Polycomb-group proteins repressthe floral activator AGL19 in the FLC-independent vernalization pathway
http://www.100md.com
基因进展 2006年第12期
Institute of Plant Sciences and Zurich-Basel Plant Science Center, Swiss Federal Institute of Technology, ETH Center, CH-8092 Zurich, Switzerland
Abstract
Polycomb-group (PcG) proteins form a cellular memory by maintaining developmental regulators in a transcriptionally repressed state. We identified a novel flowering gene that is under PcG control in Arabidopsis—the MADS-box gene AGL19. AGL19 expression is maintained at very low levels by the PcG proteins MSI1, CLF, and EMF2, and AGL19 is partly responsible for the early flowering phenotype of clf mutants. AGL19 chromatin is strongly enriched in trimethylation of Lys 27 on histone H3 (H3K27me3) but not in H3K9me2. Repressive H3K27me3 marks were reduced by decreased CLF or MSI1 levels and by prolonged cold, suggesting that the PcG proteins MSI1 and CLF repress AGL19 in the absence of cold. Ectopic expression of AGL19 strongly accelerates flowering, and agl19 mutants have a decreased response to vernalization, the promotion of flowering by prolonged cold. Epistasis analyses revealed that AGL19 works in the poorly characterized FLC-independent vernalization pathway and does not require SOC1 to function. In this pathway, prolonged cold relieves AGL19 from PcG repression by a mechanism that requires VIN3 but not VRN2. Elevated AGL19 levels activate LFY and AP1 and eventually cause flowering.
[Keywords: Polycomb proteins; histone methylation; MSI1; vernalization; FLC; SOC1]
Received December 22, 2005; revised version accepted April 13, 2006.
Polycomb-group (PcG) proteins are cellular memory modules that are thought to maintain genes in a transcriptionally inactive state after an initial repression has been established (for review, see Brock and Fisher 2005). PcG proteins were first identified in insects, but are essential for normal development in both animals and plants (for review, see Schubert et al. 2005). These proteins often form large multiprotein complexes, such as the Polycomb Repressive Complex 2 (PRC2). Drosophila PRC2 contains four core subunits: the MSI1-like protein p55, Enhancer of Zeste [E(Z)], Extra Sex Combs (ESC), and Suppressor of Zeste 12 [Su(Z)12] (Czermin et al. 2002; Müller et al. 2002). Transcriptional repression by PRC2 relies on histone methyltransferase activity of the E(Z) subunit, which preferentially catalyzes histone H3 Lys 27 (H3K27) trimethylation (for review, see Cao and Zhang 2004).
Unlike in insects, where the PRC2 members are encoded by single-copy genes, in plants homologous PRC2 subunits are often encoded by small gene families (Schubert et al. 2005), and several possible PRC2-like complexes have been proposed to exist in Arabidopsis (Chanvivattana et al. 2004; Hennig et al. 2005). However, to date, only one Arabidopsis PRC2 complex has been well characterized—the FERTILISATION INDEPENDENT SEED (FIS) complex (K?hler et al. 2003a). The FIS complex consists of MSI1, the E(Z) homolog MEDEA (MEA), the ESC homolog FERTILISATION INDEPENDENT ENDOSPERM (FIE), and likely the Su(Z)12 homolog FIS2 (K?hler et al. 2003a; Chanvivattana et al. 2004). This complex has specific functions during gametophyte and early seed development, including suppression of seed development in the absence of fertilization and repression of the MADS-box gene PHERES1 (PHE1) (Chaudhury et al. 1997; Grossniklaus et al. 1998; Ohad et al. 1999; K?hler et al. 2003b). During sporophyte development, a second PRC2 complex, the CURLY LEAF (CLF) complex, most likely represses transcription of floral homeotic genes, such as the MADS-box gene AGAMOUS (AG). The CLF complex probably consists of MSI1, the E(Z) homolog CLF, FIE, and the Su(Z)12 homolog EMBRYONIC FLOWER2 (EMF2) (Goodrich et al. 1997; Kinoshita et al. 2001; Yoshida et al. 2001; Hennig et al. 2003; Chanvivattana et al. 2004; Katz et al. 2004). The third potential PRC2-like complex is the VERNALIZATION (VRN) complex. The existence of the VRN complex was hypothesized because the Su(Z)12 homolog VRN2 is required for maintaining repression of the MADS-box gene FLOWERING LOCUS C (FLC) after vernalization and for vernalization-induced H3 methylation at the FLC locus (Chandler et al. 1996; Gendall et al. 2001; Bastow et al. 2004; Chanvivattana et al. 2004; Sung and Amasino 2004). Because FLC is a very potent repressor of flowering, epigenetic control of the floral transition has recently received considerable attention (for review, see He and Amasino 2005).
The transition to flowering is controlled by diverse environmental and developmental signals, and many genes that control flowering in Arabidopsis have been identified (for review, see Boss et al. 2004; Ausin et al. 2005). These genes have been grouped in several genetic pathways, including the photoperiod, the vernalization, and the autonomous pathways. The photoperiod pathway responds to seasonal changes in day length, while the vernalization pathway renders Arabidopsis competent to flower after prolonged exposure to low temperatures; for example, during winter (for review, see Henderson and Dean 2004; Searle and Coupland 2004). These two pathways contribute to initiate flowering of winter-annual plants in the favorable conditions of spring and summer. In contrast, the autonomous pathway promotes flowering independently of environmental signals. Both the autonomous and vernalization pathways function mainly by reducing expression of the floral repressor FLC (Michaels and Amasino 1999, 2001; Sheldon et al. 2000; Gendall et al. 2001). Repression of FLC by vernalization is initiated by VERNALIZATION INSENSITIVE3 (VIN3) and maintained by VERNALIZATION1 (VRN1) and VERNALIZATION2 (VRN2) (Gendall et al. 2001; Levy et al. 2002; Sung and Amasino 2004). When expressed, FLC prevents transcription of the pathway integrators FLOWERING LOCUS T (FT) and the MADS-box gene SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1, or AGL20) (Borner et al. 2000; Lee et al. 2000; Michaels and Amasino 2001), which combine signals from several pathways to induce flowering (for review, see Parcy 2005). Vernalization releases SOC1 from FLC repression, thus making SOC1 an important component of the FLC-dependent vernalization pathway (Lee et al. 2000; Hepworth et al. 2002; Moon et al. 2003a). Although FLC plays a central role in vernalization in Arabidopsis, flc-null mutants still respond to vernalization, demonstrating that FLC-independent mechanisms exist in the vernalization pathway (Michaels and Amasino 2001; Sung and Amasino 2004). However, the molecular nature of these FLC-independent mechanisms is poorly understood.
Among the Arabidopsis homologs of PRC2 subunits, the floral transition is not only regulated by the Su(Z)12 homolog VRN2, but also by the p55 homolog MSI4 (also called FVE) (Koornneef et al. 1991; Ausin et al. 2004; Kim et al. 2004). FVE is part of the autonomous pathway, needed for the epigenetic repression of FLC via histone deacetylation. In addition to FVE/MSI4, Arabidopsis contains four other MSI1-like proteins (MSI1–3, MSI5) (Ach et al. 1997; Hennig et al. 2003). Although MSI1-like proteins have no catalytic activity, they are often essential subunits of protein complexes controlling chromatin dynamics (for review, see Hennig et al. 2005). Arabidopsis MSI1, for instance, is vital for plant development, as msi1-null mutants are embryo lethal (K?hler et al. 2003a; Guitton et al. 2004). Strong reduction of MSI1 protein levels by cosuppression (msi1-cs plants) is not lethal but has severe impacts on development (Hennig et al. 2003). In this study, we performed transcriptional profiling of msi1-cs plants and found that the MADS-box gene AGL19 is under control of the MSI1, CLF, and EMF2 Arabidopsis PcG proteins. AGL19 is a potent floral activator, which is repressed via H3K27 trimethylation in the absence of vernalization. This work also establishes a novel role for PcG proteins as repressors of the FLC-independent vernalization pathway of Arabidopsis.
Results
AGL19 is up-regulated in msi1-cs plants
In order to identify genes affected by the severe reduction in MSI1 levels in msi1-cs plants, we performed microarray RNA profiling experiments on rosette leaves of 23-d-old wild-type and msi1-cs plants grown in long days (LD). This developmental stage was chosen because phenotypic alterations of msi1-cs rosette leaves can then first be observed (Hennig et al. 2003). Statistical analysis of the microarray data identified eight robustly down-regulated genes and 122 robustly up-regulated genes. These deregulated genes fall into diverse categories, with a large proportion being involved in histone metabolism, cell cycle, and DNA repair. The MADS-box gene AGL19 was among the strongest up-regulated genes, with a 21-fold increase in msi1-cs leaves, and was the only MADS-box gene with a significantly altered expression pattern (Fig. 1A). RT–PCR confirmed that AGL19 is usually expressed at very low levels in wild-type rosette leaves but was strongly up-regulated in msi1-cs leaves (Fig. 1B). AGL19 belongs to the type II class of MADS-box genes and is phylogenetically most closely related to AGL14 and to the floral pathway integrator SOC1/AGL20 (Becker and Theissen 2003). So far, AGL19 has not been extensively characterized, and its biological function is unknown. Because MADS-box transcription factors are often developmental key regulators, AGL19 was selected for a detailed study.
Figure 1. Molecular characterization of AGL19. (A) Signal values from microarrays testing RNA from rosette leaves of 23-d-old msi1-cs and wild-type (WT) plants grown in LD. Shown are means ± ranges of two replicates. (B) Confirmation of AGL19 up-regulation in msi1-cs using rosette leaves from 21-d-old plants grown in LD. Here and in all subsequent experiments, GAPDH was used as a control. (C) Protein blots of wild-type, 35S::MSI1 overexpressing (msi1-oe), and MSI1 cosuppression plants (msi1-cs). Protein was extracted from leaves of 20-d-old plants. (Top panel) Ten micrograms of protein per sample was subjected to immunoblotting with affinity-purified, anti-MSI1-specific antiserum (Hennig et al. 2003). (Bottom panel) Ponceau red staining of the blot is shown as a loading control. (D) Expression analysis of AGL19, SOC1, and AGL14 in LD. (S) Nine-day-old seedlings; (RL) rosette leaves from 35-d-old plants; (CL) cauline leaves from 35-d-old plants; (INF) inflorescence meristem and closed flower buds; (F) open flowers; (GS) 2-d-old germinating seeds; (R) roots. Note that PCR was done with 33 cycles for AGL19 and AGL14, but with only 30 cycles for SOC1. (E) AGL19 expression increases with plant age. RNA was extracted from seedlings grown in SD. Age was measured in days after germination (DAG). Note that PCR was done with 38 cycles for AGL19. (F) AGL19 is strongly expressed in apices. RNA was extracted from dissected 18-d-old seedlings grown in SD. (S) Total seedlings; (A) apices only; (RL) rosette leaves only. (G) Nuclear localization of AGL19-GFP. (Left) Differential interference contrast image. (Right) GFP fluorescence. Bar, 50 μm.
AGL19 is a weakly expressed transcription factor localized in the nucleus
AGL19 was originally considered as a root-specific gene (Alvarez-Buylla et al. 2000). However, as we detected AGL19 in wild-type rosette leaves using both microarrays and RT–PCR, we re-examined the expression pattern of AGL19 in different organs and compared it with that of its closest phylogenetic relatives SOC1 and AGL14 (Fig. 1D). Although the strongest expression of AGL19 was detected in roots, it was also detected in rosette leaves and seedlings and to a lesser extent in cauline leaves and flowers. This expression pattern overlaps widely with the expression of SOC1. In contrast, AGL14 transcripts were confined to roots as previously described (Alvarez-Buylla et al. 2000). In general, AGL19 transcripts increase with the age of the plant (Fig. 1E). AGL19 is enriched in apices including the shoot apical meristem and developing leaf primordia in dissected 18-d-old seedlings grown in short days (SD), while it is very weakly expressed in leaves at this developmental stage (Fig. 1F). Low gene expression levels have precluded a more detailed analysis of the AGL19 expression pattern in the apex by in situ hybridization. Localization studies using GFP fusion proteins showed that AGL19 is targeted to the nucleus (Fig. 1G). Together, AGL19, but not its close homolog AGL14, is expressed with a similar organ-specificity as the well-known floral integrator SOC1.
AGL19 is sufficient to promote flowering
To study the effects of ectopic AGL19 expression, we constructed transgenic plants containing the AGL19 cDNA under the control of the strong cauliflower mosaic virus 35S promoter (35S::AGL19) (Fig. 2A). Five independent transgenic lines were analyzed, all of which were extremely early flowering in LD conditions and produced two to four rosette leaves at the time of bolting (Fig. 2B; Table 1). In noninductive SD conditions, 35S::AGL19 plants were also early flowering but produced more leaves than in LD, and therefore, they remained responsive to photoperiod (Fig. 2C). In the late flowering fve background, the overexpression of AGL19 had the strongest effect on flowering time as 35S::AGL19 fve plants flowered similarly to 35S::AGL19 plants (Table 1). Thus, overexpression of AGL19 is sufficient to promote flowering.
Table 1 Effect of 35S::AGL19 on flowering time in LD
Figure 2. AGL19 promotes the transition to flowering. (A) Expression of AGL19 in 35S::AGL19 plants. RNA was extracted from 10-d-old seedlings grown in LD. (B) Appearance of wild-type (WT) and 35S::AGL19 plants grown for 20 d in LD. (C) Flowering time of 35S::AGL19 plants grown in LD or SD. Note that for flowering time experiments in all figures, means ± S.D. are shown for at least 14 plants.
agl19 mutants have a decreased vernalization response
To determine whether AGL19 normally controls flowering time in wild-type plants, we used two agl19 insertion alleles (agl19-1 and agl19-2), in which the AGL19 transcript was undetectable (Fig. 3A,B). The agl19 mutants flowered normally in LD conditions, but flowered slightly later than wild type in SD (Fig. 3C). Because overexpression of AGL19 caused phenotypes similar to overexpression of SOC1 (Borner et al. 2000), which is regulated by the vernalization pathway, we tested whether AGL19 was also involved in the response to vernalization. Indeed, agl19 mutants flowered considerably later than wild type after a 6-wk cold treatment (Fig. 3C,D). Thus, AGL19 is not only sufficient to promote flowering when ectopically expressed, but it is also required for the promotion of flowering in response to vernalization.
Figure 3. agl19 mutants have a decreased response to vernalization. (A) Position of T-DNA insertions for agl19-1 and agl19-2 alleles. The arrow indicates the transcription start site. The gene structure was drawn according to TIGRv5 annotation. (B) AGL19 transcripts in wild type (WT), agl19-1, and agl19-2 plants. RNA was extracted from rosette leaves of 40-d-old plants grown in LD. RT–PCR was performed with 33 cycles. (C) Effect of vernalization (Vern) on the flowering time of agl19. Wild-type (WT) and agl19-1 plants were exposed to either 2-d stratification (empty bars) or to 6-wk vernalization (hatched bars) at 4°C before being transferred to either LD or SD growing conditions. (D) Comparison of decreased vernalization response in agl19-1 and agl19-2. Flowering time in SD without vernalization (empty bars) or after a 6-wk vernalization treatment (hatched bars).
Vernalization induces AGL19 expression
Because the agl19 mutants had a reduced vernalization response, we tested whether a prolonged cold treatment affected AGL19 transcript levels. Indeed, AGL19 mRNA amounts increased considerably in response to an extended cold exposure (Fig. 4A). The same increase could be seen for SOC1 after vernalization but not for AGL19’s closest homolog, AGL14. VRN2 and VIN3 are two crucial mediators of vernalization in Arabidopsis (Chandler et al. 1996; Sung and Amasino 2004). When testing induction of AGL19 by vernalization in vrn2 and vin3 mutant plants, we found that VIN3 but not VRN2 is required for increased AGL19 expression after vernalization (Fig. 4B,C). However, VRN2 is needed to attain maximum AGL19 expression. In the absence of vernalization, FLC represses SOC1 (Lee et al. 2000). To test whether FLC represses AGL19 as well, we measured AGL19 expression in flc seedlings. Because AGL19 levels were not increased in flc but even slightly reduced, FLC is not required to maintain AGL19 in a repressed state before vernalization (Fig. 4D). Next, we tested whether AGL19 affected FLC expression. Neither loss nor strong ectopic overexpression of AGL19 influenced FLC levels considerably (Fig. 4E). After the initial repression of FLC by a vernalization cold treatment, flowering will only be promoted if FLC repression is maintained (Gendall et al. 2001; Levy et al. 2002). Mutants such as vrn1 have a reduced vernalization response, similar to agl19, due to the inability to maintain stable repression of FLC after a vernalization treatment (Levy et al. 2002). Therefore, we tested whether the decreased vernalization response observed in agl19 was also caused by a derepression of FLC. In contrast to vrn1, FLC levels remained low after vernalization in agl19, indicating that AGL19 is dispensable for the establishment and maintenance of FLC repression (Fig. 4G). Taken together, these results show that AGL19 is not repressed by FLC in wild-type plants and that the reduced vernalization response in agl19 mutants is not caused by elevated FLC levels. Thus, it is possible that AGL19 functions in an FLC-independent branch of the vernalization pathway.
Figure 4. AGL19 functions in the FLC-independent vernalization pathway. (A) AGL19, SOC1, and AGL14 expression in 10-d-old wild-type (WT) seedlings grown in SD with or without a 6-wk vernalization treatment (Vern). (B) AGL19 and SOC1 expression in 10-d-old wild-type (WT) and vin3 seedlings grown in SD with or without a 6-wk vernalization treatment. (C) AGL19 expression in 10-d-old wild-type (WT) (Landsberg erecta) and vrn2 seedlings grown in SD with or without a 6-wk vernalization treatment. (D) Effect of FLC and SOC1 on AGL19 expression. (E) Effect of AGL19 and SOC1 on FLC expression. (F) Effect of AGL19 on SOC1 expression. For D–F, RNA was extracted from 10-d-old seedlings grown in LD and analyzed by Q-PCR. (G) Expression of FLC in wild type (WT), agl19, and vrn1 after vernalization. RNA was extracted from seedlings 4, 10, and 20 d after a 2-d stratification or a 6-wk vernalization treatment. (H,I) Flowering time of wild type (WT), agl19, soc1, 35S::AGL19, and the double mutants in LD (H) or SD (I).
In order to genetically test whether AGL19 functions in the FLC-independent vernalization pathway, the vernalization response of the agl19 flc double mutant was tested. This double mutant responded less to vernalization than either single mutant (Table 2). Thus, the agl19 and flc mutants exhibit an additive rather than an epistatic relationship, confirming that AGL19 functions in an FLC-independent vernalization pathway. Similar to some previous reports (e.g., Zhang and van Nocker 2002), the flc mutant flowered later than wild type in SD after vernalization in several independent experiments. The reasons for this effect are not known, but indirect effects mediated by poorly understood FLC targets and functions (McKay et al. 2003; Edwards et al. 2006) might be involved.
Table 2 Effect of vernalization on flowering time of agl19 and flc mutants in SD
AGL19 function does not require SOC1
The FLC-dependent vernalization pathway functions mainly to control expression of the floral integrator SOC1 (Lee et al. 2000; Gendall et al. 2001; Hepworth et al. 2002; Levy et al. 2002; Moon et al. 2003a, 2005). Therefore, we next investigated whether the AGL19-containing branch of the FLC-independent vernalization pathway is integrated by SOC1 as well and whether AGL19 and SOC1 influence each other’s expression. First, we tested whether AGL19 levels affect SOC1 expression. However, neither loss of AGL19 nor increased AGL19 levels led to considerably altered SOC1 expression in wild type (Fig. 4F). Similarly, ectopic overexpression of AGL19 did not affect SOC1 in fve, which has very low SOC1 levels due to elevated FLC expression (Fig. 4F). Second, we tested whether SOC1 levels affect AGL19 expression. Loss of SOC1, which delayed flowering (Fig. 4H), reduced AGL19 levels slightly, but ectopic overexpression of SOC1, which greatly accelerated flowering (Supplementary Fig. S1), strongly repressed AGL19 (Fig. 4D). This is in contrast to AGL24, which is induced by ectopic SOC1 expression (Michaels et al. 2003). Because ectopic overexpression of SOC1 strongly promotes flowering without increasing AGL19 levels, SOC1 does not function via activation of AGL19 expression. Moreover, because ectopic overexpression of AGL19 strongly promotes flowering without increasing SOC1 levels, AGL19 does not function via activation of SOC1 expression. Rather, it is possible that SOC1 and AGL19 function in genetic parallel but cross-connected pathways.
To examine whether AGL19 and SOC1 function in the same or in parallel genetic pathways to promote flowering, the agl19 soc1 double mutant was generated. If AGL19 and SOC1 act in the same pathway, then the late-flowering phenotype of the double mutant should not transgress that of the latest single mutant. However, the agl19 soc1 double mutant flowered later than the soc1 single mutant, and this delay was more pronounced in SD (additional 15 leaves) than in LD (additional two leaves) (Fig. 4H,I), indicating that AGL19 and SOC1 act in genetically different flowering pathways. Similarly, 35S::AGL19 soc1 plants were still early flowering, but showed an additive rather than an epistatic phenotype, thus confirming that AGL19 and SOC1 probably function in parallel pathways to promote flowering (Fig. 4H,I). Nonetheless, it is possible that AGL19 and SOC1 regulate common downstream targets. Indeed, expression of the floral integrator LEAFY (LFY) and of the meristem identity gene APETALA1 (AP1) was increased in 35S::AGL19 and 35S::SOC1 seedlings (Supplementary Fig. S1).
Control of AGL19 by Polycomb-group complexes
As AGL19 expression was increased in msi1-cs plants, it may be repressed by a MSI1-containing Arabidopsis PRC2 complex, like the CLF–PRC2 complex, which functions during sporophytic development (Schubert et al. 2005). Indeed, AGL19 was similarly up-regulated in 15-d-old emf2 and clf seedlings as well as in 4-wk-old rosette leaves of clf and msi1-cs plants (Figs. 5A, 6A). Interestingly, derepression of AGL19 in clf increased with plant age (Fig. 5B). These results show that MSI1, CLF, and EMF2, three likely subunits of the CLF complex, are all required to repress AGL19 in the absence of a vernalization treatment. As AGL19 shares similarities with SOC1, we tested whether CLF represses SOC1 as well. In contrast to AGL19, SOC1 expression was not increased in clf mutants (Fig. 5C). Similarly, no differences were seen in SOC1 expression in msi1-cs leaves compared with wild-type leaves in the microarray experiments (data not shown). Together this suggests that the CLF–PRC2 complex does not regulate SOC1. Because vernalization activates AGL19, it is possible that vernalization represses MSI1, CLF, or EMF2. However, we found no considerable effect of vernalization on expression of these genes (Fig. 5D). Next, we tested whether AGL19 is a direct target of the CLF complex by chromatin immunoprecipitation (ChIP). The anti-MSI1 antiserum is not suitable for immunoprecipitation (data not shown), and no antibodies against other PRC2 subunits are available. Therefore, we used transgenic plants that express a myc-tagged FIE version. Using anti-myc antibodies, AGL19 sequences were two- to threefold enriched in precipitated chromatin from the 35S::myc-FIE line compared with wild type (Fig. 5E). These data suggest that the CLF–PRC2 complex directly binds to the AGL19 locus.
Figure 5. AGL19 is up-regulated in PcG mutants and contributes to the early flowering phenotype of clf. (A) Expression of AGL19 in clf and emf2. RNA was extracted from 15-d-old seedlings grown in LD. (B) Age dependency of AGL19 derepression in clf. RNA was extracted at 10, 15, and 33 d after germination. (C) Expression of SOC1 in clf. RNA was extracted from 10-d-old seedlings and from rosette leaves of 21-d-old plants grown in LD. (D) CLF, EMF2, and MSI1 expression in 10-d-old wild-type (WT) seedlings grown in SD with or without a 6-wk vernalization treatment (Vern). (E) ChIP using anti-myc antibodies and 10-d-old wild-type Landsberg erecta or transgenic 35S::myc-FIE seedlings grown in LD. Shown are PCR products of fragments II and III from the AGL19 locus (see Fig. 6 for details) and of an unrelated PHOSPHOFRUCTOKINASE (PFK) gene. (F) Flowering time of msi1-cs plants in SD. (G) Flowering time of wild-type (WT), agl19, clf, and clf agl19 plants in LD. Leaf numbers differ significantly between clf and clf agl19 double mutants (t-test: p < 0.001). (H) Images of 22-d-old wild-type (WT), agl19, clf, and clf agl19 plants grown in LD at 21°C. Note that all clf mutants are already flowering, while agl19 clf double mutants have only just started bolting.
Figure 6. Histone methylation pattern at the AGL19 locus. (A) Expression analysis of AGL19 in msi1-cs and clf in rosette leaves of 4-wk-old plants grown in LD. (B) Expression analysis of AGL19 in 10-d-old seedlings grown in SD with or without a 6-wk vernalization treatment (Vern). (C) Schematic representation of AGL19 showing intron–exon structure (annotated according to TIGRv5) and position of the fragments used for amplification by PCR following ChIP. The arrow indicates the transcription start site. (D) Quantification of multiplex PCR products after ChIP using anti-trimethyl-histone H3K27 (H3K27me3) antiserum on 4-wk-old rosette leaves from wild-type (WT), msi1-cs, and clf plants grown in LD. Values are shown as enrichment of H3K27me3 using AGL19 specific primers relative to PFK for wild-type (WT), msi1-cs, and clf plants. (E) ChIP experiment performed on 10-d-old wild-type (WT) seedlings grown in SD with or without a 6-wk vernalization treatment using antibodies against H3K27me3. Values represent mean ± range from two independent ChIP experiments.
Consistent with the findings that AGL19 can promote flowering and that AGL19 expression is increased in msi1-cs plants, msi1-cs plants flowered significantly earlier than wild type in SD (Fig. 5F). Because cosuppression of MSI1 occurs around an age of 2–3 wk in LD, it cannot affect flowering time any more at this stage, and msi1-cs plants were not consistently earlier flowering than wild type in LD (data not shown). Similarly, AGL19 expression is increased in clf and emf2 mutants, and both flower earlier than wild type (Yang et al. 1995; Goodrich et al. 1997), suggesting that the overexpression of AGL19 observed in these mutants could be partially responsible for their early flowering phenotype. In order to test this hypothesis, we constructed the clf agl19 double mutant. Loss of AGL19 did not abolish the clf leaf phenotype, but partially suppressed the early flowering phenotype (Fig. 5G,H). Thus, elevated AGL19 levels specifically contribute to the early flowering of clf.
MSI1 and CLF epigenetically regulate the AGL19 locus
In animals, the E(Z)-related subunit of PRC2-like complexes has H3K27 methyltransferase activity (Cao and Zhang 2004). It is likely that the Arabidopsis E(Z) homologs, such as CLF, function as methyltransferases as well to establish repressive H3K27 trimethylation (H3K27me3) at target chromatin (Schubert et al. 2005). Because the PRC2 subunits CLF and MSI1 repress AGL19, we tested whether AGL19 chromatin was enriched in H3K27me3 marks. ChIP experiments using 4-wk-old rosette leaves revealed a strong enrichment of H3K27me3 at AGL19 (Fig. 6D). Interestingly, the H3K27me3 mark did not evenly cover the entire locus, but peaked in the 5' region of the gene (Fig. 6C, PCR fragment II). In contrast to H3K27me3, the heterochromatin H3K9me2 mark, which was strongly enriched in chromatin of the Ta2 and Cinful-like transposons, was not detected in AGL19 chromatin (Supplementary Fig. S2). This indicates that in rosette leaves, the AGL19 locus normally contains H3K27me3 but not H3K9me2 marks. To test whether the H3K27me3 mark depends on MSI1 and CLF, similar experiments were performed with msi1-cs and clf plants. Expression of a PHOSPHOFRUCTOKINASE (PFK) was not changed in msi1-cs or clf plants, and PFK was thus suitable as an internal control in multiplex PCR (Fig. 6A). In msi1-cs and clf, H3K27me3 at AGL19 was greatly reduced but not completely lost (1.8-fold and threefold reduction in msi1-cs and in clf, respectively) (Fig. 6D; Supplementary Fig. S2). This decrease in H3K27me3 observed in msi1-cs and clf was also specific for the region of AGL19 covered by PCR fragment II. Thus, MSI1 and CLF are needed to establish repressive H3K27me3 marks at AGL19 chromatin in a very specific pattern, but their function is most likely partially redundant with other Arabidopsis proteins.
Vernalization prevents H3K27 trimethylation at the AGL19 locus
Our experiments have established that in standard LD conditions, AGL19 chromatin carries H3K27me3 marks that most likely repress AGL19 transcription. Because a prolonged cold treatment strongly activates AGL19 expression, we tested whether such a treatment also affected the H3K27me3 marks on AGL19 chromatin. We performed a ChIP experiment on 10-d-old wild-type seedlings grown in SD with or without a 6-wk cold treatment. PFK expression was not affected by vernalization and hence was used as an internal control for multiplex PCR (Fig. 6B). In nontreated seedlings, in which no AGL19 expression was detectable (Fig. 6B), the AGL19 locus was highly enriched for H3K27me3 marks specifically in the 5' region of the gene (Fig. 6E; Supplementary Fig. S2). This methylation pattern was similar to that seen in the previous ChIP experiment on wild-type rosette leaves (Fig. 6D). In contrast, when AGL19 expression was induced after cold treatment, this H3K27me3 mark was dramatically reduced by fivefold compared with nonvernalized seedlings (Fig. 6B,E). The overlapping pattern of H3K27me3 observed after vernalization and in the msi1-cs or clf mutants suggests that a PRC2 complex containing MSI1 and CLF is responsible for repressing AGL19 in the absence of vernalization.
Discussion
AGL19 is a novel floral activator
AGL19 has been identified as a novel floral activator in Arabidopsis, because overexpression of AGL19 greatly accelerates the transition to flowering, while loss of AGL19 delayed flowering after a vernalization treatment. AGL19 belongs to the MADS-box gene family, which contains master regulatory genes essential for diverse developmental processes (Becker and Theissen 2003). Plant MADS-box genes form several phylogenetically distinct groups, and AGL19 together with its closest homolog AGL14 (68% amino acid identity) and the well-characterized floral activator SOC1 (AGL20, 46% amino acid identity) belong to the Tm3/SOC1 clade (Becker and Theissen 2003). Although AGL19, AGL14, and AGL20 are closely related, their expression patterns differ. In particular, AGL14 is exclusively expressed in roots, while AGL19 and SOC1 are expressed in many organs of the shoot as well. Despite the prominent expression of AGL19 in roots, root growth and macroscopic root morphology did not differ between wild-type, agl19, and 35S::AGL19 plants (Supplementary Fig. S2; data not shown).
AGL19 and SOC1 both function to promote flowering, but they have diverged significantly in terms of gene regulation and flowering pathways in which they participate. The agl19 mutant flowered normally in LD and was only slightly late flowering in SD, indicating that under standard laboratory conditions, AGL19 has only a minor role in the promotion of flowering. In contrast, the soc1 mutant is significantly late flowering in both LD and SD (Borner et al. 2000; Lee et al. 2000). Because agl19 mutants could no longer adequately accelerate flowering in response to prolonged cold treatments, AGL19 functions in the vernalization response. While SOC1 integrates several genetic pathways that promote flowering, including the vernalization and photoperiod pathways (Borner et al. 2000; Lee et al. 2000; Samach et al. 2000; Hepworth et al. 2002; Moon et al. 2003a), AGL19 is mainly part of the vernalization pathway.
Similarly to ectopic expression of SOC1, ectopic expression of AGL19 did not only greatly accelerate flowering but caused defects in floral development as well (Borner et al. 2000; data not shown). The fact that only 35S::AGL19 but not agl19 plants had altered flower morphology suggests that like SOC1, AGL19 is also dispensable for flower development and appears to function specifically during the floral transition after vernalization.
AGL19 regulates the floral transition in an FLC-independent vernalization pathway
Vernalization in Arabidopsis acts mainly via repression of FLC and subsequent up-regulation of SOC1 (Lee et al. 2000; Gendall et al. 2001; Hepworth et al. 2002; Levy et al. 2002; Moon et al. 2003a). However, flc-null mutants still show a vernalization response, and thus an FLC-independent vernalization pathway must exist (Michaels and Amasino 2001). Although both AGL19 and SOC1 are up-regulated by vernalization, only SOC1 but not AGL19 is repressed by FLC. In addition, AGL19 levels had no effect on the amount of FLC transcript, and FLC remained repressed after vernalization in the agl19 mutant. Thus, AGL19 does not regulate FLC, and the reduced vernalization response of agl19 mutants is not caused by deregulation of FLC as in the vrn1 and vrn2 mutants. Because agl19 flc double mutants had an additive effect on the vernalization response, AGL19 and FLC function in genetically separate pathways. The FLC-containing pathway requires VIN3 and VRN2 (Gendall et al. 2001; Sung and Amasino 2004). The AGL19-containing pathway requires VIN3 but not VRN2. Both elevated AGL19 and SOC1 levels lead to the activation of LFY and AP1 to promote flowering.
Interestingly, even the agl19 flc double mutant showed a residual vernalization response, indicating the presence of other vernalization pathways. Such a pathway might include the MADS-box gene AGL24, which is known to mediate an FLC-independent vernalization response (Yu et al. 2002; Michaels et al. 2003). Analysis of the vernalization response of agl19 flc agl24 triple mutants will help answer this question.
Although AGL24 acts in an FLC-independent pathway, it is closely linked to SOC1, because these two genes positively regulate each other’s expression, and overexpression of one of the two genes has little effect in the absence of the other (Michaels et al. 2003). In contrast, AGL19 and SOC1 act at least partially independently, because there seems to be no or even negative cross-regulation between AGL19 and SOC1 and overexpression of AGL19 greatly accelerated flowering in a soc1 background. Both agl19 soc1 as well as 35S::AGL19 soc1 showed additive rather than epistatic interactions, indicating that SOC1 and AGL19 act in separate genetic pathways that converge to control the common downstream targets AP1 and LFY. Thus, although SOC1, AGL24, and AGL19 share several similarities, they clearly have different roles in the genetic network controlling flowering, and are all needed to coordinate a fully functional vernalization response.
AGL19 repression depends on PRC2 subunits
In the absence of vernalization, AGL19 is lowly expressed, and MSI1, CLF, and EMF2 are all required for repression of AGL19. MSI1, CLF, and EMF2 are homologs of Drosophila PRC2 subunits (Ach et al. 1997; Goodrich et al. 1997; Yoshida et al. 2001), and it was suggested that these proteins together with FIE form an Arabidopsis PRC2 complex (Chanvivattana et al. 2004). Protein–protein interaction data support this model as: MSI1 interacts with FIE, FIE interacts with CLF, and CLF interacts with EMF2 (K?hler et al. 2003a; Chanvivattana et al. 2004; Katz et al. 2004). ChIP experiments suggested that AGL19 is a direct target gene of the CLF–PRC2 complex and established that AGL19 chromatin is enriched in H3K27me3 marks, which at least in mammals and insects are set by PRC2 complexes (for review, see Cao and Zhang 2004). Arabidopsis has several homologs of PRC2 subunits and probably contains several PRC2-like complexes, which may function in different tissues or at different developmental stages (for review, see Schubert et al. 2005). PRC2-like complexes are thought to function as transcriptional repressors, and other Arabidopsis PcG target genes include PHERES1 (PHE1), AGAMOUS (AG), and FLC. MEA, FIE, and FIS2 are required to repress PHE1 (K?hler et al. 2003b); MSI1, CLF, FIE, and EMF2 are required to repress AG (Chen et al. 1997; Goodrich et al. 1997; Kinoshita et al. 2001; Hennig et al. 2003); and VRN2 is required to repress FLC (Gendall et al. 2001). Interestingly, PHE1, AG, and FLC as well as AGL19 are all MADS-box genes, supporting the notion that in contrast to metazoan PRC2 complexes, which mostly control Homeobox genes, plant PRC2 complexes often control MADS-box transcription factors. Importantly, only AGL19, but not SOC1 or AGL14, needs MSI1, CLF, and EMF2 to remain repressed in the absence of vernalization (this study; cf. also microarray data in Moon et al. 2003b). This suggests that either PRC2 repression was a feature of the Tm3/SOC1 ancestor and was lost in SOC1 and AGL14, or that AGL19 acquired PRC2 repression de novo. However, plant PRC2 complexes also have non-MADS-box targets (Katz et al. 2004), and more complete lists of primary and secondary plant and metazoan PRC2 target genes are needed for more systematic comparisons.
AGL19, which is a strong activator of flowering, is up-regulated in PRC2 mutants, and clf, fie, emf2, and msi1-cs plants are all early flowering. This suggests that CLF, EMF2, MSI1, and FIE form a PRC2 complex to control flowering time. MSI1 not only inhibits flowering by repressing AGL19, but also promotes flowering by contributing to normal SOC1 expression; this function, however, is independent of CLF (Bouveret et al. 2006). The contribution of increased AGL19 expression to the early flowering phenotype of msi1-cs plants could not be tested due to silencing of the 35S::MSI1 transgene by the AGL19 SALK-insertion allele in the 35S::MSI1 agl19 plants (data not shown). The 35S::MSI1 transgene is required to induce MSI1 cosuppression. In clf mutants, we found that AGL19 was partially responsible for the early flowering phenotype, as the agl19 clf double mutant flowered later than the clf mutant alone. However, derepression of AGL19 is not the major cause of the clf phenotype, and this is consistent with previous reports of a prominent role of ectopic AG expression for the clf phenotype (Goodrich et al. 1997). In clf mutants, expression of AG is correctly initiated in young floral meristems, but later breaks down and becomes expressed in the outer floral whorls (Goodrich et al. 1997). Similarly, PHE1 is initially normally expressed in fis mutants but is subsequently ectopically expressed in the endosperm (K?hler et al. 2003b), and FLC is initially repressed by vernalization treatments but subsequently becomes derepressed in vrn2 mutants (Gendall et al. 2001). Interestingly, the pattern of progressive AGL19 derepression in clf mutants is strikingly similar to that of other PcG targets in plants and Drosophila, where target genes are initially expressed normally but subsequently become ectopically activated (for review, see Ringrose and Paro 2004; Brock and Fisher 2005).
H3K27 trimethylation at AGL19 chromatin depends on PRC2 subunits and is regulated by vernalization
The mechanism of transcriptional repression by PRC2 complexes is not yet completely understood, but trimethylation of H3K27 appears to be of central importance in vivo (for review, see Cao and Zhang 2004). In Arabidopsis, H3K27me3 marks are found mostly in euchromatin (Lindroth et al. 2004), while H3K9me2 marks are hallmarks of silent heterochromatic genes (Soppe et al. 2002). Because AGL19 is only enriched in H3K27me3 but not in H3K9me2, repression of AGL19 is probably different from heterochromatic silencing as observed for many transposons. Instead, AGL19 appears to be a euchromatic gene that is repressed by H3K27me3. Interestingly, vernalization reduced H3K27me3 marks to lower levels than either reductions of MSI1 or loss of CLF. In the case of msi1-cs, the 5% residual MSI1 protein levels or one of the other four MSI1-like proteins in Arabidopsis could be sufficient to maintain partial H3K27me3 at the AGL19 locus. Similarly, partial redundancy of CLF and its homolog SWN has been observed (Chanvivattana et al. 2004). Thus, it is possible that functional redundancies attenuate the effects on AGL19 chromatin modification in msi1-cs and clf.
H3K27me3 was not spread uniformly across the AGL19 locus but peaked dramatically near the transcription start site. Similarly, binding of Drosophila PcG proteins seems to cluster at Polycomb-responsive elements (PREs) and promoters rather than being spread uniformly over target loci (Ringrose and Paro 2004; Brock and Fisher 2005). Consistent with this local clustering, Polycomb silencing blocks transcription initiation in Drosophila (Dellino et al. 2004). In Arabidopsis, the FIS PRC2 complex was found to bind preferentially at the PHE1 promoter (K?hler et al. 2003b), and FLC chromatin is enriched in H3 methylation at the promoter and the first large intron (Bastow et al. 2004; Sung and Amasino 2004). Taken together, these results suggest that AGL19 is repressed by a PRC2 complex containing MSI1, CLF, FIE, and EMF2 that sets H3K27me3 marks close to the transcriptional start site, similar to the situation in other PcG target genes.
Vernalization overcomes the repressive effect of the CLF–PRC2 complex. This could be carried out by active removal of the repressive H3K27me3 marks at AGL19. Because no biochemical activity that removes trimethylation is known, it is more likely that vernalization prevents the establishment and/or maintenance of H3K27me3 by the CLF–PRC2 complex during cell proliferation. This interpretation is supported by observations that vernalization often acts on dividing cells such as those present in meristems (for review, see Ausin et al. 2005). An alternative mechanism could involve exchange of methylated histones in AGL19 chromatin.
Together, the vernalization pathway in Arabidopsis contains at least two branches—the FLC-dependent and the FLC-independent branches, both of which require the regulatory effects of PcG proteins. While VRN2 functions in repressing the floral repressor FLC after vernalization, the CLF complex functions in repressing the floral activator AGL19 before vernalization. Thus, in Arabidopsis, different PcG proteins have been recruited to coordinate the vernalization response and to control the important developmental switch to reproductive growth.
Materials and methods
Plant material and growth conditions
Line msi1OEc2, which ectopically expresses MSI1 and gives rise to msi1-cs plants, was described before (Hennig et al. 2003). To construct plants expressing an epitope-tagged version of FIE, a construct encoding an N-terminal fusion of myc to FIE was cloned under the control of the 35S promoter and transformed into wild-type (Ler) plants. Line 4, which expressed the myc-FIE fusion (Supplementary Fig. S1), was used for further experiments. The two null AGL19 mutant alleles, agl19-1 (SALK_N578786) and agl19-2 (SALK_N516657) were obtained from the Nottingham Arabidopsis Seed Stock Center (NASC). The agl19-1 allele was used for all crossings and flowering time experiments unless otherwise stated. The fve-5, flc-6, vrn1-5, and clf-29 mutants are null alleles (Supplementary Fig. S1; Bouveret et al. 2006) from various collections of T-DNA insertion lines (SAIL_1167E5, SALK_41126, WiscDsLox393-396F9, and SALK_N521003, respectively) (Sessions et al. 2002; Alonso et al. 2003). All plants used in this study are in the Columbia background unless otherwise stated. Seeds of soc1-2, emf2-10, and vrn2-1 have been described (Chandler et al. 1996; Lee et al. 2000; Chanvivattana et al. 2004) and were kindly provided by I. Lee (Seoul National University, Seoul, Korea), J. Goodrich (University of Edinburgh, Edinburgh, UK), and C. Dean (John Innes Centre, Norwich, UK), respectively.
To construct plants that ectopically overexpressed AGL19 or SOC1 (35S::AGL19 and 35S::SOC1), the full-length coding sequences were amplified by PCR (for primers, see Supplementary Table S1). The cDNAs were inserted into binary destination vectors (AGL19 in pH7WG2; SOC1 in pK7WG2) (Karimi et al. 2002) downstream of the cauliflower mosaic virus (CaMV) 35S promoter. Constructs were transformed into Columbia wild-type plants, and 43 primary 35S::AGL19 transformants were obtained, of which five (35S::AGL19a–e) were selected for further study. The weak line 35S::AGL19a was used for all crossings and flowering time experiments unless otherwise stated.
For cellular localization studies, the full-length AGL19 cDNA was fused to an N-terminal GFP tag in vector pK7WGF2 (Karimi et al. 2002). AGL19_pK7WGF2 was used directly for bombardment of onion cells for transient expression assays as described previously (Gendall et al. 2001).
For measuring flowering time, seeds were plated on Murashige and Skoog (MS) medium (Duchefa), stratified for 2 d at 4°C, 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 μmol/m2·sec, 21°C ± 2°C) under LD (16 h light) photoperiods unless indicated otherwise. The flowering time was measured as the number of total rosette leaves longer than 0.5 cm at bolting for at least 14 plants. Graphs show means ± standard deviation. For the vernalization treatment, seeds were plated on MS medium and kept in continuous light for 1 d before being exposed to 4°C for 6 wk. After the vernalization treatment, plants were transferred to growth chambers (21°C ± 2°C) under LD or SD (8 h light) photoperiods as indicated.
RNA isolation and RT–PCR
RNA was extracted as previously described (Hennig et al. 2003). DNA-free RNA was reverse-transcribed using an oligo(dT) primer and SuperScript II reverse transcriptase (Invitrogen). Aliquots of the generated cDNA, which equaled 100 ng of total RNA, were used as the template for PCR with gene-specific primers (Supplementary Tables S2, S3). Q-PCR was performed in an ABI Prism 7700 Sequence Detection System (Applied Biosystems), using SYBR Green Master Mix reagent (Applied Biosystems) according to the manufacturer’s instructions. All amplification plots were analyzed with a fluorescence signal threshold of 0.1 to obtain Ct (threshold cycle) values. Experiments were performed in duplicate with error bars representing the range. Gene expression levels were normalized to an actin (ACT) control gene. The average Ct value for ACT was 23.83 (±0.18) for all templates measured in these experiments.
Microarray hybridization and evaluation
Arabidopsis wild-type (accession Columbia) and msi1OEc2 plants (Hennig et al. 2003) were grown for 23 d in growth chambers at 21°C under LD photoperiods. After identification of msi1-cs plants from the segregating msi1OEc2 population based on phenotype, rosette leaves were harvested. Material was harvested from at least 12 plants and pooled for each sample. The experiment was repeated twice to give two independent biological replicates. Affymetrix Arabidopsis ATH1 GeneChips were used in the experiment (Affymetrix). Labeling of samples, hybridizations, and measurements were performed as described (Hennig et al. 2004). Signal values were derived using the GCRMA algorithm in the statistical package R (version 1.9.1). Significantly differential gene expression was detected based on the rank-product algorithm (Breitling et al. 2004). This algorithm inherently corrects for multiple testing. Genes were considered as differentially expressed if p < 0.05.
ChIP
ChIP was performed as previously described (Bowler et al. 2004). Anti-myc 9E11 antibodies (Eurogentec), anti-dimethyl-histone H3 Lys 9 antiserum (Upstate, catalog #07-441), and anti-trimethyl-histone H3 Lys 27 antiserum (Upstate, catalog #07-449) were used for immunoprecipitation.
Immunoprecipitated DNA was analyzed by PCR using gene specific primers for AGL19 and primers against the control gene PFK (At4g04040) (Supplementary Table S4). Relative efficiencies of primer pairs were determined based on dilution series of input DNA (Supplementary Fig. S3B). Band intensities were quantified using Image J software. For quantitative displays of ChIP experiments, the values shown represent enrichment relative to PFK using the previously determined primer efficiencies. Bars represent mean and range for two independent ChIP experiments.
Acknowledgments
We thank Ihla Lee for soc1-2 mutant seeds, Justin Goodrich for emf2-10 mutant seeds, and Caroline Dean for helpful comments on the manuscript. N.S. was supported by the Roche Research Foundation. This research was supported by SNF project 3100AO-104238 to L.H., E.U. project QLG2-1999-00454 (ECCO) to W.G., a grant from the Zurich-Basel Plant Science Center to L.H. and C.K., and the Functional Genomics Center Zurich.
References
Ach R.A., Taranto P., 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.
Alonso J.M., Stepanova A.N., Leisse T.J., Kim C.J., Chen H., Shinn P., Stevenson D.K., Zimmerman J., Barajas P., Cheuk R.et al. 2003. Genome-wide insertional mutagenesis of Arabidopsis thaliana. Science 301: 653–657.
Alvarez-Buylla E.R., Liljegren S.J., Pelaz S., Gold S.E., Burgeff C., Ditta G.S., Vergara-Silva F., Yanofsky M.F. 2000. MADS-box gene evolution beyond flowers: Expression in pollen, endosperm, guard cells, roots and trichomes. Plant J. 24: 457–466.
Ausin I., Alonso-Blanco C., Jarillo J.A., Ruiz-Garcia L., Martinez-Zapater J.M. 2004. Regulation of flowering time by FVE, a retinoblastoma-associated protein. Nat. Genet. 36: 162–166.
Ausin I., Alonso-Blanco C., Martinez-Zapater J.M. 2005. Environmental regulation of flowering. Int. J. Dev. Biol. 49: 689–705.
Bastow R., Mylne J.S., Lister C., Lippman Z., Martienssen R.A., Dean C. 2004. Vernalization requires epigenetic silencing of FLC by histone methylation. Nature 427: 164–167.
Becker A. and Theissen G. 2003. The major clades of MADS-box genes and their role in the development and evolution of flowering plants. Mol. Phylogenet. Evol. 29: 464–489.
Borner R., Kampmann G., Chandler J., Gleissner R., Wisman E., Apel K., Melzer S. 2000. A MADS domain gene involved in the transition to flowering in Arabidopsis.. Plant J. 24: 591–599.
Boss P.K., Bastow R.M., Mylne J.S., Dean C. 2004. Multiple pathways in the decision to flower: Enabling, promoting, and resetting. Plant Cell 16:(Suppl.), 18–31.
Bouveret R., Sch?nrock N., Gruissem W., Hennig L. 2006. Regulation of flowering time by Arabidopsis MSI1. Development 133: 1693–1702.
Bowler C., Benvenuto G., Laflamme P., Molino D., Probst A.V., Tariq M., Paszkowski J. 2004. Chromatin techniques for plant cells. Plant J. 39: 776–789.
Breitling R., Armengaud P., Amtmann A., Herzyk P. 2004. Rank products: A simple, yet powerful, new method to detect differentially regulated genes in replicated microarray experiments. FEBS Lett. 573: 83–92.
Brock H.W. and Fisher C.L. 2005. Maintenance of gene expression patterns. Dev. Dyn. 232: 633–655.
Cao R. and Zhang Y. 2004. The functions of E(Z)/EZH2-mediated methylation of lysine 27 in histone H3. Curr. Opin. Genet. Dev. 14: 155–164.
Chandler J., Wilson A., Dean C. 1996. Arabidopsis mutants showing an altered response to vernalization. Plant J. 10: 637–644.
Chanvivattana Y., Bishopp A., Schubert D., Stock C., Moon Y.H., Sung Z.R., Goodrich J. 2004. Interaction of Polycomb-group proteins controlling flowering in Arabidopsis.. Development 131: 5263–5276.
Chaudhury A.M., Ming L., Miller C., Craig S., Dennis E.S., Peacock W.J. 1997. Fertilization-independent seed development in Arabidopsis thaliana. Proc. Natl. Acad. Sci. 94: 4223–4228.
Chen L., Cheng J.C., Castle L., Sung Z.R. 1997. EMF genes regulate Arabidopsis inflorescence development. Plant Cell 9: 2011–2024.
Czermin B., Melfi R., McCabe D., Seitz V., Imhof A., Pirrotta V. 2002. Drosophila Enhancer of Zeste/ESC complexes have a histone H3 methyltransferase activity that marks chromosomal Polycomb sites. Cell 111: 185–196.
Dellino G.I., Schwartz Y.B., Farkas G., McCabe D., Elgin S.C., Pirrotta V. 2004. Polycomb silencing blocks transcription initiation. Mol. Cell 13: 887–893.
Edwards K.D., Anderson P.E., Hall A., Salathia N.S., Locke J.C., Lynn J.R., Straume M., Smith J.Q., Millar A.J. 2006. FLOWERING LOCUS C mediates natural variation in the high-temperature response of the Arabidopsis circadian clock. Plant Cell 18: 639–650.
Gendall A.R., Levy Y.Y., Wilson A., Dean C. 2001. The VERNALIZATION 2 gene mediates the epigenetic regulation of vernalization in Arabidopsis.. Cell 107: 525–535.
Goodrich J., Puangsomlee P., Martin M., Long D., Meyerowitz E.M., Coupland G. 1997. A Polycomb-group gene regulates homeotic gene expression in Arabidopsis.. Nature 386: 44–51.
Grossniklaus U., Vielle-Calzada J.P., Hoeppner M.A., Gagliano W.B. 1998. Maternal control of embryogenesis by MEDEA, a Polycomb group gene in Arabidopsis.. Science 280: 446–450.
Guitton A.E., Page D.R., Chambrier P., Lionnet C., Faure J.E., Grossniklaus U., 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.
He Y. and Amasino R.M. 2005. Role of chromatin modification in flowering-time control. Trends Plant Sci. 10: 30–35.
Henderson I.R. and Dean C. 2004. Control of Arabidopsis flowering: The chill before the bloom. Development 131: 3829–3838.
Hennig L., Taranto P., Walser M., Sch?nrock N., Gruissem W. 2003. Arabidopsis MSI1 is required for epigenetic maintenance of reproductive development. Development 130: 2555–2565.
Hennig L., Gruissem W., Grossniklaus U., K?hler C. 2004. Transcriptional programs of early reproductive stages in Arabidopsis.. Plant Physiol. 135: 1765–1775.
Hennig L., Bouveret R., Gruissem W. 2005. MSI1-like proteins: An escort service for chromatin assembly and remodeling complexes. Trends Cell Biol. 15: 295–302.
Hepworth S.R., Valverde F., Ravenscroft D., Mouradov A., Coupland G. 2002. Antagonistic regulation of flowering-time gene SOC1 by CONSTANS and FLC via separate promoter motifs. EMBO J. 21: 4327–4337.
Karimi M., Inze D., Depicker A. 2002. GATEWAY vectors for Agrobacterium-mediated plant transformation. Trends Plant Sci. 7: 193–195.
Katz A., Oliva M., Mosquna A., Hakim O., Ohad N. 2004. FIE and CURLY LEAF Polycomb proteins interact in the regulation of homeobox gene expression during sporophyte development. Plant J. 37: 707–719.
Kim H.J., Hyun Y., Park J.Y., Park M.J., Park M.K., Kim M.D., Kim H.J., Lee M.H., Moon J., Lee I.et al. 2004. A genetic link between cold responses and flowering time through FVE in Arabidopsis thaliana. Nat. Genet. 36: 167–171.
Kinoshita T., Harada J.J., Goldberg R.B., Fischer R.L. 2001. Polycomb repression of flowering during early plant development. Proc. Natl. Acad. Sci. 98: 14156–14161.
K?hler C., Hennig L., Bouveret R., Gheyselinck J., Grossniklaus U., 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.
K?hler C., Hennig L., Spillane C., Pien S., Gruissem W., 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.
Koornneef M., Hanhart C.J., van der Veen J.H. 1991. A genetic and physiological analysis of late flowering mutants in Arabidopsis thaliana. Mol. Gen. Genet. 229: 57–66.
Lee H., Suh S.S., Park E., Cho E., Ahn J.H., Kim S.G., Lee J.S., Kwon Y.M., Lee I. 2000. The AGAMOUS-like 20 MADS domain protein integrates floral inductive pathways in Arabidopsis.. Genes & Dev. 14: 2366–2376.
Levy Y.Y., Mesnage S., Mylne J.S., Gendall A.R., Dean C. 2002. Multiple roles of Arabidopsis VRN1 in vernalization and flowering time control. Science 297: 243–246.
Lindroth A.M., Shultis D., Jasencakova Z., Fuchs J., Johnson L., Schubert D., Patnaik D., Pradhan S., Goodrich J., Schubert I.et al. 2004. Dual histone H3 methylation marks at lysines 9 and 27 required for interaction with CHROMOMETHYLASE3. EMBO J. 23: 4286–4296.
McKay J.K., Richards J.H., Mitchell-Olds T. 2003. Genetics of drought adaptation in Arabidopsis thaliana: I. Pleiotropy contributes to genetic correlations among ecological traits. Mol. Ecol. 12: 1137–1151.
Michaels S.D. and Amasino R.M. 1999. FLOWERING LOCUS C encodes a novel MADS domain protein that acts as a repressor of flowering. Plant Cell 11: 949–956.
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.
Michaels S.D., Ditta G., Gustafson-Brown C., Pelaz S., Yanofsky M., Amasino R.M. 2003. Agl24 acts as a promoter of flowering in Arabidopsis and is positively regulated by vernalization. Plant J. 33: 867–874.
Moon J., Suh S.S., Lee H., Choi K.R., Hong C.B., Paek N.C., Kim S.G., Lee I. 2003a. The SOC1 MADS-box gene integrates vernalization and gibberellin signals for flowering in Arabidopsis.. Plant J. 35: 613–623.
Moon Y.H., Chen L., Pan R.L., Chang H.S., Zhu T., Maffeo D.M., Sung Z.R. 2003b. EMF genes maintain vegetative development by repressing the flower program in Arabidopsis.. Plant Cell 15: 681–693.
Moon J., Lee H., Kim M., Lee I. 2005. Analysis of flowering pathway integrators in Arabidopsis.. Plant Cell Physiol. 46: 292–299.
Müller J., Hart C.M., Francis N.J., Vargas M.L., Sengupta A., Wild B., Miller E.L., O’Connor M.B., Kingston R.E., Simon J.A. 2002. Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell 111: 197–208.
Ohad N., Yadegari R., Margossian L., Hannon M., Michaeli D., Harada J.J., Goldberg R.B., Fischer R.L. 1999. Mutations in FIE, a WD Polycomb group gene, allow endosperm development without fertilization. Plant Cell 11: 407–416.
Parcy F. 2005. Flowering: A time for integration. Int. J. Dev. Biol. 49: 585–593.
Ringrose L. and Paro R. 2004. Epigenetic regulation of cellular memory by the Polycomb and Trithorax group proteins. Annu. Rev. Genet. 38: 413–443.
Samach A., Onouchi H., Gold S.E., Ditta G.S., Schwarz-Sommer Z., Yanofsky M.F., Coupland G. 2000. Distinct roles of CONSTANS target genes in reproductive development of Arabidopsis.. Science 288: 1613–1616.
Schubert D., Clarenz O., Goodrich J. 2005. Epigenetic control of plant development by Polycomb-group proteins. Curr. Opin. Plant Biol. 8: 553–561.
Searle I. and Coupland G. 2004. Induction of flowering by seasonal changes in photoperiod. EMBO J. 23: 1217–1222.
Sessions A., Burke E., Presting G., Aux G., McElver J., Patton D., Dietrich B., Ho P., Bacwaden J., Ko C.et al. 2002. A high-throughput Arabidopsis reverse genetics system. Plant Cell 14: 2985–2994.
Sheldon C.C., Rouse D.T., Finnegan E.J., Peacock W.J., Dennis E.S. 2000. The molecular basis of vernalization: The central role of FLOWERING LOCUS C (FLC). Proc. Natl. Acad. Sci. 97: 3753–3758.
Soppe W.J., Jasencakova Z., Houben A., Kakutani T., Meister A., Huang M.S., Jacobsen S.E., Schubert I., Fransz P.F. 2002. DNA methylation controls histone H3 lysine 9 methylation and heterochromatin assembly in Arabidopsis.. EMBO J 21: 6549–6559.
Sung S. and Amasino R.M. 2004. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 427: 159–164.
Yang C.H., Chen L.J., Sung Z.R. 1995. Genetic regulation of shoot development in Arabidopsis: Role of the EMF genes. Dev. Biol. 169: 421–435.
Yoshida N., Yanai Y., Chen L., Kato Y., Hiratsuka J., Miwa T., Sung Z.R., Takahashi S. 2001. EMBRYONIC FLOWER2, a novel Polycomb group protein homolog, mediates shoot development and flowering in Arabidopsis.. Plant Cell 13: 2471–2481.
Yu H., Xu Y., Tan E.L., Kumar P.P. 2002. Agamous-like 24, a dosage-dependent mediator of the flowering signals. Proc. Natl. Acad. Sci. 99: 16336–16341.
Zhang H. and van Nocker S. 2002. The VERNALIZATION INDEPENDENCE 4 gene encodes a novel regulator of FLOWERING LOCUS C. Plant J. 31: 663–673.(Nicole Sch?nrock, Romaric)
Abstract
Polycomb-group (PcG) proteins form a cellular memory by maintaining developmental regulators in a transcriptionally repressed state. We identified a novel flowering gene that is under PcG control in Arabidopsis—the MADS-box gene AGL19. AGL19 expression is maintained at very low levels by the PcG proteins MSI1, CLF, and EMF2, and AGL19 is partly responsible for the early flowering phenotype of clf mutants. AGL19 chromatin is strongly enriched in trimethylation of Lys 27 on histone H3 (H3K27me3) but not in H3K9me2. Repressive H3K27me3 marks were reduced by decreased CLF or MSI1 levels and by prolonged cold, suggesting that the PcG proteins MSI1 and CLF repress AGL19 in the absence of cold. Ectopic expression of AGL19 strongly accelerates flowering, and agl19 mutants have a decreased response to vernalization, the promotion of flowering by prolonged cold. Epistasis analyses revealed that AGL19 works in the poorly characterized FLC-independent vernalization pathway and does not require SOC1 to function. In this pathway, prolonged cold relieves AGL19 from PcG repression by a mechanism that requires VIN3 but not VRN2. Elevated AGL19 levels activate LFY and AP1 and eventually cause flowering.
[Keywords: Polycomb proteins; histone methylation; MSI1; vernalization; FLC; SOC1]
Received December 22, 2005; revised version accepted April 13, 2006.
Polycomb-group (PcG) proteins are cellular memory modules that are thought to maintain genes in a transcriptionally inactive state after an initial repression has been established (for review, see Brock and Fisher 2005). PcG proteins were first identified in insects, but are essential for normal development in both animals and plants (for review, see Schubert et al. 2005). These proteins often form large multiprotein complexes, such as the Polycomb Repressive Complex 2 (PRC2). Drosophila PRC2 contains four core subunits: the MSI1-like protein p55, Enhancer of Zeste [E(Z)], Extra Sex Combs (ESC), and Suppressor of Zeste 12 [Su(Z)12] (Czermin et al. 2002; Müller et al. 2002). Transcriptional repression by PRC2 relies on histone methyltransferase activity of the E(Z) subunit, which preferentially catalyzes histone H3 Lys 27 (H3K27) trimethylation (for review, see Cao and Zhang 2004).
Unlike in insects, where the PRC2 members are encoded by single-copy genes, in plants homologous PRC2 subunits are often encoded by small gene families (Schubert et al. 2005), and several possible PRC2-like complexes have been proposed to exist in Arabidopsis (Chanvivattana et al. 2004; Hennig et al. 2005). However, to date, only one Arabidopsis PRC2 complex has been well characterized—the FERTILISATION INDEPENDENT SEED (FIS) complex (K?hler et al. 2003a). The FIS complex consists of MSI1, the E(Z) homolog MEDEA (MEA), the ESC homolog FERTILISATION INDEPENDENT ENDOSPERM (FIE), and likely the Su(Z)12 homolog FIS2 (K?hler et al. 2003a; Chanvivattana et al. 2004). This complex has specific functions during gametophyte and early seed development, including suppression of seed development in the absence of fertilization and repression of the MADS-box gene PHERES1 (PHE1) (Chaudhury et al. 1997; Grossniklaus et al. 1998; Ohad et al. 1999; K?hler et al. 2003b). During sporophyte development, a second PRC2 complex, the CURLY LEAF (CLF) complex, most likely represses transcription of floral homeotic genes, such as the MADS-box gene AGAMOUS (AG). The CLF complex probably consists of MSI1, the E(Z) homolog CLF, FIE, and the Su(Z)12 homolog EMBRYONIC FLOWER2 (EMF2) (Goodrich et al. 1997; Kinoshita et al. 2001; Yoshida et al. 2001; Hennig et al. 2003; Chanvivattana et al. 2004; Katz et al. 2004). The third potential PRC2-like complex is the VERNALIZATION (VRN) complex. The existence of the VRN complex was hypothesized because the Su(Z)12 homolog VRN2 is required for maintaining repression of the MADS-box gene FLOWERING LOCUS C (FLC) after vernalization and for vernalization-induced H3 methylation at the FLC locus (Chandler et al. 1996; Gendall et al. 2001; Bastow et al. 2004; Chanvivattana et al. 2004; Sung and Amasino 2004). Because FLC is a very potent repressor of flowering, epigenetic control of the floral transition has recently received considerable attention (for review, see He and Amasino 2005).
The transition to flowering is controlled by diverse environmental and developmental signals, and many genes that control flowering in Arabidopsis have been identified (for review, see Boss et al. 2004; Ausin et al. 2005). These genes have been grouped in several genetic pathways, including the photoperiod, the vernalization, and the autonomous pathways. The photoperiod pathway responds to seasonal changes in day length, while the vernalization pathway renders Arabidopsis competent to flower after prolonged exposure to low temperatures; for example, during winter (for review, see Henderson and Dean 2004; Searle and Coupland 2004). These two pathways contribute to initiate flowering of winter-annual plants in the favorable conditions of spring and summer. In contrast, the autonomous pathway promotes flowering independently of environmental signals. Both the autonomous and vernalization pathways function mainly by reducing expression of the floral repressor FLC (Michaels and Amasino 1999, 2001; Sheldon et al. 2000; Gendall et al. 2001). Repression of FLC by vernalization is initiated by VERNALIZATION INSENSITIVE3 (VIN3) and maintained by VERNALIZATION1 (VRN1) and VERNALIZATION2 (VRN2) (Gendall et al. 2001; Levy et al. 2002; Sung and Amasino 2004). When expressed, FLC prevents transcription of the pathway integrators FLOWERING LOCUS T (FT) and the MADS-box gene SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1, or AGL20) (Borner et al. 2000; Lee et al. 2000; Michaels and Amasino 2001), which combine signals from several pathways to induce flowering (for review, see Parcy 2005). Vernalization releases SOC1 from FLC repression, thus making SOC1 an important component of the FLC-dependent vernalization pathway (Lee et al. 2000; Hepworth et al. 2002; Moon et al. 2003a). Although FLC plays a central role in vernalization in Arabidopsis, flc-null mutants still respond to vernalization, demonstrating that FLC-independent mechanisms exist in the vernalization pathway (Michaels and Amasino 2001; Sung and Amasino 2004). However, the molecular nature of these FLC-independent mechanisms is poorly understood.
Among the Arabidopsis homologs of PRC2 subunits, the floral transition is not only regulated by the Su(Z)12 homolog VRN2, but also by the p55 homolog MSI4 (also called FVE) (Koornneef et al. 1991; Ausin et al. 2004; Kim et al. 2004). FVE is part of the autonomous pathway, needed for the epigenetic repression of FLC via histone deacetylation. In addition to FVE/MSI4, Arabidopsis contains four other MSI1-like proteins (MSI1–3, MSI5) (Ach et al. 1997; Hennig et al. 2003). Although MSI1-like proteins have no catalytic activity, they are often essential subunits of protein complexes controlling chromatin dynamics (for review, see Hennig et al. 2005). Arabidopsis MSI1, for instance, is vital for plant development, as msi1-null mutants are embryo lethal (K?hler et al. 2003a; Guitton et al. 2004). Strong reduction of MSI1 protein levels by cosuppression (msi1-cs plants) is not lethal but has severe impacts on development (Hennig et al. 2003). In this study, we performed transcriptional profiling of msi1-cs plants and found that the MADS-box gene AGL19 is under control of the MSI1, CLF, and EMF2 Arabidopsis PcG proteins. AGL19 is a potent floral activator, which is repressed via H3K27 trimethylation in the absence of vernalization. This work also establishes a novel role for PcG proteins as repressors of the FLC-independent vernalization pathway of Arabidopsis.
Results
AGL19 is up-regulated in msi1-cs plants
In order to identify genes affected by the severe reduction in MSI1 levels in msi1-cs plants, we performed microarray RNA profiling experiments on rosette leaves of 23-d-old wild-type and msi1-cs plants grown in long days (LD). This developmental stage was chosen because phenotypic alterations of msi1-cs rosette leaves can then first be observed (Hennig et al. 2003). Statistical analysis of the microarray data identified eight robustly down-regulated genes and 122 robustly up-regulated genes. These deregulated genes fall into diverse categories, with a large proportion being involved in histone metabolism, cell cycle, and DNA repair. The MADS-box gene AGL19 was among the strongest up-regulated genes, with a 21-fold increase in msi1-cs leaves, and was the only MADS-box gene with a significantly altered expression pattern (Fig. 1A). RT–PCR confirmed that AGL19 is usually expressed at very low levels in wild-type rosette leaves but was strongly up-regulated in msi1-cs leaves (Fig. 1B). AGL19 belongs to the type II class of MADS-box genes and is phylogenetically most closely related to AGL14 and to the floral pathway integrator SOC1/AGL20 (Becker and Theissen 2003). So far, AGL19 has not been extensively characterized, and its biological function is unknown. Because MADS-box transcription factors are often developmental key regulators, AGL19 was selected for a detailed study.
Figure 1. Molecular characterization of AGL19. (A) Signal values from microarrays testing RNA from rosette leaves of 23-d-old msi1-cs and wild-type (WT) plants grown in LD. Shown are means ± ranges of two replicates. (B) Confirmation of AGL19 up-regulation in msi1-cs using rosette leaves from 21-d-old plants grown in LD. Here and in all subsequent experiments, GAPDH was used as a control. (C) Protein blots of wild-type, 35S::MSI1 overexpressing (msi1-oe), and MSI1 cosuppression plants (msi1-cs). Protein was extracted from leaves of 20-d-old plants. (Top panel) Ten micrograms of protein per sample was subjected to immunoblotting with affinity-purified, anti-MSI1-specific antiserum (Hennig et al. 2003). (Bottom panel) Ponceau red staining of the blot is shown as a loading control. (D) Expression analysis of AGL19, SOC1, and AGL14 in LD. (S) Nine-day-old seedlings; (RL) rosette leaves from 35-d-old plants; (CL) cauline leaves from 35-d-old plants; (INF) inflorescence meristem and closed flower buds; (F) open flowers; (GS) 2-d-old germinating seeds; (R) roots. Note that PCR was done with 33 cycles for AGL19 and AGL14, but with only 30 cycles for SOC1. (E) AGL19 expression increases with plant age. RNA was extracted from seedlings grown in SD. Age was measured in days after germination (DAG). Note that PCR was done with 38 cycles for AGL19. (F) AGL19 is strongly expressed in apices. RNA was extracted from dissected 18-d-old seedlings grown in SD. (S) Total seedlings; (A) apices only; (RL) rosette leaves only. (G) Nuclear localization of AGL19-GFP. (Left) Differential interference contrast image. (Right) GFP fluorescence. Bar, 50 μm.
AGL19 is a weakly expressed transcription factor localized in the nucleus
AGL19 was originally considered as a root-specific gene (Alvarez-Buylla et al. 2000). However, as we detected AGL19 in wild-type rosette leaves using both microarrays and RT–PCR, we re-examined the expression pattern of AGL19 in different organs and compared it with that of its closest phylogenetic relatives SOC1 and AGL14 (Fig. 1D). Although the strongest expression of AGL19 was detected in roots, it was also detected in rosette leaves and seedlings and to a lesser extent in cauline leaves and flowers. This expression pattern overlaps widely with the expression of SOC1. In contrast, AGL14 transcripts were confined to roots as previously described (Alvarez-Buylla et al. 2000). In general, AGL19 transcripts increase with the age of the plant (Fig. 1E). AGL19 is enriched in apices including the shoot apical meristem and developing leaf primordia in dissected 18-d-old seedlings grown in short days (SD), while it is very weakly expressed in leaves at this developmental stage (Fig. 1F). Low gene expression levels have precluded a more detailed analysis of the AGL19 expression pattern in the apex by in situ hybridization. Localization studies using GFP fusion proteins showed that AGL19 is targeted to the nucleus (Fig. 1G). Together, AGL19, but not its close homolog AGL14, is expressed with a similar organ-specificity as the well-known floral integrator SOC1.
AGL19 is sufficient to promote flowering
To study the effects of ectopic AGL19 expression, we constructed transgenic plants containing the AGL19 cDNA under the control of the strong cauliflower mosaic virus 35S promoter (35S::AGL19) (Fig. 2A). Five independent transgenic lines were analyzed, all of which were extremely early flowering in LD conditions and produced two to four rosette leaves at the time of bolting (Fig. 2B; Table 1). In noninductive SD conditions, 35S::AGL19 plants were also early flowering but produced more leaves than in LD, and therefore, they remained responsive to photoperiod (Fig. 2C). In the late flowering fve background, the overexpression of AGL19 had the strongest effect on flowering time as 35S::AGL19 fve plants flowered similarly to 35S::AGL19 plants (Table 1). Thus, overexpression of AGL19 is sufficient to promote flowering.
Table 1 Effect of 35S::AGL19 on flowering time in LD
Figure 2. AGL19 promotes the transition to flowering. (A) Expression of AGL19 in 35S::AGL19 plants. RNA was extracted from 10-d-old seedlings grown in LD. (B) Appearance of wild-type (WT) and 35S::AGL19 plants grown for 20 d in LD. (C) Flowering time of 35S::AGL19 plants grown in LD or SD. Note that for flowering time experiments in all figures, means ± S.D. are shown for at least 14 plants.
agl19 mutants have a decreased vernalization response
To determine whether AGL19 normally controls flowering time in wild-type plants, we used two agl19 insertion alleles (agl19-1 and agl19-2), in which the AGL19 transcript was undetectable (Fig. 3A,B). The agl19 mutants flowered normally in LD conditions, but flowered slightly later than wild type in SD (Fig. 3C). Because overexpression of AGL19 caused phenotypes similar to overexpression of SOC1 (Borner et al. 2000), which is regulated by the vernalization pathway, we tested whether AGL19 was also involved in the response to vernalization. Indeed, agl19 mutants flowered considerably later than wild type after a 6-wk cold treatment (Fig. 3C,D). Thus, AGL19 is not only sufficient to promote flowering when ectopically expressed, but it is also required for the promotion of flowering in response to vernalization.
Figure 3. agl19 mutants have a decreased response to vernalization. (A) Position of T-DNA insertions for agl19-1 and agl19-2 alleles. The arrow indicates the transcription start site. The gene structure was drawn according to TIGRv5 annotation. (B) AGL19 transcripts in wild type (WT), agl19-1, and agl19-2 plants. RNA was extracted from rosette leaves of 40-d-old plants grown in LD. RT–PCR was performed with 33 cycles. (C) Effect of vernalization (Vern) on the flowering time of agl19. Wild-type (WT) and agl19-1 plants were exposed to either 2-d stratification (empty bars) or to 6-wk vernalization (hatched bars) at 4°C before being transferred to either LD or SD growing conditions. (D) Comparison of decreased vernalization response in agl19-1 and agl19-2. Flowering time in SD without vernalization (empty bars) or after a 6-wk vernalization treatment (hatched bars).
Vernalization induces AGL19 expression
Because the agl19 mutants had a reduced vernalization response, we tested whether a prolonged cold treatment affected AGL19 transcript levels. Indeed, AGL19 mRNA amounts increased considerably in response to an extended cold exposure (Fig. 4A). The same increase could be seen for SOC1 after vernalization but not for AGL19’s closest homolog, AGL14. VRN2 and VIN3 are two crucial mediators of vernalization in Arabidopsis (Chandler et al. 1996; Sung and Amasino 2004). When testing induction of AGL19 by vernalization in vrn2 and vin3 mutant plants, we found that VIN3 but not VRN2 is required for increased AGL19 expression after vernalization (Fig. 4B,C). However, VRN2 is needed to attain maximum AGL19 expression. In the absence of vernalization, FLC represses SOC1 (Lee et al. 2000). To test whether FLC represses AGL19 as well, we measured AGL19 expression in flc seedlings. Because AGL19 levels were not increased in flc but even slightly reduced, FLC is not required to maintain AGL19 in a repressed state before vernalization (Fig. 4D). Next, we tested whether AGL19 affected FLC expression. Neither loss nor strong ectopic overexpression of AGL19 influenced FLC levels considerably (Fig. 4E). After the initial repression of FLC by a vernalization cold treatment, flowering will only be promoted if FLC repression is maintained (Gendall et al. 2001; Levy et al. 2002). Mutants such as vrn1 have a reduced vernalization response, similar to agl19, due to the inability to maintain stable repression of FLC after a vernalization treatment (Levy et al. 2002). Therefore, we tested whether the decreased vernalization response observed in agl19 was also caused by a derepression of FLC. In contrast to vrn1, FLC levels remained low after vernalization in agl19, indicating that AGL19 is dispensable for the establishment and maintenance of FLC repression (Fig. 4G). Taken together, these results show that AGL19 is not repressed by FLC in wild-type plants and that the reduced vernalization response in agl19 mutants is not caused by elevated FLC levels. Thus, it is possible that AGL19 functions in an FLC-independent branch of the vernalization pathway.
Figure 4. AGL19 functions in the FLC-independent vernalization pathway. (A) AGL19, SOC1, and AGL14 expression in 10-d-old wild-type (WT) seedlings grown in SD with or without a 6-wk vernalization treatment (Vern). (B) AGL19 and SOC1 expression in 10-d-old wild-type (WT) and vin3 seedlings grown in SD with or without a 6-wk vernalization treatment. (C) AGL19 expression in 10-d-old wild-type (WT) (Landsberg erecta) and vrn2 seedlings grown in SD with or without a 6-wk vernalization treatment. (D) Effect of FLC and SOC1 on AGL19 expression. (E) Effect of AGL19 and SOC1 on FLC expression. (F) Effect of AGL19 on SOC1 expression. For D–F, RNA was extracted from 10-d-old seedlings grown in LD and analyzed by Q-PCR. (G) Expression of FLC in wild type (WT), agl19, and vrn1 after vernalization. RNA was extracted from seedlings 4, 10, and 20 d after a 2-d stratification or a 6-wk vernalization treatment. (H,I) Flowering time of wild type (WT), agl19, soc1, 35S::AGL19, and the double mutants in LD (H) or SD (I).
In order to genetically test whether AGL19 functions in the FLC-independent vernalization pathway, the vernalization response of the agl19 flc double mutant was tested. This double mutant responded less to vernalization than either single mutant (Table 2). Thus, the agl19 and flc mutants exhibit an additive rather than an epistatic relationship, confirming that AGL19 functions in an FLC-independent vernalization pathway. Similar to some previous reports (e.g., Zhang and van Nocker 2002), the flc mutant flowered later than wild type in SD after vernalization in several independent experiments. The reasons for this effect are not known, but indirect effects mediated by poorly understood FLC targets and functions (McKay et al. 2003; Edwards et al. 2006) might be involved.
Table 2 Effect of vernalization on flowering time of agl19 and flc mutants in SD
AGL19 function does not require SOC1
The FLC-dependent vernalization pathway functions mainly to control expression of the floral integrator SOC1 (Lee et al. 2000; Gendall et al. 2001; Hepworth et al. 2002; Levy et al. 2002; Moon et al. 2003a, 2005). Therefore, we next investigated whether the AGL19-containing branch of the FLC-independent vernalization pathway is integrated by SOC1 as well and whether AGL19 and SOC1 influence each other’s expression. First, we tested whether AGL19 levels affect SOC1 expression. However, neither loss of AGL19 nor increased AGL19 levels led to considerably altered SOC1 expression in wild type (Fig. 4F). Similarly, ectopic overexpression of AGL19 did not affect SOC1 in fve, which has very low SOC1 levels due to elevated FLC expression (Fig. 4F). Second, we tested whether SOC1 levels affect AGL19 expression. Loss of SOC1, which delayed flowering (Fig. 4H), reduced AGL19 levels slightly, but ectopic overexpression of SOC1, which greatly accelerated flowering (Supplementary Fig. S1), strongly repressed AGL19 (Fig. 4D). This is in contrast to AGL24, which is induced by ectopic SOC1 expression (Michaels et al. 2003). Because ectopic overexpression of SOC1 strongly promotes flowering without increasing AGL19 levels, SOC1 does not function via activation of AGL19 expression. Moreover, because ectopic overexpression of AGL19 strongly promotes flowering without increasing SOC1 levels, AGL19 does not function via activation of SOC1 expression. Rather, it is possible that SOC1 and AGL19 function in genetic parallel but cross-connected pathways.
To examine whether AGL19 and SOC1 function in the same or in parallel genetic pathways to promote flowering, the agl19 soc1 double mutant was generated. If AGL19 and SOC1 act in the same pathway, then the late-flowering phenotype of the double mutant should not transgress that of the latest single mutant. However, the agl19 soc1 double mutant flowered later than the soc1 single mutant, and this delay was more pronounced in SD (additional 15 leaves) than in LD (additional two leaves) (Fig. 4H,I), indicating that AGL19 and SOC1 act in genetically different flowering pathways. Similarly, 35S::AGL19 soc1 plants were still early flowering, but showed an additive rather than an epistatic phenotype, thus confirming that AGL19 and SOC1 probably function in parallel pathways to promote flowering (Fig. 4H,I). Nonetheless, it is possible that AGL19 and SOC1 regulate common downstream targets. Indeed, expression of the floral integrator LEAFY (LFY) and of the meristem identity gene APETALA1 (AP1) was increased in 35S::AGL19 and 35S::SOC1 seedlings (Supplementary Fig. S1).
Control of AGL19 by Polycomb-group complexes
As AGL19 expression was increased in msi1-cs plants, it may be repressed by a MSI1-containing Arabidopsis PRC2 complex, like the CLF–PRC2 complex, which functions during sporophytic development (Schubert et al. 2005). Indeed, AGL19 was similarly up-regulated in 15-d-old emf2 and clf seedlings as well as in 4-wk-old rosette leaves of clf and msi1-cs plants (Figs. 5A, 6A). Interestingly, derepression of AGL19 in clf increased with plant age (Fig. 5B). These results show that MSI1, CLF, and EMF2, three likely subunits of the CLF complex, are all required to repress AGL19 in the absence of a vernalization treatment. As AGL19 shares similarities with SOC1, we tested whether CLF represses SOC1 as well. In contrast to AGL19, SOC1 expression was not increased in clf mutants (Fig. 5C). Similarly, no differences were seen in SOC1 expression in msi1-cs leaves compared with wild-type leaves in the microarray experiments (data not shown). Together this suggests that the CLF–PRC2 complex does not regulate SOC1. Because vernalization activates AGL19, it is possible that vernalization represses MSI1, CLF, or EMF2. However, we found no considerable effect of vernalization on expression of these genes (Fig. 5D). Next, we tested whether AGL19 is a direct target of the CLF complex by chromatin immunoprecipitation (ChIP). The anti-MSI1 antiserum is not suitable for immunoprecipitation (data not shown), and no antibodies against other PRC2 subunits are available. Therefore, we used transgenic plants that express a myc-tagged FIE version. Using anti-myc antibodies, AGL19 sequences were two- to threefold enriched in precipitated chromatin from the 35S::myc-FIE line compared with wild type (Fig. 5E). These data suggest that the CLF–PRC2 complex directly binds to the AGL19 locus.
Figure 5. AGL19 is up-regulated in PcG mutants and contributes to the early flowering phenotype of clf. (A) Expression of AGL19 in clf and emf2. RNA was extracted from 15-d-old seedlings grown in LD. (B) Age dependency of AGL19 derepression in clf. RNA was extracted at 10, 15, and 33 d after germination. (C) Expression of SOC1 in clf. RNA was extracted from 10-d-old seedlings and from rosette leaves of 21-d-old plants grown in LD. (D) CLF, EMF2, and MSI1 expression in 10-d-old wild-type (WT) seedlings grown in SD with or without a 6-wk vernalization treatment (Vern). (E) ChIP using anti-myc antibodies and 10-d-old wild-type Landsberg erecta or transgenic 35S::myc-FIE seedlings grown in LD. Shown are PCR products of fragments II and III from the AGL19 locus (see Fig. 6 for details) and of an unrelated PHOSPHOFRUCTOKINASE (PFK) gene. (F) Flowering time of msi1-cs plants in SD. (G) Flowering time of wild-type (WT), agl19, clf, and clf agl19 plants in LD. Leaf numbers differ significantly between clf and clf agl19 double mutants (t-test: p < 0.001). (H) Images of 22-d-old wild-type (WT), agl19, clf, and clf agl19 plants grown in LD at 21°C. Note that all clf mutants are already flowering, while agl19 clf double mutants have only just started bolting.
Figure 6. Histone methylation pattern at the AGL19 locus. (A) Expression analysis of AGL19 in msi1-cs and clf in rosette leaves of 4-wk-old plants grown in LD. (B) Expression analysis of AGL19 in 10-d-old seedlings grown in SD with or without a 6-wk vernalization treatment (Vern). (C) Schematic representation of AGL19 showing intron–exon structure (annotated according to TIGRv5) and position of the fragments used for amplification by PCR following ChIP. The arrow indicates the transcription start site. (D) Quantification of multiplex PCR products after ChIP using anti-trimethyl-histone H3K27 (H3K27me3) antiserum on 4-wk-old rosette leaves from wild-type (WT), msi1-cs, and clf plants grown in LD. Values are shown as enrichment of H3K27me3 using AGL19 specific primers relative to PFK for wild-type (WT), msi1-cs, and clf plants. (E) ChIP experiment performed on 10-d-old wild-type (WT) seedlings grown in SD with or without a 6-wk vernalization treatment using antibodies against H3K27me3. Values represent mean ± range from two independent ChIP experiments.
Consistent with the findings that AGL19 can promote flowering and that AGL19 expression is increased in msi1-cs plants, msi1-cs plants flowered significantly earlier than wild type in SD (Fig. 5F). Because cosuppression of MSI1 occurs around an age of 2–3 wk in LD, it cannot affect flowering time any more at this stage, and msi1-cs plants were not consistently earlier flowering than wild type in LD (data not shown). Similarly, AGL19 expression is increased in clf and emf2 mutants, and both flower earlier than wild type (Yang et al. 1995; Goodrich et al. 1997), suggesting that the overexpression of AGL19 observed in these mutants could be partially responsible for their early flowering phenotype. In order to test this hypothesis, we constructed the clf agl19 double mutant. Loss of AGL19 did not abolish the clf leaf phenotype, but partially suppressed the early flowering phenotype (Fig. 5G,H). Thus, elevated AGL19 levels specifically contribute to the early flowering of clf.
MSI1 and CLF epigenetically regulate the AGL19 locus
In animals, the E(Z)-related subunit of PRC2-like complexes has H3K27 methyltransferase activity (Cao and Zhang 2004). It is likely that the Arabidopsis E(Z) homologs, such as CLF, function as methyltransferases as well to establish repressive H3K27 trimethylation (H3K27me3) at target chromatin (Schubert et al. 2005). Because the PRC2 subunits CLF and MSI1 repress AGL19, we tested whether AGL19 chromatin was enriched in H3K27me3 marks. ChIP experiments using 4-wk-old rosette leaves revealed a strong enrichment of H3K27me3 at AGL19 (Fig. 6D). Interestingly, the H3K27me3 mark did not evenly cover the entire locus, but peaked in the 5' region of the gene (Fig. 6C, PCR fragment II). In contrast to H3K27me3, the heterochromatin H3K9me2 mark, which was strongly enriched in chromatin of the Ta2 and Cinful-like transposons, was not detected in AGL19 chromatin (Supplementary Fig. S2). This indicates that in rosette leaves, the AGL19 locus normally contains H3K27me3 but not H3K9me2 marks. To test whether the H3K27me3 mark depends on MSI1 and CLF, similar experiments were performed with msi1-cs and clf plants. Expression of a PHOSPHOFRUCTOKINASE (PFK) was not changed in msi1-cs or clf plants, and PFK was thus suitable as an internal control in multiplex PCR (Fig. 6A). In msi1-cs and clf, H3K27me3 at AGL19 was greatly reduced but not completely lost (1.8-fold and threefold reduction in msi1-cs and in clf, respectively) (Fig. 6D; Supplementary Fig. S2). This decrease in H3K27me3 observed in msi1-cs and clf was also specific for the region of AGL19 covered by PCR fragment II. Thus, MSI1 and CLF are needed to establish repressive H3K27me3 marks at AGL19 chromatin in a very specific pattern, but their function is most likely partially redundant with other Arabidopsis proteins.
Vernalization prevents H3K27 trimethylation at the AGL19 locus
Our experiments have established that in standard LD conditions, AGL19 chromatin carries H3K27me3 marks that most likely repress AGL19 transcription. Because a prolonged cold treatment strongly activates AGL19 expression, we tested whether such a treatment also affected the H3K27me3 marks on AGL19 chromatin. We performed a ChIP experiment on 10-d-old wild-type seedlings grown in SD with or without a 6-wk cold treatment. PFK expression was not affected by vernalization and hence was used as an internal control for multiplex PCR (Fig. 6B). In nontreated seedlings, in which no AGL19 expression was detectable (Fig. 6B), the AGL19 locus was highly enriched for H3K27me3 marks specifically in the 5' region of the gene (Fig. 6E; Supplementary Fig. S2). This methylation pattern was similar to that seen in the previous ChIP experiment on wild-type rosette leaves (Fig. 6D). In contrast, when AGL19 expression was induced after cold treatment, this H3K27me3 mark was dramatically reduced by fivefold compared with nonvernalized seedlings (Fig. 6B,E). The overlapping pattern of H3K27me3 observed after vernalization and in the msi1-cs or clf mutants suggests that a PRC2 complex containing MSI1 and CLF is responsible for repressing AGL19 in the absence of vernalization.
Discussion
AGL19 is a novel floral activator
AGL19 has been identified as a novel floral activator in Arabidopsis, because overexpression of AGL19 greatly accelerates the transition to flowering, while loss of AGL19 delayed flowering after a vernalization treatment. AGL19 belongs to the MADS-box gene family, which contains master regulatory genes essential for diverse developmental processes (Becker and Theissen 2003). Plant MADS-box genes form several phylogenetically distinct groups, and AGL19 together with its closest homolog AGL14 (68% amino acid identity) and the well-characterized floral activator SOC1 (AGL20, 46% amino acid identity) belong to the Tm3/SOC1 clade (Becker and Theissen 2003). Although AGL19, AGL14, and AGL20 are closely related, their expression patterns differ. In particular, AGL14 is exclusively expressed in roots, while AGL19 and SOC1 are expressed in many organs of the shoot as well. Despite the prominent expression of AGL19 in roots, root growth and macroscopic root morphology did not differ between wild-type, agl19, and 35S::AGL19 plants (Supplementary Fig. S2; data not shown).
AGL19 and SOC1 both function to promote flowering, but they have diverged significantly in terms of gene regulation and flowering pathways in which they participate. The agl19 mutant flowered normally in LD and was only slightly late flowering in SD, indicating that under standard laboratory conditions, AGL19 has only a minor role in the promotion of flowering. In contrast, the soc1 mutant is significantly late flowering in both LD and SD (Borner et al. 2000; Lee et al. 2000). Because agl19 mutants could no longer adequately accelerate flowering in response to prolonged cold treatments, AGL19 functions in the vernalization response. While SOC1 integrates several genetic pathways that promote flowering, including the vernalization and photoperiod pathways (Borner et al. 2000; Lee et al. 2000; Samach et al. 2000; Hepworth et al. 2002; Moon et al. 2003a), AGL19 is mainly part of the vernalization pathway.
Similarly to ectopic expression of SOC1, ectopic expression of AGL19 did not only greatly accelerate flowering but caused defects in floral development as well (Borner et al. 2000; data not shown). The fact that only 35S::AGL19 but not agl19 plants had altered flower morphology suggests that like SOC1, AGL19 is also dispensable for flower development and appears to function specifically during the floral transition after vernalization.
AGL19 regulates the floral transition in an FLC-independent vernalization pathway
Vernalization in Arabidopsis acts mainly via repression of FLC and subsequent up-regulation of SOC1 (Lee et al. 2000; Gendall et al. 2001; Hepworth et al. 2002; Levy et al. 2002; Moon et al. 2003a). However, flc-null mutants still show a vernalization response, and thus an FLC-independent vernalization pathway must exist (Michaels and Amasino 2001). Although both AGL19 and SOC1 are up-regulated by vernalization, only SOC1 but not AGL19 is repressed by FLC. In addition, AGL19 levels had no effect on the amount of FLC transcript, and FLC remained repressed after vernalization in the agl19 mutant. Thus, AGL19 does not regulate FLC, and the reduced vernalization response of agl19 mutants is not caused by deregulation of FLC as in the vrn1 and vrn2 mutants. Because agl19 flc double mutants had an additive effect on the vernalization response, AGL19 and FLC function in genetically separate pathways. The FLC-containing pathway requires VIN3 and VRN2 (Gendall et al. 2001; Sung and Amasino 2004). The AGL19-containing pathway requires VIN3 but not VRN2. Both elevated AGL19 and SOC1 levels lead to the activation of LFY and AP1 to promote flowering.
Interestingly, even the agl19 flc double mutant showed a residual vernalization response, indicating the presence of other vernalization pathways. Such a pathway might include the MADS-box gene AGL24, which is known to mediate an FLC-independent vernalization response (Yu et al. 2002; Michaels et al. 2003). Analysis of the vernalization response of agl19 flc agl24 triple mutants will help answer this question.
Although AGL24 acts in an FLC-independent pathway, it is closely linked to SOC1, because these two genes positively regulate each other’s expression, and overexpression of one of the two genes has little effect in the absence of the other (Michaels et al. 2003). In contrast, AGL19 and SOC1 act at least partially independently, because there seems to be no or even negative cross-regulation between AGL19 and SOC1 and overexpression of AGL19 greatly accelerated flowering in a soc1 background. Both agl19 soc1 as well as 35S::AGL19 soc1 showed additive rather than epistatic interactions, indicating that SOC1 and AGL19 act in separate genetic pathways that converge to control the common downstream targets AP1 and LFY. Thus, although SOC1, AGL24, and AGL19 share several similarities, they clearly have different roles in the genetic network controlling flowering, and are all needed to coordinate a fully functional vernalization response.
AGL19 repression depends on PRC2 subunits
In the absence of vernalization, AGL19 is lowly expressed, and MSI1, CLF, and EMF2 are all required for repression of AGL19. MSI1, CLF, and EMF2 are homologs of Drosophila PRC2 subunits (Ach et al. 1997; Goodrich et al. 1997; Yoshida et al. 2001), and it was suggested that these proteins together with FIE form an Arabidopsis PRC2 complex (Chanvivattana et al. 2004). Protein–protein interaction data support this model as: MSI1 interacts with FIE, FIE interacts with CLF, and CLF interacts with EMF2 (K?hler et al. 2003a; Chanvivattana et al. 2004; Katz et al. 2004). ChIP experiments suggested that AGL19 is a direct target gene of the CLF–PRC2 complex and established that AGL19 chromatin is enriched in H3K27me3 marks, which at least in mammals and insects are set by PRC2 complexes (for review, see Cao and Zhang 2004). Arabidopsis has several homologs of PRC2 subunits and probably contains several PRC2-like complexes, which may function in different tissues or at different developmental stages (for review, see Schubert et al. 2005). PRC2-like complexes are thought to function as transcriptional repressors, and other Arabidopsis PcG target genes include PHERES1 (PHE1), AGAMOUS (AG), and FLC. MEA, FIE, and FIS2 are required to repress PHE1 (K?hler et al. 2003b); MSI1, CLF, FIE, and EMF2 are required to repress AG (Chen et al. 1997; Goodrich et al. 1997; Kinoshita et al. 2001; Hennig et al. 2003); and VRN2 is required to repress FLC (Gendall et al. 2001). Interestingly, PHE1, AG, and FLC as well as AGL19 are all MADS-box genes, supporting the notion that in contrast to metazoan PRC2 complexes, which mostly control Homeobox genes, plant PRC2 complexes often control MADS-box transcription factors. Importantly, only AGL19, but not SOC1 or AGL14, needs MSI1, CLF, and EMF2 to remain repressed in the absence of vernalization (this study; cf. also microarray data in Moon et al. 2003b). This suggests that either PRC2 repression was a feature of the Tm3/SOC1 ancestor and was lost in SOC1 and AGL14, or that AGL19 acquired PRC2 repression de novo. However, plant PRC2 complexes also have non-MADS-box targets (Katz et al. 2004), and more complete lists of primary and secondary plant and metazoan PRC2 target genes are needed for more systematic comparisons.
AGL19, which is a strong activator of flowering, is up-regulated in PRC2 mutants, and clf, fie, emf2, and msi1-cs plants are all early flowering. This suggests that CLF, EMF2, MSI1, and FIE form a PRC2 complex to control flowering time. MSI1 not only inhibits flowering by repressing AGL19, but also promotes flowering by contributing to normal SOC1 expression; this function, however, is independent of CLF (Bouveret et al. 2006). The contribution of increased AGL19 expression to the early flowering phenotype of msi1-cs plants could not be tested due to silencing of the 35S::MSI1 transgene by the AGL19 SALK-insertion allele in the 35S::MSI1 agl19 plants (data not shown). The 35S::MSI1 transgene is required to induce MSI1 cosuppression. In clf mutants, we found that AGL19 was partially responsible for the early flowering phenotype, as the agl19 clf double mutant flowered later than the clf mutant alone. However, derepression of AGL19 is not the major cause of the clf phenotype, and this is consistent with previous reports of a prominent role of ectopic AG expression for the clf phenotype (Goodrich et al. 1997). In clf mutants, expression of AG is correctly initiated in young floral meristems, but later breaks down and becomes expressed in the outer floral whorls (Goodrich et al. 1997). Similarly, PHE1 is initially normally expressed in fis mutants but is subsequently ectopically expressed in the endosperm (K?hler et al. 2003b), and FLC is initially repressed by vernalization treatments but subsequently becomes derepressed in vrn2 mutants (Gendall et al. 2001). Interestingly, the pattern of progressive AGL19 derepression in clf mutants is strikingly similar to that of other PcG targets in plants and Drosophila, where target genes are initially expressed normally but subsequently become ectopically activated (for review, see Ringrose and Paro 2004; Brock and Fisher 2005).
H3K27 trimethylation at AGL19 chromatin depends on PRC2 subunits and is regulated by vernalization
The mechanism of transcriptional repression by PRC2 complexes is not yet completely understood, but trimethylation of H3K27 appears to be of central importance in vivo (for review, see Cao and Zhang 2004). In Arabidopsis, H3K27me3 marks are found mostly in euchromatin (Lindroth et al. 2004), while H3K9me2 marks are hallmarks of silent heterochromatic genes (Soppe et al. 2002). Because AGL19 is only enriched in H3K27me3 but not in H3K9me2, repression of AGL19 is probably different from heterochromatic silencing as observed for many transposons. Instead, AGL19 appears to be a euchromatic gene that is repressed by H3K27me3. Interestingly, vernalization reduced H3K27me3 marks to lower levels than either reductions of MSI1 or loss of CLF. In the case of msi1-cs, the 5% residual MSI1 protein levels or one of the other four MSI1-like proteins in Arabidopsis could be sufficient to maintain partial H3K27me3 at the AGL19 locus. Similarly, partial redundancy of CLF and its homolog SWN has been observed (Chanvivattana et al. 2004). Thus, it is possible that functional redundancies attenuate the effects on AGL19 chromatin modification in msi1-cs and clf.
H3K27me3 was not spread uniformly across the AGL19 locus but peaked dramatically near the transcription start site. Similarly, binding of Drosophila PcG proteins seems to cluster at Polycomb-responsive elements (PREs) and promoters rather than being spread uniformly over target loci (Ringrose and Paro 2004; Brock and Fisher 2005). Consistent with this local clustering, Polycomb silencing blocks transcription initiation in Drosophila (Dellino et al. 2004). In Arabidopsis, the FIS PRC2 complex was found to bind preferentially at the PHE1 promoter (K?hler et al. 2003b), and FLC chromatin is enriched in H3 methylation at the promoter and the first large intron (Bastow et al. 2004; Sung and Amasino 2004). Taken together, these results suggest that AGL19 is repressed by a PRC2 complex containing MSI1, CLF, FIE, and EMF2 that sets H3K27me3 marks close to the transcriptional start site, similar to the situation in other PcG target genes.
Vernalization overcomes the repressive effect of the CLF–PRC2 complex. This could be carried out by active removal of the repressive H3K27me3 marks at AGL19. Because no biochemical activity that removes trimethylation is known, it is more likely that vernalization prevents the establishment and/or maintenance of H3K27me3 by the CLF–PRC2 complex during cell proliferation. This interpretation is supported by observations that vernalization often acts on dividing cells such as those present in meristems (for review, see Ausin et al. 2005). An alternative mechanism could involve exchange of methylated histones in AGL19 chromatin.
Together, the vernalization pathway in Arabidopsis contains at least two branches—the FLC-dependent and the FLC-independent branches, both of which require the regulatory effects of PcG proteins. While VRN2 functions in repressing the floral repressor FLC after vernalization, the CLF complex functions in repressing the floral activator AGL19 before vernalization. Thus, in Arabidopsis, different PcG proteins have been recruited to coordinate the vernalization response and to control the important developmental switch to reproductive growth.
Materials and methods
Plant material and growth conditions
Line msi1OEc2, which ectopically expresses MSI1 and gives rise to msi1-cs plants, was described before (Hennig et al. 2003). To construct plants expressing an epitope-tagged version of FIE, a construct encoding an N-terminal fusion of myc to FIE was cloned under the control of the 35S promoter and transformed into wild-type (Ler) plants. Line 4, which expressed the myc-FIE fusion (Supplementary Fig. S1), was used for further experiments. The two null AGL19 mutant alleles, agl19-1 (SALK_N578786) and agl19-2 (SALK_N516657) were obtained from the Nottingham Arabidopsis Seed Stock Center (NASC). The agl19-1 allele was used for all crossings and flowering time experiments unless otherwise stated. The fve-5, flc-6, vrn1-5, and clf-29 mutants are null alleles (Supplementary Fig. S1; Bouveret et al. 2006) from various collections of T-DNA insertion lines (SAIL_1167E5, SALK_41126, WiscDsLox393-396F9, and SALK_N521003, respectively) (Sessions et al. 2002; Alonso et al. 2003). All plants used in this study are in the Columbia background unless otherwise stated. Seeds of soc1-2, emf2-10, and vrn2-1 have been described (Chandler et al. 1996; Lee et al. 2000; Chanvivattana et al. 2004) and were kindly provided by I. Lee (Seoul National University, Seoul, Korea), J. Goodrich (University of Edinburgh, Edinburgh, UK), and C. Dean (John Innes Centre, Norwich, UK), respectively.
To construct plants that ectopically overexpressed AGL19 or SOC1 (35S::AGL19 and 35S::SOC1), the full-length coding sequences were amplified by PCR (for primers, see Supplementary Table S1). The cDNAs were inserted into binary destination vectors (AGL19 in pH7WG2; SOC1 in pK7WG2) (Karimi et al. 2002) downstream of the cauliflower mosaic virus (CaMV) 35S promoter. Constructs were transformed into Columbia wild-type plants, and 43 primary 35S::AGL19 transformants were obtained, of which five (35S::AGL19a–e) were selected for further study. The weak line 35S::AGL19a was used for all crossings and flowering time experiments unless otherwise stated.
For cellular localization studies, the full-length AGL19 cDNA was fused to an N-terminal GFP tag in vector pK7WGF2 (Karimi et al. 2002). AGL19_pK7WGF2 was used directly for bombardment of onion cells for transient expression assays as described previously (Gendall et al. 2001).
For measuring flowering time, seeds were plated on Murashige and Skoog (MS) medium (Duchefa), stratified for 2 d at 4°C, 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 μmol/m2·sec, 21°C ± 2°C) under LD (16 h light) photoperiods unless indicated otherwise. The flowering time was measured as the number of total rosette leaves longer than 0.5 cm at bolting for at least 14 plants. Graphs show means ± standard deviation. For the vernalization treatment, seeds were plated on MS medium and kept in continuous light for 1 d before being exposed to 4°C for 6 wk. After the vernalization treatment, plants were transferred to growth chambers (21°C ± 2°C) under LD or SD (8 h light) photoperiods as indicated.
RNA isolation and RT–PCR
RNA was extracted as previously described (Hennig et al. 2003). DNA-free RNA was reverse-transcribed using an oligo(dT) primer and SuperScript II reverse transcriptase (Invitrogen). Aliquots of the generated cDNA, which equaled 100 ng of total RNA, were used as the template for PCR with gene-specific primers (Supplementary Tables S2, S3). Q-PCR was performed in an ABI Prism 7700 Sequence Detection System (Applied Biosystems), using SYBR Green Master Mix reagent (Applied Biosystems) according to the manufacturer’s instructions. All amplification plots were analyzed with a fluorescence signal threshold of 0.1 to obtain Ct (threshold cycle) values. Experiments were performed in duplicate with error bars representing the range. Gene expression levels were normalized to an actin (ACT) control gene. The average Ct value for ACT was 23.83 (±0.18) for all templates measured in these experiments.
Microarray hybridization and evaluation
Arabidopsis wild-type (accession Columbia) and msi1OEc2 plants (Hennig et al. 2003) were grown for 23 d in growth chambers at 21°C under LD photoperiods. After identification of msi1-cs plants from the segregating msi1OEc2 population based on phenotype, rosette leaves were harvested. Material was harvested from at least 12 plants and pooled for each sample. The experiment was repeated twice to give two independent biological replicates. Affymetrix Arabidopsis ATH1 GeneChips were used in the experiment (Affymetrix). Labeling of samples, hybridizations, and measurements were performed as described (Hennig et al. 2004). Signal values were derived using the GCRMA algorithm in the statistical package R (version 1.9.1). Significantly differential gene expression was detected based on the rank-product algorithm (Breitling et al. 2004). This algorithm inherently corrects for multiple testing. Genes were considered as differentially expressed if p < 0.05.
ChIP
ChIP was performed as previously described (Bowler et al. 2004). Anti-myc 9E11 antibodies (Eurogentec), anti-dimethyl-histone H3 Lys 9 antiserum (Upstate, catalog #07-441), and anti-trimethyl-histone H3 Lys 27 antiserum (Upstate, catalog #07-449) were used for immunoprecipitation.
Immunoprecipitated DNA was analyzed by PCR using gene specific primers for AGL19 and primers against the control gene PFK (At4g04040) (Supplementary Table S4). Relative efficiencies of primer pairs were determined based on dilution series of input DNA (Supplementary Fig. S3B). Band intensities were quantified using Image J software. For quantitative displays of ChIP experiments, the values shown represent enrichment relative to PFK using the previously determined primer efficiencies. Bars represent mean and range for two independent ChIP experiments.
Acknowledgments
We thank Ihla Lee for soc1-2 mutant seeds, Justin Goodrich for emf2-10 mutant seeds, and Caroline Dean for helpful comments on the manuscript. N.S. was supported by the Roche Research Foundation. This research was supported by SNF project 3100AO-104238 to L.H., E.U. project QLG2-1999-00454 (ECCO) to W.G., a grant from the Zurich-Basel Plant Science Center to L.H. and C.K., and the Functional Genomics Center Zurich.
References
Ach R.A., Taranto P., 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.
Alonso J.M., Stepanova A.N., Leisse T.J., Kim C.J., Chen H., Shinn P., Stevenson D.K., Zimmerman J., Barajas P., Cheuk R.et al. 2003. Genome-wide insertional mutagenesis of Arabidopsis thaliana. Science 301: 653–657.
Alvarez-Buylla E.R., Liljegren S.J., Pelaz S., Gold S.E., Burgeff C., Ditta G.S., Vergara-Silva F., Yanofsky M.F. 2000. MADS-box gene evolution beyond flowers: Expression in pollen, endosperm, guard cells, roots and trichomes. Plant J. 24: 457–466.
Ausin I., Alonso-Blanco C., Jarillo J.A., Ruiz-Garcia L., Martinez-Zapater J.M. 2004. Regulation of flowering time by FVE, a retinoblastoma-associated protein. Nat. Genet. 36: 162–166.
Ausin I., Alonso-Blanco C., Martinez-Zapater J.M. 2005. Environmental regulation of flowering. Int. J. Dev. Biol. 49: 689–705.
Bastow R., Mylne J.S., Lister C., Lippman Z., Martienssen R.A., Dean C. 2004. Vernalization requires epigenetic silencing of FLC by histone methylation. Nature 427: 164–167.
Becker A. and Theissen G. 2003. The major clades of MADS-box genes and their role in the development and evolution of flowering plants. Mol. Phylogenet. Evol. 29: 464–489.
Borner R., Kampmann G., Chandler J., Gleissner R., Wisman E., Apel K., Melzer S. 2000. A MADS domain gene involved in the transition to flowering in Arabidopsis.. Plant J. 24: 591–599.
Boss P.K., Bastow R.M., Mylne J.S., Dean C. 2004. Multiple pathways in the decision to flower: Enabling, promoting, and resetting. Plant Cell 16:(Suppl.), 18–31.
Bouveret R., Sch?nrock N., Gruissem W., Hennig L. 2006. Regulation of flowering time by Arabidopsis MSI1. Development 133: 1693–1702.
Bowler C., Benvenuto G., Laflamme P., Molino D., Probst A.V., Tariq M., Paszkowski J. 2004. Chromatin techniques for plant cells. Plant J. 39: 776–789.
Breitling R., Armengaud P., Amtmann A., Herzyk P. 2004. Rank products: A simple, yet powerful, new method to detect differentially regulated genes in replicated microarray experiments. FEBS Lett. 573: 83–92.
Brock H.W. and Fisher C.L. 2005. Maintenance of gene expression patterns. Dev. Dyn. 232: 633–655.
Cao R. and Zhang Y. 2004. The functions of E(Z)/EZH2-mediated methylation of lysine 27 in histone H3. Curr. Opin. Genet. Dev. 14: 155–164.
Chandler J., Wilson A., Dean C. 1996. Arabidopsis mutants showing an altered response to vernalization. Plant J. 10: 637–644.
Chanvivattana Y., Bishopp A., Schubert D., Stock C., Moon Y.H., Sung Z.R., Goodrich J. 2004. Interaction of Polycomb-group proteins controlling flowering in Arabidopsis.. Development 131: 5263–5276.
Chaudhury A.M., Ming L., Miller C., Craig S., Dennis E.S., Peacock W.J. 1997. Fertilization-independent seed development in Arabidopsis thaliana. Proc. Natl. Acad. Sci. 94: 4223–4228.
Chen L., Cheng J.C., Castle L., Sung Z.R. 1997. EMF genes regulate Arabidopsis inflorescence development. Plant Cell 9: 2011–2024.
Czermin B., Melfi R., McCabe D., Seitz V., Imhof A., Pirrotta V. 2002. Drosophila Enhancer of Zeste/ESC complexes have a histone H3 methyltransferase activity that marks chromosomal Polycomb sites. Cell 111: 185–196.
Dellino G.I., Schwartz Y.B., Farkas G., McCabe D., Elgin S.C., Pirrotta V. 2004. Polycomb silencing blocks transcription initiation. Mol. Cell 13: 887–893.
Edwards K.D., Anderson P.E., Hall A., Salathia N.S., Locke J.C., Lynn J.R., Straume M., Smith J.Q., Millar A.J. 2006. FLOWERING LOCUS C mediates natural variation in the high-temperature response of the Arabidopsis circadian clock. Plant Cell 18: 639–650.
Gendall A.R., Levy Y.Y., Wilson A., Dean C. 2001. The VERNALIZATION 2 gene mediates the epigenetic regulation of vernalization in Arabidopsis.. Cell 107: 525–535.
Goodrich J., Puangsomlee P., Martin M., Long D., Meyerowitz E.M., Coupland G. 1997. A Polycomb-group gene regulates homeotic gene expression in Arabidopsis.. Nature 386: 44–51.
Grossniklaus U., Vielle-Calzada J.P., Hoeppner M.A., Gagliano W.B. 1998. Maternal control of embryogenesis by MEDEA, a Polycomb group gene in Arabidopsis.. Science 280: 446–450.
Guitton A.E., Page D.R., Chambrier P., Lionnet C., Faure J.E., Grossniklaus U., 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.
He Y. and Amasino R.M. 2005. Role of chromatin modification in flowering-time control. Trends Plant Sci. 10: 30–35.
Henderson I.R. and Dean C. 2004. Control of Arabidopsis flowering: The chill before the bloom. Development 131: 3829–3838.
Hennig L., Taranto P., Walser M., Sch?nrock N., Gruissem W. 2003. Arabidopsis MSI1 is required for epigenetic maintenance of reproductive development. Development 130: 2555–2565.
Hennig L., Gruissem W., Grossniklaus U., K?hler C. 2004. Transcriptional programs of early reproductive stages in Arabidopsis.. Plant Physiol. 135: 1765–1775.
Hennig L., Bouveret R., Gruissem W. 2005. MSI1-like proteins: An escort service for chromatin assembly and remodeling complexes. Trends Cell Biol. 15: 295–302.
Hepworth S.R., Valverde F., Ravenscroft D., Mouradov A., Coupland G. 2002. Antagonistic regulation of flowering-time gene SOC1 by CONSTANS and FLC via separate promoter motifs. EMBO J. 21: 4327–4337.
Karimi M., Inze D., Depicker A. 2002. GATEWAY vectors for Agrobacterium-mediated plant transformation. Trends Plant Sci. 7: 193–195.
Katz A., Oliva M., Mosquna A., Hakim O., Ohad N. 2004. FIE and CURLY LEAF Polycomb proteins interact in the regulation of homeobox gene expression during sporophyte development. Plant J. 37: 707–719.
Kim H.J., Hyun Y., Park J.Y., Park M.J., Park M.K., Kim M.D., Kim H.J., Lee M.H., Moon J., Lee I.et al. 2004. A genetic link between cold responses and flowering time through FVE in Arabidopsis thaliana. Nat. Genet. 36: 167–171.
Kinoshita T., Harada J.J., Goldberg R.B., Fischer R.L. 2001. Polycomb repression of flowering during early plant development. Proc. Natl. Acad. Sci. 98: 14156–14161.
K?hler C., Hennig L., Bouveret R., Gheyselinck J., Grossniklaus U., 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.
K?hler C., Hennig L., Spillane C., Pien S., Gruissem W., 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.
Koornneef M., Hanhart C.J., van der Veen J.H. 1991. A genetic and physiological analysis of late flowering mutants in Arabidopsis thaliana. Mol. Gen. Genet. 229: 57–66.
Lee H., Suh S.S., Park E., Cho E., Ahn J.H., Kim S.G., Lee J.S., Kwon Y.M., Lee I. 2000. The AGAMOUS-like 20 MADS domain protein integrates floral inductive pathways in Arabidopsis.. Genes & Dev. 14: 2366–2376.
Levy Y.Y., Mesnage S., Mylne J.S., Gendall A.R., Dean C. 2002. Multiple roles of Arabidopsis VRN1 in vernalization and flowering time control. Science 297: 243–246.
Lindroth A.M., Shultis D., Jasencakova Z., Fuchs J., Johnson L., Schubert D., Patnaik D., Pradhan S., Goodrich J., Schubert I.et al. 2004. Dual histone H3 methylation marks at lysines 9 and 27 required for interaction with CHROMOMETHYLASE3. EMBO J. 23: 4286–4296.
McKay J.K., Richards J.H., Mitchell-Olds T. 2003. Genetics of drought adaptation in Arabidopsis thaliana: I. Pleiotropy contributes to genetic correlations among ecological traits. Mol. Ecol. 12: 1137–1151.
Michaels S.D. and Amasino R.M. 1999. FLOWERING LOCUS C encodes a novel MADS domain protein that acts as a repressor of flowering. Plant Cell 11: 949–956.
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.
Michaels S.D., Ditta G., Gustafson-Brown C., Pelaz S., Yanofsky M., Amasino R.M. 2003. Agl24 acts as a promoter of flowering in Arabidopsis and is positively regulated by vernalization. Plant J. 33: 867–874.
Moon J., Suh S.S., Lee H., Choi K.R., Hong C.B., Paek N.C., Kim S.G., Lee I. 2003a. The SOC1 MADS-box gene integrates vernalization and gibberellin signals for flowering in Arabidopsis.. Plant J. 35: 613–623.
Moon Y.H., Chen L., Pan R.L., Chang H.S., Zhu T., Maffeo D.M., Sung Z.R. 2003b. EMF genes maintain vegetative development by repressing the flower program in Arabidopsis.. Plant Cell 15: 681–693.
Moon J., Lee H., Kim M., Lee I. 2005. Analysis of flowering pathway integrators in Arabidopsis.. Plant Cell Physiol. 46: 292–299.
Müller J., Hart C.M., Francis N.J., Vargas M.L., Sengupta A., Wild B., Miller E.L., O’Connor M.B., Kingston R.E., Simon J.A. 2002. Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell 111: 197–208.
Ohad N., Yadegari R., Margossian L., Hannon M., Michaeli D., Harada J.J., Goldberg R.B., Fischer R.L. 1999. Mutations in FIE, a WD Polycomb group gene, allow endosperm development without fertilization. Plant Cell 11: 407–416.
Parcy F. 2005. Flowering: A time for integration. Int. J. Dev. Biol. 49: 585–593.
Ringrose L. and Paro R. 2004. Epigenetic regulation of cellular memory by the Polycomb and Trithorax group proteins. Annu. Rev. Genet. 38: 413–443.
Samach A., Onouchi H., Gold S.E., Ditta G.S., Schwarz-Sommer Z., Yanofsky M.F., Coupland G. 2000. Distinct roles of CONSTANS target genes in reproductive development of Arabidopsis.. Science 288: 1613–1616.
Schubert D., Clarenz O., Goodrich J. 2005. Epigenetic control of plant development by Polycomb-group proteins. Curr. Opin. Plant Biol. 8: 553–561.
Searle I. and Coupland G. 2004. Induction of flowering by seasonal changes in photoperiod. EMBO J. 23: 1217–1222.
Sessions A., Burke E., Presting G., Aux G., McElver J., Patton D., Dietrich B., Ho P., Bacwaden J., Ko C.et al. 2002. A high-throughput Arabidopsis reverse genetics system. Plant Cell 14: 2985–2994.
Sheldon C.C., Rouse D.T., Finnegan E.J., Peacock W.J., Dennis E.S. 2000. The molecular basis of vernalization: The central role of FLOWERING LOCUS C (FLC). Proc. Natl. Acad. Sci. 97: 3753–3758.
Soppe W.J., Jasencakova Z., Houben A., Kakutani T., Meister A., Huang M.S., Jacobsen S.E., Schubert I., Fransz P.F. 2002. DNA methylation controls histone H3 lysine 9 methylation and heterochromatin assembly in Arabidopsis.. EMBO J 21: 6549–6559.
Sung S. and Amasino R.M. 2004. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 427: 159–164.
Yang C.H., Chen L.J., Sung Z.R. 1995. Genetic regulation of shoot development in Arabidopsis: Role of the EMF genes. Dev. Biol. 169: 421–435.
Yoshida N., Yanai Y., Chen L., Kato Y., Hiratsuka J., Miwa T., Sung Z.R., Takahashi S. 2001. EMBRYONIC FLOWER2, a novel Polycomb group protein homolog, mediates shoot development and flowering in Arabidopsis.. Plant Cell 13: 2471–2481.
Yu H., Xu Y., Tan E.L., Kumar P.P. 2002. Agamous-like 24, a dosage-dependent mediator of the flowering signals. Proc. Natl. Acad. Sci. 99: 16336–16341.
Zhang H. and van Nocker S. 2002. The VERNALIZATION INDEPENDENCE 4 gene encodes a novel regulator of FLOWERING LOCUS C. Plant J. 31: 663–673.(Nicole Sch?nrock, Romaric)