Pyrimidine ring fragmentation products. Effects of lesion structure and sequence context on mutagenesis

Free radicals produce a broad spectrum of damages to DNA, a major proportion of which includes ring fragmentation and contraction products of DNA bases as well as abasic sites. In this study, the mutagenic potential of two pyrimidine ring fragmentation products, urea and α-ureidoisobuturic acid (UBA), was analyzed using the i(-d) region of the Escherichia coli lacI gene contained in a single-stranded f1-K12 phage hybrid vector. Single-stranded DNA was used so that the in vivo interactions between the damage and the DNA polymerase could be assessed in the absence of excision repair. The i(-d) region contains 20 mutable thymine sites so that 20 separate sequence contexts containing a unique lesion at a position of thymine can be analyzed simultaneously. Urea and UBA residues were uniquely introduced into f1-K12 DNA by chemical and enzymatic methods and primer extension and piperidine analysis of the damage-containing template molecules demonstrated that the potential mutable thymine sites contained randomly distributed lesions. Both fragmentation products were poorly bypassed by DNA polymerases in vitro and in the cell; although in the presence of SOS-induction, UBA was bypassed more efficiently than urea. UBA was a potent premutagenic lesion with a rate of mutation induction more than sixfold above that observed with abasic sites derived from purines. Urea residues were about as mutagenic as abasic sites derived from purines, which in turn were more mutagenic than abasic sites derived from thymine. Mutations derived from urea, UBA and abasic sites were all dependent on SOS-induction of the host cells. Since both urea and UBA were derived from DNA thymine, these data demonstrate that adenine is not routinely inserted opposite products that no longer retain the structural integrity of the pyrimidine ring. Sequence analysis showed that the mutations were targeted at thymine with 62% of the urea-derived mutations being T to C transitions and 62% of the UBA derived mutations being T to A transversions. Thus, the two fragmentation products appeared to direct specific misinsertions. The mutations were not randomly distributed over the i(-d) region for either fragmentation product and hotspots were observed for both damages. The presence of hotspots suggests that in addition to lesion structure, sequence context plays an important role in base selectivity by DNA polymerases opposite DNA lesions. Energy minimization calculations were used to model the urea and UBA lesions at two contrasting hotspot sites. In both cases, there was significant agreement between the computational and biological data sets.