CHH Methylation Islands: A Nonconserved Feature of Grass Genomes That Is Positively Associated with Transposable Elements but Negatively Associated with Gene-Body Methylation

Abstract

mCHH islands are peaks of CHH methylation that occur primarily upstream to genes. These regions are actively targeted by the methylation machinery, occur at boundaries between heterochromatin and euchromatin, and tend to be near highly expressed genes. Here we took an evolutionary perspective by studying upstream mCHH islands across a sample of eight grass species. Using a statistical approach to define mCHH islands as regions that differ from genome-wide background CHH methylation levels, we demonstrated that mCHH islands are common and associate with 39% of genes, on average. We hypothesized that islands should be more frequent in genomes of large size, because they have more heterochromatin and hence more need for defined boundaries. We found, however, that smaller genomes tended to have a higher proportion of genes associated with 5’ mCHH islands. Consistent with previous work suggesting that islands reflect the silencing of the edge of transposable elements (TEs), genes with nearby TEs were more likely to have mCHH islands. However, the presence of mCHH islands was not a function solely of TEs, both because the underlying sequences of islands were often not homologous to TEs and because genic properties also predicted the presence of 5’ mCHH islands. These genic properties included length and gene-body methylation (gbM); in fact, in three of eight species the absence of gbM was a stronger predictor of a 5’ mCHH island than TE proximity. In contrast, gene expression level was a positive but weak predictor of the presence of an island. Finally, we assessed whether mCHH islands were evolutionarily conserved by focusing on a set of 2,720 orthologs across the eight species. They were generally not conserved across evolutionary time. Overall, our data establishes additional genic properties that are associated with mCHH islands and suggests that they are not just a consequence of the TE silencing machinery.

High-quality genome and methylomes illustrate features underlying evolutionary success of oaks

“…Measured as the ratio of peak mCHH to whole-genome average mCHH, we find oaks have unusually strong 5ʹ mCHH islands (Supplementary Data 1 ), but it remains to be seen if they also contribute to boundary enforcement. As reported by Martin et al 52 among six grass species, we found the same range of significant associations between the presence of mCHH islands and various genic features. Following their GLM model, mCHH island presence was associated with lower exonic CG methylation (estimate = −0.573 ± 0.06, z = −9.1, p < 0.001), higher gene length (estimate = 2.95 ± 0.33, z = 9.1, p < 0.001), shorter distance from TSS to TE (estimate = −10.16 ± 0.97, z = −10.5, p < 0.001) and higher gene expression (estimate = 5.30 ± 1.76, z = 3.0, p < 0.01) 52 .…”

“…As reported by Martin et al 52 among six grass species, we found the same range of significant associations between the presence of mCHH islands and various genic features. Following their GLM model, mCHH island presence was associated with lower exonic CG methylation (estimate = −0.573 ± 0.06, z = −9.1, p < 0.001), higher gene length (estimate = 2.95 ± 0.33, z = 9.1, p < 0.001), shorter distance from TSS to TE (estimate = −10.16 ± 0.97, z = −10.5, p < 0.001) and higher gene expression (estimate = 5.30 ± 1.76, z = 3.0, p < 0.01) 52 . However, our importance values were lower, and it is quite possible the mCHH islands are simply the result of the type of TEs found near gene boundaries.…”

Gene Body Methylation in Plants: Mechanisms, Functions, and Important Implications for Understanding Evolutionary Processes

“…While one must always be careful that mutants can be complex and may reflect other (unknown) effects, the data again provide little support for the notion that gbM alone prevents aberrant internal transcription initiation in A. thaliana genes ( table 1 ). Interestingly, Choi et al (2020) showed that gbM and H1 play a redundant role in inhibiting aberrant reverse transcription, and Martin et al (2021) hypothesized that another epigenetic mark, CHH islands, may have some redundant function with gbM. These redundancies could explain why we did not detect any change in internal transcription start in A. thaliana gbM mutants because these redundant epigenetic marks may have been functioning in gbM mutants, thus complicating inferences about gbM effects.…”

Identification of Putative Markers of Non-infectious Bud Failure in Almond [Prunus dulcis (Mill.) D.A. Webb] Through Genome Wide DNA Methylation Profiling and Gene Expression Analysis in an Almond × Peach Hybrid Population

“…Therefore, patterns of differential DNA methylation unique to almond could indicate a possible role in NBF development, as well as a putative epigenetic mark associated with NBF, which is desirable to know and further characterize. Interestingly, the presence of CHH islands in plants has recently been described, and hypermethylation of these regions was shown in grasses to have an impact on downstream gene expression ( Martin et al, 2021 ). In addition, CHH methylation has been shown to be involved in vegetative growth.…”