Jumonji C Domain Protein JMJ705-Mediated Removal of Histone H3 Lysine 27 Trimethylation Is Involved in Defense-Related Gene Activation in Rice  

Abstract

Histone methylation is an important epigenetic modification in chromatin function, genome activity, and gene regulation. Dimethylated or trimethylated histone H3 lysine 27 (H3K27me2/3) marks silent or repressed genes involved in developmental processes and stress responses in plants. However, the role and the mechanism of the dynamic removal of H3K27me2/3 during gene activation remain unclear. Here, we show that the rice (Oryza sativa) Jumonji C (jmjC) protein gene JMJ705 encodes a histone lysine demethylase that specifically reverses H3K27me2/3. The expression of JMJ705 is induced by stress signals and during pathogen infection. Overexpression of the gene reduces the resting level of H3K27me2/3 resulting in preferential activation of H3K27me3-marked biotic stress-responsive genes and enhances rice resistance to the bacterial blight disease pathogen Xanthomonas oryzae pathovar oryzae. Mutation of the gene reduces plant resistance to the pathogen. Further analysis revealed that JMJ705 is involved in methyl jasmonate–induced dynamic removal of H3K27me3 and gene activation. The results suggest that JMJ705 is a biotic stress-responsive H3K27me2/3 demethylase that may remove H3K27me3 from marked defense-related genes and increase their basal and induced expression during pathogen infection.

INTRODUCTION

Histone modifications such as acetylation and methylation are important epigenomic information for gene regulation and genome activity. Acetylation of histone Lys residues induces permissive chromatin structure for gene activation, whereas histone Lys methylation may have a positive or a negative effect on gene expression, depending on the Lys position and the degree of methylation (Berger, 2007). For instance, dimethylation of histone H3 lysine 9 (H3K9me2) is almost exclusively associated with heterochromatin regions and is required for repression of repetitive genomic sequences in plants, whereas trimethylation of H3 lysine 27 (H3K27me3) is negatively correlated with gene expression. Conversely, trimethylation of H3 lysine 4 (H3K4me3) and H3 lysine 36 (H3K36me3) are associated with active genes (Liu et al., 2010).

Genome-wide analysis in Arabidopsis thaliana and rice (Oryza sativa) indicates that H3K4me3 is preferentially associated with actively transcribed genes (Zhang et al., 2007; Li et al., 2008; Hu et al., 2012), whereas H3K27me3 is found mostly on repressed genes (Turck et al., 2007; He et al., 2010; Hu et al., 2012). How these histone modifications affect gene expression remains unclear. It is suggested that the methylation marks are recognized and bound by specific proteins that may act as effectors to regulate transcription (Shi et al., 2006; Vermeulen et al., 2007). However, these marks can also be a consequence of a gene activation or repression process to mark and possibly to memorize gene expression states (Bonasio et al., 2010; Muramoto et al., 2010). Histone methylation marks are established by evolutionarily conserved SET-domain proteins (named after three Drosophila melanogaster genes: Su[var]3-9, Enhancer of zeste, and Trithorax, which methylate H3K9, H3K27, and H3K4, respectively). During the past few years, plant SET-domain proteins involved in H3K4me3 and H3K27me3 have been shown to play important roles in plant developmental gene expression (Pien and Grossniklaus, 2007; Liu et al., 2010; Berr et al., 2011). Changes of histone methylation patterns also occur on inducible genes under stress conditions, suggesting that histone methylation is dynamic and may play a role in inducible gene expression. For instance, H3K4me3 was found to be increased on responsive genes upon stress treatment (Kim et al., 2008a; van Dijk et al., 2010). Accordingly, Arabidopsis Trithorax1, which is involved in H3K4 trimethylation, was found to be necessary for gene induction by stress signals (Alvarez-Venegas and Avramova, 2005; Alvarez-Venegas et al., 2007; Ding et al., 2011). Conversely, H3K27me3 was removed from several inducible genes upon application of the inductive signals (Kwon et al., 2009; Kim et al., 2010), implying that active removal of this histone methylation mark may be associated with gene induction. However, the role of H3K27me3 demethylation in stress-responsive gene activation and stress tolerance has not been clarified.

Two protein groups are known to be involved in histone Lys demethylation. The Lys- specific demethylase 1 group demethylases are flavin-dependent amine oxidases that reverse monomethylated or dimethylated histone Lys (Shi et al., 2004). The Jumonji C (jmjC) group demethylases remove preferentially dimethylated and trimethylated histone lysines through ferrous ion [Fe(II)] and α-ketoglutaric acid–dependent oxidative reactions (Tsukada et al., 2006). JmjC proteins from yeast (Saccharomyces cerevisiae) and animal cells are classified into seven phylogenetic subgroups, each of which demethylates specific Lys residues (Mosammaparast and Shi, 2010). Plant jmjC proteins are generally conserved with yeast and animal homologs, while there exist a subgroup of more divergent jmjC proteins (Sun and Zhou, 2008). Members of this subgroup (i.e., Arabidopsis JMJ14, JMJ15, and JMJ18 and rice JMJ703) have been reported to be H3K4 demethylases and to regulate diverse aspects of chromatin function and plant development (Deleris et al., 2010; Lu et al., 2010; Searle et al., 2010; Yang et al., 2012a, 2012b; Chen et al., 2013; Cui et al., 2013). Conversely, the JMJD3/UTX (ubiquitously transcribed tetratricopeptide repeat, X chromosome) subgroup proteins, which exhibit H3K27 demethylase activities in mammalian cells, are not found in plants. Recent results showed that a member of the JMJD2 subgroup from Arabidopsis, known as RELATIVE OF EARLY FLOWERING6 (REF6/JMJ12) (Noh et al., 2004), can demethylate H3K27 in Arabidopsis (Lu et al., 2011). However, other JMJD2 members have been shown to demethylate H3K9me2/3 and/or H3K36me3 (Shin and Janknecht, 2007; Sun and Zhou, 2008). These results suggest that plant jmjC proteins may have diversified demethylation specificities to histone lysines.

In this work, we have studied the enzymatic activity and function of gene regulation and plant growth of a rice JMJD2 member, namely JMJ705. Our results demonstrate that JMJ705 is an H3K27 demethylase preferentially involved in biotic stress-responsive gene expression. An increase in JMJ705 expression in transgenic plants removes H3K27me3 from defense-related genes, induces their expression, and enhances plant resistance to the rice bacterial blight disease pathogen Xanthomonas oryzae pathovar oryzae (Xoo). Mutation of the gene reduced the level of resistance. Importantly, we show that the expression of JMJ705 is induced by stress signals including methyl jasmonate (MeJA) and Xoo infection and that JMJ705 is involved in MeJA-induced dynamic removal of H3K27me3 from responsive genes and their activation. The results suggest that JMJ705-mediated H3K27me3 demethylation may contribute to a more sustained activation of defense-related genes by potentiating their basal and induced expression levels during biotic stress.

RESULTS

JMJ705 Is an H3K27 Demethylase

We previously showed that the rice JMJD2 protein JMJ706 is an H3K9 demethylase (Sun and Zhou, 2008). To determine the demethylase activity of other rice JMJD2 proteins, we expressed the FLAG-hemagglutinin (HA)-tagged JMJ705 protein in transfected tobacco (Nicotiana benthamiana) cells (Figure 1A). The demethylated histones were then analyzed by protein gel blots using specific antibodies against histone H3 modification modules. The analysis revealed that the presence of JMJ705 clearly reduced the levels of H3K27me3 and H3K27me2, but not that of H3K27me1, H3K4me3, H3K9me3, or H3K36me3 (Figure 1B), suggesting that the demethylase activity of JMJ705 was specific to H3K27me3 and H3K27me2.

Figure 1.

JMJ705 Is a Histone H3K27me2/3 Demethylase.

(A) Schematic presentation of the vector 35S-JMJ705-FLAG-HA. The relative positions of JmjN, JmjC, and zinc-finger (ZnF) domains are indicated. Arrow indicates the position of the substitution mutation H244A.

(B) In vitro demethylase activity of JMJ705. Bulk histone was incubated with (+) or without (−) tobacco cell-expressed JMJ705-FLAG-HA fusion protein and analyzed by protein gel blots using antibodies against specific histone modification modules indicated on the left. The same blots were analyzed by anti-H3. JMJ705-FLAG-HA was revealed by anti-HA.

(C) In vivo histone demethylase activity of JMJ705. The 35S-JMJ705-FLAG-HA construct was transfected into tobacco leaf cells. Nuclei isolated from leaves were inspected for expression of the fusion protein (stained with anti-HA and indicated by arrows) and then examined for histone methylation levels (stained by DAPI) by using antibodies against specific histone H3 methylation modules indicated on the left. The H244A substitution mutation was tested similarly with anti-H3K27me2. At least 30 nuclei expressing JMJ705 fusion per transfection were observed and imaged. Bar = 25 μm.

(D) Histone H3K27me2/3, H3K4me3 and H3K9me3 methylation levels in wild type (ZH11, HY), JMJ705 overexpression (OX-5) and T-DNA mutant (jmj705) plants revealed by protein gel blots. Only one set of several repeated data is shown. H3 was detected as loading control. Mean signals ± sd (from three replicates) relative to the wild type (set at 1) are indicated below the bands. Significance of differences was determined using t tests. *P < 0.05; **P < 0.01.

To confirm the in vitro enzymatic activity, we transiently expressed JMJ705-FLAG-HA in tobacco leaf cells and tested the demethylase activity in vivo. Transfected cells were immunostained first with anti-HA to identify cells expressing the JMJ705 fusion protein, and then with antibodies against specific methylated histone modules. In nuclei expressing the JMJ705 fusion protein, H3K27me2 was shown to be clearly reduced, but H3K9me2, H3K4me3, H3K36me1, and H3K36me2 did not display any discernible change compared with nontransfected nuclei (Figure 1C). We tested similarly for H3K27me3, but the commercial antibodies that we tested could not produce any signal in the immunostaining assays. The Fe(II) and α-ketoglutaric acid binding residues are highly conserved in jmjC demethylases and are critical for the demethylation activity in plant cells (see Supplemental Figure 1 online; Lu et al., 2011; Chen et al., 2013). To confirm the results, we made a substitution mutation of the key Fe(II) binding residue His244 by Ala, and tested it in the in vivo assays. The substitution eliminated the H3K27me2 demethylase activity of the protein (Figure 1C). The in vitro and in vivo data together indicated that JMJ705 was a histone demethylase specific to H3K27me2/3.

JMJ705 Is Involved in Rice Defense-Related Gene Expression

RT-PCR analysis indicated that JMJ705 was expressed in all tested tissues/organs, with a relatively higher level in rice leaves, and was found to be induced by NaCl, abscisic acid, ethylene (ACC), and JA (see Supplemental Figure 2 online), suggesting that JMJ705 might be involved in stress responses. To study whether JMJ705 expression was responsive to biotic stress, we inoculated two rice varieties (IR24, a susceptible variety, and BB13, a resistant variety) with Xoo (PXO99) and measured JMJ705 transcripts at different time points. In both varieties, JMJ705 mRNA levels increased to 8- to 10-fold at 12 h after inoculation (see Supplemental Figure 2 online). This suggested that JMJ705 might be involved in biotic stress response. To study whether increased expression of JMJ705 could reduce the resting level of H3K27me3 and affect gene expression and plant growth, we produced transgenic rice plants overexpressing JMJ705-FLAG under the strong maize ubiquitin promoter in the Zhonghua11 (ZH11) background (Figure 2A). Protein gel blot analysis indicated that the overexpression of the gene resulted in a reduction of the overall H3K27me2/3 levels in the transgenic plants compared with wild type (Figure 1D). The overexpression transgenic lines displayed a leaf lesion-mimic phenotype at the mature stage (Figures 2A and 2B). The leaf lesion-mimic phenotype is usually a result of activation of biotic stress responses in plant tissues (Lorrain et al., 2003). To test this hypothesis, we analyzed the expression of biotic stress-responsive marker genes in the flag leaves of mature overexpression plants. Genes involved in oxidative stresses such as class III peroxidase genes (POX5, POX8, and POX22), JA biosynthesis and signaling pathway genes (LIPOXYGENASE [LOX], ALLENE OXIDE SYNTHASE2 [AOS2], OXOPHYTODIENOATE REDUCTASE7 (OPR7), and jasmonic acid–inducible MYB (JAMYB)], and pathogenesis-related genes (PR5 and PR10) were selected for analysis. These genes have been used as biotic stress-responsive marker genes in rice (Chittoor et al., 1997; Lee et al., 2001; van Loon et al., 2006). In the overexpression flag leaves, the tested genes showed higher expression levels than in the wild type (Figure 2C), indicating that elevated JMJ705 expression led to de-repression of the biotic stress-responsive genes under normal growth conditions. To test whether the overexpression of JMJ705 altered plant resistance to pathogens, we inoculated wild-type (ZH11) and T1 generation (segregating) plants of line 5 (17 individuals) and line 14 (13 individuals) at the mature stage with the PXO99 strain of Xoo, which causes rice blight disease. Fourteen days after inoculation, lesion areas were surveyed. In wild type and transgenic negative segregates, 35 to 45% of the leaf areas showed necrosis, whereas only 10 to 20% (for line 14) or less than 30% (for line 5) of the leaf areas in the transgenic positive segregates displayed necrosis (Figures 3A and 3B). The observations were further confirmed by inoculation with another Xoo isolate, PXO347 (see Supplemental Figure 3 online). PXO99 growth rates were significantly slower in the overexpression leaves (Figure 3C). These results suggest that elevated expression of JMJ705 induces defense-related gene expression, which may lead to enhanced plant resistance to the bacterial pathogen.

Figure 2.

JMJ705 Overexpression Produces a Leaf Lesion–Mimic Phenotype at the Mature Stage.

(A) Leaf phenotype and JMJ705 transcript levels revealed by RNA gel blot in the wild type and five transgenic lines. rRNAs are shown as loading controls.

(B) Comparison of wild-type (left) and overexpression (right) plants at the mature stage.

(C) mRNA levels of defense-related genes detected by real-time RT-PCR in wild-type (ZH11) and overexpression (OXJMJ705-5 and OXJMJ705-14) flag leaves. Bar indicates mean ± sd from three biological repeats.

[See online article for color version of this figure.]

Figure 3.

JMJ705 Overexpression Enhances Rice Resistance to the Bacterial Pathogen Xoo.

(A) and (B) Three wild-type (ZH11) plants and the T2 segregates of line 5 (OXJMJ705-5) and line 14 (OXJMJ705-14) were inoculated with the Xoo strain PXO99 that causes rice blight disease. Genotyping of the T2 plants for the presence of the transgene is shown for each line. The percentages of the leaf lesion areas were measured 14 d after inoculation. Bar indicates mean ± sd from four to five replicates for the lesion area.

(C) Leaf phenotype after PXO99 inoculation.

(D) Bacterial growth rate (log [COLONY-FORMING UNITS/leaf]) measured at 0 (2 h post inoculation) and 3 to 12 d postinoculation (dpi). Significance of bacterial growth differences between wild-type and overexpression plants was determined by Student's t tests. Bar indicates mean ± sd from five inoculated plants. *P < 0.05; **P < 0.01.

[See online article for color version of this figure.]

To study the effect of loss of function of JMJ705, we characterized a T-DNA insertion mutant line of the gene in Hwayoung (HY) background. Analysis of the genomic sequence confirmed that the T-DNA was inserted in the zinc-finger motif region of the gene (Figures 4A and 4B). A single copy of the T-DNA was detected in the mutant (Figure 4C). RT-PCR analysis indicated that the insertion truncated the full-length transcript (Figure 4D). Protein gel blot analysis did not reveal any clear change of H3K27me2/3 levels compared with the wild type (Figure 1D). It is possible that JMJ705-dependent demethylation might have little effect on the overall level of H3K27me3 in the seedlings under normal growth conditions. The mutant displayed reduced plant height and partial sterility (Figure 4E; see Supplemental Table 1 online). When challenged with the Xoo strain PXO99, the mutant plant leaves showed larger necrotic areas and faster bacterial growth than the wild type (Figures 4F and 4H), suggesting that the mutant might be more susceptible to the pathogen than the wild type.

Figure 4.

Characterization of a T-DNA Insertion Mutation of JMJ705.

(A) Schematic representation of the gene structure and position of the T-DNA insertion (open triangle). The positions of the primers used for genotyping and RT-PCR are indicated.

(B) Genotyping of nine segregates and the wild type (HY) using the two primer sets as indicated.

(C) DNA gel blot analysis of copy number of the T-DNA insertion.

(D) RT-PCR detection of JMJ705 transcripts.

(E) Mature stage and panicle phenotype comparison between the wild type (left) and mutant plant (right).

(F) Leaf lesion area (%) in three wild-type (HY) and nine T-DNA (jmj705) plants 14 d after inoculation with the Xoo strain PXO99. Bar indicates mean ± sd from four to five replicates.

(G) Leaf phenotype.

(H) Bacterial growth rate on jmj705 mutant leaves compared with HY as described in Figure 3.

[See online article for color version of this figure.]

JMJ705 Preferentially Activates H3K27me3-Marked Stress-Responsive Genes

To evaluate genome-wide effects of JMJ705 overexpression on gene expression, we used the Affymetrix microarray platform to analyze transcriptomes of the wild type (ZH11 and HY), homozygous overexpression (line 14 and line 5 combined with the same seedling number/fresh weight), and the T-DNA mutant plants 14 d after germination under normal growth conditions. Two biological repeats were performed. In the overexpression lines, 301 genes showed >2 fold upregulation and 105 genes showed >2 fold downregulation compared with wild type (ZH11) (P < 0.05) (see Supplemental Data Set 1 online). Gene ontology (GO) analysis revealed that upregulated genes were significantly enriched for the categories of stress-responsive genes (89 of 301; P < 0.001) (see Supplemental Table 2 online), while downregulated genes did not show any enrichment. In the T-DNA mutant 640 were upregulated, and 332 were downregulated (>2-fold; P < 0.05), compared with HY (see Supplemental Data Set 2 online). GO analysis revealed that upregulated genes in the mutant were also enriched for stress-responsive genes (see Supplemental Table 2 online). However, very few genes were found to be deregulated in both overexpression and T-DNA plants, implying that the effect of the mutation (which might not affect the overall levels of H3K27me2/3) and that of overexpression of the gene (which likely decreased the overall H3K27me2/3) might affect the expression of different target genes under normal growth conditions.

To study whether the overexpression of JMJ705 preferentially affected the expression of H3K27me3-marked genes, we compared the microarray data with genome-wide H3K27me3 chromatin immunoprecipitation sequencing (ChIP-seq) data that were obtained from rice seedlings grown under the same conditions (Hu et al., 2012). Because H3K4me3 and H3K27me3 are antagonist histone methylation markers for gene activity, we also analyzed H3K4me3 ChIP-seq data obtained at the same growth conditions (Hu et al., 2012). The analysis revealed that upregulated genes in the overexpression plants were significantly enriched for H3K27me3 (P = 3.596e-10), because 103 of the 301 (38.9%) upregulated genes in JMJ705 overexpression plants were marked by H3K27me3 compared with 10,831 of 56,797 genes genome-wide (19%) (see Supplemental Table 3 online). For downregulated genes, only 16 of 105 (12.5%) were marked by H3K27me3 (see Supplemental Table 3 online). Conversely, downregulated, but not upregulated, genes were found to be enriched for H3K4me3 (P = 2.2e-16) (see Supplemental Table 3 online).

Analysis of ChIP-seq read intensity (i.e., number of sequenced tags per each 5% of the genic region or per 100-bp intervals in the 2 kb upstream and downstream regions) over the deregulated genes revealed that H3K27me3 intensity was much higher in upregulated genes than the genome-wide average, whereas that of downregulated genes was close to the average level (Figure 5). The analysis indicated that the upregulated genes displayed a lower than average level of H3K4me3 in wild-type plants, whereas the downregulated genes showed a higher than average level of H3K4me3 (Figure 5). The analysis suggested that increased JMJ705 expression preferentially activated silent or underexpressed genes that were marked by a relatively high level of H3K27me3 and a relatively low level of H3K4me3.

Figure 5.

H3K27me3 and H3K4me3 ChIP-seq Intensities of Genes that Are Upregulated or Downregulated in JMJ705 Overexpression Plants Compared with the Genome-Wide Levels.

Numbers of sequenced tags (y-axis) per each 5% of the genic region (black box) or per 100-bp intervals in the 2 kb upstream and 2 kb downstream regions (line, x-axis) are shown. Arrow indicates the direction of transcription.

To validate the microarray data, we randomly selected 27 upregulated genes in the overexpression plant for real-time RT-PCR analysis (see Supplemental Table 4 online). The tested genes were clearly upregulated in the overexpression lines compared with wild type (see Supplemental Figure 4 and Supplemental Table 4 online). In addition, many of the 27 genes were found to be highly induced by MeJA (see Supplemental Figure 4 and Supplemental Table 4 online), which suggested that JA-inducible genes might be among the preferential targets of JMJ705. To study whether the upregulation of these genes was correlated to H3K27 demethylation, we performed ChIP assays to analyze H3K27me3, H3K27me2, H3K4me3, H3K9me3, and H3K36me3 levels near the transcriptional start site of 12 of the 27 genes in wild-type and overexpression plants. In the overexpression plants, H3K27me3 and H3K27me2 levels were clearly decreased in most of the tested genes, while H3K4me3, H3K9me3, and H3K36me3 levels were not clearly changed on most of the tested genes (Figure 6). This analysis suggested that upregulation of most the tested genes in the overexpression plants was correlated to demethylation of H3K27me2/3, suggesting that the removal of H3K27me2/3 might be related to the activation of the associated genes.

Figure 6.

H3K27me2/3 Removal Is Related to JMJ705-Mediated Gene Upregulation.

Chromatin fragments isolated from wild-type (ZH11) and overexpression plants (OXJMJ705) were immunoprecipitated with antibodies against H3K4me3, H3K9me3, H3K36me3, H3K27me2, and H3K27me3 as indicated, and analyzed by real-time PCR using primer sets corresponding to transcriptional start site regions of 12 upregulated genes (numbered as in Supplemental Table 4 online). Chromatin fragments of JMJ705-FLAG overexpression plants were immunoprecipitated by anti-FLAG and analyzed by PCR using the same primer sets (right bottom).

Elevated JMJ705 Enhances MeJA-Induced Gene Expression

To study whether JMJ705 plays a role in the process of JA-induced gene activation, 14-d-old seedlings of JMJ705 overexpression and wild-type (ZH11) plants were treated by MeJA (0.2 mM) for 0, 2, 4, 12, and 24 h. Transcript levels of two JA-induced genes TERPENE SYNTHASE3 (TPS3) and Os07g11739 (see Supplemental Table 4 online), and the JA-inducible marker genes PR10 and JAMYB, were determined by quantitative RT-PCR (Figure 7). Before induction, the mRNA levels of TPS3 and Os07g11739 were ∼5 to 20 times higher in the overexpression lines than in wild type, confirming the microarray data (see Supplemental Table 4 online). By contrast, the mRNA levels of JAMYB in the overexpression seedlings were not different from that in wild type seedlings, suggesting that increased JMJ705 expression had different effects on the basal expression of the two types of genes. However, increases of JAMYB mRNA were detected in the overexpression plants at the mature stage (flag leaves, Figure 2C). A possible explanation is that the increases might be due to a general activation of defense genes during necrotic phenotype production at the mature stage. The highest induction of TPS3, Os07g11739, and PR10 by MeJA occurred at 12 h after treatment, whereas that of JAMYB seemed to occur earlier. JMJ705 overexpression had no clear effect on the induction of TPS3 and Os07g11739 during the first (2 to 4) hours, but highly increased the induction levels at 12 h after treatment compared with wild type. In a lesser extent, a similar effect was observed for the induction of PR10. The induction of JAMYB was also enhanced in the overexpression plants. To confirm the results, we obtained JMJ705 RNA interference (RNAi) plants in the ZH11 background (see Supplemental Figure 5 online). The MeJA induction of the four genes was reduced in the RNAi plants compared with the wild type (Figure 7). These data indicate that JMJ705 was involved in the induction process of JA-responsive genes.

Figure 7.

JMJ705 Enhances JA Induction of Gene Expression.

mRNA levels of JA-responsive genes JAMYB, PR10, TPS3, and Os07g11739 (genes 5 and 7 in Supplemental Table 4 online) in wild type, two lines of overexpression and one line of RNAi plants (all in ZH11 background) treated by MeJA (0.2 mM) for 0 to 24 h were analyzed by quantitative RT-PCR. Relative mRNA levels are presented with wild type at 0 h set as 1. Because of great variation between induction experimental repeats, data from three biological repeats are presented individually. Insets show relative mRNA levels during the first hours of induction for TPS3 and 07g11739. Bar indicates mean ± sd from three technical repeats.

JMJ705 Is Involved in MeJA-Induced Removal of H3K27me3 from Responsive Genes

To study whether JMJ705-mediated demethylation of H3K27me3 was implicated in JA induction of gene activation, we measured H3K27me3 levels by ChIP on JA-responsive genes in wild type, JMJ705 overexpression, and RNAi 14-d-old seedlings treated with JA for 0, 8, and 12 h. The analysis revealed the following observations (Figure 8). First, before MeJA treatment, the level of H3K27me3 on TPS3 and Os07g11739 was ∼8 to 16 times higher than that on PR10 or JAMYB in the wild type, suggesting that TPS3 and Os07g11739 might be more subjected to regulation by H3K27me3. Second, in JMJ705 overexpression plants, H3K27me3 levels were clearly reduced from OsSTP3 and Os07g11739 (confirming the data in Figure 6), but not from JAMYB or PR10 before treatment. These data were consistent with the increased basal expression of OsSTP3 and Os07g11739 in the overexpression seedlings (Figure 7; see Supplemental Table 4 online), suggesting that the removal of H3K27me3 was associated with the increase of basal expression of genes. Finally, MeJA treatment reduced H3K27me3 levels from the four genes in wild type plants, suggesting that the removal of H3K27me3 was associated with MeJA-induced gene activation. The reduction of H3K27me3 was more pronounced in the overexpression plants but was attenuated in the RNAi plants at 8 or 12 h. These observations suggest that JMJ705 was implicated in MeJA-induced removal of H3K27me3.

Figure 8.

JA-Induced H3K27me3 Removal from Responsive Genes Is Dependent on JMJ705.

Anti-H3K27me3 ChIP assays were performed to analyze transcriptional start site of JA-responsive genes PR10, JAMYB, TPS3, and Os07g11739 in wild-type (ZH11) plants, overexpression (OXJMJ705), and JA-treated RNAi plants during 0, 8, and 12 h. Bar indicates mean ± sd from three biological repeats.

DISCUSSION

JMJ705 Is a Stress-Responsive H3K27me2/3 Demethylase

In this work, we have provided in vitro and in vivo evidence that the rice JMJD2 protein JMJ705 specifically demethylates histone H3K27me2/3 (Figures 1B and 1C). The results are corroborated by the observations that overexpression of the gene reduced the overall levels of H3K27me2/3 and resulted in preferential activation of H3K27me3-marked genes (Figures 1D and 5A). The stress responsiveness of JMJ705 expression and the enrichment of H3K27me3-marked stress-related genes among the upregulated genes in JMJ705 overexpression plants (Figure 5A; see Supplemental Figure 2 and Supplemental Table 2 online) suggest that JMJ705 is a histone demethylase involved in stress-responsive H3K27me3 demethylation and gene expression.

Previous results showed that a closely related Arabidopsis JMJD2 protein, REF6 (JMJ12), demethylates H3K27me2/3 (Lu et al., 2011). Mutation of REF6 delays flowering time in Arabidopsis (Noh et al., 2004), whereas its overexpression produces similar phenotypes to mutants of genes involved in H3K27me3 deposition and recognition (Noh et al., 2004; Lu et al., 2011). Although overexpression of JMJ705 did not produce any severe developmental abnormality, several important developmental regulatory genes such as OsMADS1, homeobox gene OSH71, and homeobox gene Rough Sheath1 showed clear decreases of H3K27me3 and were highly induced (see Supplemental Table 4 online). These observations, together with the reduced plant height and partial sterility phenotypes of the T-DNA mutant, suggest that JMJ705 may also be involved in H3K27me3 removal from genes involved in developmental processes.

JMJ705-Mediated H3K27me3 Removal in Stress-Inducible Gene Activation

H3K27 methylation is an important epigenetic mark involved in gene regulation in plants. Genome-wide profiling has identified ∼10 to 20% of genes that are marked by H3K27me3 in Arabidopsis and rice, depending on the analyzed plant organs or developmental stages (Turck et al., 2007; He et al., 2010; Hu et al., 2012). Unlike in animal cells, in which H3K27me3 occupies large genomic regions, in plants, H3K27me3 is mostly restricted to the transcribed region of a large number of genes with a strong bias toward the 5′ transcribed regions (Figure 5) (Roudier et al., 2011), implying that this mark may be involved in the repression of transcriptional initiation. H3K27me3 is deposited by the polycomb group–repressive complex 2 (PRC2). Resetting of this chromatin modification mark may be a necessary step for activation of the PRC2-repressed genes in plants. Because H3K27me3 is associated with many transcription factor genes that are not expressed in the examined organs, H3K27me3-mediated gene silencing is suggested to be mainly involved in developmental decision in plants (Zheng and Chen, 2011). It has been shown that reduction of H3K27me3 from the developmentally regulated gene KNUCKLES (KNU) precedes its activation, which is dependent on the transcription factor AGAMOUS (AG) (Sun et al., 2009), suggesting that the removal of H3K27me3 may be required for developmental gene activation.

However, many other genes, including stress-inducible genes, are also marked by this modification. H3K27me3 may play a role in maintaining stress-responsive genes at silent or basal expression levels under normal conditions and should be reset upon receiving inductive signals. Reducing the resting levels of H3K27me3 from marked genes may change the chromatin environment facilitating gene transcriptional activation. The data showing that overexpression of JMJ705 activates mostly H3K27me3-marked and H3K4me3-depleted stress-responsive genes suggest that JMJ705 at elevated levels removed H3K27me3 from these genes and increased their basal expression levels. The data showing that the dynamic removal of H3K27me3 from JA-responsive genes during MeJA induction was correlated with their activation and was enhanced by overexpression and impaired by RNAi of JMJ705 (Figures 7 and 8) indicate that JMJ705-mediated H3K27me3 demethylation is also involved in the defense signal-induced gene activation process. Therefore, JMJ705-mediated H3K27me3 demethylation may play a role in increasing both the basal and induced expression of defense-related genes, leading to a more sustained activation, which may possibly contribute to the eventual development of the necrotic phenotype and resistance to Xoo at the mature stage (Figures 2 and 3) observed in the overexpression plants. Because JMJ705 demethylates H3K27me2, the possibility that the phenotype is related to a reduction of H3K27me2 instead of H3K27me3 is, however, not excluded.

Histone Modification in Plant Disease Resistance

Chromatin modification and remodeling have emerged as a key process in the gene expression reprogramming of plant biotic stress responses. Histone acetylation and H3K4 and H3K36 methylation dynamics has been implicated in salicylic acid (SA)-signaling and JA-signaling pathway–mediated plant defense in Arabidopsis (Tian et al., 2005; Zhou et al., 2005; Mosher et al., 2006; Kim et al., 2008b; Wu et al., 2008; Berr et al., 2010; Jaskiewicz et al., 2011). H3K4me3 is present on the PR1 chromatin before any stimulation and its level increases upon SA treatment (Mosher et al., 2006). H3K4me3 and H3K36me3 on the downstream genes seem to be readily in place as a permissive mark providing the basal expression level of the marked genes or establishing the poised chromatin state for efficient induction when stimulated (Berr et al., 2012). In addition, histone of H3K9 and H3K14 acetylation and H3K4me3 are synergistically set during the priming of disease defense-related genes (Jaskiewicz et al., 2011). The present data provide evidence that H3K27me3 homeostasis also plays an important role for defense-related gene expression and support the hypothesis that H3K27me3 may be involved in maintaining the resting state of defense genes under normal conditions.

Histone modification genes have been reported to regulate plant defense pathways. Arabidopsis SET-domain gene8, a histone H3K36 methyltransferase, mediates induction of the JA/ethylene pathway genes in the plant defense response to necrotic fungi (Berr et al., 2010). Arabidopsis histone deacetylase HDA19 plays a positive role in basal resistance to pathogens (Zhou et al., 2005). HDA9 physically interacts with and represses the activity of two WRKY family transcription factors, WRKY38 and WRYK62, both of which act as negative regulators in plant basal defense (Kim et al., 2008b). A recent report showed that HDA19 was involved in the repression of SA-mediated defense responses in Arabidopsis (Choi et al., 2012). Recent studies in rice showed that overexpression of HDT701 (OsHDT1), a histone H4 deacetylase gene, in transgenic rice increases susceptibility to the rice pathogens Magnaporthe oryzae and Xoo. By contrast, knockdown of HDT701 in transgenic rice causes elevated levels of histone H4 acetylation and elevated transcription of defense-related genes and enhanced resistance to both M. oryzae and Xoo (Ding et al., 2012). Thus, HDT701-dependent histone H4 deacetylation plays a negative role in rice immunity to pathogen attack. The enhanced resistance conferred by overexpression of JMJ705 and the increased susceptibility in the T-DNA mutant to the bacterial pathogen Xoo suggest that JMJ705 may play a positive role in rice immunity. JMJ705 may enhance plant defense by reducing basal H3K27me3 levels over defense-related genes, thereby augmenting their basal expression and/or potentiating their higher expression upon biotic stress. Because histone acetylation and methylation play an important role in plant defense pathways, it remains unknown whether there is any functional interaction between JMJ705 and HDT701 in the interplay of histone modifications involved in rice disease resistance.

METHODS

Plant Materials

Rice cultivars (Oryza sativa spp japonica) ZH11 and HY were used in this study. The T-DNA insertion line of JMJ705 (1C-05110.L) was obtained from the Postech rice mutant database (http://www.postech.ac.kr/life/pfg/risd/). Insertion was confirmed by PCR using the primers F1 and R1 and a T-DNA specific primer 2707L1 (see Supplemental Table 5 online). The primers used for RT-PCR analysis presented in Figure 4 are listed in Supplemental Table 5 online.

Overexpression and RNAi of JMJ705 and Sequence Alignment

For the overexpression experiment, the JMJ705 full-length cDNA without a stop codon was amplified from total cDNA of ZH11 using primers OXJ5-F and OXJ5-R, and then cloned into the overexpression vector pU1301 in frame with 3 × FLAG tag at the 3′ end with KpnI and BamHI sites (Sun and Zhou, 2008). For RNAi, cDNA from 2584 bp to 3126 bp relative to the translation start site was amplified from JMJ705 cDNA using primers RiJ5-F and RiJ5-R, and then cloned into the RNAi vector pds1301 (Huang et al., 2007). The overexpression and RNAi constructs were transferred into Agrobacterium tumefaciens strain EHA105 and then transformed into ZH11 as previously described (Huang et al., 2007). For sequence alignment, we retrieved JMJ protein sequences from the ChromDB database (http://www.chromdb.org) and used ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2) for alignment.

Histone Demethylation Assays and Western Blot Analyses

For the in vivo histone demethylation assay, the JMJ705 full-length cDNA used in the overexpression event was cloned into pFA121 vector (Chen et al., 2013). For production of the substitution mutation of JMJ705, we used the Fast Mutagenesis System (Trans Gen Biotech FM111) with primers J5mut-F and J5mut-R (see Supplemental Table 5 online) according to the instructions. The JMJ705 mutant cDNA was then cloned into the vector pFA121. Tobacco (N. icotiana benthamiana) transfection, nuclei isolation, and immunostaining were described previously (Chen et al., 2013). Slides were incubated with anti-HA at 1:200 dilution (mouse, M20003M; Ab-mart) and antimethylated histone antibodies (rabbit) at 1:150 dilution, followed by Alexa Fluor 488 goat anti-mouse (green fluorescence) (A-11029; Invitrogen) and Alexa Fluor 568 goat anti-rabbit (red fluorescence) (A-11036; Invitrogen) secondary antibodies at 1:200 dilution. For presentation, native colors of fluorescence-labeled secondary antibodies were changed using confocal software, so that JMJ705-FLAG-HA protein was displayed in red and methylated histone in green. JMJ705-FLAG-HA protein was affinity purified from pFA121-JMJ705 transfected tobacco leaves with anti-FLAG M2 magnetic beads (Sigma) incubated with bulk histones (Sigma) for the in vitro histone demethylation assay according to previously described methods (Whetstine et al., 2006). For protein gel blot analysis, histone-enriched fractions were extracted from wild-type, mutant, and transgenic rice leaves as described previously (Huang et al., 2007).

Other antibodies used in this study were as follows: anti-H3K27me3 (07-449; Millipore), anti-H3K27me2 (ab24684; Abcam), anti-H3K27me1 (ab113671,;Abcam), anti-H3 (ab1791; Abcam), anti-H3K4me3 (07-473; Millipore), anti-H3K9me3 (ab8898; Abcam), anti-H3K9me2 (07-441; Millipore), H3K36me1 (ab9048; Abcam), anti-H3K36me2 (ab9049; Abcam), anti-H3K36me3 (ab9050; Abcam), and anti-FLAG (F3165; Sigma).

ChIP

The ChIP experiment was performed as described (Huang et al., 2007). Two grams of 14-d-old seedlings were harvested and crosslinked in 1% formaldehyde under vacuum. Chromatin was extracted and fragmented to 200 to 750 bp by sonication, and ChIP was performed using the following antibodies: H3K4me3, H3K9me3, H3K36me3, H3K27me2, H3K27me3, and FLAG (F3165; Sigma). The precipitated and input DNAs were then analyzed by real-time PCR with gene-specific primers listed in Supplemental Table 5 online.

Pathogen Inoculation

To examine the resistance of plants to bacterial blight disease, plants were inoculated with the Philippine Xoo strains PXO99 (race 6) and PXO347 (race 9c) at the booting stage by the leaf clipping method (Chen et al., 2002). Disease was scored (3 to 5 leaves for each plant) as the percent lesion area (lesion length/leaf length) at 2 weeks after inoculation.

For Xoo bacterial growth rate analysis, rice plants at the mature stage were inoculated by the leaf clipping method, and the bacterial population in the infected leaves was determined by counting colony-forming units. Three leaves were harvested separately from inoculated plants at 0 (2 h after inoculation), 3, 6, 9, and 12 d postinoculation. After sterilization with 75% alcohol for 1 min, samples were air-dried and ground into homogenates in 1 mL sterilized distilled water. Homogenates were serially diluted (10-fold each time) to 10⁷ with sterilized distilled water, and each dilution was then spread on potato culture medium [boiled fresh potato 30% (w/v), Suc 1.5% (w/v), Ca(NO₃)₂ 0.05% (w/v), Na₂HPO₄·12H₂O 2% (w/v), tryptone 5% (w/v), agar 2% (w/v)]. Colonies were counted after 2 to 3 d in the dark at 25°C. Average colony numbers from three samples were calculated.

Treatment with MeJA

For MeJA treatment, 14-d-old seedlings of wild-type, mutant, and transgenic rice were treated with 0.2 mM MeJA. MeJA solution was applied to leaves by spraying. Samples were collected 2, 4, 8, 12, and 24 h after treatment.

Microarray and ChIP-seq Analysis

For microarray analysis, 14-d-old seedlings of transgenic, mutant, and wild type plants, grown in one-half-strength Murashige and Skoog medium under a 16-h-light/8-h-dark cycle at 30°C for 14 d, were harvested for microarray analysis. RNA samples were extracted using TRIzol according to the manufacturer's instructions (Invitrogen). Hybridization with Affymetrix GeneChip Rice Genome Arrays was performed at CapitalBio Corporation. Two biological repeats were performed. Gene expression changes between the samples were analyzed by the AffylmGUI package from R software. For GO analysis of microarrays data, we used singular enrichment analysis (http://bioinfo.cau.edu.cn/agriGO/analysis.php), and choose P < 0.001 as the cutoff for significant GO terms.

Expression Analysis by RNA Gel Blot and RT-PCR

For RNA gel blot analysis, 15 μg total RNA samples extracted from field-grown rice leaves was separated in 1.2% (w/v) formamide-denaturing agarose gels, and then transferred to nylon membranes. Gene-specific probes were labeled with ³²P-dCTP using the Random Primer kit (Invitrogen) and hybridized to the RNA gel blots. The probe of JMJ705 was amplified from JMJ705 cDNA using primers F2 and R2 (see Supplemental Table 5 online), resulting in a fragment of 433 bp of the cDNA.

For gene expression analysis, total RNA isolation and RT-PCR were performed. Real-time PCR analysis was performed using gene-specific primers and SYBR Premix Ex Taq on a real-time PCR 7500 system (Applied Biosystems). At least three biological replicates and three technical repeats for every biological replicate were tested.

Accession Numbers

Sequence data from this article can be found in the Rice Genome Annotation Project website (http://rice.plantbiology.msu.edu/) under the following accession numbers: JMJ705 (Os01g67970), POX5 (Os07g48040), POX8 (Os07g48010), POX22 (Os07g48020), LOX (Os08g39840), AOS2 (Os03g12500), OPR7 (Os08g35740), JAMYB (Os11g45740), PR5 (Os12g43430), PR10 (Os12g36860), and TPS3 (Os07g11790). The microarray data described in this article have been deposited into the GEO database (GSE48069). The ChIP-seq data were retrieved from the GEO database (GSE30490).

Supplemental Data

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

• Supplemental Figure 1. Mutated Iron Binding Residue within the Catalytic Domain of JMJ705.

• Supplemental Figure 2. Quantitative RT-PCR Analysis of JMJ705 Transcripts in Different Tissues of ZH11 Plants and under Stress Conditions.

• Supplemental Figure 3. Inoculation of Overexpression Plants with the Xoo Strain PXO347.

• Supplemental Figure 4. Quantitative RT-PCR Verification of 27 Randomly Selected Genes from among Upregulated Genes in the Microarray of the Overexpression Plants and their Induction by MeJA.

• Supplemental Figure 5. Production of JMJ705 RNAi Plants.

• Supplemental Table 1. Phenotype Comparisons between jmj705 Mutant and the Wild-Type HY.

• Supplemental Table 2. GO Analysis of Upregulated and Downregulated Genes in JMJ705 Overexpression Plants and T-DNA Mutant Detected with Affymetrix Microarray Analysis.

• Supplemental Table 3. Number of Upregulated and Downregulated Genes and Their Enrichment for H3K27me3 and H3K4me3 in JMJ705 Overexpression and T-DNA Plants.

• Supplemental Table 4. Validation by Quantitative RT-PCR of 27 Randomly Selected Upregulated Genes from JMJ705 Overexpression Plants and Induction Tests of Their Expression by MeJA.

• Supplemental Table 5. Nucleotide Sequences of Primers Used in This Study.

• Supplemental Data Set 1. Upregulated and Downregulated Genes in JMJ705 Overexpression Plants.

• Supplemental Data Set 2. Upregulated and Downregulated Genes in the T-DNA Mutant.

Acknowledgments

We thank Shiping Wang's group for advice on Xoo inoculation and rice disease evaluation, as well as Caiguo Xu and Xianghua Li for help in field experiments and management. This work was supported by the Rice Functional Genomics 863 Key Project of the Chinese Ministry of Science and Technology (2012AA10A303), the special transgenic program of the Chinese Ministry of Agriculture (2014ZX0800902B), the Fundamental Research Funds for the Central Universities (2011PY051), and grants from the Bill and Melinda Gates Foundation.

AUTHOR CONTRIBUTIONS

T.L., X.C., X.L., Y.Z., S.Z., S.C. performed research; T.L., X.Z., D.X.Z. analyzed data; DXZ, T.L. wrote the article.

Author notes

1

Address correspondence to .

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

Some figures in this article are displayed in color online but in black and white in the print edition.

Online version contains Web-only data.

www.plantcell.org/cgi/doi/10.1105/tpc.113.118802