DNA strand-specific mutations induced by (+/-)-3 alpha,4 beta-dihydroxy- 1 alpha,2 alpha-epoxy-1,2,3,4-tetrahydrobenzo[c]phenanthrene in the dihydrofolate reductase gene.

We previously showed that the preferred mutation induced by (+/-)-3 alpha,4 beta-dihydroxy-1 alpha,2 alpha-epoxy- 1,2,3,4-tetrahydrobenzo[c]phenanthrene (BcPHDE) in the dihydrofolate reductase gene in Chinese hamster ovary cells was a purine to thymine transversion on the nontranscribed strand at the sequence 5'-RRR-3' (R is a purine and the mutated base is underlined). To determine whether the observed mutational strand specificity was due to bias in the phenotypic selection, we designed a nonsense-codon reversion assay in which a triple purine target was present on both strands and all R----T transversion mutations yielded amino acid substitutions that were compatible with dihydrofolate reductase enzyme activity. From the size of the targets, a 2:1 ratio of mutations at the purines on the nontranscribed strand was expected if the DNA strands were mutationally equivalent. We isolated a total of 66 BcPHDE-induced revertants of two mutants that carry point mutations at either the 5' or the 3' end of the gene. All reversions at the 5' end arose by substitution on the nontranscribed strand; those at the 3' end showed a strand bias that favored this strand by 7:1. For both mutants, R----T transversions accounted for 88% of all the induced base changes. Thus, in this system, mutational strand bias is independent of the selection for phenotype. The results are consistent with the model of preferential repair of the transcribed strand as proposed by others. The involvement of RNA polymerase in the selective repair recruitment is discussed.