We have constructed a single-stranded vector that contains a uniquely located cis-syn T-T cyclobutane dimer by ligating a synthetic oligomer containing this UV photoproduct into M13mp7 viral DNA linearized with EcoRI. In the absence of SOS induction, transfection of a uvrA6 mutant of Escherichia coli with this vector gave very few progeny plaques, and the data imply that a single dimer blocks replication in at least 99.5% of the molecules. In vitro photoreactivation completely abolished this inhibition. Transfection ofcells irradiated with UV at 4 JM-2 to induce the SOS response gave 27% of the number of plaques found with a dimer-free control. Nucleotide sequence analysis of 529 progeny phage showed that translesion synthesis was usually accurate: the normal sequence was found in 93% of them. Where mutations occurred, all were targeted single-nucleotide substitutions, with =90% being targeted at the 3' nucleotide of the lesion: of a total of 26 mutations, 15 were 3' T --A, 8 were 3' T --C, and 3 were 5' T -* C. No T -* G mutations were found. In addition to these results with the normal construct, data were also obtained from vectors in which the M13mp7 cloning site, which forms a hairpin in single-stranded DNA, was present 4 nucleotides on the 3' side of the T-T dimer. These hairpin-containing vectors gave a very similar mutation frequency (8% versus 7%) but altered mutation spectrum: all 12 mutations detected were 3' T -* A transversions, a difference from the previous set of data that is significant (P = 0.03).Several investigators have analyzed the spectrum of nucleotide sequence changes induced by UV in phage or plasmid genomes (1-5), chiefly with the aim of understanding the molecular processes responsible for induced mutagenesis. From this perspective, however, the interesting patterns evident in such data raise more questions than they settle: their interpretation is uncertain because ofthe heterogeneous and ill-defined nature of UV-irradiated genomes. UV induces a variety of photoproducts, and controversy exists over their relative mutagenic potential (1-10). These photoproducts are distributed quasi-randomly within and between UV-irradiated genomes, and their low frequency at any one genetic site is a major reason for the rarity of mutations. As a consequence, mutation detection or selection methods based on change in gene function must be used, a procedure that fails to identify all nucleotide sequence changes and biases the spectrum of events recovered. It also precludes estimation of the error rate of translesion synthesis. In short, UV-irradiated genomes are far from ideal substrates with which to investigate mutagenic mechanisms.One means of avoiding the problems inherent in experiments with UV-irradiated genomes is the use of artificially constructed vectors that contain a single defined photolesion located at a uniquely specified genetic site. In this report, we describe a rapid, flexible, and general method for making such vectors that was designed to meet two criteria, (i) detection o...

We have cloned the REV3 gene of Saccharomyces cerevisiae by complementation of the rev3 defect in UV-induced mutagenesis. The nucleotide sequence of this gene encodes a predicted protein of Mr 172,956 showing significant sequence similarity to Epstein-Barr virus DNA polymerase and to other members of a class of DNA polymerases including human DNA polymerase a and yeast DNA polymerase I. REV3 protein shows less sequence identity, and presumably a more distant evolutionary relationship, to the latter two enzymes than they do to each other. Haploids carrying a complete deletion of REV3 are viable. We suggest that induced mutagenesis in S. cerevisiae depends on a specialized DNA polymerase that is not required for other replicative processes. REV3 is located 2.8 centimorgans from CDC60, on chromosome XVI.The REV3 gene of bakers' yeast, Saccharomyces cerevisiae, is concerned with a recovery process in which DNA damage causes mutations within the nuclear genome. It was identified by isolation of strains displaying reduced frequencies of UV mutagenesis (28), and, compared with others selected by such a criterion (25,26,28,32,37), strains carrying rev3 mutations exhibit one of the most extreme and general deficiencies in this respect. Relative to wild type, rev3 mutants display a 96% reduction in forward mutation to auxotrophy induced by UV (31) and also display defects in UV-induced reversion of broad range of base substitution and frameshift test alleles (24,27,28). The rev3 mutant is also deficient in mutagenesis by 4-nitroquinoline-1-oxide (41) and gamma rays (34), though response to ethyl methanesulfonate and nitrous acid is normal (31, 41). Spontaneous mutation is reduced to about one-fifth of normal in rev3 mutants, suggesting that the REV3-dependent recovery process may act on spontaneously arising damage (42). It does not, however, appear to be concerned with damage in the mitochondrial genome: UV-induced mutagenesis in mitochondrial DNA is normal in rev3 mutants (40, 45). These observations suggest that the REV3 gene product may function in translesion synthesis, that is in replication on mutagen-damaged templates, though they do not point to any specific role. The REV] gene (28), which has a mutant phenotype similar to that of REV3 and was identified in the same way, has recently been shown to encode a predicted protein of Mr 112,239 with sequence similarity to the Escherichia coli umuC gene (22). The umuDC genes of E. coli also have a mutant phenotype similar to that of REV3 and encode 16,000-and 45,000-Mr proteins that are thought to enhance the processivity of DNA polymerase (1,11

We have transfected SOS-induced and uninduced cells of a uvrA6 strain of Escherichia coli with single-stranded M13mp7-based vectors that carried a single trans-syn T-T cyclobutane dimer at a unique site. Unlike constructs carrying the cis-syn isomer of this lesion, these vectors could be replicated with modest efficiency (14%) in the absence of SOS induction and therefore provided an opportunity to measure directly the influence of such induction on error rate and mutation spectrum. We found that translesion synthesis in the absence of SOS induction was remarkably accurate; only 4% of the replicated bacteriophage contained mutations, which were exclusively targeted single T deletions. In SOS-induced cells, error frequency increased to 11% and the resulting mutations included targeted substitutions and near-targeted single base additions, as well as the T deletions. Replication efficiency was 29% in these conditions. SOS induction therefore leads not only to an enhanced capacity to replicate damaged DNA but also to a marked change in mutation frequency and spectrum.Derepression of the SOS regulon in Escherichia coli results in the production of an altered replication process that copies mutagen-damaged templates with increased efficiency, although with reduced accuracy (13). Such infidelity must, at least in part, be the consequence of the distorted template itself, but it may also reflect inhibition of polymerase functions that promote replication accuracy. It has been argued that inhibition of the 3'-to-5' editing exonuclease, encoded by dnaQ, is a necessary prerequisite for translesion synthesis (12), and data indicating that the RecA protein can indeed inhibit the activity (4, 8) support this argument. Similarly, editing activity does not appear to be involved in induced mutagenesis in normal strains (6,14). It is not known, however, whether the DnaQ gene product is normally part of the altered complex that replicates damaged DNA.A direct measurement of the relative accuracy of translesion synthesis in SOS-induced and uninduced cells has not previously been possible, both because very little bypass is thought to take place without derepression of the SOS regulon and because conventional experiments with mutagen-treated DNA are not well suited for this purpose. An opportunity to make such a direct comparison arose, however, in transfection experiments with a single-stranded M13 hybrid bacteriophage vector that contained a uniquely located trans-syn T-T cyclobutane dimer. Unlike molecules carrying the cis-syn isomer of this lesion, almost none of which are replicated in uninduced cells (1), about 14% of the vector molecules carrying the trans-syn lesion were replicated in the absence of induction, thus allowing the accuracy of translesion synthesis to be determined. A similar determination can, of course, be made for this vector introduced into SOS-induced cells, in which replication efficiency in-* Corresponding author. creased to 29%. At the same time, experiments with such vectors are well suited to the purp...

No abstract

Three allelic mutations of a new yeast gene, which we have named REV7, have been isolated by testing 313 methyl methane sulfonate sensitive mutants for UV-induced reversion of a lys2 allele. Rev7 mutants are markedly deficient with respect to UV-induced reversion of lys2, are slightly sensitive to UV and appear to be in the RAD6 epistasis group for UV survival. Rev7-1, which is probably an amber mutation, does not appear to affect sporulation in homozygous diploids. The REV7 gene is located about 12 cM distal to HIS5 on chromosome IX.