Mol Genet Genomics (2002) 266: 848±857 DOI 10.1007/s00438-001-0605-x

O R I GI N A L P A P E R

T. Hryciw á M. Tang á T. Fontanie á W. Xiao

MMS1 protects against replication-dependent DNA damage in Saccharomyces cerevisiae Received: 8 August 2001 / Accepted: 4 October 2001 / Published online: 29 November 2001 Ó Springer-Verlag 2001

Abstract A series of yeast mutants were isolated that are sensitive to killing by the monofunctional DNAalkylating agent methyl methanesulfonate (MMS) but not by UV or X-radiation. We have cloned and characterized one of the corresponding genes, MMS1, and show that the mms1D mutant is dramatically sensitive to killing by MMS and mildly sensitive to UV radiation. mms1D mutants display an elevated level of spontaneous DNA damage and genomic instability. Furthermore, the mms1D cells are sensitive to killing by conditions that induce replication-dependent doublestrand breaks, such as treatment with camptothecin, and incubation of a cdc2-2 strain at the restrictive temperature. rad52D is epistatic to mms1D for MMS and camptothecin sensitivity, indicating that Mms1 acts in concert with Rad52. However, unlike mutants of the RAD52 group, mms1D cells are not sensitive to c-rays, which induce double-strand breaks independently of DNA replication. Together these results suggest a role for an Mms1- dependent, Rad52-mediated, pathway in protecting cells against replication-dependent DNA damage. Keywords Yeast á DNA replication á DNA repair á Recombination á Camptothecin

Communicated by H. Ikeda T. Hryciw1 á M. Tang á T. Fontanie á W. Xiao (&) Department of Microbiology and Immunology, University of Saskatchewan, 107 Wiggins Road, Saskatoon, SK, Canada, S7N 5E5 E-mail: [email protected] Tel.: +1-306-9664308 Fax: +1-306-9664311 Present address: Cameco MS Neuroscience Research Center, Saskatoon City Hospital, 701 Queen Street, Room 5800, Saskatoon, SK, Canada, S7K 0M7 1

Introduction Living cells su€er continual insult from exogenous and endogenous DNA-damaging agents. In a nondividing cell, the lesions per se will be of little consequence unless they interfere with transcription; however, in an actively growing cell, unrepaired DNA damage can have grave repercussions. The process of DNA replication is susceptible to inhibition by many di€erent DNA lesions. Thus, many DNA lesions, such as alkylated bases, abasic sites, thymine dimers, and protein-DNA crosslinks, can block replicating DNA polymerases (Friedberg et al. 1995). If these lesions are not repaired, or if replication cannot restart, the result is cell lethality. In addition, many mutations in genes encoding components of the DNA replication machinery also cause genetic instability. For example, Saccharomyces cerevisiae strains bearing certain mutant alleles of POL3 (Pol d), RAD27 (FEN1), POL30 (PCNA), and RFA1, RFA2, and RFA3 (RFA) all accumulate recombinogenic lesions, including single-strand and double-strand breaks (SSBs and DSBs), during S phase (Tishko€ et al. 1997; Zou and Rothstein 1997; Chen et al. 1998; Merrill and Holm 1998; Chen and Kolodner 1999). Blocked DNA replication forks are thought to be unstable (Michel 2000). In helicase mutant backgrounds in Escherichia coli, blockage of a replication fork leads to formation of a DSB, catalyzed by the RuvABC proteins (Michel et al. 1997; Seigneur et al. 1998). In wildtype cells, it is proposed that RuvAB, together with RecBCD, catalyzes replication fork repair without actually creating a DSB, possibly by annealing the newly synthesized strands (Seigneur et al. 1998). In the rDNA array of S. cerevisiae, Holliday junctions frequently form during S phase, and it is thought that this is indicative of the rescuing of blocked DNA replication forks (Zou and Rothstein 1997). The exact mechanism of this Holliday junction formation during S phase is unknown, although Rad52 is known to be required (Zou and Rothstein 1997). The error-free arm of the post-replication repair

849

pathway in yeast, de®ned by Rad6, Rad18, Mms2, Ubc13, Rad5 and PCNA (Xiao et al. 2000), is thought to facilitate error-free bypass of a replication-blocking lesion, without actually removing the o€ending lesion. This could be accomplished either by repriming DNA replication downstream of the lesion, followed by a recombination-mediated gap ®lling process, or directly with the aid of recombination proteins, as has been described in E. coli . Hence, the S phase-dependent formation of Holliday junctions observed by Zou and Rothstein (1997) is possibly a form of post-replication repair. Using DNA-damaging agents to study replicationblocking lesions can be complicated by the fact that many of these agents have pleiotropic e€ects. Methyl methanesulfonate (MMS), for example, produces predominantly 7-methylguanine and 3-methyladenine (3MeA), but also a small percentage of O6-methylguanine and O4-methylthymine (Beranek 1990). The latter two lesions cause base mispairing, whereas 3MeA blocks DNA replication. UV light induces primarily cyclobutane pyrimidine dimers and (6-4) photoproducts. Pyrimidine dimers can block DNA polymerases (Friedberg et al. 1995), but the use of UV to study this phenomenon is complicated by the fact that these lesions are eciently targeted by the nucleotide excision repair pathway, such that only a fraction of UV-induced DNA damage persists into S phase. An agent that causes only one type of DNA lesion is camptothecin (CPT). CPT inhibits the religation step of topoisomerase I (Top1) activity, the end result being a single-stranded DNA (ssDNA) nick with a molecule of Top1 protein bound to its 3¢ end (Pommier et al. 1998). This proteinbound SSB inhibits DNA replication and can be converted into a DSB by the advancing replication fork (Ryan et al. 1991; Tsao et al. 1993). Thus, CPT is an excellent reagent for studying the e€ects of DNA replication blocks in vivo. One would expect a yeast mutant that is defective exclusively in the repair of replication-dependent DNA damage to be sensitive to killing by MMS but not by c- or X-irradiation. MMS treatment directly produces 3MeA lesions and indirectly causes the accumulation of abasic sites in the genome due to the activities of the base excision repair pathway. Both 3MeA and abasic lesions inhibit DNA replication (Sagher and Strauss 1983; Larson et al. 1985) and are thus presumed to cause replication-dependent DSBs. Furthermore, such a mutant might display sensitivity to the topoisomerase I poison CPT, which also causes replication-dependent DSBs. In a search for mutants that are sensitive to simple DNA alkylating agents but show normal responses to radiation-induced DNA damage, Prakash and Prakash (1977) identi®ed mutants belonging to ®ve complementation groups designated mms1, 2, 4, 5, and 22 which are sensitive to MMS but not to UV or X-rays. Further analysis of mms5 has revealed it to be allelic to MAG1, which encodes a 3MeA DNA glycosylase (Chen et al. 1990). MMS2 has been shown to be a member of

the RAD6 post-replication repair group (Broom®eld et al. 1998). MMS4 might act as a transcriptional (co)activator to increase cellular repair capacity during S phase (Xiao et al. 1998) and mms4 is synthetically lethal in combination with sgs1 (Mullen et al. 2001). The two remaining mutants have yet to be further characterized. Here we present the molecular cloning and genetic characterization of MMS1. The mms1D mutant strain is sensitive to killing by various agents that commonly cause replication-dependent DNA strand breaks, such as MMS and CPT, and rad52D is epistatic to mms1D for MMS and CPT sensitivity. However, mms1D cells are not excessively sensitive to c-rays, and show increased recombination frequencies in an inverted-repeat assay. This sets MMS1 apart from the RAD52 epistasis group. We suggest that MMS1 acts with RAD52 to protect yeast cells from replication-dependent DNA damage, possibly SSBs or DSBs.

Materials and methods S. cerevisiae strains The yeast strains used in this study are listed in Table 1. WX15-1c was isolated from MD-1/FY86 diploid segregants in order to combine mms1-1 with ura3 for library screening. The E. coli strains DH5a (BRL, Gaithersburg, Md.) and NM522 (Pharmacia, Piscataway, N.J.) were used for molecular cloning and plasmid preparation. Yeast cell culture and transformation Yeast cells were cultured at 30°C either in a rich YPD medium or in a synthetic dextrose (SD) medium supplemented with amino acids and bases as described (Sherman et al. 1983). A dimethyl sulfoxideenhanced method (Hill et al. 1991) was used to transform intact yeast cells. For targeted integration, plasmid DNA was digested with restriction enzymes and precipitated with ethanol prior to transformation. Screening of a yeast genomic library A two-step screening protocol was followed. WX15-1c cells were ®rst transformed with a YCp50-based yeast genomic DNA library (Rose et al. 1987) obtained from Mark Rose (Princeton University, Princeton, N.J.). Ura+ transformants obtained on SD-Ura selective plates were then streaked onto YPD and YPD+0.025% MMS. Each plate carried B635 cells as a positive control and the mms1-1 mutant as a negative control. Transformants able to grow on MMS were restreaked from YPD onto MMS plates to con®rm the MMS-resistant phenotype, before being tested for plasmid cosegregation. In the plasmid cosegregation test, cells from a transformant growing on a YPD plate were used to inoculate 2 ml of YPD. A 10-ll volume of overnight culture was used to inoculate a fresh 2ml aliquot of YPD and the incubation was continued overnight. Cells were then diluted and plated onto YPD to isolate single colonies. About 100 colonies derived from each transformant were replica plated onto SD+Ura and SD-Ura. Colonies that grew on SD+Ura but not on SD-Ura were considered to have lost the plasmid. Several Ura+ and Ura- colonies from the same transformant were streaked onto a selective MMS plate. If the Uraphenotype was associated with the MMS-sensitive phenotype, MMS resistance was considered to be plasmid-borne.

850 Table 1 Yeast strains Strain

Genotype

Source

B635 MD-1 FY86 WXY348 THY172 THY173 WX15-1c DBY747 WXY344 WXY501 JC8901 WXY406 WXY394 THY170 WXY376 THY132 WXY387 THY161 6613-53a WXY433 THY109 THY175 B365-14c

MATa cyc1-115 his1 lys2 trp2 MATa cyc1-115 his1 lys2 trp2 mms1-1 MATa his3-D200 leu2D1 ura3-52 GAL+ FY86 with mms1D::LEU2 FY86 with rad52D::URA3 THY172 with mms1D::LEU2 MATa his3-D200 leu2-D1 lys2 ura3-52 mms1-1 MATa his3-D1 leu2-3,112 trp1-289 ura3-52 DBY747 with mms1D::LEU2 DBY747 with mms1D::URA3-MMS1 DBY747 with mag1D::hisG-URA3-hisG DBY747 with mms1D::LEU2 mag1D::hisG-URA3-hisG DBY747 with rad4D::hisG-URA3-hisG WXY394 with mms1D::LEU2 DBY747 with rad6D::LEU2 DBY747 with mms1D::URA3 rad6D::LEU2 DBY747 with rad52D::LEU2 DBY747 with mms1D:: hisG-URA3-hisG top1-7::LEU2 MATa ura3-52 his3-D200 leu2-3,112 hom3 gal1 can1 cdc2-2 6613-53a with rad52D::LEU2 6613-53a with mms1D::LEU2 THY109 with rad52D::URA3 MATa his3::ade2-5¢D-TRP1-ade2-n leu2-3,112 trp1-1 ura3-1, ade2-1 can1-100 B365-14c with mms1D::URA3 MATa leu2-3, 112 ura3-52 his1-7 can1 hom3-10, mms1D::LEU2 MATa leu2-3, 112 ura3-52 his1-1 CAN1 trp2 rad52D::URA3

L. Prakash L. Prakash F. Winston This study This study This study This study D. Botstein This study This study L. Samson This study Laboratory stock This study Laboratory stock This study Laboratory stock This study L. Hartwell This study This study This study L. Symington

WXY687 XS-803-2Cm1 XS-803-3Ar52

This study This study This study

Plasmids and plasmid construction Plasmid pSCP19A (Popo€ et al. 1990) carrying the apn1D::HIS3 cassette was obtained from Bruce Demple (Harvard University, Boston, Mass.). Plasmid pJC8901 (Chen et al. 1990) carrying the mag1D::hisG-URA3-hisG cassette was from Leona Samson (Harvard University). pDG38 (Gietz and Prakash 1988), containing the rad4D::hisG-URA3-hisG cassette, and pDG315 (Kang et al. 1992), containing the rad6D::LEU2 cassette, were obtained from Dan Gietz (University of Manitoba, Winnipeg, MB). Plasmid pBRDHS::LEU2 containing the rad52D::LEU2 cassette was a gift from Dennis Livingston (University of Minnesota, Minneapolis, Minn.). CB25, also known as pCT80 (Thrash et al. 1985), containing the top1-7::LEU2 cassette, was from Michael Christman (University of Virginia, Charlottesville, Va.). For construction of mms1D disruption cassettes, YEp13 (Broach et al. 1979) was used as the LEU2 donor, YDp-U (Berben et al. 1991) was used as the URA3 donor, and pNKY51 (Alani et al. 1987) was used as the hisG-URA3-hisG donor. Plasmid YEpMAG1-lacZ has been described previously (Xiao et al. 1993). Plasmid pZZ13, containing the RNR3-lacZ reporter gene (Zhou and Elledge 1992), was from Steven Elledge (Baylor College of Medicine, Houston, Tex.). pLG669z (Guarente 1983) containing CYC1-lacZ was from Leonard Guarente (MIT, Cambridge, Mass.). The multipurpose plasmids pTZ18R and pTZ19R were purchased from Pharmacia. Restriction and modifying enzymes were purchased from BRL (Bethesda, Md.) and New England Biolabs (Beverly, Mass.) and used as suggested. For deletion analysis, plasmid DNA was cleaved and the incompatible single-stranded ends were blunted either by cleaving the protruding 3¢-end using T4 DNA polymerase, or by ®lling in the 3¢-end by treatment with the Klenow fragment of E. coli DNA polymerase I before DNA ligation. To construct the mms1D::LEU2, mms1D::URA3, and mms1D::hisG-URA3-hisG cassettes, the 6.5-kb BglII-XhoI fragment from pMMS1-6 (Fig. 1) was cloned into the BamHI/SalI sites of pTZ19R to form p19R-MMS1. The 1.6-kb BamHI fragment within the MMS1 insert of p19R-MMS1 was replaced by the 2.7-kb BglII

Fig. 1 Molecular cloning and restriction mapping of the MMS1 gene. Plasmid pMMS1-6, containing a 10.8-kb insert of yeast genomic DNA, was used to generate pMMS1-6DE3, which in turn was used to generate further deletions. The bars represent the fragments that remained in the plasmid. The deletion constructs were transformed into the mms1-1 strain WX15-1c and scored for the MMS-resistant (+) or sensitive ()) phenotype on a plate containing 0.025% MMS. Restriction sites: B, BamHI; Bg, BglII; Bs, Bsu36I; E, EcoRI; EN, EcoNI; Sc, SacI; Xb, XbaI; Xh, XhoI. Also shown are the MMS1 ORF, and the organization of the mms1D::LEU2 (pWX1503), mms1D::URA3 (pWX1501), and mms1D::hisG-URA3- hisG (pWX1505) disruption cassettes fragment from YEp13 to form pWX1503 (mms1D::LEU2), by the 1.1-kb BamHI fragment from YDp-U to form pWX1501 (mms1D::URA3) or by the 3.8-kb BamHI-BglII fragment from pNKY51 to generate pWX1505 (mms1D::hisG-URA3-hisG). For MMS1 disruption pWX1503 was digested with SacI and XbaI, while pWX1501 and pWX1505 were digested with EcoRI, prior to yeast transformation.

851 DNA sequencing Fragments of pMMS1-6 were cloned into either pTZ18R or pTZ19R and various deletion constructs were made by restriction digestion. Single-stranded DNA was obtained by transforming E. coli strain NM522 with the plasmid and superinfecting with the helper phage M13-K07. The constructs were sequenced by the dideoxy chain-termination method (Sanger et al. 1977) using a T7 polymerase sequencing kit (Pharmacia). The DNA sequence of a 5.2-kb region was determined on both strands.

Cell killing by DNA damaging agents MMS, CPT, 4-nitroquinoline-N-oxide (4NQO), and 1,2;3,4-diepoxybutane (DEB) were purchased from Sigma-Aldrich (St. Louis, Mo.). Yeast cells for liquid killing were grown overnight at 30°C in 2 ml of YPD, diluted 10-fold in YPD and the incubation continued for about 4 h until a cell titer of approximately 2´107 cells/ml was reached. Cell cultures were treated with MMS for the times and at the concentrations indicated. Samples were removed, washed, diluted and plated onto YPD. For UV killing experiments, cells were diluted and plated onto YPD and then subjected to UV-irradiation in the dark at the indicated dose. For c-irradiation, cells were collected by centrifugation, resuspended in sterile water, and exposed to c-rays from a 60Co c-ray source at a dose rate of 37 rads/s. The plates were incubated for 3 days at 30°C before counting colonies. CPT was dissolved in 100% DMSO at a concentration of 4 mg/ml. It was used at a ®nal concentration of 0.25 lg/ml or 10 lg/ml in YPD bu€ered to pH 7.2 with 25 mM HEPES. For CPT sensitivity assays, yeast cells were grown overnight at 30°C in 2 ml of YPD, diluted 10-fold in YPD and the incubation continued for approximately 4 h until a cell titer of approximately 2´107 cells/ml was reached. Cells were diluted to 1´107 cells/ml, and 10-fold serial dilutions made. Aliquots (10 ll) of diluted cells were spotted on the appropriate plates, and incubated at 30°C for 2 days. The stock solution of 4NQO (10 mg/ml) was prepared in acetone. For gradient plate assays, 0.66 lg/ml 4NQO or 0.005% DEB was added to 30 ml of molten YPD agar and a bottom-layer gradient was formed by pouring the mixture into tilted, square, petri dishes. After brief solidi®cation, the petri dishes were again placed on a ¯at surface and 30 ml of molten agar in the same medium without the drug was poured onto the plates. A 0.1-ml sample was taken from an overnight culture, mixed with 0.9 ml of 1% molten agar and immediately imprinted onto freshly made gradient plates via a microscope slide. Gradient plates were incubated at 30°C for 2 days. Assay for spontaneous recombination The assay for intrachromosomal recombination between the inverted ade2 alleles in the ade2-5¢D-TRP1-ade2-n substrate was performed as described (Rattray and Symington 1994). Gene conversion, crossovers, or crossovers associated with gene conversions can all restore the ADE2 sequence. The spontaneous recombination rate was determined using the method of the median (Lea and Coulson 1948) from eleven independent colonies per strain. b-Galactosidase (b-gal) assay The b-gal assays were performed as previously described (Xiao et al. 1993). Brie¯y, 0.5 ml of overnight yeast culture was added to 2.5 ml of fresh SD selective medium, and incubation was continued for 2 h. One milliliter of culture was used to determine the cell density (OD600). The remaining cells were used for the b-gal assay, and activity was expressed in Miller units (Guarente 1983).

Results Molecular cloning and nucleotide sequence analysis of MMS1 Screening of a yeast genomic library constructed in YCp50 for a plasmid that restores wild-type MMS resistance to the mms1-1 mutant strain WX15-1c resulted in the identi®cation of three independent clones. Restriction analysis showed that they share a common insert region. Plasmid pMMS1-6, containing a 10.8-kb insert, was studied further. Deletion analysis de®ned the minimal genomic fragment necessary for complementing the MMS-sensitive phenotype to lie within the 4.6-kb EcoRISacI region (Fig. 1). DNA sequencing of the 5.2-kb insert in pMMS1-6DE3 revealed an ORF of 4224 bp, capable of encoding a 1407-amino acid, 161-kDa protein (GenBank Accession No. U14001). Its entire sequence is identical to that found in the subsequently released Saccharomyces Genome Database (ORF yPR164w). To further demonstrate that the cloned MMS1 was allelic to mms1-1, an mms1D::URA3-bearing plasmid was digested within the MMS1 sequence and used to integrate the URA3 gene into the MMS1 locus in the wild-type strain DBY747 (ura3), such that URA3 now represents the cloned MMS1 locus. This strain (WXY501) was crossed with WX15-1c (mms1-1, ura3), and the haploid segregants were analyzed for cosegregation of URA3 and MMS1. Of the 32 spores analyzed, 14 were Ura+ MMSR and 18 were Ura- MMSS, showing complete cosegregation of URA3 with MMS1. Therefore in this assay URA3 was linked to MMS1 and occupies the same locus as mms1-1. Furthermore, the mms1-1 mutation was mapped and isolated by a gaprepair technique (Orr-Weaver and Szostak 1983) and found to be a T®A transversion at nucleotide 1400 (relative to the translation initiation site), resulting in a L467®Amber nonsense mutation and a truncated 466residue Mms1 protein. Using the BLAST program (Altschul et al. 1990), we found no signi®cant similarity between Mms1 and other polypeptide sequences in the GenBank database (release 115.0). There are, however, ®ve putative transmembrane domains, suggesting that Mms1 might be a transmembrane protein. Also of note is the presence of a pyridoxal phosphate attachment site sequence shared by certain amino acid decarboxylases (Jackson 1990; Sandmeier et al. 1994). The MMS1 gene has also been identi®ed by David Botstein's group (Stanford University, Stanford, Calif.) as KIM3 (killed in mutagen) in a genome-wide screen for mutants sensitive to the DNA cross-linking agents mitomycin C or DEB (Saccharomyces Genome Database, http://genome-www.stanford.edu/Saccharomyces). Phenotypes of mms1D mutants The mms1-1 mutant was originally isolated for its enhanced sensitivity to killing by MMS, but not by UV

852

and X-rays (Prakash and Prakash 1977). We compared an mms1D null mutant with its isogenic wild-type strain (Figs. 2 and 3) and found that the mms1D mutant was indeed sensitive to MMS (see Fig. 3), mildly sensitive to UV-irradiation (Fig. 2A), and not sensitive to killing by c-radiation (Fig. 2B). The di€erence in UV sensitivities between what was observed for mms1D and what was reported for mms1-1 indicates that the mms1-1 allele might retain some activity. It is also possible that mms11 is a partial loss-of-function mutation in the MD-1 strain as a result of a low level of translational readthrough, or that the observed di€erence was simply due to the di€erent genetic backgrounds of the two strains (MD-1 versus DBY747). To further address the issue, we compared wild type and mms1D cells for sensitivity to the UV mimetic agent 4NQO. 4NQO introduces bulky DNA adducts that, like UV-induced DNA damage, are mainly removed by nucleotide excision repair (Friedberg et al. 1995). As shown in Fig. 2C, the mms1D mutant is sensitive to 4NQO in a gradient plate assay, although to a lesser degree than is the rad4D mutant, which is defective in the nucleotide excision repair pathway, or the rad52D mutant defective for recombinational repair. Fig. 2A±D Sensitivity of the mms1D mutant to varous DNAdamaging agents. A UV radiation. Strains used were: DBY747 (wild type), WXY344 (mms1D) and WXY387 (rad52D). The latter strains were isogenic to DBY747. B c-Radiation. Strains used were: FY86 (wild type), WXY348 (mms1D), THY172 (rad52D) and THY173 (mms1D rad52D). The latter strains were isogenic to FY86. C 4NQO. D DEB. Strains used in C and D were DBY747 (wt), WXY344 (mms1D), WXY394 (rad4D), and WXY387 (rad52D). All strains are isogenic to DBY747. The values shown for liquid killing (A, B) are the averages of two independent experiments. Gradient plates (C, D) were incubated for 2 days at 30°C before measuring relative growth

In addition, the mms1D mutant is also sensitive to other methylating and ethylating agents, such as 1-methyl-3-nitro-1-nitrosoguanidine, 1-ethyl-3-nitro-1nitrosoguanidine, N-nitroso-N-methylurea, and N-nitroso-N-ethylurea (data not shown), as well as to DNA cross-linking agents such as DEB (Fig. 2D). During cloning of MMS1 we noticed a slow growth phenotype for MD-1 and WX15-1c (both were mms1-1). Cell growth rates were therefore measured for the wild type and its mms1D derivative in rich medium. The doubling time for the wild type was calculated to be 1.6 h, whereas for its isogenic mms1D mutant the value was 2.4 h. Plating eciency is not a€ected by the mms1D mutation, and the strain grows equally well at 23 or 37°C. The mms1D cells have an abnormal morphology; some cells are very large and round, whereas others are hyper-elongated and display a ®lamentous-like growth pattern of colony formation on solid media (data not shown). We suspect that mms1D cells might su€er from endogenous DNA damage that alters cell cycle kinetics in these mutants, resulting in increased doubling times and an abnormal cellular morphology. Interestingly, these same morphological phenotypes are shared by

853 Fig. 3A±D Epistasis analysis of mms1D with other mutations a€ecting DNA repair. A mms1D and mag1D. B mms1D and rad4D. C mms1D and rad6D. D mms1D and rad52D. Strains used in panels A±C were: DBY747 (wt), WXY344 (mms1D), JC8901 (mag1D), WXY406 (mms1D mag1D), WXY394 (rad4D), THY170 (mms1D rad4D), WXY376 (rad6D), and THY132 (mms1D rad6D); these strains were DBY747 derivatives. The strains used in D were: FY86 (wt), WXY348 (mms1D); THY172 (rad52D); and THY173 (mms1D rad52D); these strains were FY86 derivatives. Values shown are the average of at least two independent experiments

recombination-defective mutants such as rad50D and rad52D (data not shown), suggesting that mms1D cells and mutants of the RAD52 group might su€er from similar defects. rad52D is epistatic to mms1D for MMS sensitivity Since MMS1 was initially identi®ed as an alkylation damage-speci®c gene, epistasis analysis was performed to determine whether MMS1 operates within the base excision repair pathway for the repair of DNA alkylation damage. mms1D mag1D and mms1D apn1D double mutants were created and their MMS sensitivities were compared to the corresponding single mutants. The phenotypic e€ect of mms1D appeared to be additive to both mag1D (Fig. 3A) and apn1D (not shown), defective in 3MeA DNA glycosylase and apurinic/apyrimidinic endonuclease, respectively, suggesting that MMS1 does not belong to the base excision repair pathway. Mutants for the RAD3 nucleotide excision repair pathway, the RAD6 post-replication repair pathway and the RAD52 recombination repair pathway display various degrees of

MMS sensitivity (Friedberg et al. 1995; Xiao et al. 1996). With this in mind, we carried out an MMS killing experiment to determine whether or not MMS1 operates within these pathways. As in the case of the mms1D mag1D double mutants, the e€ect of mms1D on sensitivity to MMS is additive to that of both rad4D (Fig. 3B) and rad6D (Fig. 3C), suggesting that Mms1 does not operate in the nucleotide excision or post-replication repair pathways for DNA damage resistance. Since we had diculty in creating an mms1D rad52D double mutant strain in the DBY747 background, we originally suspected that mms1D and rad52D might be synthetically lethal. We therefore tried to construct the double mutant in a di€erent genetic background to see if DBY747 contained an unknown mutation that rendered mms1D rad52D cells inviable. XS-803-3Ar52 (MATa rad52D::URA3) and XS-803-2Cm1 (MATa mms1D:: LEU2) were crossed, the diploid induced to sporulate, and the spores analyzed. Of 56 spores examined, 13 of an expected 14 Leu+ Ura+ cells were recovered, indicating that the mms1D rad52D double mutant is viable in this strain background. The double mutant was also constructed in the FY86 genetic background by two

854

rounds of transformation with appropriate disruption cassettes, and this strain was examined further. As shown in Fig. 3D, rad52D is epistatic to mms1D for MMS sensitivity; the mms1D rad52D strain is no more sensitive to MMS than is the single rad52D mutant. This was not anticipated, as RAD52-group mutants are extremely sensitive to killing by c-irradiation, whereas mms1D cells are not. Evidence for elevated rates of spontaneous DNA damage and genomic instability in mms1D cells If mms1D and rad52D mutants su€er from similar defects, then it would be anticipated that the mms1D mutation a€ects certain recombination events. We used an intrachromosomal inverted-repeat assay to measures rates of recombination between two ade2 heteroalleles (Rattray and Symington 1994). The observed recombination rate for the mms1D strain was ten-fold higher than that of the corresponding wildtype strain; the value for the mms1D strain was 2.53´10±3, as compared to 2.5´10±4 for the wild type. This observation indicates that recombinogenic lesions accumulate in the mms1D cells, and supports the hypothesis that Mms1 either prevents these lesions from arising, or plays a role in their repair. We reasoned that if mms1D cells indeed have an elevated level of genomic DNA damage, then DNA-damage-inducible genes might be activated in the absence of exogenous damaging agents. To quantify the e€ect of the mms1D mutation on the basal-level transcription of DNA-damage-inducible genes, b-gal activities of MAG1-lacZ and RNR3-lacZ transformants were analyzed in the wild type and in isogenic mms1D strains. As shown in Table 2, basal-level transcription of MAG1-lacZ and RNR3-lacZ is elevated in the mms1D mutant by two- and ®vefold, respectively. These results are consistent with our previous observations (Zhu and Xiao 1998) that RNR3 induction is more sensitive to DNA damage than is MAG1 induction. These data indicate that the mms1D mutant does indeed contain an elevated level of DNA damage in its genome. At least some of these lesions, if unrepaired, are recombinogenic, Table 2 The e€ect of mms1D mutation on the lacZ fusion gene expression

suggesting that they are, or can be processed into, DNA strand breaks. The mms1D mutation renders cells sensitive to replication-induced DNA damage Taking into account the fact that, unlike known RAD52 group mutants, mms1D is sensitive to a variety of DNAdamaging chemicals but not sensitive to ionizing radiation, we propose that MMS1 may speci®cally protect cells from replication-dependent DNA strand breaks. In order to obtain further evidence for this hypothesis, we performed two experiments to address the issue, assaying for sensitivity of the mms1D mutant to CPT and to cdc2-2-induced DNA damage. CPT is a Top1 inhibitor that traps Top1 in the cleavage complex, producing a protein-bound SSB in the DNA (Pommier et al. 1998). DNA replication can convert this single-stranded nick into a DSB (Ryan et al. 1991; Tsao et al. 1993). As shown in Fig. 4A, mms1D cells are indeed sensitive to CPT in a Top1-dependent manner. This is reminiscent of recombination-defective mutants (e.g., rad52) which are also sensitive to killing by CPT (Eng et al. 1988; Nitiss and Wang 1988). While deletion of TOP1 alleviates the CPT sensitive phenotype of the mms1D mutant, it does not a€ect MMS sensitivity (Fig. 4A). Deletion of MMS1 in the rad52D background did not further enhance the sensitivity of the rad52D mutant to CPT (Fig. 4B). Hence, rad52D is epistatic to mms1D with respect to CPT sensitivity, as it is for MMS sensitivity. DNA damage can be induced by a faulty DNA polymerase as well as by treatment with a damaging agent. POL3 (CDC2) encodes the catalytic subunit of DNA polymerase d. The cdc2-2 mutation induces DNA damage at the restrictive temperature, as indicated by the triggering of the G2/M DNA damage-dependent cell cycle checkpoint (Weinert and Hartwell 1993) and the appearance of Holliday junctions during S phase (Zou and Rothstein 1997). Deletion of either RAD52 or MMS1 alone or in combination in the cdc2-2 background decreases the survival of the cdc2-2 strain after a 4-h incubation at the non-permissive temperature to comparable levels (Table 3).

Straina

lacZ fusion construct

b-Galactosidase activity (Miller units)b

Relative induction factor

DBY747 (wt) WXY344 (mms1D) DBY747 WXY344 DBY747 WXY344

MAG1-lacZ MAG1-lacZ RNR3-lacZ RNR3-lacZ CYC1-lacZ CYC1-lacZ

0.23‹0.082 0.49‹0.086 0.64‹0.096 3.3‹0.78 9.6‹3.3 10‹2.1

1 2.1 1 5.1 1 1.1

a DBY747 (wild type) and WXY344 (mms1D) cells were transformed with YEpMAG1-lacZ (MAG1lacZ), pZZ13 (RNR3-lacZ) or pLG669z (CYC1-lacZ), b-galactosidase activity was measured, and the results are expressed in Miller units b Data represent the averages (‹standard deviation) of two (CYC1-lacZ) or three (MAG1-lacZ and RNR3-lacZ) independent experiments

855 Table 3 Survival of cdc2-2 strain derivatives after incubation at restrictive temperaturea Straina

Genotype

Survival (%)b

6613-53a THY109 WXY433 THY175

cdc2-2 cdc2-2 mms1D cdc2-2 rad52D cdc2-2 rad52D mms1D

27.2‹6.7 12.5‹1.5 11.3‹2.9 9.7‹2.2

a

Log-phase cultures growing at 23°C were shifted to 37°C for 4 h. Cells were plated and scored for survival after incubation at 23°C for 3±4 days b Data represent the averages (‹standard deviation) of three independent experiments

Fig. 4A, B mms1D cells are sensitive to replication-dependent DNA-damaging agents. A Top1-dependent camptothecin sensitivity of mms1D mutants. Ten-fold serial dilutions beginning with 107 cells/ml of DBY747 (wild type), WXY344 (mms1D), THY161 (mms1D top1), and WXY387 (rad52D) log phase cultures were spotted onto YPD (left), YPD+10 lg/ml CPT (middle), and YPD+0.01% MMS (right). All strains are isogenic to DBY747. Plates were incubated for 3 days at 30°C before photographing. B rad52D is epistatic to mms1D for camptothecin sensitivity. Ten-fold serial dilutions beginning with 107 cells/ml of FY86 (wild type), WXY348 (mms1D), THY172 (rad52D), and THY173 (mms1D rad52D) log phase cultures were spotted onto YPD (left) and YPD+0.25 lg/ml CPT (right). Plates were incubated for 2 days at 30°C before photographing. All strains are isogenic to FY86

Discussion MMS1 protects cells from accumulating DNA strand breaks We have shown that mms1D cells are sensitive to killing by a variety of DNA-damaging agents, including UV and UV-mimetic agents, methylating and ethylating agents, and DNA cross-linking agents, but are no more sensitive to c-radiation than is the wild type. This phenotype distinguishes MMS1 from all other known DNA repair pathway genes. For example, mutants that are defective in the base excision repair pathway (e.g., mag1 and apn1) are sensitive to base-damaging agents, but not to UV and c-rays; mutants in which the nucleotide excision repair pathway is a€ected (e.g., rad4) are extremely sensitive to UV and 4NQO, but display only moderate sensitivity to MMS and ionizing radiation; mutants in the major post-replication repair pathway (e.g., rad6) are sensitive to a broad range of DNAdamaging agents including MMS, UV, and ionizing radiation; and mutants for the recombination repair pathway (e.g., rad50 and rad52) are extremely sensitive to MMS and ionizing radiation, but are only moderately sensitive to UV (Friedberg et al. 1995). It was therefore surprising to ®nd that rad52D is epistatic to mms1D for MMS sensitivity. As outlined below, we believe that MMS1 acts in concert with RAD52 to repair a subset of replication-dependent DNA lesions.

Several lines of evidence allow us to conclude that recombinogenic lesions accumulate in the mms1D mutant. The mms1D mutant shares many similar phenotypes with recombination-defective mutants (e.g., rad52D). Both are sensitive to killing by DNA-alkylating agents and CPT, decrease the viability of a cdc2-2 strain after incubation at the non-permissive temperature, grow more slowly than wild-type cells, and have a similar abnormal cellular morphology. These observations suggest that mms1 and rad52 cells su€er similar types of DNA damage, possibly DSBs. However, mms1D cells are hyper-recombinant in an intrachromosomal recombination assay. Other RAD52-group mutants, such as rad50, rad51, rad54, rad55, rad57, mre11, and xrs2 (see, for example, Alani et al. 1990; Ivanov et al. 1992; Ajimura et al. 1993; McDonald and Rothstein 1994) are hyper-recombinant in certain assays, but in the assay used in this study, all RAD52group mutants tested display a decreased recombination rate (Rattray and Symington 1994, 1995; Bai and Symington 1996). This observation complicates the interpretation of the epistatic relationship between rad52D and mms1D. Replication-dependent DNA damage and MMS1 In order to reconcile the ®ndings that the mms1D mutant is sensitive to many DNA-damaging agents which can indirectly cause DSBs, is not sensitive to c-rays, and yet is hypostatic to rad52D, we propose that MMS1 may play a speci®c role in the repair of replication-dependent DNA damage. Many DNA lesions are able to inhibit DNA replication. The stalled replication fork might give rise to a replication-dependent DSB. In contrast, DSBs induced by ionizing radiation are considered to be mostly replication-independent. In this study, we examined the e€ect of deletion of MMS1 under two conditions thought to induce replication-dependent DSBs: CPT treatment, where ssDNA nicks produced in the leading strand replication template lead to DSBs; and incubation of cdc2-2 mutant cells at the restrictive temperature, which induces the formation of Holliday junctions. We suggest that Mms1 acts to prevent the formation, or repair, of replication-dependent lesions in

856

cooperation with Rad52. If Mms1 and Rad52 could independently repair the same lesion, an additive e€ect on sensitivity would be observed in the double mutant. Our epistasis analyses suggest otherwise. However, because the sensitivity of the mms1D mutant to MMS and CPT is much less than that of the rad52D mutant, Mms1 is likely to function on only a subset of the lesions induced by these agents. The hyper-rec phenotype of mms1D cells suggests that, in the absence of Mms1, at least some of these lesions will be acted upon by a Rad52-dependent, Mms1-independent recombinational repair pathway. Possible mechanism(s) of Mms1 function In E. coli, the link between replication and recombination is well established. A model has been proposed whereby Holliday junctions forming at blocked replication forks are repaired by RuvAB and RecBCD, without the production of a DSB (Seigneur et al. 1998). It is possible that in S. cerevisiae Mms1 acts with recombination proteins to repair replication-dependent DNA damage without actual DSB formation, essentially repriming DNA replication, by analogy to the situation described above for E. coli. In this case, the absence of Mms1 would lead to DSB formation and the hyper-recombinant phenotype observed due to an Mms1-independent, Rad52-dependent DSB repair pathway. Another possibility is that Mms1 operates in a Rad52-dependent pathway that speci®cally repairs replication-dependent DSBs. A DSB in the context of a replication fork probably di€ers from a replicationindependent DSB not only in the structure of the DNA break, but in the proteins found in the vicinity of the break (i.e. the replication complex). Thus it is possible that additional factors other than those encoded by genes of the RAD52 epistasis group are required for the repair of replication-dependent DSBs. The RAD6 post-replication repair pathway is thought to mediate tolerance of DNA damage acquired during S phase, but the lack of CPT sensitivity in rad6 mutants suggests that not all S phase-dependent DNA damage can be rendered tolerable by this pathway. We propose that an Mms1-mediated mechanism enables the cell to tolerate at least some of these lesions (i.e. replication-dependent DSBs). In the absence of Mms1, another Rad52-dependent pathway may repair some of these DSBs, albeit less eciently, resulting in the observed hyper-rec and MMS-sensitive phenotypes of the mms1D mutant. Isolation and characterization of a novel gene in both Schizosaccharomyces pombe and S. cerevisiae whose mutants display phenotypes reminiscent of those of mms1D was recently reported. The S. pombe mus81 mutant is sensitive to killing by UV and hydroxyurea, but not to c-irradiation (Boddy et al. 2000). For UV sensitivity, rhp51, defective in recombination repair, is epistatic to mus81. Mutating MUS81 in the cdc6-23 or

pol1-1 background, in which temperature-sensitive DNA polymerases d and a, respectively, are produced, decreases the ability of those strains to survive incubation at the restrictive temperature (Boddy et al. 2000). A mus81 culture contains elongated cells, reminiscent of S. pombe recombination-defective mutants. Thus, S. pombe mus81 cells share similar phenotypes with S. cerevisiae mms1 cells. Boddy et al. (2000) hypothesize that Mus81 is required for the resolution/repair of aberrant DNA structures that accumulate during S phase. S. cerevisiae mus81 cells are sensitive to MMS and UV, but not to c-irradiation or HO endonuclease-induced DSBs (Interthal and Heyer 2000). rad54D is epistatic to mus81D for UV sensitivity, and Mus81 physically interacts with Rad54. Interthal and Heyer (2000) propose a role for Mus81 in cooperating with recombinational proteins to repair certain lesions without recombinant formation. Given the similarities between S. pombe and S. cerevisiae mus81 strains and the S. cerevisiae mms1 strain, it will be of interest to determine whether Mms1 and Mus81 function in the same damage resistance pathway. Acknowledgements The authors wish to thank Louise Prakash for providing the mms1-1 mutant and Mark Rose for the yeast genomic library. We also thank many laboratories for plasmids and yeast strains, Carrie Walowsky for information regarding the use of camptothecin, and Stacey Broom®eld, Mahmood Chamankhah and other lab members for valuable discussions. This work was supported by a Canadian Institutes of Health Research operating grant (MOP-38104) to WX. TH was supported by a University of Saskatchewan Graduate Scholarship.

References Ajimura M, Leem SH, Ogawa H (1993) Identi®cation of new genes required for meiotic recombination in Saccharomyces cerevisiae. Genetics 133:51±66 Alani E, Cao L, Kleckner N (1987) A method for gene disruption that allows repeated use of URA3 selection in the construction of multiply disrupted yeast strains. Genetics 116:541±545 Alani E, Padmore R, Kleckner N (1990) Analysis of wild-type and rad50 mutants of yeast suggests an intimate relationship between meiotic chromosome synapsis and recombination. Cell 61:419±36 Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215:403±410 Bai Y, Symington LS (1996) A Rad52 homolog is required for RAD51-independent mitotic recombination in Saccharomyces cerevisiae. Genes Dev 10: 2025±2037 Beranek DT (1990) Distribution of methyl and ethyl adducts following alkylation with monofunctional alkylating agents. Mutat Res 231: 11±30 Berben G, Dumont J, Gilliquet V, Bolle PA, Hilger F (1991) The YDp plasmids: a uniform set of vectors bearing versatile gene disruption cassettes for Saccharomyces cerevisiae. Yeast 7:475± 477 Boddy MN, Lopez-Girona A, Shanahan P, Interthal H, Heyer WD, Russell P (2000) Damage tolerance protein Mus81 associates with the FHA1 domain of checkpoint kinase Cds1. Mol Cell Biol 20:8758±8766 Broach JR, Strathern JN, Hicks JB (1979) Transformation in yeast: development of a hybrid cloning vector and isolation of the CAN1 gene. Gene 8:121±133

857 Broom®eld S, Chow BL, Xiao W (1998) MMS2, encoding a ubiquiting-conjugating-enzyme-like protein, is a member of the yeast error-free postreplication repair pathway. Proc Natl Acad Sci USA 95:5678±5683 Chen C, Kolodner RD (1999) Gross chromosomal rearrangements in Saccharomyces cerevisiae replication and recombination defective mutants. Nature Genet 23:81±85 Chen C, Uzemo K, Kolodner RD (1998) Chromosomal rearrangements occur in S. cerevisiae rfa1 mutator mutants due to mutagenic lesions produced by double-strand-break repair. Mol Cell 2:9±22 Chen J, Der¯er B, Samson L (1990) Saccharomyces cerevisiae 3-methyladenine DNA glycosylase has homology to the AlkA glycosylase of E.coli and is induced in response to DNA alkylation damage. EMBO J 9:4569±4575 Eng WK, Faucette L, Johnson RK, Sternglanz R (1988) Evidence that DNA topoisomerase I is necessary for the cytotoxic e€ects of camptothecin. Mol Pharmacol 34:755±760 Friedberg EC, Walker GC, Siede W (1995) DNA repair and mutagenesis. ASM Press, Washington, D.C. Gietz RD, Prakash S (1988) Cloning and nucleotide sequence analysis of the Saccharomyces cerevisiae RAD4 gene required for excision repair of UV-damaged DNA. Gene 74:535±541 Guarente L (1983) Yeast promoter and lacZ fusions designed to study expression of cloned genes in yeast. Meth Enzymol 101:181±191 Hill J, Donald KAIG, Griths DE (1991) DMSO-enhanced whole cell yeast transformation. Nucleic Acids Res 19:5791 Interthal H, Heyer WD (2000) MUS81 encodes a novel helixhairpin-helix protein involved in the response to UV- and methylation-induced DNA damage in Saccharomyces cerevisiae. Mol Gen Genet 263:812±827 Ivanov EL, Korolev VG, Fabre F (1992) XRS2, a DNA repair gene of Saccharomyces cerevisiae, is needed for meiotic recombination. Genetics 132:651±664 Jackson FR (1990) Prokaryotic and eukaryotic pyridoxal-dependent decarboxylases are homologous. J Mol Evol 31:325±329 Kang XL, Yadao F, Gietz RD, Kunz BA (1992) Elimination of the yeast RAD6 ubiquitin conjugase enhances base-pair transitions and GC®TA transversions as well as transposition of the Ty element: implications for the control of spontaneous mutation. Genetics 130:285±294 Larson K, Sahm J, Shenkar R, Strauss B (1985) Methylation-induced blocks to in vitro DNA replication. Mutation Res 15:77±84 Lea DE, Coulson CA (1948) The distribution of the numbers of mutants in bacterial populations. J Genet 49:264±284 McDonald JP, Rothstein R (1994) Unrepaired heteroduplex DNA in Saccharomyces cerevisiae is decreased in RAD1 RAD52-independent recombination. Genetics 137:393±405 Merrill B, Holm C (1998) A requirement for recombinational repair in Saccharomyces cerevisiae is caused by DNA replication defects of mec1 mutants. Genetics 153:595±605 Michel B (2000) Replication fork arrest and DNA recombination. Trends Biochem Sci 25:173±178 Michel B, Ehrlich SD, Uzest M (1997) DNA double-strand breaks caused by replication arrest. EMBO J 16:430±438 Mullen JR, Kaliraman V, Ibrahim SS, Brill SJ (2001) Requirement for three novel protein complexes in the absence of the Sgs1 DNA helicase in Saccharomyces cerevisiae. Genetics 157:103±118 Nitiss J, Wang JC (1988) DNA topoisomerase-targeting drugs can be studied in yeast. Proc Natl Acad Sci USA 85:7501±7505 Orr-Weaver TL, Szostak JW (1983) Yeast recombination: the association between double-strand gap repair and crossing over. Proc Natl Acad Sci USA 80:4417±4421 Pommier Y, Pourquier P, Fan Y, Strumberg D (1998) Mechanism of action of eukaryotic DNA topoisomerase I and drugs targeted to the enzyme. Biochim Biophys Acta 1400:83±106 Popo€ SC, Spira AI, Johnson AW, Demple B (1990) Yeast structural gene (APN1) for the major apurinic endonuclease: homology to Escherichia coli endonuclease IV. Proc Natl Acad Sci USA 87:4193±4197

Prakash L, Prakash S (1977) Isolation and characterization of MMS-sensitive mutants of Saccharomyces cerevisiae. Genetics 86:33±55 Rattray AJ, Symington LS (1994) Use of a chromosomal inverted repeat to demonstrate that the RAD51 and RAD52 genes of Saccharomyces cerevisiae have di€erent roles in mitotic recombination. Genetics 138:587±595 Rattray AJ, Symington LS (1995) Multiple pathways for homologous recombination in Saccharomyces cerevisiae. Genetics 139:45±56 Rose MD, Novick P, Thomas JH, Botstein D, Fink GR (1987) A Saccharomyces cerevisiae genomic plasmid bank based on a centromere-containing shuttle vector. Gene 60:237±243 Ryan AJ, Squires S, Strutt HL, Johnson RT (1991) Camptothecin cytotoxicity in mammalian cells is associated with the induction of persistent double strand breaks in replicating DNA. Nucleic Acids Res 19:3295±3300 Sagher D, Strauss B (1983) Insertion of nucleotides opposite apurinic/apyrimidinic sites in deoxyribonucleic acid during in vitro synthesis: uniqueness of adenine nucleotides. Biochemistry 22:4518±4526 Sandmeier E, Hale TI, Christen P (1994) Multiple evolutionary origin of pyridoxal-5¢-phosphate-dependent amino acid decarboxylases. Eur J Biochem 221:997±1002 Sanger F, Nicklen S, Coulsen AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74:5463± 5467 Seigneur M, Bidnenko V, Ehrlich SD, Michel B (1998) RuvAB acts at arrested replication forks. Cell 95:419±430 Sherman F, Fink GR, Hicks J (1983) Methods in yeast genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Thrash C, Bankier AT, Barrell BG, Sternglanz R (1985) Cloning, characterization, and sequence of the yeast DNA topoisomerase I gene. Proc Natl Acad Sci USA 82:4374±4378 Tishko€ DX, Filosi N, Gaida GM, Kolodner RD (1997) A novel mutation avoidance mechanism dependent on S. cerevisiae RAD27 is distinct from DNA mismatch repair. Cell 88:253±263 Tsao YP, Russo A, Nyumuswa G, Silber R, Liu LF (1993) Interaction between replication forks and topoisomerase I-DNA cleavable complexes: studies in a cell-free SV40 DNA replication system. Cancer Res 53:5908±5914 Weinert TA, Hartwell LH (1993) Cell cycle arrest of cdc mutants and speci®city of the RAD9 checkpoint. Genetics 134:63±80 Xiao W, Singh KK, Chen B, Samson L (1993) A common element involved in transcriptional regulation of two DNA alkylation repair genes (MAG1 and MGT1) of Saccharomyces cerevisiae. Mol Cell Biol 13:7213±7221 Xiao W, Chow BL, Rathgeber L (1996) The repair of DNA methylation damage in Saccharomyces cerevisiae. Curr Genet 30: 461±468 Xiao W, Chow BL, Milo CN (1998) Mms4, a putative transcriptional (co)activator, protects Saccharomyces cerevisiae cells from endogenous and environmental DNA damage. Mol Gen Genet 257:614±623 Xiao W, Chow BL, Broom®eld S, Hanna M (2000) The Saccharomyces cerevisiae RAD6 group is composed of an error-prone and two error-free postreplication repair pathways. Genetics 155:1633±1641 Zhou Z, Elledge SJ (1992) Isolation of crt mutants constitutive for transcription of the DNA damage inducible gene RNR3 in Saccharomyces cerevisiae. Genetics 131:851±866 Zhu Y, Xiao W (1998) Di€erential regulation of two closely clustered yeast genes, MAG1 and DDI1, by cell-cycle checkpoints. Nucleic Acids Res 26: 5402±5408 Zou H, Rothstein R (1997) Holliday junctions accumulate in replication mutants via a RecA homolog-independent mechanism. Cell 90: 87±96