also induced by osmotic and salt stress, could be mediated by ABFs binding to ABREs. Because the microarray data did not indicate increased transcription of ABFs, MSI1 likely affects ABF function at the protein level, such as by interfering with ABF activity. Second, activation of genes in cluster 2, which are repressed by osmotic and salt stress, could be mediated by TCPs binding to site II elements. TCPs regulate genes involved in cell proliferation such as cyclin B1;1 (Li et al., 2005), which is strongly up-regulated in msi1-cs (Figure 2A and 2B). The repression of these genes upon osmotic or salt stress in wild-type plants probably reflects the inhibitory effect of abiotic stress on cell proliferation. In contrast to the ABFs, transcript levels of some TCPs were strongly affected in msi1-cs plants. In particular, TCP4 expression was

d


|

681

four-fold reduced while TCP17 and TCP20 expression was two-fold increased, suggesting that MSI1 affects expression of TCP-target genes via altered TCP transcription. It is likely that activation of E2F and TCP target genes in expanded leaves, which usually do not contain proliferating cells, contributes to the altered morphology of mature msi1-cs leaves.

RD20 Is a Direct Target of MSI1 Many genes with ABREs in their promoter were up-regulated in msi1-cs plants (Figures 3B and 5A). To test whether some of these genes are direct targets of MSI1 or are indirectly affected, we used a transgenic line, in which an MSI1–GFP fusion construct complements the msi1-1 null allele (Figure 6A and 6B) and the gametophytic lethal msi1-1 mutant phenotype. MSI1–GFP plants, which do not contain endogenous MSI1, developed normally and did not have any of the developmental defects of msi1-cs plants (data not shown). When using the MSI1–GFP line for chromatin immunoprecipitation (ChIP) with anti-GFP antibodies, we observed strong enrichment (4.2-fold) over wild-type for RD20, but no relevant enrichment for MYB112, At2g39030, MYB74, and MYB2 (Figure 6C and data not shown), suggesting that at least some of the osmotic stressresponsive genes are targets of MSI1.

msi1-cs Plants Have Increased Free Proline Levels and Increased Drought Tolerance Because expression of ABA- and osmotic-stress-responsive genes was increased in msi1-cs, it was possible that msi1-cs plants had an altered drought stress tolerance. To test this

Figure 6. RD20 Is a Direct Target of MSI1.

Figure 5. Analysis of Enriched Promoter Elements for Genes with Altered Expression in msi1-cs. (A) Five enriched promoter elements identified by the WeederWeb algorithm (Pavesi et al., 2006) in cluster 1 from Figure 4A resemble the ABA-responsive element (ABRE, PyACGTGGC) and ABRE-like motifs (C/G/T)ACGTG(G/T) (A/C) that are bound by ABFs. (B) One enriched promoter element identified in cluster 1 from Figure 4A resembles the CACG-like motif that is bound by NAC transcription factors. (C) Three enriched promoter elements identified in cluster 2 from Figure 4A resemble the site II motif (TGGGCC/T) that is bound by TCP-domain transcription factors.

(A) Structure of the MSI1–GFP construct introduced into the msi1-1 mutant. 2 kb upstream promoter sequences of the MSI1 gene were used to drive expression of a MSI1 cDNA–GFP fusion. The used MSI1–GFP line is homozygous for the msi1-1 allele. (B) Protein blots of wild-type (WT) and MSI1–GFP transgenic plants using 20 lg of soluble proteins extracted from 14-day-old seedlings. Endogenous MSI1 and the MSI1–GFP fusion protein were both detected by the anti-MSI1 antibody, while the anti-GFP antibody detected only the fusion protein. (C) The binding of MSI1 to possible target genes was tested in 11day-old seedlings of an MSI1–GFP line in which the MSI1–GFP fusion construct complements the msi1-1 null allele. Shown is one of two independent ChIP experiments. Numbers represent enrichment of signal in MSI1–GFP over wild-type. aGFP, anti-GFP antibody; n.d., not determined due to weak signals.

682

|

Alexandre et al.

d


hypothesis, we investigated msi1-cs plant performance under severe drought stress conditions. After 12 and 16 d without watering, msi1-cs plants were consistently less stressed than wild-type plants (Figure 7A and 7B). When watering was resumed after 17 d, only msi1-cs but not wild-type plants could be rescued (data not shown). Next, we measured free proline levels in rosette leaves of wild-type and msi1-cs non-stressed plants. Proline is thought to have an important role in osmoregulation, and proline readily accumulates in plant tissues in response to diverse stresses, most notably low water potential as the result of drought and salinity (Yoshiba et al., 1997). Free proline levels were about five-fold increased in msi1-cs (Figure 7C). Because we could not detect significant up-regulation of genes involved in proline biosynthesis on the ATH1 microarray (data not shown), proline accumulation in msi1-cs plants most likely reflects post-translational regulation. Together, increased expression of osmotic stress-responsive genes and increased free proline content can lead to an increased drought tolerance of msi1-cs plants.

overlap, however, more than 85% of genes affected in msi1-cs were not responsive to CAF-1 or the E2F–DP–RBR pathway, suggesting that, to a large extent, MSI1 acts independently of CAF-1 and the E2F–DP–RBR pathway. Unexpectedly, we failed to detect a significant enrichment for genes with H3K27me3 among the MSI1-depending genes, suggesting that functions of MSI1-containing PRC2-like complexes are only partially impaired in msi1-cs. In msi1-cs plants, expression of the two TCP transcription factor genes TCP17 and TCP20 was significantly increased. These changes in expression are most likely relevant because site II motifs, which are bound by TCP-domain transcription factors (Kosugi et al., 1995; Kosugi and Ohashi, 2002), were significantly enriched in promoters of genes up-regulated in msi1-cs. Because TCPs activate genes required for cell proliferation such as cyclin B1;1 and are known to regulate leaf curvature (Palatnik et al., 2003), the altered expression levels of TCP genes in msi1-cs may partially account for the bulged leaf phenotype.

msi1-cs Plants Have Increased Drought Stress Tolerance

DISCUSSION Transcriptional Profiling of msi1-cs MSI1 is required throughout plant development (Hennig et al., 2003b; Ko¨hler et al., 2003a; Guitton et al., 2004; Bouveret et al., 2006; Exner et al., 2006), and MSI1 can be a subunit of several protein complexes (Hennig et al., 2005). Genes whose expression was altered in plants with reduced MSI1 levels significantly overlapped with target genes of CAF-1 and the E2F–DP–RBR pathway. This is not unexpected, because MSI1 is a subunit of Arabidopsis CAF-1 and interacts with plant RBR1 physically and genetically (Ach et al., 1997; Kaya et al., 2001; Jullien et al., 2008). Despite the statistically significant

Analyses of Gene Ontology annotations revealed that many of the genes up-regulated in msi1-cs are also induced by ABA, salt stress, and osmotic stress. Salt stress consists of two major components: osmotic stress and cytotoxicity of sodium cations. Osmotic stress, which is commonly induced experimentally by salt or mannitol, lowers the soil water potential and thus mimics drought stress. Surprisingly, neither clustering nor PCA revealed any similarity between transcript profiles from msi1-cs plants and plants responding to drought stress. This is most likely due to the fact that in the set of reference experiments, cold, osmotic, and salt stresses were applied continuously for 24 h while drought stress was applied only transiently for 15 min (Kilian et al., 2007). Because genes

Figure 7. msi1-cs Plants Have Increased Free Proline Levels and Drought Tolerance. (A) Wild-type (top row) and msi1-cs (bottom row) plants after 12 d without watering. Note the wilting rosette leaves of wild-type plants in contrast to the still unstressed msi1-cs plants. (B) Wild-type (top row) and msi1-cs (bottom row) plants after 16 d without watering. Note the more severe wilting of wild-type plants compared to msi1-cs plants. (C) Free proline levels were measured in rosette leaves of non-stressed 7-week-old wild-type and msi1-cs plants.

Alexandre et al.

up-regulated in msi1-cs did not usually respond to other stresses, such as heat, UV-light, or DNA-damaging agents, MSI1 is required specifically to repress genes that are activated by reduced water availability. Although several genes that respond to osmotic and drought stress (e.g. RD20, LPT3, LPT4, AT1G80160) were upregulated in msi1-cs, many typical marker genes such as RD29A and ADH1 were not. Therefore, MSI1 specifically represses genes involved in a subset of the response to reduced water availability. At present, we do not know where MSI1 acts in the stress-signaling cascade. It is unlikely, however, that MSI1 acts far upstream, where it could affect the entire response, but rather downstream to repress a subset of stress-inducible genes. This hypothesis is supported by the finding that MSI1 can bind directly to the chromatin of the downstream stress-signaling target RD20. In addition to the activation of a subset of dehydration-responsive genes, msi-cs plants had also increased free proline levels and increased tolerance to drought, demonstrating that the altered transcriptional programs in msi1-cs are physiologically relevant. Increased drought tolerance is likely caused by effects of multiple genes, but one key gene could be AT1G80160, which is four-fold up-regulated in msi1-cs and encodes a glyoxalase I family protein. Glyoxalases function in the detoxification of methylglyoxal, which is formed as a byproduct of carbohydrate and lipid metabolism and increases strongly upon drought stress (Yadav et al., 2005). Therefore, we suggest that detoxification of toxic aldehydes through the glyoxalase pathway contributes together with increased free proline levels to the increased resistance of msi1cs against drought stress.

Epigenetic Regulation of Stress Responses Because msi1-cs plants have reduced amounts of MSI1 protein and MSI2-5 are not affected (Hennig et al., 2003b), we conclude that MSI1 functions to repress part of the drought stress response in Arabidopsis. Given the known role of MSI1 and MSI1-like proteins in chromatin dynamics (Hennig et al., 2005), it is likely that this repression has an epigenetic nature. Epigenetic control of stress responses is not without precedence (for review, see Chinnusamy et al., 2008). Histone expression and histone post-translational modifications, for instance, are altered during salt and drought stress (Waterborg et al., 1989; Ascenzi and Gantt, 1997, 1999; Kim et al., 2008). More specifically, it has been proposed that the AP2/ EREBP-type protein AtERF7 binds to the GCC-box of ABA-induced genes and recruits an AtSin3/HDA19/AtSAP18 co-repressor complex to attenuate ABA signaling (Song et al., 2005; Song and Galbraith, 2006). Our microarray data provide no evidence that MSI1 is involved in this process, because, in contrast to the ABRE, the GCC-box is not enriched among msi1-cs upregulated genes. Similar to the histone deacetylase HDA19, the histone deacetylase AtHD2C affects expression of ABA-responsive genes and tolerance to salt and drought stresses, but possibly by an indirect effect (Sridha and Wu, 2006). Of seven

d


|

683

reported target genes affected by AtHD2C, only one is changed in msi1-cs, suggesting that MSI1 affects drought stress tolerance independently of AtHD2C. Recently, it was proposed that ABA responses are modulated through the regulation of a putative SWI/SNF chromatin-remodeling complex involving SWI3B, an Arabidopsis homolog of the yeast SWI3 subunit of SWI/SNF chromatin-remodeling complexes (Saez et al., 2008). Future work will identify MSI1 protein binding partners that, together with MSI1, control expression of drought stress-inducible genes. Characterization of MSI1 function in the regulation of drought responses could therefore change our understanding of how chromatin modifications affect the response of plants to stress and open new perspectives for applied research.

METHODS Plant Material and Growth Conditions Lines MSI1OEc2 and MSI1OEa3, which ectopically express MSI1 and give rise to MSI1 co-suppression plants, were described before (Hennig et al., 2003b). The abbreviation msi1-cs refers to MSI1 co-suppression plants from line MSI1OEc2. The MSI1–GFP line was constructed by transforming a MSI1–GFP fusion construct driven by 2-kb upstream sequences of the MSI1 gene into heterozygous msi1-1 plants (Ko¨hler et al., 2003a). T2 progeny was screened for homozygous msi1-1 plants that were rescued by the MSI1–GFP fusion protein. Similar to the previously described MSI1–TAP construct (Bouveret et al., 2006), the MSI1–GFP fusion could complement the msi1-1 mutation. MSI1–GFP plants developed normally, except for a mild delay in flowering that was, however, not as severe as that observed in msi1–tap1 plants (data not shown). All plants used in this study are in the Columbia background. Seeds were plated on Murashige and Skoog (MS) medium (Duchefa, Haarlem, The Netherlands), stratified for 2 d at 4C, and grown on plates for 10 d before transfer onto soil. Plants were kept in Conviron growth chambers with mixed cold fluorescent and incandescent light (110–140 lmol m2 s, 21 6 2C) under LD (16 h light) photoperiods.

Drought Experiment Thirty-day-old wild-type and msi1-cs plants, which were raised under a non-limiting watering regime, were deprived of water and wilting of rosette leaves was monitored. After 17 d, watering was resumed and the recovery of leaf turgor was monitored visually.

RNA Isolation and RT–PCR RNA was extracted as previously described (Leroy et al., 2007). 1 lg of DNA-free RNA was reverse-transcribed using an oligo(dT) primer and RevertAid M-MuLV Reverse Transcriptase (Fermentas, St Leon Roth, Germany). Aliquots of the generated cDNA were used as template for PCR with gene-specific

684

|

Alexandre et al.

d


primers (Supplemental Tables 6 and 7). qPCR was performed in a 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), using FastStart Universal Probe Master (Rox) reagent and the Universal Probe Library set (Roche Diagnostics, Rotkreuz, Switzerland) according to manufacturers’ instructions. All amplification plots were analyzed using the 7500 Fast System SDS v1.4 Software default settings. qPCR reactions were performed in triplicate and results were analyzed as described (Simon, 2003). Gene expression levels were normalized to an ACTIN control gene.

Array Hybridization and Evaluation Experimental procedures are described according to MIAME (‘minimum information about a microarray experiment’) standards (Brazma et al., 2001). Data were deposited into the ArrayExpress database (Accession number E-MEXP-1829).

Experimental Design Transcriptional profiling of wild-type and msi1-cs rosette leaves was described before (Scho¨nrock et al., 2006a). Briefly, plants were grown for 23 d in growth chambers at 21C under LD photoperiods (16 h light) before rosette leaves were harvested from at least 12 plants and pooled for each replica. Two independent biological replicas were used for hybridization.

p , 0.05. To enrich for biologically relevant changes, only probe sets that changed at least 1.5-fold in all replicate experiments were selected. Differentially expressed genes were grouped into collapsed functional gene ontology categories (GOSLIM, obtained from www.Arabidopsis.org). Grouping according to preferred phase of cell cycle expression was based on data from Menges et al. (2003). Grouping according to phytohormone regulation was based on data from Nemhauser et al. (2006). Data for H3K27me3 target loci were from Zhang et al. (2007). The significance of enrichment was estimated based on the hypergeometric test and multiple-testing correction using the Bonferroni-correction. In addition to GOSLIM analysis, enrichment of detailed GO categories (obtained from TAIR) was tested. In this case, multiple-testing correction was according to Benjamini and Hochberg (1995), with a critical p-value of 1.0E–3. Lists of E2F–DP regulated genes were from Vandepoele et al. (2005). The transcript profiling data for the abiotic stress response were obtained from the AtGenExpress dataset (Kilian et al., 2007). Gene expression values measured on ATH1 microarrays were normalized with GCRMA; replicate measurements were averaged and normalized to time zero. Clustering was performed using SOTA (Herrero et al., 2001) in the MEV software suite (Saeed et al., 2003). Motif discovery in co-expressed genes was done using WeederWeb version 1.3 (Pavesi et al., 2006) using 500 bp of upstream sequences obtained from TAIR.

Array Design, Samples, and Hybridizations Affymetrix Arabidopsis ATH1 GeneChips were used throughout the experiment (Affymetrix, Santa Clara, CA). The exact list of probes present on the arrays can be obtained from the manufacturer’s website (www.affymetrix.com). Analysis was based upon annotations compiled by TAIR (www.Arabidopsis.org, version 2007-5-2). RNA was prepared, labeled, and hybridized to the arrays as described (Hennig et al., 2003a).

Measurements The arrays were scanned using a confocal scanner Agilent GS 2500.

Chromatin Immunoprecipitation ChIP was performed as described previously (Johnson et al., 2002). Anti-GFP antibody (catalog no. A11122, Molecular Probes) was used for immunoprecipitation. Immunoprecipitated DNA from at least two independent experiments was analyzed by PCR using gene-specific primers for RD20, MYB112, At2g39030, MYB74, and MYB2 (Supplemental Table 7). Band intensities were quantified using Image J software and enrichment relative to wild-type was calculated according to Geisberg and Struhl (2004).

Immunoblotting Normalization Signal values were derived from Affymetrix *.cel files using GCRMA (Wu et al., 2004).

Bioinformatic Analysis All data processing was performed using the statistic package R (version 2.6.2) that is freely available at www.r-project.org/ (Ihaka and Gentleman, 1996). Quality control was done using the affyQCReport package in R. In addition, we calculated coefficients of variation (cv) between replicates as a quantitative measure of data quality and consistency between replicates as described previously (Ko¨hler et al., 2003b). Differentially expressed genes were identified using the EVE algorithm in R (Wille et al., 2007). Multiple-testing correction was done according to Benjamini and Hochberg (1995). Probe sets were considered as differentially expressed when

For isolation of soluble proteins, 1 g of 14-day-old seedlings of wild-type and MSI1:GFP plants grown under LD conditions (16 h light/8 h dark) were ground with a glass pestle in 5 ml extraction buffer (40 mM Tris, pH 8, 10 mM KCl, 10 mM MgCl2, 0.5 M Saccharose, 25% Glycerol, 1 mM beta-Mercaptoethanol, 0.01% PMSF) and centrifuged for 15 min at 4C with 4000 g. 20 lg of total protein were loaded on a 10% SDS–polyacrylamid gel. Resolved proteins were blotted onto PVDF membrane. Membranes were incubated with first antibody (1:1000 dilutions of mouse a-MSI1 or mouse a-GFP) overnight at 4C and second antibody (1:5000 dilution of goat a-mouse HRP; Jackson Immuno Research) for 2 h at RT.

Measurements of Free Proline Content Seven-week-old rosette leaves from five wild-type and msi1-cs plants were pooled and ground in liquid nitrogen. 100 mg of

Alexandre et al.

powder was used for proline extraction according to Sanada et al. (1995). The resulting supernatant was diluted with two volumes of distilled water (final volume of 600 ll) before analysis with an AccQTag LC system, running on a Waters Acquity UPLC (Waters, Milford Massachusetts, USA).

SUPPLEMENTARY DATA Supplementary Data are available at Molecular Plant Online.

FUNDING This work was supported by the Swiss National Science Foundation (3100AO-116060) and ETH (TH-16/05–2) to L.H.

ACKNOWLEDGMENTS We thank Claudia Ko¨hler for helpful comments on the manuscript and Claudia Ko¨hler and Olivier Leroy (Zurich) for the MSI1–GFP line. We thank Birgit Roth Zgraggen at the Functional Genomics Center Zurich for help with the proline analysis. No conflict of interest declared.

REFERENCES Ach, R.A., Taranto, P., and Gruissem, W. (1997). A conserved family of WD-40 proteins binds to the Retinoblastoma protein in both plants and animals. Plant Cell. 9, 1595–1606. Ascenzi, R., and Gantt, J.S. (1997). A drought-stress-inducible histone gene in Arabidopsis thaliana is a member of a distinct class of plant linker histone variants. Plant Mol. Biol. 34, 629–641. Ascenzi, R., and Gantt, J.S. (1999). Molecular genetic analysis of the drought-inducible linker histone variant in Arabidopsis thaliana. Plant Mol. Biol. 41, 159–169. Ausin, I., Alonso-Blanco, C., Jarillo, J.A., Ruiz-Garcia, L., and MartinezZapater, J.M. (2004). Regulation of flowering time by FVE, a Retinoblastoma-associated protein. Nat. Genet. 36, 162–166. Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. Royal Stat. Soc. Ser. B. 57, 289–300. Bouveret, R., Scho¨nrock, N., Gruissem, W., and Hennig, L. (2006). Regulation of flowering time by Arabidopsis MSI1. Development. 133, 1693–1702. Brazma, A., et al. (2001). Minimum information about a microarray experiment (MIAME): toward standards for microarray data. Nat. Genet. 29, 365–371. Calonje, M., Sanchez, R., Chen, L., and Sung, Z.R. (2008). EMBRYONIC FLOWER1 participates in Polycomb group-mediated AG gene silencing in Arabidopsis. Plant Cell. 20, 277–291. Chanvivattana, Y., Bishopp, A., Schubert, D., Stock, C., Moon, Y.H., Sung, Z.R., and Goodrich, J. (2004). Interaction of Polycombgroup proteins controlling flowering in Arabidopsis. Development. 131, 5263–5276. Chinnusamy, V., Gong, Z., and Zhu, J.K. (2008). Abscisic acid-mediated epigenetic processes in plant development and stress responses. J. Integr. Plant Biol. 50, 1187–1195.

d


|

685

Choi, H., Hong, J., Ha, J., Kang, J., and Kim, S.Y. (2000). ABFs, a family of ABA-responsive element binding factors. J. Biol. Chem. 275, 1723–1730. Exner, V., Gruissem, W., and Hennig, L. (2008). Control of trichome branching by Chromatin Assembly Factor-1. BMC Plant Biol. 854. Exner, V., Taranto, P., Scho¨nrock, N., Gruissem, W., and Hennig, L. (2006). Chromatin assembly factor CAF-1 is required for cellular differentiation during plant development. Development. 133, 4163–4172. Geisberg, J.V., and Struhl, K. (2004). Quantitative sequential chromatin immunoprecipitation, a method for analyzing cooccupancy of proteins at genomic regions in vivo. Nucleic Acids Res. 32. Goda, H., et al. (2008). The AtGenExpress hormone- and chemicaltreatment data set: experimental design, data evaluation, model data analysis, and data access. Plant J. 55, 526–542. Guiltinan, M.J., Marcotte, W.R., Jr, and Quatrano, R.S. (1990). A plant leucine zipper protein that recognizes an abscisic acid response element. Science. 250, 267–271. Guitton, A.E., and Berger, F. (2005). Loss of function of MULTICOPY SUPPRESSOR OF IRA 1 produces nonviable parthenogenetic embryos in Arabidopsis. Curr. Biol. 15, 750–754. Guitton, A.E., Page, D.R., Chambrier, P., Lionnet, C., Faure, J.E., Grossniklaus, U., and Berger, F. (2004). Identification of new members of FERTILISATION INDEPENDENT SEED Polycomb group pathway involved in the control of seed development in Arabidopsis thaliana. Development. 131, 2971–2981. Hennig, L., Bouveret, R., and Gruissem, W. (2005). MSI1-like proteins: an escort service for chromatin assembly and remodeling complexes. Trends Cell. Biol. 15, 295–302. Hennig, L., Menges, M., Murray, J.A.H., and Gruissem, W. (2003a). Arabidopsis transcript profiling on Affymetrix genechip arrays. Plant Mol. Biol. 53, 457–465. Hennig, L., Taranto, P., Walser, M., Scho¨nrock, N., and Gruissem, W. (2003b). Arabidopsis MSI1 is required for epigenetic maintenance of reproductive development. Development. 130, 2555–2565. Herrero, J., Valencia, A., and Dopazo, J. (2001). A hierarchical unsupervised growing neural network for clustering gene expression patterns. Bioinformatics. 17, 126–136. Ihaka, R., and Gentleman, R. (1996). R: a language for data analysis and graphics. J. Comput. Graph. Stat. 5, 299–314. Johnson, L., Cao, X., and Jacobsen, S. (2002). Interplay between two epigenetic marks: DNA methylation and histone H3 lysine 9 methylation. Curr. Biol. 12, 1360–1367. Jullien, P.E., Mosquna, A., Ingouff, M., Sakata, T., Ohad, N., and Berger, F. (2008). Retinoblastoma and its binding partner MSI1 control imprinting in. Arabidopsis. PLoS Biol. 6. Kaya, H., Shibahara, K., Taoka, K., Iwabuchi, M., Stillman, B., and Araki, T. (2001). FASCIATA genes for Chromatin Assembly Factor1 in Arabidopsis maintain the cellular organization of apical meristems. Cell. 104, 131–142. Kenzior, A.L., and Folk, W.R. (1998). AtMSI4 and Rbap48 WD-40 repeat proteins bind metal ions. FEBS Lett. 440, 425–429. Kilian, J., Whitehead, D., Horak, J., Wanke, D., Weinl, S., Batistic, O., D’Angelo, C., Bornberg-Bauer, E., Kudla, J., and Harter, K. (2007).

686

|

Alexandre et al.

d


The AtGenExpress global stress expression data set: protocols, evaluation and model data analysis of UV-B light, drought and cold stress responses. Plant J. 50, 347–363.

Palatnik, J.F., Allen, E., Wu, X., Schommer, C., Schwab, R., Carrington, J.C., and Weigel, D. (2003). Control of leaf morphogenesis by microRNAs. Nature. 425, 257–263.

Kim, H.J., et al. (2004). A genetic link between cold responses and flowering time through FVE in Arabidopsis thaliana. Nat. Genet. 36, 167–171.

Pavesi, G., Mereghetti, P., Zambelli, F., Stefani, M., Mauri, G., and Pesole, G. (2006). MoD tools: regulatory motif discovery in nucleotide sequences from co-regulated or homologous genes. Nucleic Acids Res. 34, W566–W570.

Kim, J.M., To, T.K., Ishida, J., Morosawa, T., Kawashima, M., Matsui, A., Toyoda, T., Kimura, H., Shinozaki, K., and Seki, M. (2008). Alterations of lysine modifications on the histone H3 N-tail under drought stress conditions in Arabidopsis thaliana. Plant Cell Physiol. 49, 1580–1588. Kim, S.Y., and Volsky, D.J. (2005). PAGE: parametric analysis of gene set enrichment. BMC Bioinformatics. 6, 144. Ko¨hler, C., and Villar, C.B. (2008). Programming of gene expression by Polycomb group proteins. Trends Cell Biol. 18, 236–243. Ko¨hler, C., Hennig, L., Bouveret, R., Gheyselinck, J., Grossniklaus, U., and Gruissem, W. (2003a). Arabidopsis MSI1 is a component of the MEA/FIE Polycomb group complex and required for seed development. EMBO J. 22, 4804–4814. Ko¨hler, C., Hennig, L., Spillane, C., Pien, S., Gruissem, W., and Grossniklaus, U. (2003b). The Polycomb-group protein MEDEA regulates seed development by controlling expression of the MADS-box gene PHERES1. Genes Dev. 17, 1540–1553. Kosugi, S., and Ohashi, Y. (2002). DNA binding and dimerization specificity and potential targets for the TCP protein family. Plant J. 30, 337–348. Kosugi, S., Suzuka, I., and Ohashi, Y. (1995). Two of three promoter elements identified in a rice gene for proliferating cell nuclear antigen are essential for meristematic tissue-specific expression. Plant J. 7, 877–886. Leroy, O., Hennig, L., Breuninger, H., Laux, T., and Ko¨hler, C. (2007). Polycomb group proteins function in the female gametophyte to determine seed development in plants. Development. 134, 3639–3648. Leyser, H.M., and Furner, I.J. (1992). Characterisation of three shoot apical meristem mutants of Arabidopsis thaliana. Development. 116, 397–403. Li, C., Potuschak, T., Colon-Carmona, A., Gutierrez, R.A., and Doerner, P. (2005). Arabidopsis TCP20 links regulation of growth and cell division control pathways. Proc. Natl Acad. Sci. U S A. 102, 12978–12983.

Reinholz, E. (1966). Radiation induced mutants showing changed inflorescence characteristics. Arabid. Inf. Serv. 3, 19–20. Saeed, A.I., et al. (2003). TM4: a free, open-source system for microarray data management and analysis. Biotechniques. 34, 374–378. Saez, A., Rodrigues, A., Santiago, J., Rubio, S., and Rodriguez, P.L. (2008). HAB1–SWI3B interaction reveals a link between abscisic acid signaling and putative SWI/SNF chromatin-remodeling complexes in Arabidopsis. Plant Cell. 20, 2972–2988. Sanada, Y., Ueda, H.R., Kuribayashi, K., Andoh, T., Hayashi, F., Tamai, N., and Wada, K. (1995). Novel light–dark change of proline levels in halophyte (Mesembryanthemum crystallinum L.) and glycophytes (Hordeum vulgare L. and Triticum aestivum L.) leaves and roots under salt stress. Plant Cell Physiol. 36, 965–970. Scho¨nrock, N., Bouveret, R., Leroy, O., Borghi, L., Ko¨hler, C., Gruissem, W., and Hennig, L. (2006a). Polycomb-group proteins repress the floral activator AGL19 in the FLC-independent vernalization pathway. Genes Dev. 20, 1667–1678. Scho¨nrock, N., Exner, V., Probst, A., Gruissem, W., and Hennig, L. (2006b). Functional genomic analysis of CAF-1 mutants in Arabidopsis thaliana. J. Biol. Chem. 281, 9560–9568. Simon, P. (2003). Q-Gene: processing quantitative real-time RT–PCR data. Bioinformatics. 19, 1439–1440. Song, C.P., and Galbraith, D.W. (2006). AtSAP18, an orthologue of human SAP18, is involved in the regulation of salt stress and mediates transcriptional repression in Arabidopsis. Plant Mol. Biol. 60, 241–257. Song, C.P., Agarwal, M., Ohta, M., Guo, Y., Halfter, U., Wang, P., and Zhu, J.K. (2005). Role of an Arabidopsis AP2/EREBP-type transcriptional repressor in abscisic acid and drought stress responses. Plant Cell. 17, 2384–2396. Song, J.J., Garlick, J.D., and Kingston, R.E. (2008). Structural basis of histone H4 recognition by p55. Genes Dev. 22, 1313–1318.

Menges, M., Hennig, L., Gruissem, W., and Murray, J.A. (2003). Genome-wide gene expression in an Arabidopsis cell suspension. Plant Mol. Biol. 53, 423–442.

Sridha, S., and Wu, K. (2006). Identification of AtHD2C as a novel regulator of abscisic acid responses in Arabidopsis. Plant J. 46, 124–133.

Michaels, S.D., and Amasino, R.M. (2001). Loss of FLOWERING LOCUS C activity eliminates the late-flowering phenotype of FRIGIDA and autonomous pathway mutations but not responsiveness to vernalization. Plant Cell. 13, 935–941.

Tran, L.S., Nakashima, K., Sakuma, Y., Simpson, S.D., Fujita, Y., Maruyama, K., Fujita, M., Seki, M., Shinozaki, K., and YamaguchiShinozaki, K. (2004). Isolation and functional analysis of Arabidopsis stress-inducible NAC transcription factors that bind to a drought-responsive cis-element in the early responsive to dehydration stress 1 promoter. Plant Cell. 16, 2481–2498.

Munemasa, S., Oda, K., Watanabe-Sugimoto, M., Nakamura, Y., Shimoishi, Y., and Murata, Y. (2007). The coronatine-insensitive 1 mutation reveals the hormonal signaling interaction between abscisic acid and methyl jasmonate in Arabidopsis guard cells: specific impairment of ion channel activation and second messenger production. Plant Physiol. 143, 1398–1407. Nemhauser, J.L., Hong, F., and Chory, J. (2006). Different plant hormones regulate similar processes through largely nonoverlapping transcriptional responses. Cell. 126, 467–475.

van den Heuve, S., and Dyson, N.J. (2008). Conserved functions of the pRB and E2F families. Nat. Rev. Mol. Cell Biol. 9, 713–724. Vandepoele, K., Vlieghe, K., Florquin, K., Hennig, L., Beemster, G.T., Gruissem, W., Van de Peer, Y., Inze, D., and De Veylder, L. (2005). Genome-wide identification of potential plant E2F target genes. Plant Physiol. 139, 316–328.

Alexandre et al.

Waterborg, J.H., Harrington, R.E., and Winicov, I. (1989). Differential histone acetylation in alfalfa (Medicago sativa) due to growth in NaCl: responses in salt stressed and salt tolerant callus cultures. Plant Physiol. 90, 237–245.

d


|

687

Yamaguchi-Shinozaki, K., and Shinozaki, K. (2006). Transcriptional regulatory networks in cellular responses and tolerance to dehydration and cold stresses. Annu. Rev. Plant Biol. 57, 781–803.

Wille, A., Gruissem, W., Bu¨hlmann, P., and Hennig, L. (2007). EVE (external variance estimation) increases statistical power for detecting differentially expressed genes. Plant J. 52, 561–569.

Yoshiba, Y., Kiyosue, T., Nakashima, K., Yamaguchi-Shinozaki, K., and Shinozaki, K. (1997). Regulation of levels of proline as an osmolyte in plants under water stress. Plant Cell Physiol. 38, 1095–1102.

Wu, Z., Irizarry, R.A., Gentleman, R., Murillo, F.M., and Spencer, F. (2004). A model based background adjustment for oligonucleotide expression arrays. John Hopkins University, Dept of Biostatistics Working Paper.

Yoshida, N., Yanai, Y., Chen, L., Kato, Y., Hiratsuka, J., Miwa, T., Sung, Z.R., and Takahashi, S. (2001). EMBRYONIC FLOWER2, a novel Polycomb group protein homolog, mediates shoot development and flowering in Arabidopsis. Plant Cell. 13, 2471–2481.

Yadav, S.K., Singla-Pareek, S.L., Ray, M., Reddy, M.K., and Sopory, S.K. (2005). Methylglyoxal levels in plants under salinity stress are dependent on glyoxalase I and glutathione. Biochem. Biophys. Res. Commun. 337, 61–67.

Zhang, X., Clarenz, O., Cokus, S., Bernatavichute, Y.V., Pellegrini, M., Goodrich, J., and Jacobsen, S.E. (2007). Whole-genome analysis of Histone H3 Lysine 27 trimethylation in Arabidopsis. PLoS Biol. 5, 1026–1035.