[CANCERRESEARCH57,2715—2720,July1. l997J A Mutation in the MSH5 Gene Results in Alkylation Tolerance1 Sonya Bawa and Wei Xiao2 Department of Microbiology, University of Saskatchewan, 107 Wiggins Road, Saskatoon, Saskatchewan S7N 5E5, Canada ABSTRACT results in the accumulation of single-stranded nicks in DNA, which are ultimately lethal to the cell. This hypothesis is attractive not only DNA methylating agents such as N-methyl-N'-nitro-N-nitrosoguanidine because it provides an alternative to 06 MeG/O4 MeT genotoxicity (MNNG)are potent carcinogens;their carcinogeniceffectis mainlydue to but also because genetic defects in several human MMR genes have the effect of production of O'-methylguanine (06 MeG) on DNA. 06 MeG is not only mutagenic but also toxic to the cell because Mer/Mex cells been linked to hereditary nonpolyposis colon cancer as well as other unable to remove O@MeG are very sensitive to killing by MNNG. It has types of cancers (9—14).However, the abortive MMR hypothesis been proposed that repeated futile mismatch correction of 06 MeG should be viewed with caution because the current supporting cvi containing bp is responsible for the genotoxicity of the O@MeG lesion and dence with mammalian cells is not conclusive and a convenient that loss of mismatch repair activity results in cellular tolerance to 06 mammalian system is not available to vigorously test the hypothesis. MeG, but the hypothesis has not been proved. We used yeast as a model The MMR system consisting of Escherichia coli MutS and MutL to test this hypothesis and found that chromosome deletion of any known homologues has been extensively studied recently in yeast and human nuclear mitotic mismatch repair genes, including MUll, MSH2, MSH3, cells (reviewed in Refs. 14 and 15). Saccharomyces cerevisiae con MSH6, and PMSJ, did not rescue mgtlA 06 MeG DNA repair methyl tains six MutS homologues (Msh) and three MutL homologues. Mshl transferase-deficient cells from killing by MNNG. A large number of is responsible for mitochondrial DNA repair (16). Msh2, Mlhl, and mgtTh, MNNG-tolerant revertants were isolated, among which one cell line, XS-14, has been found to carry a mutated allele of the MSH5 gene. Pmsl are required for the correction of various mismatches because The mutation also affected spore survival but did not increase the spon mutation in any of the three genes results in a markedly enhanced taneous mutation rate. We further demonstrated that a mutated form of spontaneous mutation rate or instability of simple repeats, which is not the MSHS gene, mshs—14, not the msh5@-null mutation, is responsible for further increased in the double mutants (16—18).In contrast, Msh3 the cellular tolerance to MNNG in XS-14 cells. This observation offers an and Msh6 appear to form alternative pathways to correct 2—4-bp alternative model that may reconcile seemingly contradictory observa insertions and deletions or l-bp mismatches, respectively (19, 20). lions of yeast and mammalian cells. Msh4 (21) and MshS (22) have been shown to facilitate meiotic reciprocal recombination but are believed to play no role in the actual INTRODUCTION process of mismatch correction. The abortive MMR hypothesis has been tested in yeast, which is The genome of the cell is dynamic and reacts readily with physical relevant to mammalian systems in that all MTase and MMR proteins and chemical agents, resulting in modifications to its molecular struc are highly conserved between the two. It was found that none of the ture. DNA alkylating agents are present as one of the largest classes msh2iX, msh3& mlhM, or pms1@-null mutations could rescue mgt1@s of environmental chemical carcinogens; some of the alkylating agents strains from MNNG-induced killing and that MMR defects were not are also produced endogenously as a consequence of cellular meta required for the selection of MNNG-tolerant clones (23). The conflict bolic processes. Challenging cells with S@l-type methylating agents between the observations of yeast and human cells suggests that either such as MNNG3 and N-methyl-N-nitrosourea produces mutagenic and the abortive MMR hypothesis is not universal or 06 MeG genotox carcinogenic lesions like @6MeG and (to a lesser extent) O@ MeT, icity is the result of recognition/repair of 06 MeG by a mechanism which, if not corrected by the DNA repair MTases, pair with thymine other than MTase that interferes with DNA replication and results in and guanine, respectively, resulting in transition mutagenesis (1, 2) cell death. The former possibility seems unlikely, because in the and carcinogenesis in mammals (3). absence of MTase the lesion is equally toxic in all organisms studied In the absence of Mlase, bacterial, yeast, and mammalian cells to date (3—5,24). It also remains possible that 06 MeG toxicity is due become extremely sensitive not only to mutagenesis but also to killing to the involvement of yet to be identified MutHLS homologues or to by MNNG, which cannot be explained simply on the basis of induced the specific mutations of known MutHLS homologous genes. mutation rates (3—5).Itis thus evident that persistence of @6MeGand We report here an attempt to further investigate the mechanism of possibly O@MeT lesions is genotoxic, although the mechanism of cell alkylation toxicity and its relationship to MMR. We have isolated and death is not yet clear. The MMR system has been implicated in the characterized a large number of MTase-deficient, MNNG-tolerant cell processing of 06 MeG-containing bp into lethal lesions, because lines; one of the extensively studied mutants appears to carry a recent studies with MerTh'Mex mammalian cells lacking MTase mutation in the MSH5 gene that is associated with the MNNG resist activity (6, 7) suggest that alkylation tolerance is accompanied by loss ant phenotype. of normal MMR activities. This correlation was predicted by the abortive MMR hypothesis (8), which suggested that repeated futile attempts to correct mispairs during replication of a damaged template MATERIALS AND METHODS Received 11/18/96; accepted 4/29/97. Plasmids and Yeast Transformation. Plasmid pWX1149 carries the The costs of publication of this article were defrayed in part by the payment of page MGTJ gene in a single-copy YCp vector. YCp-MLH1 (17) was from Dr. M. charges. This article must therefore be hereby marked advertisement in accordance with Liskay (Oregon Health Sciences University, Portland, OR), p11-2 (YCp 18 U.S.C. Section 1734 solely to indicate this fact. MSH2; Ref. 25) was obtained from Dr. R. Kolodner (Harvard Medical School, I This research was supported by National Cancer Institute of Canada Grant NC1C007412(to W. X.). W. X. is a Research Scientist of the National Cancer Institute of Cambridge, MA), pWBK3-PMSl(YCp-PMS 1; Ref. 26) from Dr. W. Kramer Canada. (Georg-August University, Gottingen, Germany), YCp-MSH3 (27) and YCp 2 To whom requests for reprints should be addressed. Phone: (306) 966-4308; Fax: MSH6 from Dr. 0. Crouse (Emory University, Atlanta, GA) and YCp-MSH4 (306) 966-4311; E-mail: [email protected]. (21) was from Dr. S. Roeder (Yale University, New Haven, CT). Plasmid 3 The abbreviations used are: MNNG, N-methyl-N'-nitro-N-nitrosoguanidine; 0° MeG, 0°-methylguanine;0' MeT, 0―-methylthymine;MMR,mismatch repair; MTase, pNHl89—2 (22) was a gift from Dr. N. Hollingsworth (State University of 0°MeG/04 MeT DNA repair methyltransferase; Msh, MutS homologue. New York, Stony Brook, NY). A 4.5-kb BamHI/HindIII fragment containing 2715 Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1997 American Association for Cancer Research. MSHS MUTATION AND ALKYLATION TOLERANCE strainsStrainGenotypeSourceWXY9IO2MATa Table I S. cerevisiae 54XS-803-2CMATa ura3-52 trpl-M3 leu2-@I mgtM ::LEU2 GAL@Ref. GietzaXS-803-3AMATa leu2-3.112 ura3-52 hisl-2 hom3-lO canirD. Gietz2CrngtI@sLike leu2-3,112 ura3-52 his/-I trp2 canisD. study3AnzgtlL@&Like XS-803-2C but with mgtM::LEU2This study2Crngtl@unlhLXLike XS-803-3A but with mgtM ::LEU2This study2Cmsh5Like XS-803-2C but with mgtM::LEU2 mlhM:: URA3This study2C1ngtl@?nsh5Like XS-803-2C but with msh5 :: URA3This study3Amsh5Like XS-803-2C but with mgtliX::LEU2 msh5::URA3This study3Aingth@sinsh5Like XS-803-3A but with msh5::URA3This study2Cmsh6@Like XS-803-3A but with mgt1@::LEU2 msh5:: URA3This study2C,ngtI@sinsh6@iLike XS-803-2C but with ,nsh6/s ::hisG-URA3-hisGThis study2CmgzI@ansh5@sLike XS-803-2C but with mgtlts ::LEU2 mshóts::hisG-URA3-hisGThis studyXS-l4MNNG-resistant XS-803-3A but with mgtI@::LEU2 msh5@::hisGThis mutant of 2CmgtIiX ::LEU2This study “Universityof Manitoba. Canada. the entire MSH5 gene was cloned into plasmids YCplac33 and YIplac2l 1 (28) onine to score Hom@ revertants from hom3—1Omutants and on YPD for total to form YCp-MSH5 and YIp-MSH5, respectively. Yeast transformation was as cell survival. described previously (23). Plasmid pmgtl@::LEU2 (5) was used to create Quantitation of Cellular Glutathione. Crude yeast cell extract was pre mglU@ strains. pmsh5@::hisG-URA3-hisG was constructed by removing a pared as described previously (32). Glutathione content was estimated by the 0.84-kb CIaI-EcoRV fragment from the MSH5 coding region and inserting a method of Hissin and Hinf (33). 3.8-kb BamHI-Bg(II hisG-URA3-hisG fragment from pNKY5 I (29) at the unique Bg!1l site. Strains and Cell Culture. The S. cerevisiae strains used in this study are RESULTS listed in Table 1. All mgtI@::LEU2 mutants were created as described (5). The The msh6@ Mutant Does Not Rescue mgtTh from Kiffing by msh6@smutants were obtained by SphI+EcoRI digestion of plasmid MNNG. Our previous results had shown that deletion of any of the pmsh6@::hisG-URA3-hisG, constructed by Dr.