Virus-encoded glycosyltransferases hypermodify DNA with diverse glycans.

Enzymatic modification of DNA nucleobases can coordinate gene expression, nuclease protection, or mutagenesis. We recently discovered a clade of phage-specific cytosine methyltransferase (MT) and 5-methylpyrimidine dioxygenase (5mYOX) enzymes that produce 5-hydroxymethylcytosine (5hmC) as a precursor for enzymatic hypermodifications on viral genomes. Here, we identify phage MT- and 5mYOX-associated glycosyltransferases (GTs) that catalyze linkage of diverse sugars to 5hmC nucleobase substrates. Metavirome mining revealed thousands of biosynthetic gene clusters containing enzymes with predicted roles in cytosine sugar hypermodification. We developed a platform for high-throughput screening of GT-containing pathways, relying on the Escherichia coli metabolome as a substrate pool. We successfully reconstituted several pathways and isolated diverse sugar modifications appended to cytosine, including mono-, di-, or tri-saccharides comprised of hexoses, N-acetylhexosamines, or heptose. These findings expand our knowledge of hypermodifications on nucleic acids and the origins of corresponding sugar-installing enzymes.

Phage-encoded ten-eleven translocation dioxygenase (TET) is active in C5-cytosine hypermodification in DNA.

TET/JBP (ten-eleven translocation/base J binding protein) enzymes are iron(II)- and 2-oxo-glutarate-dependent dioxygenases that are found in all kingdoms of life and oxidize 5-methylpyrimidines on the polynucleotide level. Despite their prevalence, few examples have been biochemically characterized. Among those studied are the metazoan TET enzymes that oxidize 5-methylcytosine in DNA to hydroxy, formyl, and carboxy forms and the euglenozoa JBP dioxygenases that oxidize thymine in the first step of base J biosynthesis. Both enzymes have roles in epigenetic regulation. It has been hypothesized that all TET/JBPs have their ancestral origins in bacteriophages, but only eukaryotic orthologs have been described. Here we demonstrate the 5mC-dioxygenase activity of several phage TETs encoded within viral metagenomes. The clustering of these TETs in a phylogenetic tree correlates with the sequence specificity of their genomically cooccurring cytosine C5-methyltransferases, which install the methyl groups upon which TETs operate. The phage TETs favor Gp5mC dinucleotides over the 5mCpG sites targeted by the eukaryotic TETs and are found within gene clusters specifying complex cytosine modifications that may be important for DNA packaging and evasion of host restriction.