DNA methylation is an essential epigenetic modification responsible for the regulation of gene expression, suppression of transposable elements, and maintenance of genomic stability. It primarily occurs at CpG sites in vertebrates and is catalyzed by DNA methyltransferases (DNMTs). In mammals, DNMT3A and DNMT3B are the main de novo methyltransferases, with DNMT3L serving as a regulatory cofactor. A more recently discovered DNMT, DNMT3C, emerged through duplication of DNMT3B in muroid rodents. DNMT3C plays a critical role in silencing young retrotransposons in the male germline.

The main goal of this study was to characterize the flanking sequence preferences of DNMT3C and investigate their evolutionary significance and biological implications, particularly their relationship with retrotransposon silencing. Using Deep Enzymology, the authors mapped the methylation activity of DNMT3C across thousands of DNA sequence contexts. They found that DNMT3C shows a distinct preference for cytosines at positions -2 and -1 relative to the target CpG site—markedly different from DNMT3A and DNMT3B, which prefer T at -2.

Comparison with DNMT3A and DNMT3B revealed that:

• DNMT3C uniquely prefers C at -2 and -1 positions.
• Preferences at +1 to +4 positions were more similar to DNMT3B.

This unique sequence specificity was confirmed using synthetic DNA substrates designed with flanking sequences optimized for DNMT3C and DNMT3B. Through structural and mutational analysis, the study identified two critical amino acid residues in DNMT3C—C543 and V547—that determine its flanking sequence preferences. These residues are located in the catalytic domain and influence how DNMT3C interacts with DNA:

• C543 (instead of N656 in DNMT3B) likely disrupts a DNA contact in the major groove.
• V547 (instead of A660 in DNMT3B) is bulkier and alters the geometry of the DNA binding loop.

Mutating these residues to their DNMT3B counterparts (C543N/V547A) fully reversed DNMT3C’s preferences to resemble DNMT3B, indicating their central role in defining DNMT3C specificity. These residues are evolutionarily conserved in muroids, underscoring their functional importance.

Using whole-genome bisulfite sequencing data from mouse ES cells engineered to express only DNMT3C (with all other DNMT and TET enzymes knocked out), the authors demonstrated a strong correlation (r = 0.83) between DNMT3C’s flanking sequence preferences and actual methylation patterns in the genome. CpG sites with CCCG or CGGG contexts (matching DNMT3C’s preferred flanks) showed the highest methylation, directly linking enzyme specificity to biological methylation outcomes.

Next, the study analyzed sequences of retrotransposons targeted by DNMT3C, including L1Md_A, L1Md_T, L1Md_Gf, and ERVK elements like IAPEz. They found a significant enrichment of CpG sites in CCCG or CGGG contexts in the promoter regions and full-length sequences of these elements. In contrast, retrotransposons not targeted by DNMT3C lacked such sequence enrichment. This supports the idea that DNMT3C has co-evolved with its retrotransposon targets—acquiring sequence preferences that match their CpG context, thereby enhancing methylation and silencing efficacy.

This study provides mechanistic insights into how DNA methyltransferases evolve substrate specificity to meet biological demands. DNMT3C exemplifies how protein sequence variation and selective pressure from transposon activity can lead to the emergence of specialized enzymes with dedicated roles. DNMT3C is a specialized DNA methyltransferase with unique sequence preferences for cytosines at -2 and -1 positions. This specificity is mediated by two conserved residues and aligns with the sequence features of young retrotransposons, which are its biological targets in the mouse germline. The study offers a compelling example of molecular co-evolution between an epigenetic enzyme and transposable elements, advancing our understanding of genome defense mechanisms.