Common alleles ofCMT2andNRPE1are major determinants ofde novoDNA methylation variation inArabidopsis thaliana

Abstract

DNA cytosine methylation is an epigenetic mark associated with silencing of transposable elements (TEs) and heterochromatin formation. In plants, it occurs in three sequence contexts: CG, CHG, and CHH (where H is A, T, or C). The latter does not allow direct inheritance of methylation during DNA replication due to lack of symmetry, and methylation must therefore be re-established every cell generation. Genome-wide association studies (GWAS) have previously shown that CMT2 and NRPE1 are major determinants of genome-wide patterns of TE CHH-methylation. Here we instead focus on CHH-methylation of individual TEs and TE-families, allowing us to identify the pathways involved in CHH-methylation simply from natural variation and confirm the associations by comparing them with mutant phenotypes. Methylation at TEs targeted by the RNA-directed DNA methylation (RdDM) pathway is unaffected by CMT2 variation, but is strongly affected by variation at NRPE1, which is largely responsible for the longitudinal cline in this phenotype. In contrast, CMT2-targeted TEs are affected by both loci, which jointly explain 7.3% of the phenotypic variation (13.2% of total genetic effects). There is no longitudinal pattern for this phenotype, however, because the geographic patterns appear to compensate for each other in a pattern suggestive of stabilizing selection. Author SummaryDNA methylation is a major component of transposon silencing, and essential for genomic integrity. Recent studies revealed large-scale geographic variation as well as the existence of major trans-acting polymorphisms that partly explained this variation. In this study, we re-analyze previously published data (The 1001 Epigenomes), focusing on de novo DNA methylation patterns of individual TEs and TE families rather than on genome-wide averages (as was done in previous studies). GWAS of the patterns reveals the underlying regulatory networks, and allowed us to comprehensively characterize 1/17 trans-regulation of de novo DNA methylation and its role in the striking geographic pattern for this phenotype. 1 DNA cytosine-methylation (DNA methylation) is an epigenetic mark associated with 2 diverse molecular functions, such as silencing of transposable elements (TEs) and 3 heterochromatin formation. The majority of plant methylation is found in TEs, and 4 there are three types of DNA methylation contexts: CG and CHG, both of which are 5 symmetric, and CHH, which is not (H is A, T, or C). CG-methylation (mCG) and 6 CHG-methylation (mCHG) can be maintained in a semi-conservative manner during 7 DNA replication by DNA METHYLTRANSFERASE 1 (MET1 ) and 8 CHROMOMETHYLASE 3 (CMT3 ), respectively, whereas CHH-methylation (mCHH) 9 must be re-established every cell generation, presumably by one of two de novo 10 pathways, one involving CHROMOMETHYLASE 2 (CMT2 ), the other RNA-directed 11 DNA methylation (RdDM) [1-3]. CMT2 preferentially methylates heterochromatic 12 non-CG cytosines that are marked by H3 Lys9 (H3K9) di-and tri-methylation [4, 5], 13 whiles RdDM involves small RNAs...