| Abstract: | ![]()
DNA methylation is important for the regulation of gene expression and the silencing of transposons in plants. Here we present genome-wide methylation patterns at single-base pair resolution for cassava (Manihot esculenta, cultivar TME 7), a crop with a substantial impact in the agriculture of subtropical and tropical regions. On average, DNA methylation levels were higher in all three DNA sequence contexts (CG, CHG, and CHH, where H equals A, T, or C) than those of the most well-studied model plant Arabidopsis thaliana. As in other plants, DNA methylation was found both on transposons and in the transcribed regions (bodies) of many genes. Consistent with these patterns, at least one cassava gene copy of all of the known components of Arabidopsis DNA methylation pathways was identified. Methylation of LTR transposons (GYPSY and COPIA) was found to be unusually high compared with other types of transposons, suggesting that the control of the activity of these two types of transposons may be especially important. Analysis of duplicated gene pairs resulting from whole-genome duplication showed that gene body DNA methylation and gene expression levels have coevolved over short evolutionary time scales, reinforcing the positive relationship between gene body methylation and high levels of gene expression. Duplicated genes with the most divergent gene body methylation and expression patterns were found to have distinct biological functions and may have been under natural or human selection for cassava traits.DNA methylation plays an important role in the regulation of the expression of genes and the maintenance of transposable element (TE) silencing. In contrast to animals, in which methylation is often restricted to the CG context, plants exhibit robust methylation in every possible context CG, CHG (H is A, T, or C), and CHH. Previous research has identified different pathways responsible for the maintenance and establishment of DNA methylation patterns. In Arabidopsis thaliana, METHYLTRANSFERASE1 (MET1), a homolog of mammalian Dnmt1, mainly maintains methylation at the CG context, whereas CHROMOMETHYLASE3 (CMT3) mainly maintains CHG methylation. DOMAINS REARRANGED METHYLTRANSFERASE2 (DRM2) and CHROMOMETHYLASE2 (CMT2) maintain CHH methylation in the chromosome arms and pericentromeric regions, respectively (1–3). On the other hand, establishment of DNA methylation is performed by DRM2 through a complex pathway termed RNA-directed DNA methylation (RdDM) (4).To date, the majority of our knowledge about DNA methylation is derived from the model plant Arabidopsis. These studies have allowed the identification of different components involved in different methylation pathways, the genome-wide identification of methylation patterns, and the study of effects of DNA methylation on gene expression. The knowledge acquired from Arabidopsis can now be used as the basis for investigations of methylation in agronomically important plants. However, thus far very few crop species have been subjected to detailed DNA methylation studies (5). Cassava (Manihot esculenta) is cultivated for its starch-rich tuberous roots and is one of the world’s most important staple crops, especially in tropical America, Africa, and Asia (6). Cassava is a source of carbohydrates for nearly a billion people, but it is especially important for a large portion of Africa, where it serves as a subsistence crop because of its ability to tolerate drought and grow on poor soils, conditions unsuitable for rice and maize (6, 7). The genome sequence of cassava has been described recently with an estimated genome size of roughly 760 million base pairs (7). We have used bisulfite sequencing (BS-seq) to examine DNA methylation in cassava at single-base pair resolution. Broadly, the pattern of DNA methylation of both protein-coding genes and TEs is similar to other plants, although DNA methylation levels in cassava are higher than those in Arabidopsis. LTR retrotransposons, such as GYPSY and COPIA, tend to be more heavily methylated than other TEs. Interestingly, differentially expressed gene pairs derived from the last genome duplication tend to show differential gene body methylation, with the highly expressed paralogs displaying significantly higher gene body methylation. We also find that the most differentially gene body-methylated paralogs have distinct biological functions compared with genes that have maintained similar gene body methylation patterns. |
