Analyses of methylomes of upland and lowland switchgrass (Panicum virgatum) ecotypes using MeDIP-seq and BS-seq

Abstract

Background: Switchgrass is a crop with many desirable traits for bioenergy production. Plant genomes have high DNA methylation levels throughout genes and transposable elements and DNA methylation is known to play a role in silencing transposable elements. Here we analyzed methylomes in two switchgrass genotypes AP13 and VS16. AP13 is derived from a lowland ecotype and VS16, typically considered drought-tolerant, is derived from an upland ecotype, both genotypes are tetraploid (2n = 4× = 36).

Results: Methylated DNA immunoprecipitation-sequencing (MeDIP-seq) and bisulfite-sequencing (BS-seq) were used to profile DNA methylation in genomic features of AP13 and VS16. The methylation patterns in genes and transposable elements were similar to other plants, however, overall CHH methylation levels were comparatively low. Differentially methylated regions (DMRs) were assessed and a total of 1777 CG-DMRs, 573 CHG-DMRs, and 3 CHH-DMRs were detected between the two genotypes. TEs and their flanking regions were higher than that of genic regions. Different types of TEs had different methylation patterns, but the two LTRs (Copia and Gypsy) were similarly methylated, while LINEs and DNA transposons typically had different methylation patterns. MeDIP-seq data was compared to BS-seq data and most of the peaks generated by MeDIP-seq were confirmed to be highly methylated by BS-seq.

Conclusions: DNA methylation in switchgrass genotypes obtained from the two ecotypes were found similar. Collinear gene pairs in two subgenomes (A and B) were not significantly differentially methylated. Both BS-seq and MeDIP-seq methodologies were found effective. Methylation levels were highest at CG and least in CHH. Increased DNA methylation was seen in TEs compared to genic regions. Exploitation of TE methylations can be a viable option in future crop improvement.

Keywords: MeDIP-Seq; Methylome; Panicum virgatum; Switchgrass; Whole genome bisulfite-sequencing.

Fig. 1

Circos plot of gene density, TE density, and methylation levels of CG, CHG and CHH context across each chromosome of switchgrass genome. Each MeDIP-seq replicate and BS-seq sample, by context, are displayed. The color key for MeDIP-seq rings indicates the density of peak numbers in each sample, not the methylation levels. The methylation signals detected by MeDIP are combinations of CG, CHG and CHH methylation. In BS-Seq each context is represented separately

Fig. 2

Global methylation levels, distribution, and across genomic features for CG, CHG and CHH contexts. a Methylation levels are shown for AP13 and VS16, derived from BS-seq data. b Percentages are shown for AP13 and VS16, derived from BS-seq data. c Violin plot of methylation level in different features

Fig. 3

Meta-plots of DNA methylation level across genes and TEs. AP13 and VS16 methylation levels are shown for 2 kb upstream and downstream regions, as well as a) in the gene body and b) in the TE body

Fig. 4

Methylation patterns of different types of TEs for AP13 (a-c) and VS16 (d-f). a and d show CG methylation levels, b and e show CHG methylation levels, and c and f show CHH methylation levels in Copia, Gypsy, LINE, and DNA TEs

Fig. 5

Heat maps (a-d) and Violin-boxplots (e-h) of DNA methylation levels of CG- and CHG-DMRs

Fig. 6

Distribution of DMRs based on genomic features. Asterisks indicate genomic features with overrepresentation in DMRs

Fig. 7

Distribution of types of TEs that were associated with DMRs

Fig. 8

Distribution of peaks based on genomic features

Fig. 9

Heat map and Violin-boxplots representing DNA methylation levels for common (a and d), hypermethylated peaks and differential peaks (b, c, e and f) in CG, CHG and CHH contexts

Fig. 10

Heat map of MeDIP-Seq signals for hypermethylated-DMRs (a) and hypomethylated-DMRs (b). The color key for indicates the density of peak numbers in each sample, not the methylation levels