The replication of double-stranded plasmids containing a single N-2-acetylaminofluorene (AAF) adduct located in a short, heteroduplex sequence was analyzed in Saccharomyces cerevisiae. The strains used were proficient or deficient for the activity of DNA polymerase (REV3 and rev3⌬, respectively) in a mismatch and nucleotide excision repair-defective background (msh2⌬ rad10⌬). The plasmid design enabled the determination of the frequency with which translesion synthesis (TLS) and mechanisms avoiding the adduct by using the undamaged, complementary strand (damage avoidance mechanisms) are invoked to complete replication. To this end, a hybridization technique was implemented to probe plasmid DNA isolated from individual yeast transformants by using short, 32 P-end-labeled oligonucleotides specific to each strand of the heteroduplex. In both the REV3 and rev3⌬ strains, the two strands of an unmodified heteroduplex plasmid were replicated in ϳ80% of the transformants, with the remaining 20% having possibly undergone prereplicative MSH2-independent mismatch repair. However, in the presence of the AAF adduct, TLS occurred in only 8% of the REV3 transformants, among which 97% was mostly error free and only 3% resulted in a mutation. All TLS observed in the REV3 strain was abolished in the rev3⌬ mutant, providing for the first time in vivo biochemical evidence of a requirement for the Rev3 protein in TLS.

No abstract

SummaryTranslesion synthesis (TLS) in Saccharomyces cerevisiae requires at least Rev1p and polymerase z (Pol z), a complex of the Rev3 polymerase and its accessory factor Rev7p. Although their precise role(s) are poorly characterized, in vitro studies suggest that each protein contributes to TLS in a manner dependent on the particular lesion and surrounding DNA sequence. In the present study, strand segregation analysis is used to attempt to identify the role(s) of the Rev1 and Rev7 proteins during TLS. This assay uses double-stranded plasmids containing a genetic marker opposite to a replication blocking lesion (N-2-acetylamino¯uorene; AAF) to measure TLS quantitatively and qualitatively in vivo. The AAF adduct is localized within a repetitive sequence in a manner that allows the formation of misaligned primer±template replication intermediates. Elongation from a misaligned intermediate ®xes a frameshift mutation (slipped TLS), while extension of the correctly aligned lesion terminus yields error-free (nonslipped) TLS. The results indicate that there is a strong requirement for Rev7p during Pol z-mediated TLS measured in vivo. Furthermore, Rev1p is needed only for non-slipped TLS; slipped TLS remains ef®cient in its absence, revealing a previously uncharacterized Rev1p activity similar to Escherichia coli UmuDC function. Speci®cally, this activity is required for elongation from a correctly aligned lesion terminus.