Genome-wide analysis of mono-, di- and trimethylation of histone H3 lysine 4 in Arabidopsis thaliana

Abstract

Background: Post-translational modifications of histones play important roles in maintaining normal transcription patterns by directly or indirectly affecting the structural properties of the chromatin. In plants, methylation of histone H3 lysine 4 (H3K4me) is associated with genes and required for normal plant development.

Results: We have characterized the genome-wide distribution patterns of mono-, di- and trimethylation of H3K4 (H3K4me1, H3K4me2 and H3K4me3, respectively) in Arabidopsis thaliana seedlings using chromatin immunoprecipitation and high-resolution whole-genome tiling microarrays (ChIP-chip). All three types of H3K4me are found to be almost exclusively genic, and two-thirds of Arabidopsis genes contain at least one type of H3K4me. H3K4me2 and H3K4me3 accumulate predominantly in promoters and 5' genic regions, whereas H3K4me1 is distributed within transcribed regions. In addition, H3K4me3-containing genes are highly expressed with low levels of tissue specificity, but H3K4me1 or H3K4me2 may not be directly involved in transcriptional activation. Furthermore, the preferential co-localization of H3K4me3 and H3K27me3 found in mammals does not appear to occur in plants at a genome-wide level, but H3K4me2 and H3K27me3 co-localize at a higher-than-expected frequency. Finally, we found that H3K4me2/3 and DNA methylation appear to be mutually exclusive, but surprisingly, H3K4me1 is highly correlated with CG DNA methylation in the transcribed regions of genes.

Conclusions: H3K4me plays widespread roles in regulating gene expression in plants. Although many aspects of the mechanisms and functions of H3K4me appear to be conserved among all three kingdoms, we observed significant differences in the relationship between H3K4me and transcription or other epigenetic pathways in plants and mammals.

Figure 1

Distribution of H3K4me in the Arabidopsis genome. (a) Chromosomal distribution of H3K4me. Top row: the total length of repetitive sequences (y-axis, left-side scale) and number of genes per 100 kb (y-axis, right-side scale). Bottom panels: chromosomal distribution of H3K4me1, H3K4me2 and H3K4me3. X-axis: chromosomal position; y-axis: the total length of genomic regions containing H3K4me1, H3K4me2 and H3K4me3 per 100 kb, respectively. Arrows indicate the heterochromatic knob on chromosome 4. (b) Local distribution of H3K4me1, H3K4me2, H3K4me3, other epigenetic marks (DNA methylation, H3K9me2, H3K27me3, nucleosome density, small RNAs) and transcription activity in an approximately 40-kb euchromatic region on chromosome 1. Repetitive sequences are shown as filled red boxes on top. Individual genes are shown in open red boxes (arrows indicate direction of transcription; filled light blue boxes, exons; light blue lines, introns). Distribution of H3K4me on the gene labeled by a red asterisk is enlarged and shown in detail at the bottom.

Figure 2

Distribution of H3K4me relative to genes. (a) Distribution of H3K4me at the 5' and 3' ends of genes. 'Isolated' genes are divided into four groups according to their length (see text for details). Genes belonging to each length group were aligned at the transcription start sites, and the percentage of genes containing H3K4me in their promoters or 5' ends is determined at 200-bp intervals (left y-axis). Similarly, genes belonging to each length group were aligned at the end of transcribed regions, and the percentage of genes containing H3K4me in their 3' ends or downstream flanking regions is determined at 200-bp intervals (right y-axis). The first and last 500 bp, 1 kb, 1.5 kb and 2 kb are shown for genes that are 1 to 2 kb, 2 to 3 kb, 3 to 4 kb and >4 kb in length, respectively. (b) Distribution of H3K4me across genes. Each gene (thick horizontal bar) was divided into 20 intervals (5% each interval), and the 1-kb regions upstream and downstream of each gene (thin horizontal bars) were divided into 50-bp intervals. The percentage of genes with H3K27me3 in each interval was graphed (y-axis). (c) Relationship between gene length and H3K4me. Genes are divided into eight categories according to the combination of H3K4me (see text for details), and the percentage of genes within each length group that are associated with a particular combination of H3K4me is shown (y-axis). (d) Length distribution of genes associated with different combinations of H3K4me. X-axis: gene length in kb (200 bp per bin); y-axis: the percentage of genes associated with a particular combination of H3K4me that are of the corresponding length. A small number of genes longer than 8 kb are not shown.

Figure 3

Genes with different expression levels and patterns are associated with different combinations of H3K4me. (a) Distribution of expression levels of genes associated with different combinations of H3K4me. X-axis: gene expression level determined in a previous study (log₂ scale) [42]. Y-axis: the percentage of genes with corresponding H3K4me combination and expression level. (b) The degree of tissue-specific expression of genes associated with different combinations of H3K4me, as measured by entropy (x-axis). Y-axis: the percentage of genes with corresponding H3K4me combination and entropy values.

Figure 4

Comparisons of H3K4me accumulated in wild-type Arabidopsis (Wt, green) and the met1 mutant (light brown). Left: chromosome-level changes in H3K4me, showing the ectopic accumulation of H3K4me in the pericentromeric heterochromatin. Chromosome 5 is shown as an example (Wt, green; met1, light brown). X-axis: chromosome position; y-axis: the percentage of H3K4me on chromosome 5 in the corresponding region (in 100 kb bins). Right: local changes in DNA methylation, H3K4me and transcription in a euchromatic region (top right) and a heterochromatic region (bottom right) on chromosome 5. The five genes shown in the euchromatic region likely encode cellular proteins and their expression patterns are unaffected in the met1 mutant. These are (from left to right): At5g56210, WPP-DOMAIN INTERACTING PROTEIN 2; At5g56220, nucleoside-triphosphatase; At5g56230, prenylated rab acceptor (PRA1) family protein; At5g56240, unknown protein. The six genes shown in the heterochromatic region are all transposon-encoded genes. These are (from left to right): At5g32925, CACTA-like transposase; At5g32950, CACTA-like transposase, At5g32975, similar to En/Spm-like transposon protein; At5g33000, Transposable element gene; At5g33025, gypsy-like retrotransposon; At5g33050, gypsy-like retrotransposon. Note that the overexpression of At5g32950 and At5g33050 is associated with ectopic accumulation of H3K4me.