Skip to main content


  • Research article
  • Open Access

Genome-wide high-resolution mapping of DNA methylation identifies epigenetic variation across embryo and endosperm in Maize (Zea may)

  • 1, 2,
  • 2,
  • 2,
  • 2,
  • 2,
  • 2,
  • 2,
  • 2,
  • 1Email author and
  • 2Email author
Contributed equally
BMC Genomics201516:21

  • Received: 10 November 2014
  • Accepted: 24 December 2014
  • Published:



Epigenetic modifications play important roles in plant and animal development. DNA methylation impacts the transposable element (TE) silencing, gene imprinting and expression regulation.


Through a genome-wide analysis, DNA methylation peaks were characterized and mapped in maize embryo and endosperm genome, respectively. Distinct methylation level was observed across maize embryo and endosperm. The maize embryo genome contained more DNA methylation than endosperm. Totally, 985,478 CG islands (CGIs) were identified and most of them were unmethylated. More CGI shores were methylated than CGIs in maize suggested that DNA methylation level was not positively correlated with CpG density. The promoter sequence and transcriptional termination region (TTR) were more methylated than the gene body (intron and exon) region based on peak number and methylated depth. Result showed that 99% TEs were methylated in maize embryo, but a large portion of them (34.8%) were not methylated in endosperm. Maize embryo and endosperm exhibit distinct pattern/level of methylation. The most differentially methylated region between embryo and endosperm are CGI shores. Our results indicated that DNA methylation is associated with both gene silencing and gene activation in maize. Many genes involved in embryogenesis and seed development were found differentially methylated in embryo and endosperm. We found 41.5% imprinting genes were similarly methylated and 58.5% imprinting genes were differentially methylated between embryo and endosperm. Methylation level was associated with allelic silencing of only a small number of imprinting genes. The expression of maize DEMETER-like (DME-like) gene and MBD101 gene (MBD4 homolog) were higher in endosperm than in embryo. These two genes may be associated with distinct methylation levels across maize embryo and endosperm.


Through MeDIP-seq we systematically analyzed the methylomes of maize embryo and endosperm and results indicated that the global methylation status of embryo was more than that of the endosperm. Differences could be observed at the total number of methylation peaks, DMRs and specific methylated genes which were tightly associated with development of embryo and endosperm. Our results also revealed that many DNA methylation regions didn’t affect transcription of the corresponding genes.


  • DNA methylation
  • Maize
  • Embryo
  • Endosperm
  • Transposable element
  • Imprinting gene
  • MeDIP-seq


DNA methylation, a conserved epigenetic mechanism involved in many important biological processes, is associated with gene silencing, X chromosome inactivation in females, and maintenance of genomic integrity in eukaryotes [1-3]. DNA methylation protects against transposon proliferation and impacts genomic imprinting [4-6].

Similar to mammalian genomes, DNA methylation in plant genomes predominantly occurs at CpG site. This is maintained by METHYLTRANSFERASE1 (MET1), a homolog of DNA methyltransferase1 (Dnmt1). In addition, plants DNA methylation occurs at CpHpG and CpHpH sites, and is maintained by CHROMOMETHYLASE3 (CMT3) [7-9]. In plant, DNA glycosidase subfamily including DEMETER (DME) and Repressor of Silencing 1 (ROS1) could mediate demethylation [10,11]. Although the methylated cytosine contexts in animals and plants are different, DNA methylation is conserved in both TEs and genes. In animals, plants and fungi, the active genes are generally unmethylated, while TEs are heavily methylated. However, green algae have an unusual pattern of methylation compared to other eukaryotes. More methylation was detected in exons to compare with TEs regions [12-14].

There is strong evidence that DNA methylation in promoter region represses gene expression [15-20]. In rice, methylation of transcriptional termination region (TTR) showed stronger repression effect on gene expression to compare with promoter methylation [21]. Results showed that gene-body methylation were positively associated with gene expression [22-24]. However, DNA methylation in the first exon was found to associate with gene silencing [25,26].

Bisulfite sequencing has been used to detect the methylated cytosines [27]. This method is very accurate to find the methylated cytosine of the individual locus. However, it is difficult to explore genome-wide methylation. MSAP (methylation sensitive amplification polymorphism) has been used to explore the genome-wide methylation, but it can only detect few methylation fractions and is limited by the types of enzymes used. Sequencing-based and microarray-based high-throughput detection of DNA methylation approaches are widely used in genome-wide methylation studies. DNA methylation microarray, DNA immunoprecipitation combined with high-throughput sequencing (MeDIP-seq) and bisulfite libraries construction combined with high-throughput sequencing are proved to be efficient [4,21,28-30]. These approaches have been used to discover global methylation dynamics in different plant species including Arabidopsis [4,7,30,31], sorghum [32] and rice [21,33].

Several studies reported the differential DNA methylated regions which were correlated with variable gene expression within the examined tissues [21,34]. Results showed that the difference of DNA methylation could only account for a limited extend of gene expression variation among plant vegetative tissues [21,35].

DNA methylation is very important for plant embryogenesis and seed development. Abnormal embryo methylation causes defect in embryogenesis, such as abnormal of cell division, embryo apical domain aberrance and reduction of viability [8]. In seed plants, gene imprinting occurs in endosperm [31,36-38]. Studies showed that the expression of only a small portion of imprinting genes was correlated with DNA methylation in Arabidopsis [31,39,40]. The expression level of maize imprinting genes was much higher in endosperm than in embryo [39,41]. TEs exhibited toxic effects on genome, and embryo represses parasitic TEs to prevent damage of genome during seed development. DNA methylation on TEs is an important way to repress TEs [42]. TEs silencing relies mainly on RNAi pathway directed methylation, and siRNA is the major mediator for CpHpH DNA methylation [22,43]. Large amount of small RNA was accumulated in rice endosperm. However, the CpHpH methylation level of endosperm is quite low, while the embryo CpHpH methylation level is rather high. The major methylation in rice endosperm is CpHpH, suggesting RNAi pathway does not participate in endosperm DNA methylation. The endosperm derived small RNAs could be transported to embryo where they mediate TEs silencing by DNA methylation [7,31,33,44].

In plants, endosperm DNA was hypomethylated in various sequence contexts. For example, rice endosperm CG methylation is about 93% of the methylation rate in embryos. CHG and CHH methylation is by 2–5 folds lower in endosperm compared to embryo. In Arabidopsis, CG methylation of most loci is demethylated in endosperm [12]. In Arabidopsis ecotype Col-gl and Ler, thousands of genes exhibited higher level of methylation in embryo than that in endosperm [31].

Previous studies on DNA methylation in embryo and endosperm using DNA methylation microarray, immunoprecipitation and MSAP have assessed only a small portion of tissue-specific DNA methylation variations in maize genome [29,31,35,37,45,46]. In the current study, we used Solexa MeDIP-seq to profile the methylomic landscape across embryo and endosperm, comparing the differences of their methylation modes. Many genes that were differentially methylated between embryo and endosperm were identified.


Methylomic profiling of embryo and endosperm in maize

We generated a total of 2,748,497,900 bp of DNA immunoprecipitation sequencing (MeDIP-seq) data from maize endosperm and 2,807,090,100 bp data from maize embryo. From endosperm and embryo, 53,541,909 and 54,639,671 clean reads (average length 50 bp) were obtained, respectively. More than 96% of MeDIP-seq reads were aligned (mapped) on maize genome in each tissue (Table 1). Figure 1 showed the chromosomal distribution of DNA methylation read of maize endosperm and embryo.
Table 1

Solexa MeDIP-seq data


Total reads

Clean reads

Mapped reads

Mapping ratio











Figure 1
Figure 1

Chromosomal distribution of DNA methylation read of maize endosperm and embryo. Each chromosome was split in 10 k wide windows and the methylated read count was calculated for each window in embryo and endosperm. Y-axis is the read count mapped in each window.

The mapped reads were used in scanning the methylation peak with Model-based analysis of chIP-seq (MACS) (version1.4). A total number of 115,599 methylation peaks (diffScore ≥ 50, p ≤ 1e-5, diffScore = −10*LOG10pvalue) from endosperm and 353,232 methylation peaks from embryo were identified (Figure 2). DNA methylation peak number shows the popularity of methylation in genome. More DNA methylation peaks mean more loci in genome are methylated.
Figure 2
Figure 2

Distribution of DNA methylation peaks in chromosomes. A, Distribution of DNA methylation peaks in maize embryo and endosperm chromosomes. B, The number of methylation peaks per Mb in maize embryo and endosperm chromosomes.

Characterization of methylated DNA regions

We analyzed the methylation status of CpG islands (CGIs), CGI shores (spanning 2,000 bp up-and down-stream of each CGI) as well as other locations in the genome. CpG island in maize genome was identified using CpG report software (EMBOSS: The default parameters are as following: the minimum length is 200 bp, minimum observed/expected value is 0.6, the minimum percentage of CpG content is 50% ( In total, 985,478 CGIs were identified in this study. Additional file 1: File S1a, Additional file 2: File S1b, Additional file 3: File S1c lists all CpG islands of whole maize genome. The locations of CpG islands on chromosomes, the length of the island and the observed/expected value were also provided. 2000 bp up- or down-stream of CpG island was considered to be the CpG shore [28]. If a CpG shore was overlapped with the methylation region identified in this study, we considered that this CpG shore was methylated.

There were 108,441 methylated CGIs and 214,787 methylated CGI shores in embryo, and 26,009 methylated CGIs and 67,483 methylated CGI shores in endosperm. In both embryo and endosperm, less CGIs were methylated compared to CGI shores (Figure 3A). Most CG islands were unmethylated (only 11% of CGIs in embryo and 2.6% CGIs in endosperm were methylated), which was in agreement with the methylation profiles in human and pig [28,47,48]. Methylation status of embryo and endosperm was analyzed in promoter, exon, intron, transcriptional termination region, 5’ UTR, 3’ UTR and coding sequences (Additional file 4: Table S1a, Additional file 5: Table S1b). Gene regions that overlapped with DNA methylation peaks were considered methylated. Methylated promoters and transcriptional termination region (TTR) could be classified into four types by CpG content: high CpG content promoters (HCPs), low CpG content promoters (LCPs), high CpG content TTRs (HCTTRs) and low CpG content TTRs (LCTTRs) as previously described [28,48,49].
Figure 3
Figure 3

Distribution of DNA methylated peaks in gene regions. Distribution of DNA methylation peaks in CGIs, CGI shores and other regions (A), Distribution of DNA methylation peaks in TTRs, promoters, exons and introns (B), Distribution of DNA methylation peaks in HCP, LCP, HCTTR and LCTTR (C). Distribution of DNA methylation peaks in CDs, 5’UTR and 3’UTR (D).

We detected 16,835 methylated promoters and 16,758 methylated TTRs in maize endosperm genome, while 44,488 methylated promoters and 41,715 methylated TTRs in maize embryo were identified. The DNA methylation occurred more frequently in promoter and TTR regions than other gene regions (Figure 3B).

In embryo, 12,313 methylated HCPs, 42,790 methylated LCPs, 10,628 methylated HCTTRs and 39,982 methylated LCTTRs were detected. Endosperm contained much less methylated HCPs (3,284), LCPs (16,835), HCTTRs (3,351) and LCTTRs (16,757) (Figure 3C). We also analyzed the methylation levels within introns and exons in embryo and endosperm. There were 17,475 methylated exons and 15,703 methylated introns in embryo, and 8,388 methylated exons, 7,970 methylated introns in endosperm. Figure 4 shows that the promoter and TTR were more methylated than the gene body (intron and exon) on average.
Figure 4
Figure 4

DNA methylation level in gene body, promoter and transcription termination region.

The transposase genes that overlapped with DNA methylation peak were considered methylated TEs. In maize seed, we detected 184 methylated TEs which contain MuDR transposon, gypsy-type retrotransposon, copia sub-class retrotransposon, CACTA sub-class transposon and other types of transposons by BLASTX against nr protein database. Only two of these TEs were not methylated in embryo, while 64 of them were not methylated in endosperm (Additional file 6: Table S2).

Characterization of Differential Methylated Region (DMR)

We identified DNA methylation regions using a newly developed method. Any peak detected in embryo overlapping with peaks in endosperm, we will select the genomic region covering them as one DNA methylation region. If a peak detected in embryo (or endosperm) doesn’t overlap with any peak from endosperm (or embryo), we also considered the genomic region covering the peak to be a DNA methylation region in embryo or endosperm. Thus, 381,221 DNA methylation regions were identified in this study.

The read number of each methylated region from embryo or endosperm was used to calculate the normalized log2 value (log2 ratio of read number of embryo versus endosperm) and test p-value using the DEGseq R package. If normalized log2 value >0 (or read number from endosperm in the methylation region =0) and p < 0.001, the methylated region was considered up-methylated in embryo, and down-methylated in endosperm. If normalized log2 value < 0(or read number from embryo in the methylation region =0) and p < 0.001, the methylated region was considered up-methylated in endosperm, and down-methylated in embryo.

Among the identified 381,221 methylated regions, 238,088 regions were differentially methylated between embryo and endosperm (Additional file 7: Table S3). Totally, 175,337 and 62,751 differentially methylated regions (DMRs) were up-methylated in embryo and endosperm, respectively.

More DMRs were located in CGI shore to compare with CGI. In promoter and TTR, we found more DMRs to compare with intron and exon. In embryo, more up-methylated DMRs were distributed in promoters and TTRs than intron and exon, while more down-methylated DMRs were distributed in promoter and TTR in endosperm. In embryo, more up-methylated DMRs could overlap with promoter, TTR, intron and exon to compare with endosperm (Figure 5). The overlapped up-methylated DMRs of some promoters, TTRs, introns and exons were plotted using the Integrative Genomics Viewer (IGV) software (Figure 6).
Figure 5
Figure 5

Distribution of DMRs in maize gene regions. A, Distribution of DMRs in TTRs, promoters, exons and introns. B, Distribution of up-methylated and down-methylated DMRs in embryo TTRs, promoters, exons and introns.

Figure 6
Figure 6

Promoter, TTR, intron and exon overlapped up-methylated DMRs in embryo. Some gene regions overlapped with the up-methylated DMRs in embryo, such as GRMZM2G136067 promoter (A), GRMZM2G343767 TTR (B), GRMZM2G100344 exon (C) and GRMZM2G093344 intron (D). Red boxes were the DMRs that were up-methylated in embryo and down-methylated in endosperm. The arrow on gene structure represents the direction of transcription.

Function annotation of the methylated genes

Function annotation of the methylated genes was carried out by BLASTX against non-redundant (nr) protein database. Information from proteins with the highest similarity to the given methylated gene was used to annotate the gene function. The encoding proteins of the methylated genes were further compared with KEGG (Kyoto Encyclopedia of Genes and Genomes). Totally, 15,958 methylated genes were annotated by BLASTX analysis. Some ribosomal protein, storage protein, LEA (late embryogenesis abundant) protein encoding genes and imprinting genes were found methylated. Many genes encoding transcription factors, such as WRKY, SBP, NAC, MYB, bZIP families were methylated (Additional file 8: Table S4).

The methylated genes were predicted to be involved in 118 metabolic pathways based on KEGG database. Many methylated genes encoding proteins involved in chromatin structure and DNA synthesis, cell cycle regulation, nitrogen metabolism, fatty acid synthesis and elongation, starch and sugar metabolism, amino acid metabolism, protein metabolism, brassinosteriod biosynthesis, tricarboxylic acid cycle pathway, hormone metabolism and signal transduction pathways (Additional file 9: Figure S1-9). These results indicated that DNA methylation was involved in a wide range of biological processes.

Differentially methylated genes in maize embryo and endosperm

Of the 15,958 annotated methylated genes, 296 in embryo, and 7,735 in endosperm were de-methylated. Many genes were only methylated in embryo but not in endosperm and many genes are more methylated in embryo than endosperm (all DMRs that the gene contained is up-regulated in embryo than in endosperm). Some of these genes may contribute to epigenetic inheritance and reprogramming across generations, for example, gene encoding DNA-3-methyladenine glycosylase I, gene encoding histone deacetylase, lysine-specific histone demethylase and histone-arginine methyltransferase. Some genes are associated with cell differentiation and vascular development, for example, the ZF-HD-type transcription factor, TCP-1/cpn60 chaperonin family protein and TCP family protein. Some genes are involved in hormone metabolism and signal transduction pathways, such as ABI3/VP1 type transcription factor, auxin response factor (ARF), Aux/IAA family, GRAS family transcription factor, B3 DNA binding domain family protein, GID1-like gibberellin receptor, and BRASSINOSTEROID INSENSITIVE 1-associated receptor kinase 1. Gene involved in cell fate determination, GeBP type transcription factor [50], was identified. WOX family proteins, key regulators of embryo development, were detected in this study. Genes encoding key enzymes in starch synthesis, for example, starch synthase I, starch branching enzyme IIb and granule-bound starch synthase precursor were detected differentially methylated. Dicer-like (DCL) and Argonaute (AGO), key enzymes of small RNA biogenesis pathway were found to be more methylated in embryo than in endosperm. bHLH and MADS-box transcription factor also showed more methylation in embryo than in endosperm (Additional file 10: Table S5).

Gene Ontology (GO) enrichment of differentially methylated genes

All methylated genes were annotated based on GO annotation. Additional file 11: Table S6 shows the GO categories and function of methylated genes.

To better understand the potential function of the differential methylated genes, GO functional classification of these genes was performed by Blast2GO program. Fisher’s exact test p-values were calculated for overrepresentation of the differential methylated genes (genes contained DMRs) in all GO categories. GO terms with p < 0.05 were considered as significant enriched. Totally, 97 GO terms were significantly enriched, with 37 in cellular component, 19 in molecular function and 41 in biological process. In biological process, the most significantly enriched differentially methylated genes are related to photosynthesis, electron transport chain, and respiratory electron transport chain. In cellular component, the most significantly enriched genes are related to chloroplast, photosynthetic membrane, and thylakoid part. While, in molecular function, the most significantly enriched genes are involved in quinone binding, NADH dehydrogenase (quinone) and NADH dehydrogenase (ubiquinone) activity (Table 2).
Table 2

GO function analysis of the enriched differential methylated genes


GO term



biological process







electron transport chain




respiratory electron transport chain




generation of precursor metabolites and energy




oxidative phosphorylation




ATP synthesis coupled electron transport




energy derivation by oxidation of organic compounds




cellular respiration




photosynthesis, light reaction




photosynthetic electron transport chain




photosynthetic electron transport in photosystem II




cellular metabolic process




biosynthetic process




cellular biosynthetic process




cellular process




energy coupled proton transport, down electrochemical gradient




ATP synthesis coupled proton transport




plasma membrane ATP synthesis coupled proton transport




transcription, DNA-dependent




RNA biosynthetic process




purine nucleoside triphosphate biosynthetic process




purine ribonucleoside triphosphate biosynthetic process




cellular macromolecule biosynthetic process




macromolecule biosynthetic process




ribonucleoside triphosphate biosynthetic process




nucleoside triphosphate biosynthetic process




monovalent inorganic cation transport




purine ribonucleotide biosynthetic process




hydrogen transport




proton transport




mitochondrial ATP synthesis coupled electron transport




ATP biosynthetic process




purine nucleotide biosynthetic process




mitochondrial electron transport, NADH to ubiquinone




gene expression




purine-containing compound biosynthetic process




metabolic process



cellular component







photosynthetic membrane




thylakoid part




thylakoid membrane




plastid thylakoid membrane




chloroplast thylakoid membrane




chloroplast thylakoid




plastid thylakoid








organelle subcompartment




photosystem II reaction center








photosystem II








light-harvesting complex




organelle part




intracellular organelle part




proton-transporting two-sector ATPase complex, proton-transporting domain




macromolecular complex




proton-transporting ATP synthase complex, coupling factor F(o)




mitochondrial ribosome




proton-transporting ATP synthase complex




proton-transporting two-sector ATPase complex








chloroplast part




mitochondrial part




ribosomal subunit




plastid part




mitochondrial large ribosomal subunit








protein complex




organellar ribosome




mitochondrial matrix




small ribosomal subunit




plasma membrane-derived thylakoid




plasma membrane-derived thylakoid photosystem II




ribonucleoprotein complex




mitochondrial proton-transporting ATP synthase complex




integral to membrane




membrane part




cytoplasmic part



molecular function



quinone binding




NADH dehydrogenase (quinone) activity




NADH dehydrogenase (ubiquinone) activity




NADH dehydrogenase activity




rRNA binding




DNA-directed RNA polymerase activity




RNA polymerase activity




oxidoreductase activity, acting on NADH or NADPH, quinone or similar compound as acceptor




oxidoreductase activity, acting on NADH or NADPH




electron transporter, transferring electrons within the cyclic electron transport pathway of photosynthesis activity




structural molecule activity




hydrogen ion transporting ATP synthase activity, rotational mechanism




structural constituent of ribosome




ribonucleoside binding




nucleoside binding




hydrogen ion transmembrane transporter activity




monovalent inorganic cation transmembrane transporter activity




nucleotidyltransferase activity




ribulose-bisphosphate carboxylase activity



Methylation and transcriptional repression of imprinting genes

Previous studies demonstrated that some imprinting genes in endosperm were associated with DNA methylation [31,39-41]. In maize endosperm, 17.3% of the analyzed imprinting genes showed differential methylation between the two parental alleles [39,41]. We examined the correlation between expression of imprinting genes and DNA methylation level in embryo and endosperm. We analyzed the methylation mode of 176 imprinting genes identified by Zhang [39]. Only 65 out of the 176 imprinting genes were detected to be methylated in embryo or endosperm. Thirty eight of the methylated genes were only methylated in embryo, suggesting that these imprinting genes were not associated with allelic silencing in endosperm. Five of the methylated genes were only methylated in endosperm, suggesting that these imprinting genes were not associated with allelic silencing in embryo.

Based on the transcription data [37], we found 25 of the methylated genes showed much higher expression level in endosperm than in embryo (endosperm/embryo average RPKM >1.5), among them, the majority (17/25) of genes were more methylated or only methylated in embryo, four of them were more methylated in endosperm or only methylated in endosperm and the methylation level of rest four genes was similar between embryo and endosperm.

15 of the methylated genes showed much higher expression level in embryo than in endosperm (embryo/endosperm average RPKM >1.5), among them, one of the 15 genes was more methylated in endosperm, 6 were more methylated in embryo, and the rest 8 genes showed similar methylation level between embryo and endosperm.

The rest 25 methylated genes showed similar expression level between endosperm and embryo, among them, ten of the 25 genes were more methylated in embryo, the rest 15 genes showed similar methylation level between embryo and endosperm. Totally, among the 65 imprinting genes which were methylated in embryo or endosperm, 41.5% were methylated similarly between embryo and endosperm, while 58.5% were differentially methylated (Additional file 12: Table S7). These results showed that only a small portion of the imprinting genes were regulated by DNA methylation.

A maize DME-like gene (GRMZM2G123587) and a MBD4-like gene (GRMZM5G847045) were identified in this study. The DME-like gene encodes 5-methylcytosine DNA glycosylase, and MBD4-like gene encodes methylation-binding domain 101 protein. Maize DME-like (GenBank: AFW71475.1) is homologous to Arabidopsis DME (AAM77215.1, identifies = 64%, E-value = 0) and contains HhH-GPD base excision DNA repair protein domain. Maize MBD4-like NP_001105172.1 is homologous to Arabidopsis MBD4 (NP_191862.1, identifies = 47%, E-value = 2e-37) and contains methyl-CpG-binding domain. DME participates in demethylation of the maternal genome in endosperm [10,11]. Overexpression of MBD4/AID gene caused bulk genome de-methylation in zebra fish [9]. The expression levels of DME-like and MBD4-like gene were all higher in maize endosperm (Figure 7). Therefore, high transcriptional activity of these two genes could be associated with the low methylation level in endosperm.
Figure 7
Figure 7

Different expression of DME-like gene, MBD4-like gene, SBEIIB gene and GBSS gene in maize embryo and endosperm.


Our data showed the different pattern of DNA methylation between maize embryo and endosperm. Embryo contained more DNA methylation peaks to compare with endosperm in each chromosome.

More DNA methylation peaks were located in CGI shores compared to CGIs and other gene regions, which is in agreement with the results from human. CGI shore also contained more DMRs than CGI. More DMRs were distributed in promoters and TTRs than introns and exons. Many studies showed that LCPs were more methylated than HCPs [49], we found the same result in maize. In maize, we found that TTRs methylation may have similar function to promoter methylation, which is in agreement with previous study [21]. We found that LCTTRs contained more DNA methylation peaks than the HCTTRs. More DMRs were up-methylated in embryo than in endosperm. This methylation may cause tissue-specific expression of genes between embryo and endosperm. Our results showed that 58.5% imprinting genes were differentially methylated between embryo and endosperm. Methylation level could be negatively or positively correlated with the expression level of imprinting genes. Methylation level was associated with allelic silencing of only a small number of imprinting genes.

Starch is a major component of maize endosperm and comprises two different forms of carbohydrate polymers: a linear amylose and a branched amylopectin. Our results showed that the genes encoding starch synthase I, starch branching enzyme IIb (SbeIIb) and granule-bound starch synthase precursor (GBSS) were more methylated in embryo than in endosperm. The expression levels of these genes were much higher in endosperm than in embryo, which was consistent with previous transcriptome data [37]. Both GBSS and starch synthase I showed embryo specific DNA methylated, and the methylated regions were located in promoter. SbeIIb gene promoter was methylated only in embryo and intron was more methylated in embryo than in endosperm. This is in agreement with previous studies that SbeII and GBSS promoters are endosperm-specific promoters [51,52]. Tissue-specific promoters may contain tissue-specific cis-elements, for example, RY motif [51,52] or other elements, for example, GGATCC palindrome, which could be recognized by DNA methyltransferases [53]. GGATCC was detected in maize SbeIIb gene promoter. It is possible that the promoters of the above described three genes were de-methylated in endosperm, and promoter methylation of these genes in embryo was associated with their transcription repression. In addition, genes in starch synthesis and metabolism, such as isoamylase-type starch debranching enzyme ISO3, starch branching enzyme IIa, starch synthase IIIb-1, starch phosphorylase and starch binding domain containing family protein genes were all methylated mainly in promoter region only in embryo. These genes were hypermethylated in embryo and hypomethylated in endosperm.

Storage protein zein and oleosin genes were found to be more methylated in embryo than in endosperm. Zein gene expression level was higher in endosperm than in embryo [37]. The 15 kDa beta zein, 22 kDa alpha zein1, 22 kDa alpha zein 4 and 22 kDa alpha zein 5 encoding genes were methylated specifically in embryo. The DMR of 15 kDa beta zein encoding gene was located in TTR, while the DMRs of other three genes all located in exon. The methylated level and the expression level of zeins were negatively correlated. Oleosins are key components of oil body. In maize seed, the expression level of 16 kDa oleosin gene and Zm-II oleosin gene was higher in embryo compared to endosperm. However, our data clearly showed that these two genes were methylated in embryo but not in endosperm. The methylation was located in promoter and TTR regions. The methylation level of these oleosin genes was not negatively correlated with their expression.

WUSCHEL-related homeobox (WOX) transcription factor is necessary for cell division that forms the apical domain of embryo. In maize embryo, the WOX8 promoter and TTR was hypermethylated, however, this gene was highly expressed in embryo. In embryos of Arabidopsis met1-6 mutant, expression level of de-methylated WOX8 is lower than the hypermethylated WOX8 in wild-type embryo [8]. This result suggested that methylation could enhance WOX8 expression in embryo. Gene methylation could repress or active gene expression, and sometimes gene methylation may not correlate with transcription [21].

A hypothesis is that TE silencing is through the RNAi pathway. Small RNAs may be transported from endosperm to embryo where they lead to siRNA or miRNA-mediated methylation of TEs. The evidence is that abundant TE-derived small RNAs were accumulated in endosperm in Arabidopsis, but extremely low levels of CHH methylation occurs in endosperm. In contrast, high CHH methylation was detected in embryo [54]. In other words, the link between RNAi and DNA methylation may be weakened in endosperm and the small RNAs could be taken away. In our study, TEs were more methylated in embryo than in endosperm, possibly due to the imported small RNA. If this is true, it could explain, at least in part, the hypomethylation of endosperm. We identified a maize DME-like gene and a MBD4-like gene, and found that their expression levels were higher in maize endosperm than in embryo. DME and MBD4 could mediate strong bulk genome DNA demethylation. Therefore, the differentially expressed DME-like and MBD101 genes could be a possible reason for the differential methylation patterns in embryo and endosperm.


Through MeDIP-seq we systematically analyzed the methylomes of maize embryo and endosperm and results indicated that the global methylation status of embryo was more than that of the endosperm. Differences could be observed at the total number of methylation peaks, DMRs and specific methylated genes which were tightly associated with development of embryo and endosperm. Our results also revealed that many DNA methylation regions didn’t affect transcription of the corresponding genes.


Tissue collection and genomic DNA extraction

Maize endosperm and embryos were collected from B73 ears 14 days after self-pollination (DAP). The genomic DNA of two tissues was extracted by CTAB method. Tissues were ground with liquid nitrogen and 0.1 g powder was transferred into 2.0 ml tube containing 600 μl 2% CTAB solution (65°C) and 1 μl 10 μM RNase. Sample was incubated in 65°C water bath for 20 min, then mixed with 300 μl chloroform and 300 μl Tris saturated Phenol. The sample was centrifuged at 12000 rpm for 10 min at 4°C. Supernatant was transferred into a new tube and mixed with 600 μl chloroform, centrifuged at 12000 rpm for 10 min at 4°C. Supernatant was mixed with 2 volume of 100% ethanol for DNA precipitation. After centrifugation, the liquid was discarded and the precipitated DNA was washed with 70% ethanol. Dried the DNA briefly and then dissolved in 30 μl double distilled water.

MeDIP libraries construction and sequencing

DNA was sheared using the Bioruptor sonicator (Diagenode). End reparation, base addition and adaptor ligation were performed using Methyl-Seq 1 Kit. Adaptor-ligated sheared DNA was immunoprecipitated by 5-methylcytidine antibody with Magnetic Methylated DNA Immunoprecipitation kit (Diagenode) to construct the MeDIP libraries. Each MeDIP library was subjected to high-throughput sequencing by Illummina solexa HiSeq2000 platform.

Sequencing quality control and reads processing

Sequencing data was analyzed by data collection software. Quality control was performed using FastQC ( Clean reads were generated by using fastx software (version: 0.0.13) (

Genome mapping was performed using Bowtie2 (version: 2.0.5) software with default parameters [55] to generate the BAM files. The plant chloroplast genome has no methylation activity [56], so we only accounted the nuclear methylation reads detected in genome.

The peaks were detected from the BAM files using MACS (version: 1.4) [57]. Peaks with p-value = < 1e-5 were selected for further analysis.

The MeDIP-seq data from this study have been submitted to the NCBI Gene Expression Omnibus ( under accession no. GSE58549 (

Classfication of methylated genomic regions

Based on maize genome database transcript annotation and our methylation data, we detected the methylated (overlapped regions with DNA methylation peaks) exon, intron, promoter, CD, 3’ UTRs, 5’UTR and TTR in maize. Methylated promoters and TTRs were classified into four types: high CpG content promoters (HCPs), low CpG content promoters (LCPs), high CpG content TTRs (HCTTRs) and low CpG content TTRs (LCTTRs) according to CpG content as previously described [49].

CGIs and CGI shores identification

CGIs and CGI shores (2000 bp up- and down-stream of CGI) of maize genome were predicted by new CpG report software (Version: EMBOSS: ( The default parameters are as following: the minimum length is 200 bp, minimum observed/expected value is 0.6, the minimum percentage of CpG content is 50%.

Identification of DME-like gene and a MBD101 gene

We employed the protein sequences of Arabidopsis DME gene (AT5G04560.1) and MBD4 gene (AT3G63030.1) as queries to identify the most possible maize DME and MBD4 gene in genome database using BLASTp program (E-value <10). Sequences with the smallest E-value and the highest identities were considered DME and MBD4 homologs. A search on the Pfam database was performed to confirm the sequences.

RT-PCR analyses

Total RNA was prepared using Trizol agent (TaKaRa, Dalian, China) according to the manufacturer’s instructions. For reverse transcription, after DNase I treatment, the first-strand cDNA was synthesized with an oligo (dT) primer using a PrimeScript™ first-strand cDNA synthesis kit (D6110A; TaKaRa, Dalian, China). Equal amounts of RT products were used to perform subsequent PCR amplification. Primers used to amplify DME-like were 5’-CACAAACCCAGGAAACGGAG-3’ and 5’-ACCACCCCAACCCCAATG-3’. Primers used to amplify MBD4-like were 5’-AACATACCAAAGCCTCCACCA-3’ and 5-TGCCTCCAGAAACTTATCCACA -3’. Primers used to amplify the control, Actin 1, were 5’-GGGATTGCCGATC GTATGAG-3’ and 5’-GAGCCACCGATCCAGACACT-3’. Primers used to amplify SBEIIB were 5’-ACACCGGCCTCTTCTTAACTC-3’ and 5’-CTCGCCCTCAGGAACCAT-3’. Primers used to amplify GBSS were 5’-CTGAGCCTCAACAACAACCC-3’ and 5’-TGTAGATGCCGTGGGACTG-3’.




This work is supported by National Natural Science Foundation of China (31000720, 31101427), the Initial Special Research for 973 Program of China (2012CB126313), Shandong Province Taishan Scholar Foundation (tshw20100416), Shandong Province Germplasm Innovation and Utilization Project and grants from Shandong Province (2012ZHZXIA0418, 2008GG10002002, ZR2010CQ008, 201103023).

Authors’ Affiliations

Agricultural College, Anhui Agricultural University, Hefei, 230036, PR China
Bio-Tech Research Center, Shandong Academy of Agricultural Sciences; Shandong Provincial Key Laboratory of Crop Genetic Improvement, Ecology and Physiology, Jinan, 250100, PR China


  1. Klose RJ, Bird AP. Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci. 2006;31(2):89–97.PubMedView ArticleGoogle Scholar
  2. Feinberg AP. Phenotypic plasticity and the epigenetics of human disease. Nature. 2007;447(7143):433–40.PubMedView ArticleGoogle Scholar
  3. Heard E, Disteche CM. Dosage compensation in mammals: fine-tuning the expression of the X chromosome. Genes Dev. 2006;20(14):1848–67.PubMedView ArticleGoogle Scholar
  4. Zhang XY, Yazaki J, Sundaresan A, Cokus S, Chan SW-L, Chen HM, et al. Genome-wide high-resolution mapping and functional analysis of DNA methylation in Arabidopsis. Cell. 2006;126(6):1189–201.PubMedView ArticleGoogle Scholar
  5. Goll MG, Bestor TH. Eukaryotic cytosine methyltransferases. Annu Rev Biochem. 2005;74:481–514.PubMedView ArticleGoogle Scholar
  6. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev. 2002;16:6–21.PubMedView ArticleGoogle Scholar
  7. Hsieh TF, Ibarra CA, Silva P, Zemach A, Eshed-Williams L, Fischer RL, et al. Genome-wide demethylation of Arabidopsis endosperm. Science. 2009;324(5933):1451–4.PubMed CentralPubMedView ArticleGoogle Scholar
  8. Xiao WY, Custard KD, Brown RC, Lemmon BE, Harada JJ, Goldberg RB, et al. DNA methylation is critical for Arabidopsis embryogenesis and seed viability. Plant Cell. 2006;18(4):805–14.PubMed CentralPubMedView ArticleGoogle Scholar
  9. Rai K, Huggins IJ, James SR, Karpf AR, Jones DA, Cairns BR. DNA demethylation in zebrafish involves the coupling of a deaminase, a glycosylase, and gadd45. Cell. 2008;135(7):1201–12.PubMed CentralPubMedView ArticleGoogle Scholar
  10. Zhu JK. Active DNA demethylation mediated by DNA glycosylases. Annu Rev Genet. 2009;43:143–66.PubMed CentralPubMedView ArticleGoogle Scholar
  11. Martínez-Macías MI, Qian W, Miki D, Pontes O, Liu Y, Tang K, et al. A DNA 3 phosphatase functions in active DNA demethylation in arabidopsis. Mol Cell. 2012;45(3):357–70.PubMed CentralPubMedView ArticleGoogle Scholar
  12. Zemach A, McDaniel IE, Silva P, Zilberman D. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science. 2010;328(5980):916–9.PubMedView ArticleGoogle Scholar
  13. Feng S, Cokus SJ, Zhang X, Chen PY, Bostick M, Goll MG, et al. Conservation and divergence of methylation patterning in plants and animals. PANS. 2010;107(19):8689–94.View ArticleGoogle Scholar
  14. Lee TF, Zhai JX, Meyers BC. Conservation and divergence in eukaryotic DNA methylation. Proc Natl Acad Sci U S A. 2010;107(20):9027–8.PubMed CentralPubMedView ArticleGoogle Scholar
  15. Bell JT, Pai AA. DNA methylation patterns associate with genetic and gene expression variation in HapMap cell lines. Genome Biol. 2011;12:R10.PubMed CentralPubMedView ArticleGoogle Scholar
  16. Baek D, Jiang JF. Regulated AtHKT1 gene expression by a distal enhancer element and DNA methylation in the promoter plays an important role in salt tolerance. Plant Cell Physiol. 2011;52(1):149–61.PubMedView ArticleGoogle Scholar
  17. Saze H, Sasaki T, Kakutani T. Negative regulation of DNA methylation in plants. Epigenetics. 2008;3(3):122–4.PubMedView ArticleGoogle Scholar
  18. Tan MP. Analysis of DNA methylation of maize in response to osmotic and salt stress based on methylation-sensitive amplified polymorphism. Plant Physiol Biochem. 2010;48(1):21–6.PubMedView ArticleGoogle Scholar
  19. Curradi M, Izzo A, Badaracco G. Molecular mechanisms of gene silencing mediated by DNA methylation. Mol Cell Biol. 2002;22(9):3157–73.PubMed CentralPubMedView ArticleGoogle Scholar
  20. Tolley BJ, Woodfield H. Light-regulated and cell-specific methylation of the maize PEPC promoter. J Exp Bot. 2012;63(3):1381–90.PubMed CentralPubMedView ArticleGoogle Scholar
  21. Li X, Zhu JD, Hu FY, Ge S, Ye MZ, Xiang H, et al. Single-base resolution maps of cultivated and wild rice methylomes and regulatory roles of DNA methylation in plant gene expression. BMC Genomics. 2012;13:300.PubMed CentralPubMedView ArticleGoogle Scholar
  22. Tran RK, Zilberman D, Bustos CD, Ditt RF, Henikoff JG, Lindroth AM, et al. Chromatin and siRNA pathways cooperate to maintain DNA methylation of small transposable elements in Arabidopsis. Genome Biol. 2005;6(11):R90.PubMed CentralPubMedView ArticleGoogle Scholar
  23. Zilberman D, Gehring M, Tran RK, Ballinger T, Henikoff S. Genome-wide analysis of Arabidopsis thaliana DNA methylation uncovers an interdependence between methylation and transcription. Nat Genet. 2007;39(1):61–9.PubMedView ArticleGoogle Scholar
  24. Simmen MW, Leitgeb S, Charlton J, Jones SJ, Harris BR, Clark VH, et al. Nonmethylated transposable elements and methylated genes in a chordate genome. Science. 1999;283(5405):1164–7.PubMedView ArticleGoogle Scholar
  25. Brenet F, Moh M, Funk P, Feierstein E, Viale AJ, Socci ND, et al. DNA methylation of the first exon is tightly linked to transcriptional silencing. PLoS One. 2011;6(1):e14524.PubMed CentralPubMedView ArticleGoogle Scholar
  26. Kass SU, Pruss D, Wolfee AP. How does DNA methylation repress transcription? Science. 1997;13(11):444–9.Google Scholar
  27. Henderson IR, Chan SR, Cao XF, Johnson L, Jacobsen SE. Accurate sodium bisulfite sequencing in plants. Epigenetics. 2010;5(1):47–9.PubMed CentralPubMedView ArticleGoogle Scholar
  28. Davies MN, Volta M, Pidsley R, Lunnon K, Dixit A, Lovestone S, et al. Functional annotation of the human brain methylome identifies tissue-specific epigenetic variation across brain and blood. Genome Biol. 2012;13(6):R43.PubMed CentralPubMedView ArticleGoogle Scholar
  29. Eichten SR, Swanson-Wagner RA. Heritable epigenetic variation among maize inbreds. PLoS Genet. 2011;7(11):e1002372.PubMed CentralPubMedView ArticleGoogle Scholar
  30. Shen HS, He H, Li J, Chen W, Wang XC, Guo L, et al. Genome-wide analysis of DNA methylation and gene expression changes in two Arabidopsis ecotypes and their reciprocal hybrids. Plant Cell. 2012;24(3):875–92.PubMed CentralPubMedView ArticleGoogle Scholar
  31. Gehring M, Bubb KL, Henikoff S. Extensive demethylation of repetitive elements during seed development underlies gene imprinting. Science. 2009;324(5933):1447–51.PubMed CentralPubMedView ArticleGoogle Scholar
  32. Zhang MS, Yan HY, Zhao N, Lin XY, Pang JS, Xu KZ, et al. Endosperm-specific hypomethylation, and meiotic inheritance and variation of DNA methylation level and pattern in sorghum (Sorghum bicolor L.) inter-strain hybrids. Theor Appl Genet. 2007;115(2):195–207.PubMedView ArticleGoogle Scholar
  33. Zemach A, Kim MY, Silva P, Rodrigues JA, Dotson B, Brooks MD, et al. Local DNA hypomethylation activates genes in rice endosperm. Proc Natl Acad Sci U S A. 2010;107(43):18729–34.PubMed CentralPubMedView ArticleGoogle Scholar
  34. Zhang MS, Xu CM, Wettstein DV, Liu B. Tissue-specific differences in cytosine methylation and their association with differential gene expression in sorghum. Plant Physiol. 2011;156(4):1955–66.PubMed CentralPubMedView ArticleGoogle Scholar
  35. Eichten SR, Vaughn MW, Hermanson PJ, Springer NM. Variation in DNA methylation patterns is more common among maize inbreds than among tissues. Plant Genome. 2012;6(2):1–10.Google Scholar
  36. Huh JH, Bauer MJ, Hsieh TF, Fischer RL. Cellular programming of plant gene imprinting. Cell. 2008;132(5):735–44.PubMedView ArticleGoogle Scholar
  37. Waters AJ, Makarevitch I. Parent-of-origin effects on gene expression and DNA methylation in the maize endosperm. Plant Cell. 2011;23(12):4221–33.PubMed CentralPubMedView ArticleGoogle Scholar
  38. Luo M, Taylor JM, Spriggs A, Zhang H, Wu X, Russell S, et al. A genome-wide survey of imprinted genes in rice seeds reveals imprinting primarily occurs in the endosperm. PLoS Genet. 2011;7(6):e1002125.PubMed CentralPubMedView ArticleGoogle Scholar
  39. Zhang M, Zhao HN, Xie SJ. Extensive, clustered parental imprinting of protein-coding and noncoding RNAs in developing maize endosperm. Proc Natl Acad Sci U S A. 2011;108(50):20042–7.PubMed CentralPubMedView ArticleGoogle Scholar
  40. Gutiérrez-Marcos JF, Costa LM, Biderre-Petit C, Khbaya B, O’Sullivan DM, Mark Wormald M, et al. Maternally expressed gene1 is a novel maize endosperm transfer cell-specific gene with a maternal parent-of-origin pattern of expression. Plant Cell. 2004;16(5):1288–301.PubMed CentralPubMedView ArticleGoogle Scholar
  41. Jahnke S, Scholten S. Epigenetic resetting of a gene imprinted in plant embryos. Curr Biol. 2009;19(19):1677–81.PubMedView ArticleGoogle Scholar
  42. Slotkin RK, Martienssen R. Transposable elements and the epigenetic regulation of the genome. Nature. 2007;8(4):272–85.Google Scholar
  43. Slotkin RK, Vaughn M, Borges F, Tanurdzić M, Becker JD, Feijó JA, et al. Epigenetic reprogramming and small RNA silencing of transposable elements in pollen. Cell. 2009;136(3):461–72.PubMed CentralPubMedView ArticleGoogle Scholar
  44. Wu L, Zhou HY, Zhang QQ, Zhang JG, Ni FR, Liu C, et al. DNA methylation mediated by a microRNA pathway. Mol Cell. 2010;38(3):465–75.PubMedView ArticleGoogle Scholar
  45. Lauria M, Rupe M, Guo M, Kranz E, Pirona R, Viotti A, et al. Extensive maternal DNA hypomethylation in the endosperm of zea mays. Plant Cell. 2004;16(2):510–22.PubMed CentralPubMedView ArticleGoogle Scholar
  46. Lu YL. Analysis of DNA methylation in different maize tissue. J Genet Genomics. 2008;35(1):41–8.PubMedView ArticleGoogle Scholar
  47. Doi A, Park IH, Wen B, Murakami P, Aryee MJ, Irizarry R, et al. Differential methylation of tissue-and cancer-specific CpG island shores distinguishes human induced pluripotent stem cells, embryonic stem cells and fibroblasts. Nat Genet. 2009;41(12):1350–3.PubMed CentralPubMedView ArticleGoogle Scholar
  48. Li MZ, Wu HL, Luo ZG, Xia YD. An atlas of DNA methylomes in porcine adipose and muscle tissues. Nat Commun. 2013;3:850.View ArticleGoogle Scholar
  49. Weber M, Hellmann I, Stadler MB, Ramos L, Pääbo S, Rebhan M, et al. Distribution, silencing potential and evolutionary impact of promoter DNA methylation in the human genome. Nat Genet. 2007;39(4):457–66.PubMedView ArticleGoogle Scholar
  50. Curaba J, Herzog M, Vachon G. GeBP, the first member of a new gene family in Arabidopsis, encodes a nuclear protein with DNA-binding activity and is regulated by KNAT1. Plant J. 2003;33(2):305–17.PubMedView ArticleGoogle Scholar
  51. Kluth A, Sprunck S, Becker D, Lörz H, Lütticke S. 5’deletion of a gbss1 promoter region from wheat leads to changes in tissue and developmental specificities. Plant Mol Biol. 2002;49(6):669–82.PubMedView ArticleGoogle Scholar
  52. Miao HM, Fleming JE, LÜ DB, Han JF. Evaluation and characterization of an endosperm-specific SBeIIa promoter in wheat. Chinese Sci Bull. 2004;49(6):579–85.View ArticleGoogle Scholar
  53. Zinoviev VV, Yakishchik SI, Evdokimov AA, Malygin EG, Hattman S. Symmetry elements in DNA structure important for recognition/methylation by DNA [amino]-methyltransferases. Nucleic Acids Res. 2004;32(13):3930–4.PubMed CentralPubMedView ArticleGoogle Scholar
  54. Mosher RA, Melnyk CW, Kelly KA, Dunn RM, Studholme DJ, Baulcombe DC. Uniparental expression of PolIV-dependent siRNAs in developing endosperm of Arabidopsis. Nature. 2009;460(7252):283–6.PubMedView ArticleGoogle Scholar
  55. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009;10:R25.PubMed CentralPubMedView ArticleGoogle Scholar
  56. Fojtova M, Kovarik A, Matyasek R. Cytosine methylation of plastid genome in higher plants. fact or artefact? Plant Sci. 2001;160:585–93.PubMedView ArticleGoogle Scholar
  57. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, et al. Model-based Analysis of ChIP-Seq (MACS). Genome Biol. 2008;9(9):R137.PubMed CentralPubMedView ArticleGoogle Scholar


© Wang et al.; licensee Biomed Central. 2015

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.