Biosynthesis of glycosylated 5-hydroxycytosine in the DNA of diverse viruses

Characterizing viral strategies for circumventing cellular defenses is critical to understanding the biology and therapeutic application of bacteriophages as antimicrobials. Many bacteriophages synthesize complex modifications (i.e. hypermodifications) to the nucleobases of their virion DNA in order to circumvent the endonuclease-based defenses of their hosts. To date, most hypermodified pyrimidines in viruses are synthesized via group transfer to a pre-existing hydroxyl moiety in DNA during the later stages of lytic development leading up to packaging of the viral chromosome into the capsid. Typically, these occur at hydroxymethylcytosine or hydroxymethyluracil. We find mono- and poly-arabinosylated cytidines completely replacing canonical cytidine in the native DNA of the Escherichia phage RB69 and Acinetobacter baumannii phage DLP3. In both cases, an arabinosyl moiety is connected directly to the pyrimidine through an ether linkage, with the bridging oxygen originating from 5-hydroxycytosine (5hoC) in DNA. The 5hoC is synthesized as a mononucleotide by a virally encoded thymidylate synthase homolog. We find diverse virally encoded thymidylate synthase homologs, including the Rhizobium phage RL38J1, are capable of producing 5hoC, suggesting this base modification is a starting point for glycodiverse cytosine derivatives. Characterizing these cytosine derivatives adds to our understanding of viral anti-defense mechanisms.