Figure 2. DNMT3A shows selectivity for residues flanking the target CpG site at positions −2 and +2.

The sequences flanking the most (panels A, D) and least (panels B, E) methylated sites for DNMT3A were extracted and aligned as described in the text. Residues that were enriched at any position in the 4 base-pairs surrounding the target CpG site on each side (x-axis) are represented in the Logos format whereby sequence enrichment at each position is indicated by the size of each letter in bits (y-axis; 2 reflects perfect conservation while 0 reflects a random distribution). The corresponding P-values measuring the enrichment of a given residue over the input DNA sequence are indicated for positions showing significant preference. Information was derived either from our episomal assay in vivo (panels A and B) or using purified DNMT3A complexes in vitro (panels D and E). Panel C corresponds to the entire set of CpG sites analyzed here – these sites showed no intrinsic sequence enrichment.