Arabidopsis replacement histone variant H3.3 occupies promoters of regulated genes

Abstract

Background: Histone variants establish structural and functional diversity of chromatin by affecting nucleosome stability and histone-protein interactions. H3.3 is an H3 histone variant that is incorporated into chromatin outside of S-phase in various eukaryotes. In animals, H3.3 is associated with active transcription and possibly maintenance of transcriptional memory. Plant H3 variants, which evolved independently of their animal counterparts, are much less well understood.

Results: We profile the H3.3 distribution in Arabidopsis at mono-nucleosomal resolution using native chromatin immunoprecipitation. This results in the precise mapping of H3.3-containing nucleosomes, which are not only enriched in gene bodies as previously reported, but also at a subset of promoter regions and downstream of the 3' ends of active genes. While H3.3 presence within transcribed regions is strongly associated with transcriptional activity, H3.3 at promoters is often independent of transcription. In particular, promoters with GA motifs carry H3.3 regardless of the gene expression levels. H3.3 on promoters of inactive genes is associated with H3K27me3 at gene bodies. In addition, H3.3-enriched plant promoters often contain RNA Pol II considerably upstream of the transcriptional start site. H3.3 and RNA Pol II are found on active as well as on inactive promoters and are enriched at strongly regulated genes.

Conclusions: In animals and plants, H3.3 organizes chromatin in transcribed regions and in promoters. The results suggest a function of H3.3 in transcriptional regulation and support a model that a single ancestral H3 evolved into H3 variants with similar sub-functionalization patterns in plants and animals.

Figure 1

Genome-wide profiling of H3.3 in Arabidopsis. (A) H3.3-YFP is targeted to the nucleus. Confocal images of nuclei from Arabidopsis roots showing H3.3-YFP (green) and a DNA-counterstain with DRAQ5 (red). Open arrowheads mark heterochromatic chromocenters. Scale bars: 2 μm. (B) Genome-wide profiles of H3.3 enrichment (red), transposable element (TE; yellow) and gene density (green). Bars mark centromeric and pericentric heterochromatin. (C) Nucleosome positions were detected by fitting the smoothed ChIP-chip signal (black line with squares) to a parabola model (green dashed line). The position of the maximum of each fitted parabola peak was regarded as the center of a nucleosome (cross). (D) Deconvolution of H3.3 enrichment at positions of nucleosomes (black) into a background (green) and a specific (blue) component using a mixture model. The red area indicates nucleosomes that have higher than 75% probability belonging to the specific component. (E) Comparison of observed and expected distribution of H3.3 nucleosomes across different genomic features (that is, protein coding gene, pseudogene, transposable element gene, transposable element, microRNA (miRNA), tRNA, non-coding RNA (ncRNA)).

Figure 2

H3.3 is enriched at transcribed genes and has a strong 3′ bias. (A) Genes with at least one H3.3 nucleosome (right) have higher transcript abundance than genes without H3.3 (left). Horizontal bars indicate median values. Expression data are from [22] and in logarithmic scale. (B,C) H3.3 nucleosome (B) and total nucleosome (C) occupancy were measured as H3.3-YFP-ChIP and histone-ChIP signals, respectively. Metagene plots across gene bodies (blue bar) were constructed between -3 kb and +3 kb. Genes were grouped according to transcript abundance from low (red) to high (violet). (D) Validation of H3.3 at promoters by ChIP-qPCR. 5′ Upstream regions of 10 genes that had H3.3-enrichment at promoters in the ChIP-chip experiment were tested. ACTIN7 and a CINFUL-like locus served as controls. Error bars represent standard error of mean, n = 3. GFP, green fluorescent protein.

Figure 3

H3.3 presence coincides with RNA Polymerase II distribution. (A) H3.3 enrichment scores of promoter-associated and TTS-associated H3.3-enriched nucleosomes. Horizontal bars indicate median values. (B) Metagene plots across gene bodies (blue bar) were constructed between -3 kb and +3 kb for H3.3 signals normalized to histone signals. Here and in subsequent metagene plots, positions plotted on the x-axis are relative to annotated transcription start site and TTSs and given in kilobases for upstream and downstream sequences and as percentages for gene bodies. Genes were grouped according to the presence of H3.3-enriched nucleosomes in the promoter (red), close to the TTS (green) or both (blue). (C) Immunocytological detection (upper) of H3.3 (left) and RNA Polymerase II (Pol II; center). An overlay is shown at the right. Scale bars: 5 μm. The diagram (lower) shows H3.3 (green) and Pol II (red) signals along a section through the nucleus (yellow line in the overlay image). (D) Metagene plot of Pol II across gene bodies (blue bar) between -3 kb and +3 kb. Pol II ChIP-chip data are from [23]. Genes were grouped according to the presence of H3.3-enriched nucleosomes in the promoter (red), close to the TTS (green) or both (blue).

Figure 4

Genes with H3.3-enriched nucleosomes in the promoter have weak expression but strong regulation. (A) Transcript abundance for all genes (white), genes with H3.3-enriched nucleosomes in the promoter (red), close to the TTS (green) or both (blue). Expression data are from [22] and in logarithmic scale. (B) Expression entropy for all genes (white), genes with H3.3-enriched nucleosomes in the promoter (red), close to the TTS (green) or both TTS (blue). Expression entropy is based on data from [26]. Horizontal bars indicate median values.

Figure 5

Genes with H3.3-enriched nucleosomes in the promoter are marked by H3K27me3. Metagene plot of H3K27me3 for all repressed genes (black) and for repressed genes with promoter H3.3 (red) across gene bodies (blue bar) between -3 kb and +3 kb. H3K27me3 ChIP-chip data are from [29].

Figure 6

DNA mCG methylation at promoters does not coincide with H3.3 enrichment. (A) Metagene plot of mCG across gene bodies (blue bar) between -3 kb and +3 kb. mCG data are from [40]. Genes were grouped according to the presence of H3.3-containing nucleosomes in the promoter (red), close to the TTS (green) or both (blue). (B) H3.3, mCG, Pol II and H3K36me2 profiles around TTS. Signals were plotted from the TSS up to 1 kb proximal to the TTS for 5,000 random sampled genes ordered by their gene body lengths. For genes that are longer than 5 kb, only 5 kb of the gene body are shown. Epigenome profiles are represented by a heat map color code with red and blue representing highest and lowest values, respectively. White dashed lines indicate the TTS positions. Data are from [23,29,40]. (C) Metagene plot of H3.3 signal (left) and mCG (right) between -2 kb and +1 kb of the TSS. mCG data are from [40]. Genes were grouped according to the presence of H3.3-containing nucleosomes at the promoter (red), close to the TTS (green) or both (blue). Grey areas indicate -800 to -200 bp of the TSS.

Figure 7

Promoter cis-regulatory sequences affect local H3.3 incorporation. (A-C) Metagene plots of H3.3 signals across gene bodies (blue bar) between -3 kb and +3 kb for genes with promoters that carry TATA elements (A), GA elements (B) or neither TATA nor GA elements (coreless) (C). Genes were first grouped according to transcript abundance, and genes with promoters of each type are presented separately. (B) H3.3 signals at promoters with GA elements (red) or promoters of the top 25% most strongly expressed genes (blue).

Figure 8

DNA at H3.3-enriched nucleosomes is highly accessible. (A,B) Metagene plots of inaccessibility of DNA around nucleosomes without (A) or with (B) H3.3 between -600 bp and +600 bp. Blue bars, nucleosome covered region. Data are from [22]. Shown are mean (black and red) and 95% confidence intervals (grey).