Rice cytosine DNA methyltransferases – gene expression profiling during reproductive development and abiotic stress Rita Sharma 1, *, R. K. Mohan Singh 1, *, Garima Malik 2 , Priyanka Deveshwar 1 , Akhilesh K. Tyagi 1 , Sanjay Kapoor 1 and Meenu Kapoor 2 1 Interdisciplinary Centre for Plant Genomics and Department of Plant Molecular Biology, University of Delhi South Campus, New Delhi, India 2 University School of Biotechnology, Guru Gobind Singh Indraprastha University, Kashmere Gate, Delhi, India Introduction DNA methylation is an epigenetic modification that regulates key developmental processes, e.g. X chromo- some inactivation, genomic imprinting, gene expression and protection of genomes from invading transposons, retrotransposons and viruses [1]. The methylation of cytosine residues is catalysed by a class of proteins Keywords DNA methyltransferase; epigenetics; methylation; microarray; rice (Oryza sativa) Correspondence S. Kapoor, Interdisciplinary Centre for Plant Genomics and Department of Plant Molecular Biology, University of Delhi South Campus, Benito Juarez Road, New Delhi 110021, India Fax: +91 11 24115095 Tel: +91 9811644436 E-mail: kapoors@south.du.ac.in M. Kapoor, University School of Biotechnology, Guru Gobind Singh Indraprastha University, Kashmere Gate, Delhi 110006, India Fax: +91 11 23900111 Tel: +91 11 23900237 E-mail: kapoorsk@genomeindia.org *These authors contributed equally to this work (Received 21 June 2009, revised 21 August 2009, accepted 1 September 2009) doi:10.1111/j.1742-4658.2009.07338.x DNA methylation affects important developmental processes in both plants and animals. The process of methylation of cytosines at C-5 is catalysed by DNA methyltransferases (MTases), which are highly conserved, both struc- turally and functionally, in eukaryotes. In this study, we identified and characterized cytosine DNA MTase genes that are activated with the onset of reproductive development in rice. The rice genome (Oryza sativa L. subsp. japonica) encodes a total of 10 genes that contain the highly con- served MTase catalytic domain. These genes have been categorized into subfamilies on the basis of phylogenetic relationships. A microarray-based gene expression profile of all 10 MTases during 22 stages ⁄ tissues that included 14 stages of reproductive development and five vegetative tissues together with three stresses, cold, salt and dehydration stress, revealed spe- cific windows of MTase activity during panicle and seed development. The expression of six methylases was specifically ⁄ preferentially upregulated with the initiation of floral organs. Significantly, one of the MTases was also activated in young seedlings in response to cold and salt stress. The molecular studies presented here suggest a greater role for these proteins and the epigenetic process in affecting genome activity during reproductive development and stress than was previously anticipated. Abbreviations BAH, bromo-adjacent homology; CMT, chromomethyltransferase; Dnmt2, DNA methyltransferase 2; DRM, domains rearranged methyltransferase; HMM, hidden Markov model; GCRMA, GC robust multi-array average; MTase, methyltransferase; MET1, DNA methyltransferase 1; NCBI, National Center for Biotechnology Information; SAM, shoot apical meristem; TIGR, The Institute of Genomic Research; QPCR, quantitative PCR. FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS 6301 known as DNA methyltransferases (MTases). These proteins in prokaryotes and eukaryotes possess a catalytic domain with conserved motifs that are arranged in a specific order, thus reflecting their origin from a common ancestor. In eukaryotes, this domain is associated with N-terminal extensions having a vari- able number of domains conserved across different members of MTase subclasses [2]. This structural arrangement facilitates the interaction of MTases with a variety of cellular proteins [3]. The diverse types of MTase that exist in eukaryotes are known to be involved in two basic kinds of methylation activities: maintenance methylation and de novo methylation. When a methyl group is added to cytosines of a hemimethylated DNA after DNA replication, the pro- cess is referred to as maintenance methylation. Methyl- ation of cytosines in nonmethylated DNA is referred to as de novo methylation. This process is responsible for establishing new methyl patterns that are then maintained by maintenance MTases. In plants, four main subfamilies of MTase have been identified: domains rearranged methyltransferase (DRM), DNA methyltransferase 1 (MET1), DNA methyltransferase 2 (Dnmt2) and chromomethyltrans- ferase (CMT) [4–7]. DRMs are similar to human Dnmt3, which is required for establishing methylation patterns during development. In Arabidopsis, three DRM genes are known, AtDRM1, AtDRM2 and At- DRM3. AtDRM1 and AtDRM2 are required for de novo methylation of cytosines in all sequence con- texts, CpG, CpNpG and at asymmetric positions [8]. The MET1 class encodes MTases resembling the ani- mal Dnmt1. In Arabidopsis, the MET1 proteins are encoded by a small multigene family and these pro- teins function predominantly as maintenance methylas- es, methylating CpG residues [4]. MET1-like genes have been identified in a variety of plant species, such as pea, maize, tobacco and Brassica [9–12]. Dnmt2 has been reported in yeast, Drosophila, animals and plants, suggesting an ancestral origin and its involvement in essential cellular functions. Unlike other MTases, Dnmt2s have broader substrate specificity. The human Dnmt2 also acts as an RNA MTase and has been shown to specifically methylate cytosine 38 in the anti- codon loop of tRNA(Asp) [13]. CMTs are plant-spe- cific MTases that are characterized by the presence of a chromo (chromatin organization modifier) domain and a bromo-adjacent homology (BAH) domain in the N-terminal region. In Arabidopsis, three CMTs have been reported and AtCMT3 is known to control meth- ylation at CNG and CNN sites in Arabidopsis [3,5]. There is significant variability in the types and numbers of MTases utilized by various organisms. Although a single methylase has been identified in many protozoa, insects and fungi, more than 10 genes are known in higher plants [2]. The availability of complete genome sequences of numerous eukaryotes provides an unprecedented opportunity to explore the diversity among this class of protein in various organ- isms and to study their roles in regulating critical developmental processes at the whole genome level. In the present study, an attempt was made to gain insight into the expression pattern of rice DNA MTases during panicle and seed development. Utilizing an in-house generated microarray-based gene expression dataset, expression profiles of all rice MTases were compiled from stages ⁄ tissues of vegetative, panicle and seed development, together with seedlings subjected to desiccation, cold and salt stress conditions. Wherever possible, expression profiles of rice genes were compared with expression patterns of Arabidopsis MTases avail- able in public databases. Results Identification of rice cytosine-specific DNA MTases Putative DNA MTases were searched in the Rice Genome Annotation Project (http://rice.plantbiology. msu.edu), formerly at The Institute of Genomic Research (TIGR) using the keyword ‘cytosine methyl- transferase’ and the hidden Markov model (HMM) corresponding to the DNA MTase domain as described in Pfam (http://www.pfam.sanger.ac.uk/). The domains were searched using hmmer, version 2.3.2 (http:// hmmer.janelia.org/). All proteins with an E-value above e )10 were considered as putative MTases. The values for putative DRMs were lower than the other MTases due to rearrangement of conserved motifs in their catalytic domains (Fig. S1). These MTases were further verified by other search tools (see below). In total, 10 nonredun- dant loci were identified from HMM analysis (Table 1). Eight of these loci have previously been reported [2]. We identified two additional loci corresponding to putative CMTs and de novo MTase. The length of the open read- ing frames encoded by all putative MTases varied from 1152 to 4584 bp and the derived proteins ranged from 383 to 1527 amino acids (Table 1). The presence of conserved domains was also verified using the Simple Modular Architecture Research Tool (http://smar- t.embl-heidelberg.de/) and the National Center for Bio- technology Information (NCBI) conserved domain database (http://www.ncbi.nlm.nih.gov/Structure/cdd/ cdd.shtml). All 10 MTases possessed the MTase domain with conserved motifs at the carboxyl terminus. Most of Rice cytosine DNA methyltransferases R. Sharma et al. 6302 FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS Table 1. Structural characteristics of DNA MTase genes identified in rice. Serial no. Gene name Chromatin database identification Accession numbers of gene models Locus identification UniProtKB accession number Coordinates (5¢–3¢) ORF l ength (bp) Protein No. of introns KOME clone accession number Location on pseudomolecules Rice Genome Annotation Project (Release 5) Length (amino acids) pI Molecular mass (Da) 1 OsCMT1 DMT701 12003.m06707 LOC_Os03g12570 Q8H854 6636960–6647579 3102 1033 5.23 116072.1183 20 AK242575 3 2 OsCMT3 DMT704 12010.m03581 LOC_Os10g01570 Q8SBB4 366495–361696 2724 907 5.32 101404.1216 19 AK112062 10 3 OsCMT2 DMT703 12005.m05793 LOC_Os05g13790 B9FJ44 7627439–7625715 3930 1309 6.04 143748.8881 19 AK109728 5 4 OsDNMT2 DMT705 12001.m10527, 12001.m150650, 12001.m150651 LOC_Os01g42630 B9EY12 24576129–24573991 1152 383 6.65 43290.1533 10 AK111502 1 5 OsMET1-1 DMT702 12003.m34704 LOC_Os03g58400 Q10C15 33218057–33219960 4584 1527 5.75 170855.1565 15 AK108034 3 6 OsMET1-2 DMT707 12007.m05306 LOC_Os07g08500 Q7PC77 4388309–4389958 4458 1486 14 AK111461 7 7 OsDRM3 DMT710 12005.m04966, 12005.m64178 LOC_Os05g04330 Q6AUQ7 1945379–1951545 2043 680 5.47 76128.4635 10 AK063482 5 8 OsDRM2 DMT706 12003.m05744, 12003.m34976, 12003.m34977, 12003.m34978 LOC_Os03g02010 Q10SU5 611179–614559 1794 597 5.14 66721.097 7 AK065147 3 9 OsDRM1b a DMT708 12012.m04183 LOC_Os12g01800 B9G900 482987–485738 1674 557 5.47 62537.1692 7 AK065147 12 10 OsDRM1a a DMT709 12011.m04382 LOC_Os11g01810 Q2RBJ4 427376–431591 1695 564 5.36 63354.2207 7 AK065147 11 a Genes located in segmentally duplicated region. R. Sharma et al. Rice cytosine DNA methyltransferases FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS 6303 the proteins also had amino terminal extensions with dif- ferent types of domain, such as ubiquitin-associated domain (UBA), BAH or chromo domains depending on the type of MTase. Rice MTases were classified into known subfamilies and the unnamed loci were assigned names on the basis of their phylogenetic relatedness to corresponding members of the MTase subfamilies in humans and other plants (see below). In total, three CMT (OsCMT1–3), one OsDnmt2 showing similarity to vertebrate Dnmt2, four DRM (OsDRM1a, OsDRM1b, OsDRM2 and OsDRM3) having similarity with human Dnmt3 and two MET1 genes (OsMET1-1 and OsMET1- 2) were identified on the basis of these searches (Table 1). Phylogeny and genome distribution In order to analyse evolutionary relatedness of rice proteins, complete amino acid sequences of all MTases were aligned with corresponding proteins from maize, Arabidopsis, Chlamydomonas, mouse, human, Drosoph- ila and yeast using clustalx 1.83 [14]. A bootstrapped phylogenetic tree was generated using the neighbour- joining method. This revealed the existence of rice proteins that belonged to all four subclasses of MTase known in plants (Fig. 1). Rice contains a single puta- tive OsDnmt2 gene whose product shows 80% simi- larity to maize Dnmt2, 67% to Arabidopsis AtDnmt2 and 38% to yeast PMT1 proteins at the amino acid level. These genes code for small proteins similar to prokaryotic MTases that possess only the MTase domain with conserved motifs (Fig. 2). Four loci having highly similar amino acid sequences with corresponding Arabidopsis AtDRMs were identified in rice. The products of LOC_ Os12g01800 and Loc_Os11g01810, respectively, were found to be >90 and 60% identical to Arabidopsis AtDRM1. Hence, these genes have been named OsDRM1a and OsDRM1b (Table 1). All rice DRMs exhibited characteristic altered arrangement of con- served motifs in the MTase domain with motifs I and X juxtaposed, similar to DRMs of Arabidopsis, Zea mays and Nicotiana (Figs 2 and S1). OsDRM3 showed greater similarity to Arabidopsis AtDRM3 in the rearranged MTase domain than to AtDRM2 and AtDRM1 in this subclass. Phylogenetic analysis showed that DRM3 genes of rice and Arabidopsis sepa- rated from the lineage that gave rise to DRM2 and DRM1 earlier in evolution and evolved independently. Among the CMTs, three genes were identified in rice that showed significant similarity to the corresponding Arabidopsis genes at both the nucleotide and the protein level (more than 60%) (Fig. 1). Rice CMTs have been named OsCMT1–3 on the basis of their similarity to corresponding proteins in Arabidopsis and similarity in the expression profiles in both the plant species (see below). OsCMT1–3 are characterized by the presence of a BAH domain in the N-terminal half and a conserved chromo domain inserted between motifs I and IV in the MTase domain (Figs 2 and S2A). OsCMT2 differs from the other two proteins in encoding an additional stretch of 193 amino acids ahead of the BAH domain in the amino terminal region (Fig. S2A). The chromo domain of rice CMTs possesses conserved aromatic amino acids, tyrosine at position 412 (Y412) and tryptophan at position 409 (W409). However, a phenylalanine was observed to be present in all plant proteins at position 382 in place of a tyrosine at the same position in animal proteins, with Arabidopsis LHP1 being an exception (Fig. S2B). We have identified two closely related puta- tive MET1-type genes, OsMET1-1 and OsMET1-2,in rice, confirming the two previous reports [17,18]. Both genes encode a conserved BAH domain and a C-termi- nal catalytic domain similar to the animal Dnmt1 proteins (Figs 1 and 2). The putative DNA MTase from the green alga, Chlamydomonas reinhardtii, did not possess any of the conserved regulatory domains identi- fied in other MTases. At least three of the MTases in this alga, CrDMT3407, CrDMT3403 and CrDMT3408, appear to have evolved independently from the com- mon ancestor that gave rise to the higher plants and humans de novo MTases. Localization of all rice MTase genes on pseudomolecules showed unbiased genomic distribution. Except for OsDRM1a and OsDRM1b, which were observed to be located in segmentally dupli- cated regions of chromosomes 11 and 12, all the other genes were observed to be located in unique regions specific to each chromosome (Table 1). Gene expression analysis To gain insight into the developmental windows during which rice MTase genes are expressed, spatial and tem- poral expression patterns of these genes were analysed during different developmental stages ⁄ tissues. For this purpose, an in-house generated microarray expression dataset prepared by using 57 k Affymetrix GeneChip Ò rice genome arrays was utilized, as mentioned previously (Affymetrix Inc., Santa Clara, CA, USA). The develop- mental stages ⁄ tissues of rice plants used in the present investigation are summarized in Table S1. The rice gen- ome array contains a total of 57 381 probe sets, of which 55 515 correspond to 51 279 rice transcripts and the rest, 1866, are hybridization, poly A and housekeeping controls. In our dataset, the average intensity values of the nonrice control probe sets were found to be < 10. Rice cytosine DNA methyltransferases R. Sharma et al. 6304 FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS Therefore, the value ‘10’ was taken as the cut-off to dis- tinguish between expressed and nonexpressed genes in a particular tissue ⁄ developmental stage. On the basis of signal intensities obtained for rice transcripts, three genes, OsCMT1, OsDRM1a and OsDRM1b, did not express in any of the 22 stages ⁄ tissues analysed (Table S2). The remaining six genes showed specific ⁄ preferential enhancement in transcript abundance at the onset of reproductive development as compared to their counterparts in Arabidopsis (Table S3). These genes also showed moderate to low level expression in vegetative tissues (Fig. 3). OsMET1- 1 transcripts accumulated at low level (£ 5; amounting to no expression) in most of the stages analysed, except P1-I to P1-III, where the average expression value was 11 in all three stages. The OsMET1-2 Fig. 1. Phylogeny of rice cytosine DNA MTases. An unrooted, neighbour-joining tree was constructed by alignment of total protein sequences from rice (Os), Arabidopsis (At), Zea mays (Zm), Nicotiana (Nt), yeast (Sp), Chlamydomonas reinhardtii (Cr), Drosophila (Dm), human (Hs) and mouse were downloaded from the chromatin database. The names of proteins mentioned for each organism indicate the chromatin database identification. Putative maintenance MTase from rice OsMET1-1 and OsMET1-2 cluster with other functionally character- ized maintenance MTases from Arabidopsis and other organisms and are shown connected by blue lines. Putative rice CMTs and similar methylases from other plants are connected by green lines. Evolutionarily conserved DNMT2 proteins from animals and plants, including rice, are shown with purple lines. De novo MTases from animals and plants, including, rice are shown connected by brown lines. The scale bar shows 0.1 amino acid substitutions per site. R. Sharma et al. Rice cytosine DNA methyltransferases FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS 6305 transcripts accumulated at higher than the cut-off value of 10 in all the samples analysed, with the high- est transcript accumulation observed in the SAM (shoot apical meristem), the early stages of panicle ini- tiation (P1-I and P1-II) and the early seed development stages (Tables S2 and S4). In comparison with leaf tissues (mature leaf and Y leaf), there was an 87-, 77- and 81-fold increase (P £ 0.005) observed in the tran- script accumulation of OsMET1-2 in the SAM, early panicle and early seed stages, respectively. In response to abiotic stress conditions, more than a two-fold reduction (P £ 0.05) in the transcript levels of this gene was observed in both salt and dehydration stress, whereas, 3 h cold stress did not show any significant effect. OsDnmt2, the only representative of the Dnmt2 subclass, was found to express at moderate levels in all the stages ⁄ tissues analysed. Similar to OsMET1-2, the OsDnmt2 transcripts also accumulated at higher levels in the SAM and early stages (P1-I–III) of panicle development, but the enhancement in transcript levels was only about two-fold. The OsDnmt2 transcript levels were significantly reduced in seed stages (Tables S1–S4; more than two-fold downregulation at P £ 0.05). The two DRM genes, OsDRM2 and -3, exhibited similar expression profiles and significant expression levels throughout rice development. Differ- ential expression analysis revealed that both the genes were upregulated by approximately two-fold in the SAM and early stages of panicle development in com- parison with any of the vegetative stages (mature leaf, Y leaf, root and seedling) used in this study. A signifi- cant downregulation (up to 4.8- and 2.5-fold for OsDRM2 and -3, respectively) in their accumulation was also observed during late seed development stages (Table S4). Two Arabidopsis counterparts of the rice DRM genes also expressed throughout development, but, interestingly, in contrast to that in rice, AtDRM2 and -3 showed up to four-fold enhancement in tran- script accumulation during seed development stages and showed higher expression in vegetative tissues as compared with rice genes (Fig. 3). For the two CMT genes, OsCMT2 and -3, highly differential and con- trasting profiles were observed. OsCMT2 showed mod- erate expression with specific downregulation (up to two-fold at P £ 0.001) in the SAM and early panicle stages (P1-I, II, III, P1 and P2) (Tables S2 and S4). However, OsCMT3, which expressed at very low levels (average normalized expression value of 24) during late panicle and seed development stages, showed up to 140-fold enhancement in the SAM and stages of early panicle development. The Arabidopsis AtCMT3 gene also exhibited a similar reduction in transcript accumu- lation during seed development, but unlike OsCMT3, it expressed at significantly higher levels in vegetative tissues. Interestingly, transcripts of OsCMT2 accumu- lated in seedlings in response to cold and salt stress, whereas the transcriptional activity of this gene was unaffected under dehydration conditions (Table S3). In contrast, OsCMT3 showed approximately a six- and four-fold reduction in transcript accumulation in rice Fig. 2. Schematic representation of conserved domains and motifs identified among proteins of each subclass of rice MTases. Blue rectan- gles represent the conserved motifs in the catalytic MTase domains, whereas the conserved domains in the N-terminal regulatory region are represented by red, yellow and purple rectangles. The numbers on the right denote the length of each protein. Rice cytosine DNA methyltransferases R. Sharma et al. 6306 FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS seedlings subjected to salt and dehydration stress, respectively (Table S2). Statistically significant and differential expression profiles of three genes, OsCMT3, OsDRM3 and OsMET1-2, were further validated by quantitative PCR (QPCR). The transcript accumulation patterns for all three genes observed by QPCR were similar to those observed from the microarray analysis (Fig. 4). Tran- scripts for OsCMT3 and OsDRM3 showed maximum accumulation in the SAM and early stages of panicle initiation (P1-I, P1-II and P1-III) and thereafter their levels declined in mature organs and during seed forma- tion. Transcripts for OsMET1-2, on the other hand, showed the highest accumulation during the S2 stage of seed development, which coincides with the onset of differentiation of embryonic organs such as the SAM and the radicle 3–4 days after pollination (Table S1). Discussion In this investigation, putative cytosine DNA MTase genes encoded in the rice genome were identified and characterized. Transcript profiling of these genes dur- ing the transition of plants from the vegetative to the reproductive phase has provided weighted insight into the developmental timing of the activity of these genes. The specific activation of OsCMT2 in seedlings sub- jected to cold and salt stress and the downregulation of OsCMT3 expression under salt and drought stress is interesting and provides the foundation for further in depth analysis to study the possible involvement of these proteins in affecting overall genomic activity under abiotic stress. The first part of our analysis identified 10 putative cytosine DNA MTases that have sequence similarities Fig. 3. Microarray-based expression analysis of DNA MTase genes of rice and Arabidopsis. The expression profiles of rice genes were analy- sed in vegetative tissues (Y leaf, mature leaf and roots), SAM, six stages of panicle development (P1–P6), three substages of P1 (P1-I, P1-II and P1-III), five stages of seed development (S1–S5) and under three abiotic stress conditions (cold, salt and dehydration). For Arabidopsis, the expression profiles of two vegetative stages (leaf and root), five stages of flower development (F1–F5), five stages of silique develop- ment and three stress treatments (cold, salt and dehydration) were compiled. The colour bar represents Log 2 expression values. A cluster dendrogram is shown on the left-hand side of the expression maps. R. Sharma et al. Rice cytosine DNA methyltransferases FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS 6307 with known de novo and maintenance MTases of eukaryotes encoded in the completely sequenced rice genome. Phylogenetic analysis with genes in other organisms revealed that the DNA methylation system in rice utilizes the same basic set of MTases functional in eukaryotes. However, duplications and multiplicity in gene numbers was observed in DRM and OsMET1 subfamilies in rice. In the case of duplicated genes, it is usually observed that either both the genes remain active and function redundantly in an organism or the regulatory role of these genes is divided or is shared between the two part- ners equally or one of the gene member functions prominently while the other either weakly complements the function of the main gene or becomes a pseudogene. In the case of OsDRM1, it was observed that the conse- quence of duplication of this gene has probably resulted in inactivity of both OsDRM1a and OsDRM1b,asno transcripts could be detected for either of these genes during the stages of development analysed in this study. It is also plausible that OsDRM1a and OsDRM1b could be expressed during a much narrower develop- mental window and in specific tissues of floral organs, as we analysed the expression of these genes in whole panicles that were developmentally segregated on the basis of their total length. The expression of Arabidop- sis AtDRM1 was also very weak during floral organ ini- tiation and in mature seeds. In the case of OsMET1, two previous reports have shown expression of both OsMET1-1 and OsMET1-2 genes by RT-PCR and ribonuclease protection assay in actively dividing tissues from various vegetative and floral organs studied in japonica variety cv. Nipponbare and Taipei 309, although transcripts of OsMET1-2 were observed to accumulate between 7- and 12-fold higher than those of OsMET1-1 in both the cultivars [17,18]. In our micro- array experimental set-up, expression signals for OsMET1-1 transcripts were below the cut-off limit of 10 in most of the stages ⁄ tissues analysed, except in early stages of panicle development, whereas the expression of OsMET1-2 varied from moderate to high levels spe- cifically during early stages of floral organ initiation and in mature seeds. Interestingly, however, the pattern of transcript accumulation as seen from the values of signal intensities for both OsMET1-1 and OsMET1-2 overlapped in vegetative organs and at all stages of panicle and seed development (Table S2). It can there- fore be speculated that OsMET1-2 possibly functions as the major maintenance MTase. The putative functions of genes can be inferred from a comparison of their gene expression profiles with that of known genes in other organisms. The expres- sion pattern of OsCMT3 at different developmental stages showed similarity with the expression of AtCMT3 in young actively dividing tissues of roots and during the formation of inflorescence and floral meristems and the initiation of floral organs (P1-I and P1-II in rice and F1 and F2 in Arabidopsis). AtCMT3 is known to regulate DNA methylation by interacting with the N-terminal tail of histone H3, which is simultaneously methylated at lysine 9 and lysine 27 by kryptonite, a histone H3 lysine 9 MTase and an unknown protein [3,19]. All the rice CMTs share simi- larity in their N-terminal regulatory domains (BAH Fig. 4. QPCR results for three selected genes, OsCMT3, OsDRM3 and OsMET1-2, and their correlation with microarray data. The y-axis represents raw expression values obtained from the microarray analysis in the form of relative transcript abundance. The x-axis depicts developmental stages of panicle and seeds as described in Table S1. Standard error bars are shown for data obtained using both techniques. Rice cytosine DNA methyltransferases R. Sharma et al. 6308 FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS and chromo) and in possessing a highly conserved cat- alytic MTase domain at the C-terminus. Also, the chromo domain of all rice proteins contains the highly conserved aromatic amino acids that are also present in Arabidopsis and maize CMT and HP1 and the poly- comb group proteins from humans and Drosophila (Fig. S2B) [20,21]. These amino acids are known to form a methyl ammonium recognition cage in the complex that interacts with the methylated histone peptides. Amino acids on either side of these aromatic amino acids were also observed to be highly conserved, suggesting that the rice proteins can also adopt higher order structures similar to other proteins and possibly interact with methylated lysines in a manner similar to AtCMT3. Although the observations presented in this study strongly suggest that the subsets of cytosine DNA MTases for performing putative de novo and mainte- nance methylation activities during reproductive devel- opment are largely conserved across monocot and dicot species, finer differences in terms of specific genes targeted by these proteins probably occur in a species- specific manner. Further investigation into the roles of these proteins in development and the biochemical mechanisms underlying these activities could extend our knowledge and understanding of the role of DNA methyl- ation in programming reproductive development in rice. Materials and methods Identification of genes, chromosomal localization and phylogenetic analysis Cytosine DNA MTases encoded in the rice genome were identified using search tools as described previously [15,16]. Briefly, a name search and the HMM analysis were used to identify cytosine DNA MTases encoded in the rice genome. The specific sequences were downloaded from the Rice Genome Annotation Project (http://rice.plantbiology.msu. edu) and an HMM profile was generated using hmmer 2.1.1 software (http://hmmer.wust.edu). This resulting pro- file was then used to search the proteome database of rice available at TIGR using the Basic Local Alignment Search Tool with the filter off option. The details of encoded pro- teins [length, isoelectric point (pI), molecular mass] were obtained from the Rice Genome Annotation Project (http://rice.plantbiology.msu.edu) and chromatin database (http://www.chromdb.org). Conserved domains in encoded proteins were identified by using the Simple Modular Architecture Research Tool version 3.4 or the NCBI con- served domain database. The newly identified genes were named on lines of nomenclature used for the genes identified previously and on the basis of their phylogenetic relationship with other members of the same subfamily. Chromosomal localization of all rice genes was determined from the 5¢ and 3¢ coordinates of each gene mentioned in TIGR. A phylogenetic analysis was performed after constructing the tree using the neighbour-joining method followed by bootstrap analysis using 1000 replicates [14]. Plant material Plant tissues for all the vegetative, panicle and seed stages ⁄ tis- sues were collected from field-grown rice plants (Oryza sati- va indica var. IR64). Abiotic stress treatments, namely salt, drought and cold, were given to rice seedlings as described previously [15]. Briefly, salt stress was given to 7-day-old seedlings by transferring them to NaCl solution (200 mm) for 3 h; for cold stress, the seedlings were kept at 4 °C for 3 h; drought stress was induced by drying the plants on tissue paper and spreading them on a Whatmann 3 mm sheet for 3 h. Seven-day-old seedlings with roots submerged in water for 3 h were used as controls for all stress treatments. Microarray hybridization and analysis of data Affymetrix GeneChip Ò rice genome arrays representing  49 824 rice transcripts were used to prepare a compendium of transcriptome profiles of 22 stages of vegetative and repro- ductive development and stress response in rice [15]. The microarray analysis of 17 stages has been described previ- ously and the data were deposited in the Gene Expression Omnibus database at the NCBI under the series accession numbers GSE6893 and GSE6901. In the present investiga- tion, five more stages, namely Y leaf, SAM, P1-I, P1-II and P1-III were included to analyse gene expression patterns dur- ing early stages of panicle differentiation. Sixty-six cell inten- sity files generated by GeneChip Ò operating software were further analysed [14,15]. Expression data for cytosine MTase genes were extracted using the gene locus indentifications listed in Table 1. Wherever more than one probe set was available for one gene, the probe set designed from the 3¢ end was given preference. A differential expression analysis was performed by taking a mature leaf as the reference to identify genes expressing at more than the two-fold level in various stages of reproductive development (panicle and seed), with P £ 0.05. Similarly, to identify stress-induced genes, a differ- ential expression analysis was performed with no correction applied and P £ 0.05. Expression analysis of Arabidopsis MTases Expression data for Arabidopsis, derived using Affymetrix GeneChip Ò ATH1 genome arrays from stages comparable with those used for rice, were downloaded from the Gene Expression Omnibus database at the NCBI under the series accession numbers GSE5620, GSE5621, GSE5623, R. Sharma et al. Rice cytosine DNA methyltransferases FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS 6309 GSE5624, GSE5629, GSE5630, GSE5631, GSE5632 and GSE5634. The downloaded files were imported using the microarray analysis software, array assist, and normalized using the algorithm GCRMA and then Log 2 transformed. A further analysis leading to the generation of heat maps was performed in a manner identical to that for rice [15]. QPCR analysis Real time PCR reactions were carried out using RNA sam- ples isolated from the stages ⁄ tissue similar to those used for the microarray analysis. The primers for amplification were designed using primer express, version 2.0 (PE Applied Biosystems, Foster City, CA, USA), preferentially from the 3¢ end of the gene. The specificity of each primer was checked using the Basic Local Alignment Search Tool of the NCBI. Two biological replicates were taken and for each three technical replicates were carried out. Actin was used as the endogenous control. RT-PCR was performed using the ABI Prism 7000 Sequence Detection System and software (PE Applied Biosystems) [14,15]. The data were normalized to facilitate the profile matching that obtained from the microarrays and bar charts were plotted using Microsoft excel. Acknowledgements A senior research fellowship awarded by the Council for Scientific and Industrial Research (CSIR) to R.S., P.D., M.S. and a junior research fellowship to G.M. are acknowledged. This work has been funded by the Department of Science and Technology and the Department of Biotechnology, Government of India. References 1 Colot V & Rossignol JL (1999) Eukaryotic DNA methylation as an evolutionary device. BioEssays 21, 402–411. 2 Ponger L & Li WH (2005) Evolutionary diversification of DNA methyltransferases in eukaryotic genomes. Mol Biol Evol 22, 1119–1128. 3 Lindroth AM, Shultis D, Jasencakova Z, Fuchs J, John- son 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. 4 Finnegan EJ, Peacock WJ & Dennis ES (1996) Reduced DNA methylation in Arabidopsis thaliana results in abnormal plant development. Proc Natl Acad Sci USA 93, 8449–8454. 5 Henikoff S & Comai L (1998) A DNA methyltransferase homolog with a chromodomain exists in multiple poly- morphic forms in Arabidopsis. Genetics 149, 307–318. 6 Cao X, Springer NM, Muszynski MG, Phillips RL, Kaeppler S & Jacobsen SE (2000) Conserved plant genes with similarity to mammalian de novo DNA meth- yltransferases. Proc Natl Acad Sci USA 97, 4979–4984. 7 Finnegan EJ & Kovac KA (2000) Plant DNA meth- yltransferases. Plant Mol Biol 43, 189–201. 8 Cao X & Jacobsen SE (2002) Locus-specific control of asymmetric and CpNpG methylation by the DRM and CMT3 methyltransferase genes. Proc Natl Acad Sci USA 99(Suppl. 4), 16491–16498. 9 Pradhan S, Cummings M, Roberts RJ & Adams RL (1998) Isolation, characterization and baculovirus- mediated expression of the cDNA encoding cytosine DNA methyltransferase from Pisum sativum. Nucleic Acids Res 26, 1214–1222. 10 Papa CM, Springer NM, Muszynski MG, Meeley R & Kaeppler SM (2001) Maize chromomethylase Zea methyltransferase2 is required for CpNpG methylation. Plant Cell 13, 1919–1928. 11 Fujimoto R, Sasaki T & Nishio T (2006) Characteriza- tion of DNA methyltransferase genes in Brassica rapa. Genes Genet Syst 81, 235–242. 12 Kim HJ, Yano A, Wada Y & Sano H (2007) Properties of a tobacco DNA methyltransferase, NtMET1 and its involvement in chromatin movement during cell divi- sion. Ann Bot (Lond) 99, 845–856. 13 Hermann A, Schmitt S & Jeltsch A (2003) The human Dnmt2 has residual DNA-(cytosine-C5) methyltransfer- ase activity. J Biol Chem 278, 31717–31721. 14 Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F & Higgins DG (1997) The CLUSTAL_X windows interface: flexible strategies for multiple sequence align- ment aided by quality analysis tools. Nucleic Acids Res 25, 4876–4882. 15 Kapoor M, Arora R, Lama T, Nijhawan A, Khurana JP, Tyagi AK & Kapoor S (2008) Genome-wide identi- fication, organization and phylogenetic analysis of Dicer-like, Argonaute and RNA-dependent RNA polymerase gene families and their expression analysis during reproductive development and stress in rice. BMC Genomics 9, 451. 16 Arora R, Agarwal P, Ray S, Singh AK, Singh VP, Tyagi AK & Kapoor S (2007) MADS-box gene family in rice: genome-wide identification, organization and expression profiling during reproductive development and stress. BMC Genomics 8, 242. 17 Teerawanichpan P, Chandrasekharan MB, Jiang Y, Narangajavana J & Hall TC (2004) Characterization of two rice DNA methyltransferase genes and RNAi-medi- ated reactivation of a silenced transgene in rice callus. Planta 218, 337–349. 18 Yamauchi T, Moritoh S, Johzuka-Hisatomi Y, Ono A, Terada R, Nakamura I & Iida S (2008) Alternative splicing of the rice OsMET1 genes encoding Rice cytosine DNA methyltransferases R. Sharma et al. 6310 FEBS Journal 276 (2009) 6301–6311 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... Table S1 Developmental stages ⁄ organs of rice plants analysed Table S2 GCRMA normalized expression values obtained for cytosine DNA MTase genes of rice by using microarray data Table S3 GCRMA normalized expression values for cytosine DNA MTase genes of Arabidopsis by using microarray data in the public domain (Gene Expression Omnibus, http://www.ncbi.nlm.nih.gov/geo/) Table S4 Differential expression. .. Nature 416, 10 3–1 07 21 Min J, Zhang Y & Xu RM (2003) Structural basis for specific binding of Polycomb chromodomain to histone H3 methylated at Lys 27 Genes Dev 17, 182 3–1 828 Supporting information The following supplementary material is available: Fig S1 Comparison of DRMs of rice Arabidopsis, Zea mays and Nicotiana Fig S2 Comparison of CMTs of rice, Arabidopsis and Zea mays Rice cytosine DNA methyltransferases. ..R Sharma et al maintenance DNA methyltransferase J Plant Physiol 165, 177 4–1 782 19 Jackson JP, Lindroth AM, Cao X & Jacobsen SE (2002) Control of CpNpG DNA methylation by the KRYPTONITE histone H3 methyltransferase Nature 416, 55 6–5 60 20 Nielsen PR, Nietlispach D, Mott HR, Callaghan J, Bannister A, Kouzarides T, Murzin AG, Murzina... http://www.ncbi.nlm.nih.gov/geo/) Table S4 Differential expression analysis of rice MTase genes This supplementary material can be found in the online version of this article Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset... online delivery, but are not copy-edited or typeset Technical support issues arising from supporting information (other than missing files) should be addressed to the authors FEBS Journal 276 (2009) 630 1–6 311 ª 2009 The Authors Journal compilation ª 2009 FEBS 6311 . Rice cytosine DNA methyltransferases – gene expression profiling during reproductive development and abiotic stress Rita Sharma 1, *,. the expression pattern of rice DNA MTases during panicle and seed development. Utilizing an in-house generated microarray-based gene expression dataset, expression