Inter-individual variation in DNA methylation is largely restricted to tissue-specific differentially methylated regions in maize

Abstract

BackgroundVariation in DNA methylation across distinct genetic populations, or in response to specific biotic or abiotic stimuli, has typically been studied in leaf DNA from pooled individuals using either reduced representation bisulfite sequencing, whole genome bisulfite sequencing (WGBS) or methylation sensitive amplified polymorphism (MSAP). The latter represents a useful alterative when sample size is large, or when analysing methylation changes in genomes that have yet to be sequenced. In this study we compared variation in methylation across ten individual leaf and endosperm samples from maize hybrid and inbred lines using MSAP. We also addressed the methodological implications of analysing methylation variation using pooled versus individual DNA samples, in addition to the validity of MSAP compared to WGBS. Finally, we analysed a subset of variable and non-variable fragments with respect to genomic location, vicinity to repetitive elements and expression patterns across leaf and endosperm tissues.ResultsOn average, 30% of individuals showed inter-individual methylation variation, mostly of leaf and endosperm-specific differentially methylated DNA regions. With the exception of low frequency demethylation events, the bulk of inter-individual methylation variation (84 and 80% in leaf and endosperm, respectively) was effectively captured in DNA from pooled individuals. Furthermore, available genome-wide methylation data largely confirmed MSAP leaf methylation profiles. Most variable methylation that mapped within genes was associated with CG methylation, and many of such genes showed tissue-specific expression profiles. Finally, we found that the hAT DNA transposon was the most common class II transposable element found in close proximity to variable DNA regions.ConclusionsThe relevance of our results with respect to future studies of methylation variation is the following: firstly, the finding that inter-individual methylation variation is largely restricted to tissue-specific differentially methylated DNA regions, underlines the importance of tissue-type when analysing the methylation response to a defined stimulus. Secondly, we show that pooled sample-based MSAP studies are methodologically appropriate to study methylation variation. Thirdly, we confirm that MSAP is a powerful tool when WGBS is not required or feasible, for example in plant species that have yet to be sequenced.