We are using the simple eukaryote Saccharomyces cervisiae as a source of highly purified replication and repair proteins with which to study the factors that control mutation rates and the accuracy of DNA synthesis. Yeast offers the important advantage over higher eukaryotes of a good genetic system to dissect DNA metabolic processes. We have begun by determining, in the M13mp2 mutagenesis assay, the accuracy of in vitro DNA synthesis by yeast DNA polymerases I and II, without accessory proteins. DNA polymerases I, the replicative enzyme for the nuclear genome, has been purified free of DNA primase subunit(s) and the beta sub-unit. This enzyme also lacks associated proofreading exonuclease activity. The accuracy of this form of Pol I is insufficient to account for the low spontaneous mutation rates observed in vivo. Pol I produces base substitution errors at a frequency that varies from 1 in 2000 to less than 1 in 60,000 nucleotides polymerized, depending on the composition and site of the mispair. Frameshift errors are also produced, at a frequency of 1 in 2000 to less than 1 in 75,000 nucleotides polymerized, again depending on the site. Pol I also produces deletion errors at high frequency. Most, but not all, of these mutants can be explained by either simple or complex variations of direct repeat mutagenesis. The fidelity of a more complex form of Pol I and the effects on fidelity of accessory proteins, including single-stranded DNA binding proteins, are currently being examined. Yeast DNA polymerase II, a single polypeptide containing associated 3' exonuclease activity, is much more accurate than Pol I for base substitutions, frameshifts and deletion errors. Variations in reaction conditions which diminish 3' exonuclease activity also decrease fidelity, clearly demonstrating that proofreading by this enzyme normally functions to reduce errors.
