Limits...
Widespread natural variation of DNA methylation within angiosperms

View Article: PubMed Central - PubMed

ABSTRACT

Background: DNA methylation is an important feature of plant epigenomes, involved in the formation of heterochromatin and affecting gene expression. Extensive variation of DNA methylation patterns within a species has been uncovered from studies of natural variation. However, the extent to which DNA methylation varies between flowering plant species is still unclear. To understand the variation in genomic patterning of DNA methylation across flowering plant species, we compared single base resolution DNA methylomes of 34 diverse angiosperm species.

Results: By analyzing whole-genome bisulfite sequencing data in a phylogenetic context, it becomes clear that there is extensive variation throughout angiosperms in gene body DNA methylation, euchromatic silencing of transposons and repeats, as well as silencing of heterochromatic transposons. The Brassicaceae have reduced CHG methylation levels and also reduced or loss of CG gene body methylation. The Poaceae are characterized by a lack or reduction of heterochromatic CHH methylation and enrichment of CHH methylation in genic regions. Furthermore, low levels of CHH methylation are observed in a number of species, especially in clonally propagated species.

Conclusions: These results reveal the extent of variation in DNA methylation in angiosperms and show that DNA methylation patterns are broadly a reflection of the evolutionary and life histories of plant species.

Electronic supplementary material: The online version of this article (doi:10.1186/s13059-016-1059-0) contains supplementary material, which is available to authorized users.

No MeSH data available.


a Patterns of methylation across conserved non-coding sequences (CNS) for mCG (blue), mCHG (green), and mCHH (maroon). b Percentage of genes with mCHH islands 2 kb upstream or downstream. c Association of upstream mCHH islands with gene expression. Genes are divided into not-expressed (NE) and quartiles of increasing expression. ** indicates a difference in proportion from the fourth quartile at p < 0.01. d Patterns of upstream mCHH islands for mCG (blue), mCHG (green), and mCHH (maroon)
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC5037628&req=5

Fig6: a Patterns of methylation across conserved non-coding sequences (CNS) for mCG (blue), mCHG (green), and mCHH (maroon). b Percentage of genes with mCHH islands 2 kb upstream or downstream. c Association of upstream mCHH islands with gene expression. Genes are divided into not-expressed (NE) and quartiles of increasing expression. ** indicates a difference in proportion from the fourth quartile at p < 0.01. d Patterns of upstream mCHH islands for mCG (blue), mCHG (green), and mCHH (maroon)

Mentions: Outside of the gene body, DNA methylation might have an impact on gene expression through the DNA methylation of neighboring transcription factor binding sites (TFBS) or other regulatory elements. To date, there is limited in vivo evidence of such effects in plants, although the recent example of repressor of silencing 1 (ROS1) hints at this possibility [84, 85]. In vitro evidence also supports the possibility of DNA methylation inhibiting and in some cases, promoting, transcription factor binding [86]. Conserved non-coding sequences contain many important regulatory elements, including TFBS [87, 88]. We identified CNS regions for a sample of species across the phylogeny and plotted DNA methylation levels (Fig. 6a, Additional file 3: Figure S24). DNA methylation in all three contexts was depleted across these regions, compared to outside. Locations of CNS regions were defined as either proximal (within 1 kbps), distal (>1 kbps), within untranslated regions (UTR), or within introns. Similar patterns were observed for CNS regions whether they were located proximally or distally to a gene (Additional file 3: Figure S24). UTR and intronic CNS sequences do show elevated levels of mCG in comparison, which might result from elevated mCG levels across the gene bodies of gbM genes. In Z. mays, high mCHH is enriched in the upstream and downstream regions of highly expressed genes and are termed mCHH islands [36, 37]. We identified mCHH islands 2 kb upstream and downstream of annotated genes for each species, finding that the percentage of genes with such regions varied considerably across species (Fig. 6b). Although some species other than Z. mays also show an association between mCHH islands and gene expression, many showed no such association, indicating no universal causal relationship between the two (Fig. 6c, Additional file 3: Figure S25). As has been observed previously in Z. mays, mCG and mCHG levels are generally higher on the distal side of the mCHH island to the gene (Fig. 6d, Additional file 3: Figure S26) [37]. However, this difference in DNA methylation level is much less pronounced in most other species as compared to Z. mays (Additional file 3: Figure S26). It is thought that these differences in DNA methylation on proximal versus distal sides of mCHH islands mark euchromatin-heterochromatin boundaries [37]. Indeed, mCHH islands are often associated with transposons [36, 37], however, there was no correlation found between the total number of repeats in the genome and the number of genes with mCHH islands (Additional files 1 and 3: Table S3 and Figure S27a). When correlated to the percentage of genes with repeats 2 kb upstream or downstream, both upstream and downstream mCHH islands are correlated (upstream p value = 4.5 × 10–5, downstream p value = 9.3 × 10–4) (Additional files 1 and 3: Table S3 and Figure S27b). While there was a correlation between the total repeat content and the percentage of upstream and downstream repeats (upstream p value = 1.3 × 10–2, downstream p value = 1.6 × 10–3), there were numerous outlying species which may explain the lack of correlation between mCHH islands and total repeat content (Additional files 1 and 3: Table S3 and Figure S27c). This supports a hypothesis that transposon distribution as opposed to transposon load alone is critical in shaping the epigenome.Fig. 6


Widespread natural variation of DNA methylation within angiosperms
a Patterns of methylation across conserved non-coding sequences (CNS) for mCG (blue), mCHG (green), and mCHH (maroon). b Percentage of genes with mCHH islands 2 kb upstream or downstream. c Association of upstream mCHH islands with gene expression. Genes are divided into not-expressed (NE) and quartiles of increasing expression. ** indicates a difference in proportion from the fourth quartile at p < 0.01. d Patterns of upstream mCHH islands for mCG (blue), mCHG (green), and mCHH (maroon)
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC5037628&req=5

Fig6: a Patterns of methylation across conserved non-coding sequences (CNS) for mCG (blue), mCHG (green), and mCHH (maroon). b Percentage of genes with mCHH islands 2 kb upstream or downstream. c Association of upstream mCHH islands with gene expression. Genes are divided into not-expressed (NE) and quartiles of increasing expression. ** indicates a difference in proportion from the fourth quartile at p < 0.01. d Patterns of upstream mCHH islands for mCG (blue), mCHG (green), and mCHH (maroon)
Mentions: Outside of the gene body, DNA methylation might have an impact on gene expression through the DNA methylation of neighboring transcription factor binding sites (TFBS) or other regulatory elements. To date, there is limited in vivo evidence of such effects in plants, although the recent example of repressor of silencing 1 (ROS1) hints at this possibility [84, 85]. In vitro evidence also supports the possibility of DNA methylation inhibiting and in some cases, promoting, transcription factor binding [86]. Conserved non-coding sequences contain many important regulatory elements, including TFBS [87, 88]. We identified CNS regions for a sample of species across the phylogeny and plotted DNA methylation levels (Fig. 6a, Additional file 3: Figure S24). DNA methylation in all three contexts was depleted across these regions, compared to outside. Locations of CNS regions were defined as either proximal (within 1 kbps), distal (>1 kbps), within untranslated regions (UTR), or within introns. Similar patterns were observed for CNS regions whether they were located proximally or distally to a gene (Additional file 3: Figure S24). UTR and intronic CNS sequences do show elevated levels of mCG in comparison, which might result from elevated mCG levels across the gene bodies of gbM genes. In Z. mays, high mCHH is enriched in the upstream and downstream regions of highly expressed genes and are termed mCHH islands [36, 37]. We identified mCHH islands 2 kb upstream and downstream of annotated genes for each species, finding that the percentage of genes with such regions varied considerably across species (Fig. 6b). Although some species other than Z. mays also show an association between mCHH islands and gene expression, many showed no such association, indicating no universal causal relationship between the two (Fig. 6c, Additional file 3: Figure S25). As has been observed previously in Z. mays, mCG and mCHG levels are generally higher on the distal side of the mCHH island to the gene (Fig. 6d, Additional file 3: Figure S26) [37]. However, this difference in DNA methylation level is much less pronounced in most other species as compared to Z. mays (Additional file 3: Figure S26). It is thought that these differences in DNA methylation on proximal versus distal sides of mCHH islands mark euchromatin-heterochromatin boundaries [37]. Indeed, mCHH islands are often associated with transposons [36, 37], however, there was no correlation found between the total number of repeats in the genome and the number of genes with mCHH islands (Additional files 1 and 3: Table S3 and Figure S27a). When correlated to the percentage of genes with repeats 2 kb upstream or downstream, both upstream and downstream mCHH islands are correlated (upstream p value = 4.5 × 10–5, downstream p value = 9.3 × 10–4) (Additional files 1 and 3: Table S3 and Figure S27b). While there was a correlation between the total repeat content and the percentage of upstream and downstream repeats (upstream p value = 1.3 × 10–2, downstream p value = 1.6 × 10–3), there were numerous outlying species which may explain the lack of correlation between mCHH islands and total repeat content (Additional files 1 and 3: Table S3 and Figure S27c). This supports a hypothesis that transposon distribution as opposed to transposon load alone is critical in shaping the epigenome.Fig. 6

View Article: PubMed Central - PubMed

ABSTRACT

Background: DNA methylation is an important feature of plant epigenomes, involved in the formation of heterochromatin and affecting gene expression. Extensive variation of DNA methylation patterns within a species has been uncovered from studies of natural variation. However, the extent to which DNA methylation varies between flowering plant species is still unclear. To understand the variation in genomic patterning of DNA methylation across flowering plant species, we compared single base resolution DNA methylomes of 34 diverse angiosperm species.

Results: By analyzing whole-genome bisulfite sequencing data in a phylogenetic context, it becomes clear that there is extensive variation throughout angiosperms in gene body DNA methylation, euchromatic silencing of transposons and repeats, as well as silencing of heterochromatic transposons. The Brassicaceae have reduced CHG methylation levels and also reduced or loss of CG gene body methylation. The Poaceae are characterized by a lack or reduction of heterochromatic CHH methylation and enrichment of CHH methylation in genic regions. Furthermore, low levels of CHH methylation are observed in a number of species, especially in clonally propagated species.

Conclusions: These results reveal the extent of variation in DNA methylation in angiosperms and show that DNA methylation patterns are broadly a reflection of the evolutionary and life histories of plant species.

Electronic supplementary material: The online version of this article (doi:10.1186/s13059-016-1059-0) contains supplementary material, which is available to authorized users.

No MeSH data available.