Antagonism of Ultraviolet-Light Mutagenesis by the Methyl-Directed Mismatch-Repair System of Escherichia coli

Corresponding author: John B. Hays, Department of Environmental and Molecular Toxicology, Oregon State University, ALS 1007, Corvallis, OR 97331-7301., haysj{at}bcc.orst.edu (E-mail)

Communicating editor: P. L. FOSTER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Previous studies have demonstrated that the Escherichia coli MutHLS mismatch-repair system can process UV-irradiated DNA in vivo and that the human MSH2·MSH6 mismatch-repair protein binds more strongly in vitro to photoproduct/base mismatches than to "matched" photoproducts in DNA. We tested the hypothesis that mismatch repair directed against incorrect bases opposite photoproducts might reduce UV mutagenesis, using two alleles at E. coli lacZ codon 461, which revert, respectively, via CCC CTC and CTT CTC transitions. F' lacZ targets were mated from mut⁺ donors into mutH, mutL, or mutS recipients, once cells were at substantial densities, to minimize spontaneous mutation prior to irradiation. In umu⁺ mut⁺ recipients, a range of UV fluences induced lac⁺ revertant frequencies of 4–25 x 10^-8; these frequencies were consistently 2-fold higher in mutH, mutL, or mutS recipients. Since this effect on mutation frequency was unaltered by an Mfd^- defect, it appears not to involve transcription-coupled excision repair. In mut⁺ umuC122::Tn5 bacteria, UV mutagenesis (at 60 J/m²) was very low, but mutH or mutL or mutS mutations increased reversion of both lacZ alleles roughly 25-fold, to 5–10 x 10^-8. Thus, at UV doses too low to induce SOS functions, such as Umu₂'D, most incorrect bases opposite occasional photoproducts may be removed by mismatch repair, whereas in heavily irradiated (SOS-induced) cells, mismatch repair may only correct some photoproduct/base mismatches, so UV mutagenesis remains substantial.

IN most prokaryotes, and in all eukaryotes examined, highly conserved protein systems that recognize DNA mismatches and certain DNA lesions play critical roles in maintenance of genetic stability. These long-patch mismatch-repair systems decrease DNA replication error rates 100- to 1000-fold, by recognizing and correcting base/base and (insertion/deletion)-loopout mismatches that escape proofreading by DNA polymerase (KORNBERG and BAKER 1992 ). Genetic analyses, and subsequent comprehensive biochemical characterization in vitro (LAHUE et al. 1989 ), have elucidated the mismatch-repair pathway in Escherichia coli. Homodimers of E. coli MutS protein bind preferentially to mismatches; then, MutS and MutL homodimers activate MutH protein to specifically nick unmethylated DNA strands at the nearest adenine-hemimethylated d(GATC) sites, during the interval before adenines in newly replicated d(GATC) sequences are methylated. The activation/nicking process most likely involves a protein/DNA translocation and search process driven by ATP hydrolysis (ALLEN et al. 1997 ). Thus, MutH directs incision and subsequent excision to the nascent DNA strand specifically, so replication errors are corrected rather than fixed. In a reconstituted system (LAHUE et al. 1989 ), excision requires UvrD protein (DNA helicase II) and either 3'-5' or 5'-3' single-stranded DNA (ssDNA) endonuclease, depending on the relative orientations of the mismatch and the nicked d(GATC) site (COOPER et al. 1993 ). The replicative polymerase (E. coli DNA polymerase III) fills the excision gap. Recombination involving partially diverged DNA sequences is antagonized by mismatch-repair systems, which presumably recognize mismatches in heteroduplex joints (FEINSTEIN and LOW 1986 ; RAYSSIGUIER et al. 1989 ; PETIT et al. 1991 ; WORTH et al. 1994 ). Accumulating evidence now suggests processing by mismatch-repair systems of DNA molecules containing various DNA lesions. These include ultraviolet-light (UV) photoproducts (FENG et al. 1991 ; FENG and HAYS 1995 ; MU et al. 1997 ; WANG et al. 1999 ), O⁶meG residues (KAT et al. 1993 ; DUCKETT et al. 1996 ), cisplatin G-G intrastrand cross-links (DUCKETT et al. 1996 ), adriamycin (DRUMMOND et al. 1996 ) and acetyl-aminofluorene and aminofluorene (AAF/AF) adducts (LI et al. 1996 ), and S⁶-methylthioguanine/base mismatches (SWANN et al. 1996 ). MutS and MutL activities or their homologs are required for efficient transcription-coupled nucleotide excision repair of cyclobutane pyrimidine dimers (CPDs) in E. coli (MELLON and CHAMPE 1995 ) and human cells (MELLON et al. 1996 ).

The fate of nonreplicating UV-irradiated phage chromosomes in E. coli deficient in nucleotide excision repair (Uvr^-) has provided direct in vivo evidence for processing of photoproduct-containing DNA by mismatch-repair proteins. Elevation of homologous recombination, from nearly undetectable frequencies to as much as 10%, and physical breakdown that resulted in duplex DNA breaks and ssDNA gaps and in loss of biological activity required MutS, MutL, and MutH functions and adenine-undermethylated d(GATC) sites (FENG et al. 1991 ). Further work (FENG and HAYS 1995 ) strongly implicated helicase and ssDNA exonuclease activities involved in mismatch repair (COOPER et al. 1993 ). Studies of photoproduct binding by hMutS, the human homolog of MutS, have demonstrated specific binding to a variety of mismatched CPDs and [6-4] photoproducts, but not to matched photoproducts (MU et al. 1997 ; WANG et al. 1999 ).

Mismatch repair, targeted to mismatched photoproducts generated by translesion synthesis during the interval that d(GATC) sites on nascent strands remained unmethylated, would excise incorrect bases rather than the photoproducts. For such a process to antagonize mutagenesis efficiently, the subsequent ssDNA-gap-filling DNA synthesis would have to insert the correct base opposite the template photoproduct that originally provoked the repair process. Correct insertion seems the usual result of synthesis past CPDs in phage ssDNA in E. coli (BANERJEE et al. 1988 ; JIANG and TAYLOR 1993 ) and in human nuclear extracts replicating UV-irradiated SV40-origin plasmids (CARTY et al. 1993 ; THOMAS and KUNKEL 1993 ). Alternatively, stalled mismatch-repair resynthesis tracts could be extended by recombinational template-switching or gap-filling mechanisms. Finally, failed mismatch-repair resynthesis could lead to DNA degradation, the net result being elimination of mutant chromosomes. In any event, a prediction of these mutagenesis-antagonism models is that UV mutagenesis should be enhanced in E. coli mutS, mutL, and mutH mutants, beyond increases in spontaneous mutation.

To test this prediction we analyzed UV-induced reversion of two E. coli lacZ codon-461 alleles, constructed by CUPPLES and MILLER 1989 to revert to lac⁺ only by CCC CTC or CTT CTC transitions. The targets are thus 3' pyrimidines in potential photoproduct sites. To minimize culture-to-culture fluctuations in spontaneous mutant frequencies, we mated the F' lacZ targets into mut^- cells just before irradiation, when cell densities were already substantial. We find a consistent twofold increase in UV mutagenesis in mismatch-repair-deficient (umu⁺) bacteria and also hitherto unsuspected substantial UmuC-independent UV mutagenesis, which is readily detectable only in mut^- cells.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains and plasmids:
E. coli K-12 strains employed in the mutation studies are described in Table 1. Transduction with P1 phage employed standard techniques (MILLER 1972 ). Transposon-insertion alleles encoding cat, Tn5, and Tn10 were selected on Luria broth (LB) plates containing, respectively, chloramphenicol (35 µg/ml), kanamycin (50 µg/ml), or tetracycline (15 µg/ml). MutS^-, MutH^-, and MutL^- phenotypes were identified by increased spontaneous resistance to rifampicin (100 µg/ml) in LB plates, and the UmuC^- phenotype was recognized by increased UV sensitivity. Plasmid pMQ315 contains the mutS⁺-encoding BglII restriction fragment of E. coli DNA, inserted into the BamHI site of plasmid pBR322 (WU and MARINUS 1994 ).

Media and antibiotics:
M9 minimal media contained 6 g/liter K₂HPO₄ plus 3 g/liter KH₂PO₄ (pH 7.0), 1 g/liter (19 mM) NH₄Cl, 0.1 mM CaCl₂, 1 mM MgSO₄, and 0.001% thiamine, plus glucose (0.2%), glycerol (0.2%), or lactose (0.2%). LB plates have been described (MILLER 1972 ). Plates were solidified with 1.5% agar (glucose plates; Difco, Detroit) or 1.5% noble agar (lactose plates; United States Biochemical, Swampscott, MA).

Growth, mating, and irradiation of bacteria:
For mutagenesis experiments, bacteria were streaked from frozen-glycerol cultures onto LB plates, and single colonies were used, after no more than 36 hr of growth, to inoculate glycerol-minimal-medium cultures, typically 30 ml, containing 60 µg/ml proline. (Inoculation with older colonies resulted in significantly higher final levels of spontaneous lac⁺ revertants in cultures of mut^- strains.) After overnight growth at 37°, cultures of various F^- (mut⁺ or mut^-) phr::cat (lac-pro) recipients and mut⁺ F' (pro⁺ lacI^- lacZ^-) donors were both diluted 1.5-fold with fresh medium, or grown without dilution, to ~3.5 x 10⁸ cells/ml, mixed together, and incubated at 37° with very gentle swirling. (All cultures of F^- recipients contained 60 µg/ml proline.) After a 2-hr mating period, mixtures were washed with glycerol-minimal medium containing chloramphenicol, but no proline, resuspended to 1 x 10⁸ cells/ml in the same medium, and grown for 6–8 hr, to select for phr::cat F' (pro⁺ lacI^- lacZ^-) transconjugants. At the end of selective growth, transconjugants typically represented 80% of the total colony-forming units in the mating mixture, as determined by selective vs. nonselective plating.

For UV irradiation, mixtures were diluted with glycerol-minimal medium plus chloramphenicol to yield 2.5 x 10⁸ transconjugants/ml, and 10-ml aliquots were added to uncovered 10-cm plastic petri dishes and irradiated at 2 W/m², using 254-nm lamps attenuated by window screen. Unirradiated aliquots were used as controls. Lamp fluences were checked using a Spectronics DRC-100X meter and/or an International Light IL1700 radiometer. To determine surviving fractions, irradiated and unirradiated control cells were spread immediately on glucose-minimal plates containing chloramphenicol, and colonies were scored after 48 hr incubation. Under these conditions, survival frequencies of (unmated) mutS-, mut⁺-, and mutS⁺-overproducing (mutS⁺⁺⁺) bacteria (strains LH2519, LH3179, and LH2536, respectively) were very similar to one another: ~25, 17, 10, 5, and 1.3%, at 20, 30, 45, 60, and 90 J/m², respectively.

Analysis of mutation in UV-irradiated bacteria:
For initial screening of all six lacZ461 alleles for effects of mismatch repair on UV mutagenesis, mut⁺ and mutS201::Tn5 F' (lacZ) strains were grown as described above without mating and were irradiated and analyzed. In all other experiments, irradiated and unirradiated transconjugant cells were diluted with equal volumes of glycerol-minimal medium containing chloramphenicol and were incubated at 37° with shaking. Cultures were incubated in minimal medium rather than broth to reduce carryover of trace nutrients onto the selective plates, and glycerol was used rather than glucose to ensure maximal expression of the lacZ gene. Cells were harvested by centrifugation after 8–10 hr (logarithmic growth was fully restored by 4 hr), resuspended in 0.2 vol of M9 minimal salts, diluted appropriately, and spread on glucose-minimal/chloramphenicol plates to score total (transconjugant) bacteria, or cells were spread directly on scavenged lactose-minimal/chloramphenicol plates to score revertants and were incubated for 48 hr at 37°. [Plates were scavenged 1 day before initiation of experiments by spreading 5 x 10⁷ bacteria of strain LH3302 (phr::cat transductant of lac strain FC755) and incubating overnight at 37°. This prevents further mutagenesis on the plates (see RESULTS).] Throughout the text, "UV-induced revertant frequencies" implies that lac⁺ revertant frequencies were determined for parallel unirradiated cultures and were subtracted from apparent frequencies for UV-irradiated cells. Since unirradiated cultures grew (increased in turbidity) about four times as well as irradiated cultures during the 8-hr postirradiation period, this subtraction may overcorrect slightly for background spontaneous mutant frequencies in the irradiated cells. Revertant frequencies determined after as much as 24 hr of postirradiation growth in liquid medium were the same as 8-hr frequencies.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Design and evaluation of experimental approaches:
Our aims were (i) to ensure that most UV mutagenesis (here reversion of lacZ^- alleles to lac⁺) took place during a defined period of postirradiation growth in liquid culture, rather than on selective plates, and (ii) to minimize fluctuations in the background spontaneous mutant frequencies in cultures of mismatch-repair-defective (mut^-) bacteria.

To determine the time required for recovery from UV irradiation and concomitant mutagenesis, we incubated irradiated cells in glycerol-minimal medium and measured total cell masses (turbidity units; Figure 1) and numbers of viable bacteria (data not shown). After a 4-hr recovery period these both began to increase exponentially. However, frequencies of lac⁺ revertants after recovery could not be determined unequivocally by spreading on lactose-minimal plates, even when cells were washed several times: the number of visible Lac⁺ colonies increased every day for at least 5 days, and their number was not proportional to the number of cells spread. These problems were most severe for mut^- bacteria. We suspected that trace carbon sources in the lactose-minimal plates were supporting limited slow growth by lacZ^- bacteria, the final numbers of cells being more dependent on the amounts of trace nutrients than on the initial numbers of cells, as observed previously (ZIEG and KUSHNER 1977 ). We also suspected that lac⁺ revertants continued to arise during this slow growth. To eliminate trace nutrients in the plates, we used purified ("noble") agar, high-purity (glucose-free) lactose and scavenged the plates by spreading with 5 x 10⁷ lac bacteria and incubating overnight, just before UV mutagenesis experiments. Under these conditions, the number of Lac⁺ colonies arising was directly proportional to the number of cells spread, and did not change during 2–7 days of incubation on plates.

View larger version (18K): In this window In a new window Download PPT slide   Figure 1. Time course of UV mutagenesis and recovery from irradiation. Cultures of phr::cat F' (lacZ461-6) bacteria (strain LH3140) were grown and irradiated to 60 J/m2 in glycerol-minimal medium, diluted to a turbidity of 15 Klett units (~1 x 108 cfu/ml, for unirradiated bacteria), and shaken at 37°. Data correspond to means of turbidity measurements (•) for two independent experiments. Measurements of frequencies of UV-induced lac+ revertants () were essentially as described in MATERIALS AND METHODS; data correspond to means for four independent experiments, with standard deviations indicated.

The low frequency of lac⁺ revertants among cells that were spread immediately on selective plates (Figure 1) indicates that the technique mostly scores mutants that arise during postirradiation incubation in liquid culture. The maximum in revertant frequency after ~4 hr suggests that mutation fixation is nearly complete by then, but that chromosome replication and/or segregation of lac⁺ and lacZ^- alleles into daughter cells has not occurred yet. The subsequent 50% decrease in apparent revertant frequency indicates that few additional mutations arise during subsequent rounds of replication. Segregation of parental and daughter strands and cells after mutation of DNA strands would thus double the number of survivors without increasing the number of revertants, as suggested previously (BRIDGES and MUNSON 1968 ). After 6 hr of growth in culture, numbers of both lac⁺ and total cells increase exponentially in parallel, providing a stable mutant frequency (Figure 1 and data not shown).

To screen for effects of mismatch-repair deficiency on UV mutagenesis via different transition and transversion pathways, we employed bacteria encoding on respective F' episomes the six lacZ codon-461 alleles constructed by CUPPLES and MILLER 1989 ; the designations lacZ461-1, lacZ461-2 ... lacZ461-6 used here correspond to their strains CC101, CC102 ... CC106. Each different allele reverts to a (GAG) codon for glutamate, the only amino acid compatible with ß-galactosidase activity (CUPPLES and MILLER 1988 ), by a different single base substitution. Table 2 shows lac⁺ revertant frequencies (corrected for spontaneous frequencies in the same cultures) induced by a UV dose of 45 J/m², in mut⁺ and mutS201 bacteria bearing each of the six F'(lacZ) episomes. For the two lacZ alleles where target bases are not in potential dipyrimidine-photoproduct sites, the revertant frequencies (x 10⁸) in mut⁺ and mutS201 derivatives were 48 vs. 56 for lacZ461-4, and 14 vs. 16 for lacZ461-5 (sense-strand GCG GAG and GTG GAG transversions, respectively; Table 2, lines 7–10). Thus, neither of these (presumably untargeted) transversions showed a mismatch-repair effect. The two other transversion alleles showed low UV-induced reversion frequencies, although target bases were in potential photoproduct sites, and low mismatch-repair effects: frequencies (x 10⁸) in mut⁺ and mutS201 derivatives were 2.8 vs. 4.5 for lacZ461-1 and 1.6 vs. 2.0 for lacZ461-3 (sense-strand T TAG T GAG and T CAG T GAG reversions, respectively; Table 2, lines 1 and 2, 5 and 6). However, where the mutation targets were 3' bases in dipyrimidines, reverting via transitions, revertant frequencies (x 10⁸) were high and were consistently twofold higher in mutS201 derivatives: 73 vs. 131 for lacZ461-2 and 71 vs. 159 for lacZ461-6 (template strand CCC CTC and CTT CTC reversions, respectively; Table 2, lines 3 and 4, 11 and 12). The lacZ461-2 and lacZ461-6 alleles were employed in all subsequent experiments.

Revertant frequencies among unirradiated mutS F' (lacZ461-2) and mutS F' (lacZ461-6) bacteria fluctuated considerably from culture to culture, perhaps reflecting spontaneous reversion relatively early in colony isolation or culture growth (before irradiation). When the spreads in the respective sets of revertant frequencies observed for a particular strain were measured by calculating the standard deviations from the mean, large culture-to-culture fluctuations in spontaneous mutation frequencies for mut^- strains were apparent. Thus, in the experiments shown in Table 2, the standard deviations of the sets of spontaneous-reversion frequencies for mutS201 strains bearing F'(lacZ461-1) ... F'(lacZ461-6) episomes were, respectively, 94, 110, 36, 63, 89, and 88% of the corresponding means. In other experiments, with umuC122 mutL96 F'(lacZ461-6) bacteria (strain LH3290), the set of spontaneous revertant frequencies (x 10⁸) combined from several experiments (total n = 16), showed a mean of 50 and a standard deviation of 88, 176% of the mean. In subsequent experiments, therefore, we propagated parallel cultures of F^- mutS bacteria or other recipients of interest, and mut⁺ umuC122 (F' lacZ^-) donors, and mated F⁺ to F^- cultures for 2 hr. Selection for 6–8 hr yielded ~80% pro⁺ chloramphenicol-resistant transconjugants. These were irradiated, grown out, and plated, with continued selection, which increased the transconjugant fraction to 90–95%. When umu⁺ mut^- recipients were mated with F'(lac-Z461-6) donors (see Figure 2 and Figure 3), the set of all spontaneous revertant frequencies (x 10⁸) of transconjugant cultures (n = 26 for all cultures of all experiments with mutS, mutL, and mutH recipients) showed a mean of 3.3 and standard deviation 1.0, only 30% of the mean. For F'(lacZ461-2) donors, the set of all spontaneous revertant frequencies (x 10⁸) for all umu⁺ mut^- transconjugant cultures (n = 7) showed mean 3.9 and standard deviation 1.0, only 25% of the mean. In experiments with umuC122 mut^- recipients (see Figure 4), the set of all transconjugant spontaneous revertant frequencies (x 10⁸) for lacZ461-6 donors (n = 17) showed mean 2.9, standard deviation 1.5 (52%), and for lacZ461-2 donors (n = 17), mean 2.7, standard deviation 1.0 (37%).

View larger version (16K): In this window In a new window Download PPT slide   Figure 2. Dependence of mutant frequency in umu+ bacteria on UV dose. Growth and mating of umu+ phr::cat (lac-pro) recipients with F' (lacZ461-6 pro+) donors (strain LH3305), selection for transconjugants, UV irradiation to indicated fluences, further growth and plating, and analysis of UV-induced lac+ revertant frequencies were as described in MATERIALS AND METHODS. For each UV fluence, each open, filled, or crossed symbol corresponds to an independent experiment employing typically three parallel cultures of a particular strain: (open, solid, and crossed circles) mutS97, strain LH2519; (open, solid, and crossed squares) mut+, strain LH3279; (open, solid, and crossed triangles) mutS+++, strain LH23305. (For crossed and solid symbols at 45 J/m2 there were only one or two independent cultures.) Values represent means for the multiple independent cultures for each strain. Some symbols at each UV fluence are off-set for clarity. Solid lines are drawn through means of values corresponding to circles, and dashed lines are drawn through means of triangle values. To estimate standard errors of the mean (SEMs), the set of all values for a particular strain and UV dose, for all cultures and experiments, were combined and analyzed, since experiment-to-experiment variations were not consistently markedly different from culture-to-culture variations. Revertant frequency means (± SEM) (x 108) for mutS97, mut+, and mutS+++ bacteria, respectively, were as follows: 20 J/m2, 14.4 (± 0.9), 5.7 (± 0.2), 3.4 (± 0.6); 30 J/m2, 28.0 (± 3.4), 10.2 (± 0.9), (7.1 ± 1.6); 45 J/m2, 55.0 (± 1.0), 24.4 (± 4.2), 20.5 (± 2.7). For the set of all data for unirradiated mutS97 bacteria (n = 14), mean (SEM) was 3.9 (± 0.3) x 10-8 (frequencies for unirradiated mut+ and mutS+++ bacteria were negligible).

View larger version (15K): In this window In a new window Download PPT slide   Figure 3. UV-induced mutation in umu+ bacteria. Indicated umu+ phr::cat (lac-pro) recipients and mut+ donors of F' (lacZ-pro+) episomes were grown and mated, and transconjugants were selected, irradiated to 60 J/m2, grown 8 hr, and scored for frequencies of UV-induced lac+ revertants, as described in MATERIALS AND METHODS. For each indicated genotype, each individual symbol refers to an independent experiment, in which mutant frequencies for two to three parallel cultures of mut+ recipients (strain LH3179) were, respectively, compared to frequencies for two to three parallel cultures of one particular mutant recipient: mutS (•), strain LH2519; mutL (), LH3183; mutH (), LH2520; mutS+++ (), LH2536. [mut+ mfd recipients (strain LH2548) were compared to mutS mfd () recipients (strain LH2549).] For each independent experiment (independent symbol), relative mutant frequency equals mean of mutant frequencies for the mut- (or mutS+++) cultures divided by the mean for the mut+ cultures. Expected reversion pathways for the two F' (lacZ- pro+) donors employed are indicated: C T, lacZ461-2 (strain LH3303); T C, lacZ461-6 (strain LH3305). For the experiments employing mut- vs. mut+ bacteria (solid symbols), mean (± SEM) UV-induced revertant frequencies (x 108), of transconjugants for sets of pooled data from all cultures in all repetitions of each particular experiment, were as follows for mut+ and mut- strains, respectively: mutS C T, 24.3 (± 1.3) and 13.4 (± 0.3); mutS T C, 28.0 (± 3.4) and 10.2 (± 0.9); mutL T C, 24.8 (± 1.3) and 12.0 (± 0.3); mutH T C, 23.9 (± 1.4) and 11.6 (± 1.0); mutS mfd T C, 30.4 (± 5.3) and 19.2 (± 1.4). Corresponding frequencies (± SEM) (x 108) for pooled data for unirradiated mut- bacteria were 3.3 (± 0.4), 4.7 (± 0.3), 3.4 (± 0.4), 4.1 (± 0.4), and 4.4 (± 0.4). Corresponding mean revertant frequencies (± SEM) x 108 for mutS+++ C T and mutS+++ T C were 12.6 ± 1.3 and 7.1 ± 1.6.

View larger version (19K): In this window In a new window Download PPT slide   Figure 4. UV-induced mutation in umuC122 bacteria. Indicated umuC122 phr::cat (lac-pro) recipients and mut+ donors of F' (lacZ461-2 pro+) [strain LH3303, open bars (C T reversion)] or F' (lacZ461-6 pro+) [strain LH3305, shaded bars (T C reversion)] were grown and mated, and transconjugants were selected, irradiated to 60 J/m2, grown 8 hr, and scored for frequencies of lac+ revertants, as described in MATERIALS AND METHODS. Individual symbols correspond to individual cultures, and different symbol shapes (, •, ), respectively, correspond to different independent experiments each employing two or three cultures of the particular indicated strains in parallel. Recipient genotypes indicated: mut+, strain LH3265; mutS, LH2534; mutL, LH3269; mutH, LH2535. For the experiments employing mut- bacteria (bars 3–8), respective mean (± SEM) UV-induced and spontaneous transconjugant revertant frequencies (x 108) were as follows: mutS C T, 5.0 (± 0.5) and 2.6 (± 0.1); mutS T C, 5.9 (± 0.1) and 2.8 (± 0.6); mutL C T, 5.0 (± 0.4) and 2.7 (± 0.4); mutL T C, 5.6 (± 1.0) and 3.3 (± 0.5); mutH C T, 5.8 (± 1.3) and 1.7 (± 0.5); mutH T C, 5.9 (± 0.8) and 2.9 (± 0.8).

UV mutagenesis in umu⁺ mut^- bacteria:
We used the mating technique to compare reversion of lacZ461-6 mutations in mutS vs. mut⁺ transconjugants at three UV fluences (Figure 2). Revertant frequencies increased with fluence in both strains, but averaged 2.6-fold higher in mutS97 than in mut⁺ bacteria, in good agreement with the mutS/mut⁺ ratio of 2.2 obtained in experiments with single cultures of F' (lacZ) bacteria (no mating; see above). In mut⁺ bacteria harboring a mutS⁺-encoding plasmid, there was a small trend to even lower revertant frequencies.

Figure 3 shows the effects of mismatch-repair deficiencies on lac461-2 and lacZ461-6 reversions at a single UV fluence of 30 J/m². Revertant frequencies for mutH, mutL, and mutS bacteria were again consistently twice as high as for mut⁺. Both the MutS^- and MutL^- effects might reflect reductions in transcription-coupled nucleotide excision repair (MELLON and CHAMPE 1995 ), since the lacZ gene should be derepressed in these lacI-defective bacteria growing in glycerol. However, in strains defective in Mfd activity, which normally couples transcription blockage to template-strand-specific excision repair (SELBY et al. 1991 ; SELBY and SANCAR 1993 ), a mutS mutation again increased lacZ461-6 T C reversion, to 2.2 times the mut⁺ value (Figure 4, triangles). The Mfd^- phenotype was confirmed by lower survival at 30 J/m² (5% vs. 18% for Mfd⁺) and by transduction of the mfd::Tn5 marker into strain WU3610 and verification of loss of the mutation-frequency-decline phenotype (WITKIN 1966 ).

UV mutagenesis in UmuC^- Mut^- bacteria:
The defining phenotype of umuC mutations is a drastic decrease in UV mutagenesis (KATO and SHINOURA 1977 ; STEINBORN 1978 ). Figure 4 shows that UV-induced reversions of lacZ461-2 and lacZ461-6 alleles in umuC::Tn5 mut⁺ bacteria were almost undetectable (Figure 4), but that inactivation of mismatch repair unmasked substantial UmuC-independent UV mutagenesis: UV-induced frequencies of both T C and C T reversion induced by 30 J/m² were roughly 5 x 10^-8, in mutS, mutL, and mutH bacteria. These experiments were repeated, using 100 plates each to analyze revertant frequencies among irradiated umuC122 mut⁺ recipients (strain LH3266) containing F' lacZ461-2 or F' lacZ461-6 episomes. On the basis of 135 and 121 total Lac⁺ colonies, respectively, revertant frequencies were 4.0 x 10^-9 and 2.7 x 10^-9; the respective revertant frequencies for umuC122 mutS::Tn10 recipients (strain LH2534) containing the same episomes were 10.7 x 10^-8 and 7 x 10^-8, a 25-fold increase in each case. In earlier experiments not using the mating technique, mutL umuC122 (F' lacZ461-6) bacteria showed wide fluctuations in spontaneous reversion frequency (see above). Here we analyzed UV mutagenesis only for 7 (of 17) cultures that showed spontaneous revertant frequencies of <10 x 10^-8 before irradiation; the mean (standard deviation) revertant frequency was 6.7 (± 2.2) x 10^-8. UV irradiation of each of these cultures to 60 J/m² increased revertant frequencies; the mean was 12.1 (± 6) x 10^-8. The apparent UV-specific component, 5.4 x 10^-8, was thus in good agreement with values obtained using the mating technique (see above). Revertant frequencies for mut⁺ umuC122 (F' lacZ461-6) bacteria (no mating) were again <5 x 10^-9.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The experiments described here support a hypothesis that mismatch-repair systems antagonize UV mutagenesis by excising incorrect bases inserted in nascent strands opposite photoproducts during the course of DNA replication, and replacing them, some or most of the time, with the correct bases. This hypothesis was motivated by observations that human MSH2 · MSH6 protein bound specifically to DNA containing a mismatched cyclobutane pyrimidine dimer (CPD, Py<>Py), e.g., T<>T/AG, or a mismatched pyrimidine-(6-4')-pyrimidinone photoproduct ([6-4] photoproduct, Py[6-4]Py), e.g., T[6-4]T/AG, but not to DNA containing T<>T/AA or T[6-4]T/AA pairs (MU et al. 1997 ; WANG et al. 1999 ). Other explanations for these results, especially for the twofold difference between mut⁺ and mut^- bacteria seen in umu⁺ strains, might be more efficient excision repair in mut⁺ bacteria, or faster replication (relative to repair) in mut^- cells. DNA replication might in principle be faster if not delayed by binding and/or processing of mismatched bases or photoproducts by mismatch-repair proteins, but we are not aware of any reports of faster replication in mut^- bacteria. We think it is unlikely that the antagonism of UV mutagenesis by mismatch repair described here simply reflects removal, by MutSL-assisted transcription-coupled excision repair, of premutagenic photoproducts at the CCC and TT targets in the lacZ461-2 and lacZ461-6 template strands, because we observed MutH-dependent and Mfd-independent antagonism of mutagenesis, in contradistinction to the requirements for transcription-coupled repair. Furthermore, the previously reported MutHLS-dependent recombination of repressor-blocked UV-irradiated phage DNA occurred in excision-repair-deficient (Uvr^-) bacteria (FENG et al. 1991 ). Clearly, it would be of interest to analyze mismatch-repair effects on UV mutagenesis in Uvr^- bacteria. UV mutagenesis might occur in Uvr⁺ bacteria during replication, or by insertion of an incorrect nucleotide opposite a nearby second UV photoproduct during DNA synthesis to fill a gap created by excision repair of a first photoproduct in the opposite strand. In the latter case, mismatch repair might prevent fixation of mutations during subsequent excision repair of the second photoproduct. In any event, the constant ratio between UV-induced revertant frequencies for mut^- and mut⁺ bacteria seen in Figure 2 as the UV dose increases seems most consistent with a constant efficiency of mismatch repair.

Mismatch repair provoked by photoproduct/base mismatches, as by base/base mismatches, would be expected only in the vicinity of replication forks, where unmethylated d(GATC) sites were still present (LAHUE et al. 1989 ). In unreplicated DNA, therefore, Py<>U/PuG mismatches resulting from the relatively rapid deamination of cytosines in CPDs (PENG and SHAW 1996 ) would not be processed, whereas Py<>U/PuG mismatches arising in newly replicated DNA might be "corrected" to Py<>U/PuA, actually fixing C T transitions. Also, if progress of replication forks past template Py<>C photoproducts depended on deamination to Py<>U, so as to facilitate insertion of an adenine nucleotide (TESSMAN et al. 1992 ), the resulting Py<>U/PuA moiety would not be recognized by MutS (WANG et al. 1999 ). Thus, if deamination of cytosines in CPDs played a major role in UV mutagenesis in the experiments described here, mismatch repair might antagonize UV-induced T C transitions more efficiently than C T. However, the apparent mutation-removal efficiencies were approximately the same for UV-induced mutation of the two alleles tested, roughly 50% in umu⁺ and 96% in umuC122 bacteria. Therefore, most C T reversions as measured here in mut^- lacZ461-2 bacteria would appear not to have arisen by cytosine-deamination mechanisms, but rather by insertion during DNA replication of adenines opposite photoproducts containing cytosines [perhaps their imino tautomers (PERSON et al. 1974 ; JIANG and TAYLOR 1993 )]. (Although [6-4] photoproducts may be responsible for some or all of the MutHLS-sensitive UV mutagenesis described here, cytosine deamination is not enhanced in these photoproducts.)

In comparisons among different studies of UV mutagenesis, the specific circumstances under which UV-induced mutations arise and are fixed may be important. In our experiments irradiated cells were diluted and shaken in liquid glycerol-minimal medium, under conditions such that parallel unirradiated cultures grew exponentially. Irradiated cultures resumed exponential growth after ~4 hr, at which time fixation of mutations appeared complete (Figure 1). In the scavenged lactose-minimal plates trace carbon sources appeared not to be available to the lacZ^- bacteria for even limited growth and further mutagenesis, since very few Lac⁺ colonies arose when bacteria were plated immediately after irradiation. Most mutations appeared to arise 2–4 hr after irradiation, although the cultures did not show appreciable growth until 6 hr (Figure 1). Other workers (CROWLEY and HANAWALT 1998 ) have shown that for excision-repair-proficient bacteria growing in glucose-minimal medium, removal from the genome overall of CPDs induced by 40 J/m² of 254-nm irradiation is essentially complete within 60 min; repair of [6-4] photoproducts is complete in 20 min. Thus, factors other than photoproduct removal would appear to be rate limiting here for resumption of growth and colony-forming proficiency.

Mismatch repair might reduce UV mutagenesis by removing incorrect bases inserted opposite photoproducts during translesion synthesis and/or correcting mismatches from "untargeted" SOS mutagenesis (KUNZ and GLICKMAN 1984 ; CHRISTENSEN et al. 1985 ; MILLER 1985 ; CAILLET-FAUQUET and MAENHAUT-MICHEL 1988 ). Several considerations argue against a hypothesis that most of the UV mutagenesis seen here in mut^- umuC122 bacteria is untargeted.

1. CAILLET-FAUQUET and MAENHAUT-MICHEL 1988 reported no significant difference between the frequencies of spontaneous reversion to rifampicin resistance of umuC122 mutS bacteria that were recA⁺ or those that carried the recA730 mutation, known to cause an SOS-like activation of RecA protein that increases (untargeted) spontaneous mutation ("SOS mutagenesis"), whereas a recA270 mutation in umu⁺ mutS/L backgrounds increased the mutation frequency 2.5-fold. This is consistent with other reports that SOS mutagenesis requires UmuC function.

2. FIJALKOWSKA et al. 1997 observed that a recA730 mutation increased transversion frequencies (reversions of lacZ461-4 and lacZ461-5 mutations) much more than transition frequencies (lacZ461-2 and lacZ461-6 reversions), respectively, 2–6-fold vs. 20–70-fold, in both mut⁺ and mutS bacteria (both umuC⁺). There is no evidence to suggest that this preference might be reversed in umuC122 bacteria.

3. Previous studies suggest that untargeted mutations amount to only 3–10% of all UV-induced mutations, including those at both putative photoproduct and nonphotoproduct sites (KUNZ and GLICKMAN 1984 ; CHRISTENSEN et al. 1985 ; MILLER 1985 ).

Thus, although we cannot rule out some contribution of untargeted events to the greatly increased UV-induced transition frequencies seen in mut^- derivatives of umuC122 bacteria, it seems likely that much of this reflects photoproduct-targeted events. In umu⁺ bacteria, the greater-than-first-power dependence of mutant frequency on dose seen in both mut⁺ and mutS derivatives (Figure 2) argues in favor of targeted mutagenesis in both strains, since untargeted mutation would be expected to level off at higher doses, once SOS induction was maximal.

Some previous descriptions of UmuC-independent UV mutagenesis in mismatch-repair-proficient bacteria represent special cases not applicable here. Fuchs and co-workers (NAPOLITANO et al. 1997 ) have described SOS-dependent, but UmuC-independent, frameshift mutagenesis at certain hotspots, targeted by 2-acetylaminofluorene adducts, that involves slip mispairing at mononucleotide or dinucleotide repeats at these sequences. The UVM pathway, studied extensively by HUMAYUN 1998 , specifically causes mutations by inserting thymine nucleotides opposite etheno-cytosine residues. Although inducible by other kinds of DNA damage, it is independent of RecA and UmuD₂'C.

Lawrence and co-workers (CHRISTENSEN et al. 1988 ) observed substantial UmuC-independent UV mutagenesis when they spread cultures of mut⁺ uvrA6 umuC122 bacteria, after post-UV-irradiation growth, on nonselective plates, and identified lacI^- mutants among other survivors by colony color. However, when they spread the same cultures on scavenged selective (phenyl-galactosidase) plates, similar to the procedure used here, very few colonies (lacI^- mutants) were seen. On nonselective plates, mismatch-repair proteins might have declined during colony growth, making cells (umuC122) Mut^- phenocopies, as has been reported for stationary-phase (FENG et al. 1996 ) or carbon-starved (HARRIS et al. 1997 ) bacteria. [P. L. Foster has argued against a hypothesis of low mismatch repair under starvation conditions, however (FOSTER 1999 ).] Alternatively, some of these irradiated uvrA umuC122 bacteria might have become Mut^- phenocopies eventually during incubation on the nonselective plates, due to saturation of mismatch repair (SCHAAPER and RADMAN 1989 ; CUPPLES et al. 1990 ) by uracil/base mismatches arising from deamination of unrepaired CPD cytosines (WANG et al. 1999 ).

Another source of UmuC-independent UV mutagenesis might be the product of the dinB (dinP) gene (OHMORI et al. 1995 ), which maps close to lacZ and would thus be present in the F' episomes used here. Although the DinB protein is a UmuC homolog, it is not known whether it interacts with UmuD protein, which should be present at some level in umuC122 bacteria. It would be of interest to analyze UV mutagenesis in umuCD mutS bacteria.

Why is antagonism of UV mutagenesis apparently more complete in umuC122 than in umu⁺ bacteria? One possibility is that the high numbers of misinsertions in umu⁺ bacteria (corresponding to revertant frequencies of ~50 x 10^-8 in mut^- umu⁺ cells irradiated to 45 J/m²) saturate the mismatch-repair system, as observed previously in other contexts (SCHAAPER and RADMAN 1989 ; CUPPLES et al. 1990 ). By this hypothesis, the revertant frequencies in umuC122 mut^- cells at 60 J/m² (~5 x 10^-8) would correspond to misinsertion levels below the saturation threshold. However, the apparent constant efficiency of antagonism of UV mutagenesis by mismatch repair in umu⁺ bacteria, as fluences increase from 20 to 45 J/m² (Figure 2), argues against this. Since MutS protein shows lower affinity for photoproduct/base mismatches than for corresponding base/base mismatches (WANG et al. 1999 ; H. WANG and J. B. HAYS, unpublished data), high concentrations of MutS might be needed for efficient antagonism of UV mutagenesis. However, a multicopy plasmid encoding the mutS⁺ gene had little effect on mutant frequencies (Figure 2 and Figure 3); simultaneous overproduction of MutS, MutL, and MutH activities has not been tested. A second explanation would be that one class of photoproducts, perhaps CPDs, is responsible for UV mutagenesis in umuC^- bacteria, whereas both CPDs and [6-4] photoproducts might target mutations in umu⁺ bacteria. Only mutations targeted by one class might be subject to mismatch repair. However, both T-T CPDs and T-T [6-4] photoproducts mismatched opposite A-G are recognized by MutS or its human homolog hMutS, and neither photoproduct opposite A-A is recognized (H. WANG and J. B. HAYS, unpublished data; WANG et al. 1999 ); i.e., there is no evidence for MutS/hMutS preference for one mismatched photoproduct over the other. A third explanation might be that in UV-irradiated umu⁺ bacteria, DNA resynthesis associated with mismatch repair triggered by photoproducts might itself result in new misinsertions, opposite photoproducts or at nonphotoproduct sites, by DNA polymerase III (LAHUE et al. 1989 ) associated with UmuD'₂C and RecA protein.

In bacteria subjected to UV fluences insufficient to induce the SOS response, occasional photoproduct/base mismatches might arise from DNA replication past UV photoproducts, perhaps facilitated by low-level UmuC-independent constitutive activities that allowed error-prone translesion synthesis. If so, these mismatches would appear to be almost all corrected by the MutHLS system. In heavily DNA-damaged SOS-induced bacteria, however, where a burst of mutagenesis to increase genetic variability has been hypothesized to promote species survival (ECHOLS 1982 ), inefficient correction of photoproduct/base mismatches, because of error-prone mismatch-repair resynthesis or for other reasons, might be of less concern. One testable prediction of this model is that UV mutagenesis in mut^- strains independent of UmuC function should not require other SOS functions. Another prediction, that MutS-like proteins specifically bind to mismatched but not to matched photoproducts, has been confirmed for both human hMutS protein (MU et al. 1997 ; WANG et al. 1999 ) and E. coli MutS protein (H. WANG and J. B. HAYS, unpublished observations).

	ACKNOWLEDGMENTS

We thank Claire Cupples, Patricia Foster, Martin Marinus, Jeffrey Miller, and Roger Woodgate for providing bacterial strains, Rick Bockrath, Chris Lawrence, and Roger Woodgate for advice on the manuscript, and Cliff Pereira, Oregon State University, for valuable help with statistical analyses. This work was supported by American Cancer Society grant RPG-96-074-03-CNE to J.B.H. This is technical report 11552 from the Oregon Agricultural Experiment Station.

Manuscript received April 13, 1999; Accepted for publication October 1, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLEN, D. J., A. MAKHOV, M. GRILLEY, J. TAYLOR, and R. THRESHER et al., 1997  MutS mediates heteroduplex formation by a translocation mechanism. EMBO J. 14:4467-4470.

BANERJEE, S. K., R. B. CHRISTENSEN, C. W. LAWRENCE, and J. E. LECLERC, 1988  Frequency and spectrum of mutations produced by a single cis-syn thymine-thymine cyclobutane dimer in a single-stranded vector. Proc. Natl. Acad. Sci. USA 85:8141-8145[Abstract/Free Full Text].

BRIDGES, B. A. and R. J. MUNSON, 1968  Evidence for the mechanism of base change mutation by ultraviolet light in a strain deficient in excision-repair. Proc. R. Soc. Lond. Ser B 171:213-226[Medline].

CAILLET-FAUQUET, P. and G. MAENHAUT-MICHEL, 1988  Nature of the SOS mutator activity: genetic characterization of untargeted mutagenesis in Escherichia coli.. Mol. Gen. Genet. 213:491-498[Medline].

CARTY, M. P., J. HAUSER, A. S. LEVINE, and K. DIXON, 1993  Replication and mutagenesis of UV-damaged DNA templates in human and monkey cell extracts. Mol. Cell. Biol. 13:533-547[Abstract/Free Full Text].

CHRISTENSEN, J. R., J. E. LECLERC, P. V. TATA, R. B. CHRISTENSEN, and C. W. LAWRENCE, 1988  UmuC function is not essential for the production of all targeted lacI mutations induced by ultraviolet light. J. Mol. Biol. 203:635-641[Medline].

CHRISTENSEN, R. B., J. R. CHRISTENSEN, I. KOENIG, and C. W. LAWRENCE, 1985  Untargeted mutagenesis induced by UV in the lacI gene of Escherichia coli.. Mol. Gen. Genet. 201:30-34[Medline].

COOPER, D. L., R. S. LAHUE, and P. MODRICH, 1993  Methyl-directed mismatch repair is bidirectional. J. Biol. Chem. 268:11823-11829[Abstract/Free Full Text].

CROWLEY, D. J. and P. C. HANAWALT, 1998  Induction of the SOS response increases the efficiency of global nucleotide excision repair of cyclobutane pyrimidine dimers, but not 6-4 photoproducts, in UV-irradiated Escherichia coli.. J. Bacteriol. 180:3345-3352[Abstract/Free Full Text].

CUPPLES, C. G. and J. H. MILLER, 1988  Effects of amino acid substitutions at the active site in Escherichia coli ß-galactosidase. Genetics 120:637-644[Abstract/Free Full Text].

CUPPLES, C. G. and J. H. MILLER, 1989  A set of lacZ mutations in Escherichia coli that allow rapid detection of each of the six base substitutions. Proc. Natl. Acad. Sci. USA 86:5345-5349[Abstract/Free Full Text].

CUPPLES, C. G., M. CABRERA, C. CRUZ, and J. H. MILLER, 1990  A set of lacZ mutations in Escherichia coli that allow rapid detection of specific frameshift mutations. Genetics 125:275-280[Abstract].

DRUMMOND, J. T., A. ANTHONEY, R. BROWN, and P. MODRICH, 1996  Cisplatin and adriamycin resistance are associated with MutL and mismatch repair deficiency in an ovarian tumor cell line. J. Biol. Chem. 271:19645-19648[Abstract/Free Full Text].

DUCKETT, D. R., J. T. DRUMMOND, A. I. H. MURCHIE, J. T. REARDON, and A. SANCAR et al., 1996  Human MutS recognizes damaged DNA basepairs containing O⁶-methylguanine, O⁴-methylthymine, or the cisplatin-d(GPG) adduct. Proc. Natl. Acad. Sci. USA 93:6443-6447[Abstract/Free Full Text].

ECHOLS, H., 1982  Mutation rate: some biological and biochemical considerations. Biochemie 64:571-575[Medline].

ELLEDGE, S. J. and G. C. WALKER, 1983  Proteins required for ultraviolet light and chemical mutagenesis: identification of the products of the umuC locus of Escherichia coli.. J. Mol. Biol. 164:175-192[Medline].

FEINSTEIN, S. I. and K. B. LOW, 1986  Hyper-recombining recipient strains in bacterial conjugation. Genetics 113:13-33[Abstract/Free Full Text].

FENG, G., H.-C. T. TSUI, and M. WINKLER, 1996  Depletion of the cellular amounts of MutS and MutH methyl-directed mismatch-repair proteins in stationary-phase Escherichia coli K-12 cells. J. Bacteriol. 178:2388-2396[Abstract/Free Full Text].

FENG, W.-Y. and J. B. HAYS, 1995  DNA structures generated during mismatch repair of nonreplicating phage DNA in Escherichia coli: requirements for helicase, exonuclease, RecF and RecBCD functions. Genetics 140:1175-1186[Abstract].

FENG, W.-Y., E. LEE, and J. B. HAYS, 1991  Recombinagenic processing of UV-light photoproducts in nonreplicating phage DNA by the Escherichia coli methyl-directed mismatch repair system. Genetics 129:1007-1020[Abstract].

FIJALKOWSKA, I. J., R. L. DUNN, and R. M. SCHAAPER, 1997  Genetic requirements and mutational specificity of the Escherichia coli SOS mutator activity. J. Bacteriol. 179:7435-7445[Abstract/Free Full Text].

FOSTER, P. L., 1999  Are adaptive mutations due to a decline in mismatch repair? The evidence is lacking. Mutat. Res. 436:179-184[Medline].

HARRIS, R. S., G. FENG, K. J. ROSS, R. SIDHU, and C. THULIN et al., 1997  Mismatch repair protein MutL becomes limiting during stationary-phase mutation. Genes Dev. 11:2426-2437[Abstract/Free Full Text].

HUMAYUN, M. Z., 1998  SOS and Mayday: multiple inducible mutagenic pathways in Escherichia coli. Mol. Microbiol. 30:905-910[Medline].

JIANG, N. and J.-S. TAYLOR, 1993  In vivo evidence that UV induced C T mutations at dipyrimidine sites could result from the replicative bypass of cis-syn cyclobutane dimers or their deamination products. Biochemistry 32:472-481[Medline].

KAT, A., W. G. THILLY, W.-H. FANG, M. J. LONGLEY, and G.-M. LI et al., 1993  An alkylation-tolerant mutator cell line is deficient in strand-specific mismatch repair. Proc. Natl. Acad. Sci. USA 90:6424-6428[Abstract/Free Full Text].

KATO, T. and Y. SHINOURA, 1977  Isolation and characterization of mutants of Escherichia coli deficient in induction of mutations by ultraviolet light. Mol. Gen. Genet. 156:121-131[Medline].

KORNBERG, A., and T. A. BAKER, 1992 DNA Replication. W. H. Freeman and Company, New York.

KUNZ, B. A. and B. W. GLICKMAN, 1984  The role of pyrimidine dimers as premutagenic lesions: a study of targeted vs. untargeted mutagenesis in the lacI gene of Escherichia coli. Genetics 106:347-364[Abstract/Free Full Text].

LAHUE, R. S., K. G. AU, and P. MODRICH, 1989  DNA mismatch correction in a defined system. Science 245:160-164[Abstract/Free Full Text].

LI, G.-M., H. WANG, and L. J. ROMANO, 1996  Human MutS specifically binds to DNA containing aminofluorene and acetylaminofluorene adducts. J. Biol. Chem. 271:24084-24088[Abstract/Free Full Text].

MELLON, I. and G. N. CHAMPE, 1995  Products of DNA mismatch repair genes mutS and mutL are required for transcription-coupled nucleotide excision repair of the lactose operon in Escherichia coli.. Proc. Natl. Acad. Sci. USA 93:1292-1297[Abstract/Free Full Text].

MELLON, I., D. K. RAJPAL, M. KOI, C. R. BOLAND, and G. N. CHAMPE, 1996  Transcription-coupled repair deficiency and mutations in human mismatch repair genes. Science 272:557-560[Abstract].

MILLER, J. H., (Editor), 1972 Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MILLER, J. H., 1985  Mutational specificity of ultraviolet light. J. Mol. Biol. 182:45-68[Medline].

MU, D., M. TURSUN, D. R. DUCKETT, J. T. DRUMMOND, and P. MODRICH et al., 1997  Recognition and repair of compound DNA lesions (base damage and mismatch) by human mismatch repair and excision repair systems. Mol. Cell. Biol. 17:760-769[Abstract].

NAPOLITANO, R. L., I. B. LAMBERT, and R. P. P. FUCHS, 1997  SOS factors involved in translesion synthesis. Proc. Natl. Acad. Sci. USA 94:5733-5738[Abstract/Free Full Text].

OHMORI, H., E. HATADA, Y. QIAO, M. TSUJI, and R. FUKUDA, 1995  dinP, a new gene in Escherichia coli, whose product shows similarities to UmuC and its homologs. Mutat. Res. 347:1-7[Medline].

PENG, W. and B. R. SHAW, 1996  Accelerated deamination of cytosine residue in UV-induced cyclobutane dimers leads to CCTT transitions. Biochemistry 35:10172-10181[Medline].

PERSON, S., W. MCCLOSKEY, W. SNIPES, and R. BOCKRATH, 1974  Ultraviolet mutagenesis and its repair in Escherichia coli strain containing a nonsense codon. Genetics 78:1035-1049[Abstract/Free Full Text].

PETIT, M. A., J. DIMPFL, M. RADMAN, and H. E. ECHOLS, 1991  Control of large chromosomal duplications in E. coli by the mismatch repair system. Genetics 129:327-332[Abstract].

RAYSSIGUIER, C., D. S. THALER, and M. RADMAN, 1989  The barrier to recombination between Escherichia coli and Salmonella typhimurium is disrupted in mismatch-repair mutants. Nature 342:396-401[Medline].

SCHAAPER, R. M. and M. RADMAN, 1989  The extreme mutator effect of Escherichia coli mutD5 results from saturation of mismatch repair by excessive DNA replication errors. EMBO J. 8:3511-3516[Medline].

SELBY, C. P. and A. SANCAR, 1993  Molecular mechanism of transcription-repair coupling. Science 260:53-58[Abstract/Free Full Text].

SELBY, C. P. and A. SANCAR, 1995  Structure and function of transcription-repair coupling factor. I. Structural domains and binding properties. J. Biol. Chem. 270:4882-4889[Abstract/Free Full Text].

SELBY, C. P., E. M. WITKIN, and A. SANCAR, 1991  Escherichia coli mutant deficient in "mutation frequency decline" lacks strand-specific repair: in vitro complementation with purified coupling factor. Proc. Natl. Acad. Sci. USA 88:11574-11578[Abstract/Free Full Text].

STEINBORN, G., 1978  Uvm mutants of Escherichia coli K12 deficient in UV mutagenesis. I. Isolation of uvm mutants and their phenotypical characterization in DNA repair and mutagenesis. Mol. Gen. Genet. 165:87-93[Medline].

SWANN, P., T. R. WATERS, D. C. MOULTON, Y.-Z. XU, and M. EDWARDS et al., 1996  Role of postreplicative DNA mismatch repair in cytotoxic action of thioguanine. Science 273:1109-1111[Abstract].

TESSMAN, I., S.-K. LIU, and M. A. KENNEDY, 1992  Mechanism of SOS mutagenesis of UV-irradiated DNA: mostly error-free processing of deaminated cytosine. Proc. Natl. Acad. Sci. USA 89:1159-1163[Abstract/Free Full Text].

THOMAS, D. C. and T. A. KUNKEL, 1993  Replication of UV-irradiated DNA in human cell extracts—evidence for mutagenic bypass of pyrimidine dimers. Proc. Natl. Acad. Sci. USA 90:7744-7753[Abstract/Free Full Text].

WANG, H., C. W. LAWRENCE, G.-M. LI, and J. B. HAYS, 1999  Specific binding of human MSH2·MSH6 mismatch-repair protein heterodimers to DNA incorporating thymine- or uracil-containing UV-light photoproducts opposite mismatched bases. J. Biol. Chem. 274:16894-16900[Abstract/Free Full Text].

WITKIN, E. M., 1966  Radiation-induced mutations and their repair. Science 252:1345-1353.

WORTH, L. J., S. CLARK, M. RADMAN, and P. MODRICH, 1994  Mismatch repair proteins MutS and MutL inhibit RecA-catalyzed strand transfer between diverged DNAs. Proc. Natl. Acad. Sci. USA 91:3238-3241[Abstract/Free Full Text].

WU, T. H. and M. G. MARINUS, 1994  Dominant-negative mutations in the mutS gene of Escherichia coli.. J. Bacteriol. 176:5393-5400[Abstract/Free Full Text].

YAJIMA, H., M. TAKO, S. YASUHIRA, J. H. ZHAO, and C. ISCHII et al., 1995  A eukaryotic gene encoding an endonuclease that specifically repairs DNA damaged by ultraviolet light. EMBO J. 14:2393-2399[Medline].

ZIEG, J. and S. R. KUSHNER, 1977  Analysis of recombination between two partially deleted lactose operons of Escherichia coli K-12. J. Bacteriol. 131:123-132[Abstract/Free Full Text].

This article has been cited by other articles:

				H. Park, K. Zhang, Y. Ren, S. Nadji, N. Sinha, J.-S. Taylor, and C. Kang Crystal structure of a DNA decamer containing a cis-syn thymine dimer PNAS, December 10, 2002; 99(25): 15965 - 15970. [Abstract] [Full Text] [PDF]