Ligation of highly modified bacteriophage DNA

After digestion by TaqI or nicking by DNAase I, five highly modified bacteriophage DNAs were tested as substrates for T4 DNA ligase. The DNAs used were from phages T4, XP12, PBS1, SP82, and SP15, which contain as a major base either glucosylated 5-hydroxymethylcytosine, 5-methylcytosine, uracil, 5-hydroxymethyluracil, or phosphoglucuronated, glucosylated 5-(4′,5′-dihydroxypentyl)uracil, respectively. The relative ability of cohesive-ended TaqI fragments of these DNAs and of normal, λ DNA to be ligated was as follows: $\text{λ}\text{DNA}\text{=}\text{XP}\text{12}\text{DNA}\text{>}\text{SP}\text{82}\text{DNA}\text{?}\text{nonglucosylated}\text{T}\text{4}\text{DNA}\text{>}\text{T}\text{4}\text{DNA}\text{=}\text{PBS}\text{1}\text{DNA}\text{?}\text{SP}\text{15}\text{DNA}$. TaqI-T4 DNA fragments were also inefficiently ligated by Escherichia coli DNA ligase. However, annealing-independent ligation of DNAase I-nicked T4, PBS1, and λ DNAs was equally efficient. We conclude that the poor ligation of TaqI fragments of T4 and PBS1 DNAs was due to the hydroxymethylation (and glucosylation) of cytosine residues at T4's cohesive ends and the substitution of uracil residues for thymine residues adjacent to PBS1's cohesive ends destabilizing the annealing of the restriction fragments. Only SP15 DNA with its negatively charged, modified base was unable to serve as a substrate for T4 DNA ligase in an annealing-independent reaction; therefore, its modification directly interfered with enzyme binding or catalysis.