Flanking sequence preference modulates de novo DNA methylation in the mouse genome

Mammalian de novo DNA methyltransferases (DNMT) are responsible for the establishment of cell-type-specific DNA methylation in healthy and diseased tissues. Through genome-wide analysis of de novo methylation activity in murine stem cells we uncover that DNMT3A prefers to methylate CpGs followed by cytosines or thymines, while DNMT3B predominantly methylates CpGs followed by guanines or adenines. These signatures are further observed at non-CpG sites, resembling methylation context observed in specialised cell types, including neurons and oocytes. We further show that these preferences result from structural differences in the catalytic domains of the two de novo DNMTs and are not a consequence of differential recruitment to the genome. Molecular dynamics simulations suggest that, in case of human DNMT3A, the preference is due to favourable polar interactions between the flexible Arg836 side chain and the guanine that base-pairs with the cytosine following the CpG. By exchanging arginine to a lysine, the corresponding side chain in DNMT3B, the sequence preference is reversed, confirming the requirement for arginine at this position. This context-dependent enzymatic activity provides additional insights into the complex regulation of DNA methylation patterns.