Determining the conservation of DNA methylation in Arabidopsis

A high- resolution map of DNA methylation in Arabidopsis has recently been generated using high-throughput sequencing of bisulfite-converted DNA. This detailed profile measures the methylation state of most of the cytosines in the Arabidopsis genome, and allows us for the first time to address questions regarding the conservation of methylation across duplicated regions of the genome. To address these questions we measured the degree to which methylation is conserved in both duplicated genes and duplicated non-coding regions of the genome. Methylation is controlled by different mechanisms and methyltransferases depending on the genomic location. Methylation in genes occurs primarily at CG sites and is controlled by the maintenance methyltransferase MET1. In contrast, an RNAi mediated methylation pathway that leads to de novo methylation of asymmetric CHH sites along with CG and CHG sites by the methyltransferase DRM2, drives methylation at tandem and inverted repeats. We find that the cytosine methylation profile is strongly preserved between duplicated genes and repeat regions. The highest level of conservation can be found at CG sites in genes and CHH sites in repeat regions. By constructing substitution matrices between aligned genes we see that methylated cytosines often pair with thymines, which may be explained by the spontaneous deamination of methyl-cytosine to thymine. Despite this observation, we find that methylated cytosines are less often paired with other nucleotides than non-methylated cytosines within gene bodies indicating that they may play an important functional role.