Mismatch Repair Mutants in Yeast Are Not Defective in Transcription-Coupled DNA Repair of UV-Induced DNA Damage

Kevin S. Sweder,* Richard A. Verhage,t David J. Crowley," Gray F. Grouse,* Jaap Brouwert and Philip C. Hanawalt" *Department of Biological Sciences, Stanford University, Stanford, California 94305-5020, fLaboratory of Molecular Genetics, Leiden Znstitute of Chemistry, Gorlaeus Laboratories, University of Leiden, 2300 RA Leiden, The Netherlands and $Department of Biology, Emory University, Atlanta, Georgia 30322 Manuscript received December 4, 1995 Accepted for publication April 8, 1996

ABSTRACT Transcription-coupled repair, the targeted repairof the transcribed strandsof active genes, is defective in bacteria, yeast, and human cells carrying mutations in mfd, RAD26 and ERCC6, respectively. Other factorsprobably are also uniquely involved in transcription-repair coupling. Recently, a defect was described in transcription-coupled repair forEscherichia coli mismatch repair mutants andhuman tumor cell lines with mutations in mismatch repair genes. We examined removal of UV-induced DNA damage in yeast strains mutated in mismatch repair genes in an effort to confirm a defect in transcription- coupled repair in this system. In addition, we determined the contributionof the mismatch repair gene MSHZ to transcription-coupled repair in the absence of global genomic repair using rad7A mutants. We also determined whether the RadZGindependent transcription-coupled repair observedin rad26A and rad7A rad26A mutants depends on MSH2 by examining repair deficienciesof rad26A mh2Aand rad7A rad26A mh2Amutants. We found no defectsin transcriptioncoupled repaircaused by mutations in the mismatch repair genesMSH2, MLHI, PMSI, and MSH3. Yeast appears to differ from bacteria and human cells in the capacity for transcription-coupled repair in a mismatch repair mutant background.

ULTIPLE mechanisms for processing damaged and RAD16 genes are essential for this mode of repair M DNA haveevolved in essentiallyall organisms. of chromatin in vivo (VERHAGEet al. 1994) and in vitro In the yeast Saccharomyces cermisiae, for example, these (see ref.39 of VERHACE et al. 1996). A second class include the nucleotide and base excision repair path- of nucleotide excision repair acts specifically for the ways, the mismatch repair pathway, and photoreactiva- removal of DNA damage from the transcribed strand tion, as well as lesion tolerance modes such as recombi- of activegenes. This specialized form of nucleotide exci- nation. On the basis of epistasis analyses, it is believed sion repair, called transcription-coupled repair, re- that these pathways operate largely independently from quires most of the same proteins required for genomic one another. However, recent reports indicate that a nucleotide excision repair (SWEDER1994). The pres- number of repair proteins function in more than one ence of W-induced DNA damage in the transcribed of these pathways. For example, Rad1 and Rad10 operate strand of an expressed gene blocks the progression of in both nucleotide excision repair and recombination transcribing RNA polymerase complexes in vivo (SAUER- BIER and HERCULES1978) and vitro (SELBYand (FISHMAN-LOBELLand HABER 1992; IVANOVand HABER in 1995). Thus,it is possible that nucleotide excision repair SANCAR1990; DONAHUEet al. 1994). Such impeded proteins can interact with proteins from other repair RNA polymerase complexes may be perceived as sub strates for the nucleotide excision repair pathway. The pathways and influence those other processes. result of recognition of stalled RNA polymerase com- Nucleotide excision repair is an ubiquitous process plexes may be the preferential repairof the transcribed wherein repair proteins recognize DNA lesions and ef- strands of active genes over that of the nontranscribed fect dual DNA incisions resulting in removal of damage strands of active genes and nontranscribed regions of as part of an oligonucleotide. Subsequently, DNA poly- thegenome (i.e., transcription-coupled repair). Defi- merases, ligases, and accessory factors fill in the re- ciencies in transcription-coupled repair result in repair sulting gapped duplex DNA to regenerate the intact of the transcribed strands of active genes that is no duplex. There are at least two subpathways of nucleo- greater than that of the nontranscribed strands or the tide excision repair in vivo. One operates globally genome overall. Mutations in the mfd gene of Eschm'cl~ia throughout an organism's genome. In yeast the RAD7 coli, the RAD26 gene of S. cermisiae, and the ERCCG (CSB) and CSA genes of humans result in deficiencies Corresponding author: Kevin S. Sweder, Laboratory €or Cancer Re- search, Department of Chemical Biology, College of Pharmacy, Rut- in transcription-coupled repair (VENEMAet al. 1990; gers, The State University of New Jersey, Frelinghuysen Rd., Piscata- SELBYet nl. 1991; VAN HOFFENet al. 1993; VAN GOOLet way, NJ 088554789, E-mail: [email protected] nl. 1994; HENNINGet al. 1995).

Genetics 143 3127-1135 (July, 1996) 1128 K. S. Sweder et al.

Recently another repair pathway, mismatch repair, dependent upon MSH2, by examining the repair defi- was found to influencethe process of transcription- ciencies of rad26A msh2A and rad7A rad26A msh2A coupled repair in E. coli (MELLONand CHAMPE1996). mutants. Unlike the results reported for theE. coli mis- As its name implies, mismatch repair is a pathway for match repair mutants,we found that mutationsin yeast the removal of mismatched base pairs in duplex DNA. mismatch repair genes had no discernable effect upon In E. coli the recognition and removal of mismatches is either transcription-coupled repair or global nucleotide accomplished by the MutS, MutL, and MutH proteins excision repair of UV-induced DNA damage. (MODRICH 1991). MutS binds to the mismatched bases, while MutH binds to a nearby hemimethylated GATC MATERIALSAND METHODS site. The MutS, MutL, and MutH proteins then form Media, plasmids, and strains: All media were prepared as a multiprotein complex that leads to a single-strand described by SHERMANet al. (1986). WD medium is 1% yeast cleavage atthe hemimethylated site. Excision of a extract/2% Bacto-peptone (Difco)/2% glucose. Synthetic stretch of nucleotides either 3’ or 5’ from the nick is glucose medium (SD) is 2% glucose/0.67% Bacto-yeast nitro- followed by resynthesis employing DNA polymerase 111 gen base without vitamins (Difco) supplemented with the ap and associated factors. When MELLONand CHAMPE propriate amino acids and bases. Agar (1.5%) was added to media for plates. Yeast strains used in this study are listed in (1996) examined transcription-coupled repair in E. coli Table 1. Plasmid pKS212 is a Bluescript vector (Stratagene) mismatch repair mutants mutS and mutL, they observed containing the internal1.0-kb EcoRI-XhoI from RPB2 (SWEDEK a lack of transcription-coupled repair similar to the re- and HANAWALT1992). RPB2 encodes the second largest sub- pair deficiency observed for mfd mutants. A deficiency unit of RNA polymerase 11. Plasmid pKS212 was linearized by in transcription-coupled repair was not observed in cleaving with XhoI or EcoRI and was incubated with rNTPs and T7RNA polymerase or T3RNA polymerase, respectively, mutH mutants. Thus, mutations in mutS and mutL ap- under conditions recommended by the manufacturer to gen- pear to influence transcription-coupled repair in vivo. erate strand-specific RNA probes for RPB2. Alternatively, However, SELsYandSmcm (1995) reported that MutS strand-specific single-stranded DNA probes were used as de- and MutL proteins had no effect on transcription-cou- scribed previously (VERHAGEet al. 1994). Haploid strains YSH1475(msh2A) and YSH1476(MSH2) pled repair in vitro. were isolated by sporulation of RKYl476 (REENAN and KO- Eukaryotes alsopossess a general mismatch repair 1,ODNER 1992) under selection for uracil and histidine auxo- pathway that is performed by proteins that arehomolo- trophy and lysine and leucine prototrophy as described gous to the E. coli MutS and MutL proteins. Mutations (SHERMANet al. 1986). The mutator phenotype of the mis- in the genes encoding human homologuesof MutS and match repair strains, except for the W303-1Bderived ones, used in this study was confirmed by determining themutation MutL, hMSH2, hPMS2 and hMLH1, result in genomic frequency on complete synthetic medium lacking arginine instability of nucleotide repeats and are associated with and containing 60 bg/ml of L-canavanine. severaltypes of cancer(FISHEL et al. 1993; LEACH et Disruption of mismatch repair genes in yeast: All strains al. 1993; BRONNERet al. 1994; NICOWDESet al. 1994; designated with a GCY were derived from SJR231 (DATTAet PAPADOPOULOSet al. 1994; BOER et al. 1995). Mutations al. 1996) by lithiumacetate mediated transformation. GCYl27(msh2A) was constructed by two-step gene replace- in the yeast MSH2, MSH3, PMSl, and MLHl genes re- ment (ROTHSTEIN 1991). Strains were transformed with lin- sult ingenomic instabilityof dinucleotiderepeats earized URAkontaining plasmid p306m2RID (mh2A), and (STRAND et al. 1993, 1995). Ura+transformants were selected (DATTA et al. 1996). To determine whether mismatch repair influences p306m2RID contains MSH2 with an internal PuuII-XbaI frag- transcription-coupled repair in yeastas it appears to ment (2.4 kb) deleted. Excision of the integrated plasmids was selected for by resistance to 5-fluoroorotic acid (5-FOA) . do in E. coli, we examined the ability of several yeast Ura- yeast were assayed for amutator phenotype. GCYl07( m- mismatch repair mutants to perform transcription-cou- h3A : : hi&) was constructed by one-step gene disruption with pled repair, Therates and extentof repair for the indi- linearized plasmid pEN33 (DATTAet al. 1996).Plasmid pEN33 vidual strands within the actively transcribing RPB2 contains an mh3A:: hisGURA3-hisG allele. Loss of URA3 by gene were examined in yeast strains deficient in mis- recombination of direct hisG repeats was selected for by resistance to 5-FOA.GCYl21 (msh2A mshM:: hi@ was con- match repair. In addition,we determined the contribu- structed by disrupting MSH2 in the msh3A:: hisG strain. tion of the mismatch repair gene MSH2 to transcrip- Disruption of MSH2 in rad7A, rad26A, and rad7A rad26A tion-coupled repair in the absence of global genomic strains was accomplished by recombination (ROTHSTEIN repair. By using rad7A mutants, which lack the “back- 1983) using the 9.5-kb SpeI fragment of pII2-TnlOLUK 7-7 ground” of global repair of transcribed and nontran- (REENAN and KOLODNEK1992). Transformants were selected for uracil prototrophy. DNA isolated from the transformant scribed sequences foundin repair-proficient strains was digested with Cld, electrophoresed, immobilized on ny- (VERHAGEet al. 1994), we were able to investigate tran- lon membranes and hybridized with a 1.4kb radioactively scription-coupled repair uniquely in a mismatch repair labeled SpeI-NdeI fragment of MSH2 to confirm that the chro- deficient strain. It was recently demonstratedthat mosomal MSH2 gene was disrupted. strains lacking RAD26 (and global genomic repair) are Growth and UV irradiation of yeast cells: The growth and irradiation of strains was as previously described (SWEDERand still capable of residual transcription-coupled repair HANAWALT1994; VEKHAGE et al. 1994).All manipulations were (VERHAGEet al. 1996). Here we determined whether performed under yellow light to preclude photoreactivation. this Rad26-independent transcription-coupled repair is Briefly, exponentially growing cultures were collected by cen- Excision Repair in Mismatch Mutants 1129

TABLE 1 Mismatch deficient strains used in this report

Strain Genotype Source Genotype Strain or reference GCYl 00 MATa ura3 (NheZ) his3A200 ade2-101 (same as SJR231) A. DATTA~ GCYl27 Same as GCYl00 except msh2A This study' GCYl07 Same as GCYlOO except msh3A::hisG This study' GCYl2l Same as GCMOO except msh2Amsh3A::hisG This study' MW3317-21A MATa trpl ura3-52 ade2 add-AKpnI hom3-10 hid-KpnI met4 met13 YGSCb and R. M. LISKAY MW3317-21AmlhlA Same as MW3317-21A except mlhlA::URA3 R. M. LISKAV MW3317-21AmlhApmslASame MW3317-21A as except mlhlA::URA3 pmslA R. M. LISKAV MW3317-21ApmslAMW3317-21A as Same except pmslA YGSCb and R. M. LISKAY YSH 1475ura3-52 msh2::TnlOLUK7-7his4 (haploid derivedfrom RKYl476) E. ALANI~ YSH 1476ura3-52 MSH2his4 (haploid derivedfrom RKYl476) W303-1 B MATa ho ade2-1 trpl-1 leu2-3,112 canl-100 his3-11,15ura3-1R. ROTHSTEIN' M GSC 104 Same as W303-1B as SameMGSC104 stockfexcept Lab rad7A::LEU2 M GSC 151 Same as W303-1B as MGSC151Same except rad7A::LEU2msh2A::URAjr study' This M GSC 102 Same as W303-1B as SameMGSC102 stockgexcept Lab rad26A::HZSj M GSC 152 Same as W303-1B as MGSC152Same except rad26A::HZSjmsh2A::URA3 study'This M GSC lO6 Same as W303-1B as MGSClO6Same except rad7A::LEU2rad26A::HISjLab stockh MGSC153Sameas W3031B except rad7A::LEU2rad26A::HIS3 msh2A::URA3 study'This

References: a DATTA et al. (1996); WILLIAMSONet al. (1985); KRAMER et al. (1989); PROLLAet al. (1994b); REENAN and KOLODNER(1992); ROTHSTEIN1983; ~VERHAGEet al. (1994); VAN G~OLet al. (1994); VERHAGEet al. (1996); Constructed as described in MATERIALS AND METHODS. YGSC, Yeast Genetic Stock Center/University of California at Berkeley. trifugation and resuspended in ice-cold phosphate-buffered hybridized with probes for the transcribed or nontran- saline (PBS) at 1 X lo7 cells/ml.Shaking cell suspensions scribedstrand of RPB2 in GCYl00, GCYl27, and (-0.2 cm deep to ensure a uniform UV dose to all cells) were irradiated with predominantly 254 nm UV light at 0.33J/m2/ GCYl07 are shown in Figure 1. The important feature sec using a Westinghouse IL782-30 germicidal lamp or at 3.5 to noteis the ratio of the hybridizationsignal in the T4 J/m2/sec using a Philips T UV 30 W lamp. The cells were endonuclease V-treated sample to that of the mock- collected by centrifugation after irradiation and either lysed treated sample within each time point. immediately or resuspended in their original growth media Repair of the transcribed strandof RPB2 is very rapid at 30". Cells were incubated for various times to allow DNA repair and then lysed. in the repair-proficient parent strainGCYlOO as shown Isolation of yeast DNA: Cells were digested with Zymolyase in Figure 1. By 5 min after UV irradiation -30% of the 100T, and DNA was isolated as described (SWEDERand HANA CPDs had been removed from the transcribed strand WALT 1994). Alternatively, DNA was isolated as described by and within 30 min 70% of CPDs had been removed. In SHERMANet al. (1986) and purified on CsCl gradients (SAM- a mismatch repairdeficient strain, GCYl27, containing BROOK et al. 1989). Strand-specific analysisof frequency of cyclobutane pyrimi- a deletion of the MSH2 gene, repair of the transcribed dine dimers (CPDs): The incidence of CPDs in a particular strand of RPB2 was not detectably different from that restriction fragment was determined by methods previously in the repair-proficient parent strain. There were 25- developed (BOHR et al. 1985; MELLONet al. 1987; SWEDERand 30% of the CPDs removed 5 min after UV irradiation HANAWALT 1994). Membranes were prehybridized for at least 2 hr, then hybridized with strand-specific RNA probes made (Figure 1). Similarresults were obtainedin a yeast from pKS212. Autoradiographic signal intensitieswere quan- strain, GCYl07, carrying a deletion in another gene tified using a Hewlett-Packard Scanjet I1 and analysed using involved in mismatch repair (msh3A) (Figure 1). the application NIH Image 1.59. Alternatively, membranes Membranes were stripped and hybridized with radioac- were hybridized with strand-specific DNA probes. The signal tive RNA probes specific for the nontranscribed strand intensities were then quantified by use of a Betascope 603 blot analyser (Betagen). of WB2 (Figure 1). The data obtained by scanning the autoradiographs from these experiments and others are presented graphically in Figure 2. Mismatch repair mu- RESULTS tants GCYl27 and GCYl07 repaired the nontranscribed Transcription-coupled repairin mismatch repair mu- strand of RPB2as well as the parentstrain GCYlOO (Figure tants We examined the ability of yeast mismatch re- 24). Mismatch repair mutant YSH1475( mh2A) and repair-deficient strains to perform transcription-coupled pairproficient strain YSH1476(MSH2) also displayed nucleotide excision repair. We measured removal of faster repair of the transcribed strand of RPB2 than thatof CPDs from the individual strands of the RPB2 gene in the nontranscribed strand (Figure2A). YSH1475 (mh2A) the yeast strains listed in Table 1. Cells were irradiated, and BH1476(MSH2) are from a different genetic back- and DNA was isolated as described in MATERIALS AND ground than that of the GCY strains (Table 1). Repair of METHODS and the figure legends. Membranes thatwere the nontranscribed strand in these strains is similar to 1130 K. S. Sweder rt al.

Time (min) 100 0 5 15 30 60 - + - + - + - + - + T4endoV 80 '0 0 5 transcribed 0 E 60 MSH2 MSH3 a nontranscrlbed 0 a 40 0 1 8 20 transcrlbed 0 nontranscrlbed 0 10 20 30 40 50 60 70 '1 Time (min)

transcribed

msh3A nontranscrlbed 6 :l

FIGURE:1.-Removal of CPDs from the RI'B2 gene is proficient in mismatch repairdeficient yeast. Exponentially growing cultures of GCYl00 (MSH2 MSH3),GCM27 (mth2A).and GCYl07 (mrh3A) at 30" were irradiated with 30 J/m' of LTV radiation and incubated in growth medium forthe times indicated. DNA prepared from the cells was digested with restriction endonucleases. Restricted DNA was digested with T4 endonuclease V or mock-treated and then electrophoresed 0 10 20 30 40 50 60 70 through a 1.O% alkaline agarose.After transfer to Hvbond N' Time (min) membrane DNA was hybridized with RNA probes or ssDNA probesspecific for the transcribed or the nontranscribed FIGURE2.-Single and double mismatch repair mutants are proficient at transcriptioncoupled repair of UV-induced dam- strand of the RPB2 gene. The autoradiograms show the .i.S kb RruI-RwII restriction fragment. age.Repair of transcribed strands was determined from the measured incidences of CPDs in the PtmI-RmII restriction fragments from RPB2. Shown is a graphical representation of the that of several other repair-proficient strains reported pre- repair at the times indicated. (A) Time course for removal of viously (SNYDERand HANAMMJ 1992, 1994). CPDsfrom the RPBZ at 30" in GCYlOO (MSH2MSH3)-tran- We also examinedrepair of RPB2 in pmlA and scribedstrand (W), -nontranscribedstrand (0);GCYlO7 mllzlA mismatch repair-deficient strains from a third (mrh3A)-transcribedstrand (A) -nontranscribedstrand (A); GCM27 (ms/z2A)-tranxribed strand(0) -nontranscribed strand genetic background, MW3317-21A (Table 1). Pmsl and (0);GCYl21(mh2A msh3A)-transcribed strand(e), -nontran- Mlhlproteins form a heterotrimericcomplex with scribed strand (0);k5H1475(msh2A)-transcribed strand (+) Msh2 to initiate mismatch repair in yeast (PKOLIAet al. -nontranscribed strand (-- + --); kSH1476(MSH2)-transcribed 1994a,b). Results are shown in Figure 2B. Repair of the strand (X), -nontranscribed strand (- - X - -). (B) Time course transcribed strand of RPB2 is very rapid in the repair- for removal of CPDs from the RPBZ at 30" in MW3317-21A (PMSI MLHI)-transcribed strand (B),-nontranscribed strand proficient MW3317-21A andthe mismatchrepair- (0);hW3317-21ApmslA (pmtlA)-transcribed strand (A), deficientmutants, MW3317-21ApmslA(pmslA) and -nonuanscribedstrand (A); MW3317-2lAmlhlA (mlhIA)- MW3317-2lAmlhlA(mllzlA). Repairof thenontran- transcribed strand (O),-nonWanscribed strand (0);"3317- scribed strand of RPB2 is slower than that of the tran- PlAmlhlApmslA (dMA pmslA)-transcribedstrand (+), scribed strand in theMWS317-21A strains. We note that -nontranscribed strand ( 0 ). all strains derived from MW3317-21A display a some- what slower rate of repair of the transcribed strand of Effectof double mutations in mismatch repair RPB2 than the repair observedin the otherstrains used genes: We examined theremoval of CPDs from theRPB2 in this study. However,1 hr after W irradiation the gene in yeast strains possessing various combinations of extent of repair of the transcribed strand of RPB2 is mutations inseveral mismatch repair genes. We observed the same as that observed in other genetic backgrounds no measurabledecrease in repair of the transcribed used in this study. Hence, single mismatch repair mu- strand of RPB2 in any of these double mutants. As can tants exhibited faster repairof the transcribed strand of be seen in Figure 2, A and B, GCYl21(msh2Amh3A) RPB2 than the nontranscribed strand,ie., transcription- and MW3317-21A-mlhlA pmslA(m1hlA pnslA) all ex- coupled repair. hibit the same rapid repair of the transcribed strand of Excision Repair in Mismatch Mutants 1131 Mutants Mismatch in Repair Excision

RPB2 as that observed in the parent strains GCYlOO and MW3317-21A. StrainGCYl21 removed -50% of the CPDs from the transcribed strand of RPB2within 15 min 80 after UV irradiation (Figure 2A). By 30 min, -85-90% -0 0> 0 of the CPDs had been removed from the transcribed 60 strand ofRPB2. The mismatch repair double mutant E E (mlhlApmslA) derived from MW3317-21A also exhibited n n 40 proficient repair of the transcribed strand of RPB2similar 0 to the repair observed in the parent strain and the pmslA and mlhlA single mutants (Figure 2B). 20 We then stripped the membranes and rehybridized them with RNA probes specific for the nontranscribed 0 strand of RPB2. As shown in Figure 2, A and B, repair 040 20 60 80 100 120 140 of the nontranscribed strand of RPB2 is proficient but Time (rnin) slower than that of the transcribed strand of RPB2. Spe- FIGURE3.-Mismatch repair mutantS are proficient at tran- cifically, all strains removed -10% of CPDs from the scriptionqoupled repair of UV-induced damage. Exponen- nontranscribed strand of RPB2 in the first 15 min after tially growing cells were irradiated with 70 J/mz of UV radia- UV irradiation. By 30 min -30-40%of CPDs were tion and incubated in growth medium for the times indicated. removed from the nontranscribed strand of RPB2. Al- Repair of transcribed strands was determined from the mea- most 60% ofthe CPDs were removed from the nontran- sured incidences of CPDs in the Pod-Pod1 restriction fragments from RPB2. Shown is a graphical representation of the scribed strand ofRPB2 in most strains by 1 hr after repairat the times indicated. Time course forremoval of UV irradiation. Strains with mutations in two mismatch CPDs from the RPB2 at 30" in W303- 1B (RAD7 MSH2)-tran- repair genes areproficient at transcription-coupled nu- scribedstrand (W), -nontranscribedstrand (0);MGSC151 cleotide excision repair. (rud7Amsh2A)-transcribed strand (A), -nontranscribed strand ; MGSC104 (rud7A)-transcribed strand (O),-nom Repair in mismatch mutants lacking global nucleotide (A) transcribed strand (0). excisionrepair: We next examined transcriptioncoupled repair in an msh2A mutant that also hasa disruption mutation in RAD7. The RAD7gene is essential for global triple mutant, MGSC153(rad7Arad26Amsh2A), that is nucleotide excision repair in vivo (VERHAGEet al. 1994). deficient in transcriptioncoupled repair, lacks global We constructed the double mutant MGSC151 (rad7A genomic repair and mismatch repair. msh2A) (ie, lacking globalgenomic repair and mismatch We examined transcriptioncoupled repair in these dis repair) and analyzed repair of both strands of the RPB2 ruption mutants. The results of the repair experiments gene. No differences in repair were observed between forMGSC102(rud26A) and MGSC152(rud26Amsh2A) MGSC104(rad7A) and MGSC151(rad7Amsh2A) (Figure are shown graphically in Figure 4. We observed that the 3). The nontranscribed strand shows little or no repair rud26Amsh2A double mutant is no more deficient in re- during the first 2 hr after UV irradiation. However, in pair of either strand of RPB2 than is the rad26A single both mutants repair of the transcribed strand isvery mutant (Figure 4). Disruption ofMSH2 did not result rapid, like that of the repair-proficient parent strain in a further deficiency in transcriptioncoupled repair or W3031B (VAN GOOLet al. 1994). Twenty minutes after affect global nucleotide excision repair. UV irradiation 68% ofCPDs were removed from the The resultsof the repairexperiments for MGSClO6 transcribed strand (Figure 3). Transcriptioncoupled re- (rad7Arad26A) and MGSC153(rad7Arad26Amh2A) are pair, in the absence of global nucleotide excision repair, shown graphically in Figure 5. The residual transcription- is unaffected in a mismatch repairdeficient mutant. coupled repair observed in rud26A mutants is even more Repair in mismatch mutants deficient in transcrip- evident in these strains, due to the additional mutation tioncoupled repair: Previously, it was shown that in (rad7A) that eliminates global nucleotide excision repair rad26A mutants repair of the transcribed strand of ac- (VERHAGEet al. 1996).Deficiencies in transcriptioncou- tive genes is reduced to a level that is similar to, but pled repair are more readily detected in the absence of slightly higher than, that of the nontranscribed strand global repair. Therefore, we investigated whether MSH2 (VAN GOOLet al. 1994). This residual transcription-cou- was responsible for the residual transcriptioncoupled re- pled repair is even more evident in rad7Arad26A mu- pair observed in the yeast strain lacking transcriptioncou- tants that are also defective in global nucleotide exci- pled repair and global nucleotide excision repair (Figure sion repair(VERHAGE et al. 1996). We sought to 5). Disruption of MSH2 in the rad7Arud26A genetic back- determine if mismatch repair was responsible for the ground had no effect upon the residual repair of the tran- residual transcription-coupled repair observed in these scribed strand of RPB2 (Figure 5). In addition, repair of mutants. We constructed MGSC152(rad26Amsh2A) the induced GAL7 gene inrad7A, rad26A or rad7A that is deficient in transcriptioncoupledrepair and rad26A strains was unaffected by disruption of the MSH2 lacks mismatch repair. In addition, we constructed a gene (data not shown). Results of the experiments demon- 1132 K. S. Swc:der et ul.

100

80 '0 2> >al 0 0 E E 60 K K n n n n 40 0 0 8 8 20

0 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 Time (min) Time(min) FIGURE4.-Mismatch repairmutants are proficient at FIGURE5.-Disruption of MSH2 does not abolish the resid- global genomic repair ofUV-induced damage. Exponen- ual repair of UV-induced damage observed in rud7Arad26A tially growing cells at 30" were irradiated with 70 J/m' of mutants. Exponentially growing cells were irradiated with70 UV radiation and incubated in growth medium for the timesJ/m' of UV radiation and incubated in growth medium for indicated.Repair of transcribedstrands was determined the times indicated. Repair of transcribed strands was deter- from the measured incidences of CPDs in the PuuI-PvuII mined from the measured incidences of CPDs in the PuuI- restriction fragments fromRPB2. Shown is a graphical rep- PuuII restriction fragments from RPB2. Shown is a graphical resentation of the repair at thetimes indicated. Time course representation of the repair at the times indicated.Time for removal of CPDs from the RPB2 in W303-1B (RAD26 course for removal of CPDs from the RPB2 at 30" in W303- MSH2)-transcribed strand(m) ; -nontranscribed strand(0) ; 1B (RAD7 RAD26 MSH2)-transcribed strand (W), -nontran- MGSC152 (rad26Amsh2A)-transcribed strand (A),-non- scribedstrand (0);MGSC153 (rud7Arud26Amsh2A)-tran- transcribedstrand (A); MGSC102 (rud26AMSH2)-transcribedstrand (A),-nontranscribed strand (A); MGSClO6 scribed strand (O),-nontranscribed strand (0). (rud7Arud26A)-transcribed strand (O), -nontranscribed strand (0). strate that the residual RadZGindependent transcription- coupled repair activity is not dependent on Msh2. exhibit asignificant amount of repair of the transcribed strand that is independent of RAD26. Disruption of DISCUSSION MSH2in rad7Arad26A mutants didnot result in further We demonstrate here that yeast strains carrying dis- reduction of repair of the transcribed strand of RPBB. ruption mutations inmismatch repair genes exhibitno Similarly, disruption of MSH2in rad7A rad26A mutants deficiency in transcription-coupled repair ofUV-in- did not change the rateor level ofrepair in the induced duced DNA damage. Deletion mutation of MSH2, GAL7 gene observed in rad7A rad26A mutants (data MSH3, MLHl, or €"SI yielded strains fully competent not shown). The RadZGindependent transcription-cou- for transcription-coupled repair. Furthermore, double pled repair activity that was recently observed in yeast mutants in mismatch repair genes were also proficient is therefore not dependent on Msh2 (VERHAGEet al. at transcription-coupled repair. By using mismatch re- 1996). Thus, any interaction between mismatch repair pair-deficient rad7A mutants, we have shown that mis- and nucleotide excision repair following W irradiation match repair did not the influence transcription-cou- appears to be absent in the yeast strains we investigated. pled nucleotide excision repair in the absence of global Based upon the original epistasis analyses (reviewed nucleotide excision repair. Furthermore, DNA repair in HAYNESand KUNZ1981 and GAME1983) it was antici- levels in yeast strains deficient in transcription-coupled pated that therewould be little interaction between the repair were unaffected by disruption of the MSH2 gene. mismatch repair pathway and the nucleotide excision The small difference between repair of both strands in repair pathway.However, MCGRAW and MAFUNUS a rud26A mutant that is probably due to some residual (1980) reportedthat methyltransferasedeficient mis- transcription-coupled repair activity is still present in match repairdeficient E. coli strains (dam-3 mutS and rad26A mh2A doublemutants. We concludethat dam-3 mutL) were only as UV-sensitive as the wild-type global genomic repair of UV-induced DNA damage is strains (and less UV-sensitive than the dam-3 strains). not affected by disruption of MSH2 since no effect was The lack of UV sensitivity they observed may indicate a observed on the level of repair of nontranscribed DNA role for mismatch repair proteins in repair of UV-in- of mismatch repairmutants. Finally, we examined duced DNA damage. In addition, HAS and colleagues rad7A rud26A double mutants lacking global genomic proposed that proteinsof the E. coli MutHLS may recog- repair and partly defective in transcription-coupled re- nize UV-photoproducts and initiate recombination pair activity (VERHAGEet al. 1996). Previously, VERHAGE (FENGet al. 1991; FENGand HAYS 1995). Recently, SW- et al. (1996) demonstrated that rad7A rad26A mutants DALIS et ul. (1994) reported that the mutation spectrum Excision Repair in Mismatch MutantsMismatch in Repair Excision 1133 induced by 5methylcytosine in the human HPRT gene malian genomes. Disruption of a yeast mismatch repair indicates a strand bias in the repair of the mismatches gene may have a less profound effect upon expression generated by deamination of 5-methylcytosine.The mu- relative to disruption of a mismatch repairgene in tations occur at sites indicating the original "CpGs were mammalian cells. Demonstration thatproteins can overwhelmingly (17/19) in the nontranscribed strand function both in a repair pathway and in transcription of HPRT. Thus, repair of mismatches generated by de- is well established (FEAVERet al. 1993; SCHAEFFERet al. amination of 5-methylcytosine appears to take place in 1993,1994; DRAFWNet al. 1994; SWEDERand HANAWALT a transcriptioncoupledmanner in human cells, i.e., 1994; WANGet al. 1994). Second, the binding specifici- preferential repair of the transcribed strand by the mis ties of the mismatch repair proteins vary from prokary- match repair pathway. Alternatively, apyrimidinic sites ote to eukaryote. For example, the bacterial Mu6 pro- or nicks generated at the sitesof alkylated cytosine tein binds to one to three nucleotide loops, binds less could act as transcription-pausing or transcription-ar- well to four nucleotide loops, and exhibits little or no resting lesions (FLAILIEEand VERLY 1985; ZHOU and binding to five nucleotide loops (PARKERand MARINUS DOETSCH1993), which might be recognized by the pro- 1992).In contrast, the purified yeastMsh2 protein teins of the human transcriptioncoupled nucleotide binds strongly to palindromic insertions and large loop excision repair pathway. insertions, and binds moderately to smaller loop inser- A connection between transcription-coupled nucleo- tions and mismatches (MIRET et al. 1993; ALANI et al. tide excision repair and mismatch repair was also seen 1995). Purified human MSH2 protein has similar bind- by an analysis of transcriptioncoupled repair in E. coli ing affinities (FISHELet al. 1994a,b). However, the active mismatch repair mutants. MELLONand CWPE (1996) form of Msh2 involved in human and yeast mismatch found no preferential repair of the transcribed strand repair in vivo may require other factors (DRUMMONDet of the lac operon in E. coli strains with a mutation in al. 1995; KOLODNER1995). The in vivomismatch repair either the mutS or mutL genes. In addition, they ob- complex in yeast and humans may differ in composition served a measurable decrease in repair of the nontran- and binding specificity. scribed strand of lac. It remains to be determinedif the In addition to our results presented here, in uivo evi- influence of mismatch repair upon nucleotide excision dence for differences in substrate specificities among repair in E. coli is a direct effect or an indirect effect. mismatch repair proteins from E. coli, yeast, and human Studies using an in vitro nucleotide excision repair sys- cells comes from mutagenesis and survival studies using tem and purified proteins revealed no effect of mis- the alkylating agent Nmethyl-N-nitro-Nnitrosoguani- match repair proteins upon transcriptioncoupled re- dine (MNNG) (XMO et al. 1995). MNNG produces a pair (SELBYand SANCAR1995). Additional evidence for number of Oalkyl lesions, mainly@methylguanine and a connection between transcription-coupled nucleotide @-methylthymine. Yeast strains containing mutations in excision repair and mismatch repair comes from studies mismatch repair genes and the gene encoding o6-meth- using human tumor cell lines. Strand-specific DNA re- yltransferase, a suicideenzyme that removesmethyl pair assays in human tumor cell lines with mutated mis- groups from guanine and thymine, were grown in the match repair genes (HNPCC etc.) revealed a lack of presence of MNNG. Following exposure to MNNG, cell transcription-coupled repair (MELLON et al. 1996). survival and mutation frequencies were determined. Mis Transcription-coupled repair was restored toone of the match repair mutants displayed spontaneous mutation mismatch repairdeficient tumorcell lines (HCT116) by frequencies that were -20- to 50-fold higher than the transfer of chromosome 3 (MELLONet al. 1996). Thus, mutation frequencies determined for repair-proficient bacterial mismatch repairmutants andhuman mis- and methyltransferase-deficient strains (XIAo et al. match repairdeficient cell lines examined thus farlack 1995). Disruption of individual yeast mismatch repair transcriptioncoupled nucleotide excision repair. genes had no effect upon survival in the presence of There are several possibleexplanations for thediffer- MNNG when compared to repair-proficient strains or ences observed for transcriptioncoupled nucleotide ex- methyltransferasedeficient strains, ie, there was no in- cision repair in mismatch repairdeficient cells from creased alkylation resistance. This is in marked contrast different organisms. First, the effect of mismatch repair to the alkylation-resistantphenotype of mismatch repair- upon transcription-coupled repair in bacteria and hu- deficient E. coli dam-3 and human mismatch repair mu- mans may be indirect. Mutations in genes encoding tants (JONES and WAGNER1981; GOLDMACHER et al. mismatch repair proteinsmay alter expression of a wide 1986; BRANCHet al. 1993, 1995; KAT et al. 1993). variety of genes. In prokaryotes and eukaryotes tram A third possible explanation for thelack of influence scriptioncoupled repair of a DNA sequence requires of mismatch repair upon transcriptioncoupled repair passage of the transcription complex through the DNA of W-induced DNA damage is that there may be a sequence. Therefore, an overall lowering of transcrip- hitherto uncharacterized protein in yeast that has a dif- tion levels resulting from mutations in mismatch repair ferent substrate-binding specificity than that of Msh2 genes might reduce transcriptioncoupled repair. The and can function in the absence of Msh2, Mlhl, Pmsl, yeast genome is intensely transcribed relative to mam- and Msh3, i.e., there is functional redundancy. There 1134 K. S. Sweder et al. exist several open reading frames in yeast that are pre- DRUMMOND,J. T., G.-M. LI, M. J. LONGLEYand P. MODRICH,1995 Isolation of an hMSH2-pl60 heterodimer that restores DNA mis- dicted to encode proteins that sharesignificant homol- match repair to tumor cells. Science 268: 1909-1912. ogy to Msh2. One of these open reading frames is very FEAVER,W. J., J. Q. SVEJSTRUP, BARDWELL, L. A. J. BARDWELI.,S. BURA- similar to the human G/T Binding Protein (GTBP or TOWSKI et al., 1993 Dual roles of a multiprotein complex from S. cerevisiae in transcription and DNA repair. Cell 75 1379- p160) ( DRUMMONDet al. 1995; PALOMBOet al. 1995; 1387. PAPADOPOULOSet al. 1995).The human MutS mismatch FENG, W.-Y., and J. B. HAYS,1995 DNA structures generated during repair complex contains two subunits, hMSH2 and recombinationinitiated by mismatch repair of UV-irradiated nonreplicating phage DNA in Escherichia coli: requirements for GTBP, that have different DNA substrate binding pref- helicase, exonucleases, and RecF and RecBCD functions. Genet- erences. Based upon in vivo phenotypes of human cell ics 140: 1175-1186. lines harboring mutations in either of the genes encod- FENG,W.-Y., E. LEEand J. B. HAYS, 1991 Recombinagenic processing of UV-light photoproducts in nonreplicating phageDNA by the ing GTBP and hMSH2, it appears that hMSH2 requires Escherichia coli methyldirected mismatch repair system. Genetics GTBP to recognize G/T mismatches and single nucleo- 129: 1007-1020. tide insertions or deletions. In other words, GTBPis FISHEL,R., M. K. LESCOE,M. R. RAO, N. G. COPELAND,N. A. JENKINS et al., 1993 The human mutator gene homolog MSH2 and its necessary for hMSH2 recognition of more subtle DNA association with hereditary nonpolyposis colon cancer. Cell 75: distortions due to G/T mismatches and single nucleo- 1027-1038. tide insertions or deletions. If yeast possess a GTBP- FISHEI.,R., A. EWEL,S. LEE, M. K. LESCOEand J. GRIFRTH,1994a Binding of mismatched microsatellite DNA sequences by the like protein, GTBPmay recognize UV-induced DNA human MSH2 protein. Science 266: 1403-1405. damage, obviating the needfor Msh2, and thereby func- FISHEL, R., A. EWELand M. K. LESCOE,1994b Purified human MSH2 tion in transcription-coupled repair. Such a possibility protein binds to DNA containing mismatched nucleotides. Can- cer Res. 54: 5539-5542. remains to be tested. FISHMAN-LOBELL,J., and J. E. HABER, 1992 Removal of nonhomolo- gous DNA ends in double-strand break recombination: the role The authors thankL. LOMMEI.and C. A. SMITHfor helpful discus- of the yeast ultraviolet repair gene MI.Science 258: 480-484. sions and critical reading of the manuscript; I. MELLONfor helpful FIAMEE,P. A., and W. G. VERLY,1985 Action of intact AP (apurinic/ discussions and communicationof results beforepublication; E. apyrimidinic) sites and AP sites associated with breaks on the AIANI,R. M. LISKAYand R. KOLODNERfor their generosity in provid- transcription of T7 coliphage DNA by Escherichia coli RNA poly- ing some ofthe mismatch repair strains. This work was supported by merase. Biochem. J. 229 173-181. Postdoctoral Training Grant T32 AR07422 from the National Insti- GAME,J. C., 1983 Radiation-sensitive mutants and repair in yeast, tute of Arthritis and Musculoskeletal and Skin Diseases and by an pp. 109-137 in YeastGenetics: Fundamental and Applied Aspects, Outstanding Investigator Award CA44349 from the National Cancer edited by J. F. T. SPENCER,D. M. SPENCERand A. R. W. SMITH. Institute. R.A.V. and J.B. are supported by the J. A. Cohen Institute Springer-Verlag, New York. GOLDMACHER,V. S., R. A. J. CuZrCK and W. G.THIILY, 1986 Isolation for Radiopathology and Radiation Protection (IRS), project 4.2.13. and partial characterization of human cell mutants differing in sensitivity to killing and mutation by methylnitrosourea and N- LITERATURE CITED methyl-N’-nitro-N-nitrosoguanidine.J. Biol. Chem. 261: 12462- 12471. ALANI, E.,N. W. CHI and R. KOLODNER,1995 The Saccharomyces HAYNES, R.H., and B.A. KUNZ,1981 DNA repair and mutagenesis cerevisiae Msh2 protein specifically binds to duplex oligonucleo- in yeast, pp. 371-414 in The Molecular Biology ofthe Yeast Saccharo- tides containing mismatched DNA base pairs and insertions. myces: Life Cycle and Inheritance, edited by J. N. STRATHERN,E. W. Genes Dev. 9: 234-247. JONES and J. R. BROACH.Cold Spring Harbor Laboratory Press, BOHR,V. A,, C. A. SMITH,D. S. OKUMOTOand P. C. HANAWALT,1985 Cold Spring Harbor, Ny. DNA repair in an active gene: removal of pyrimidine dimers HENNING,K. A,, L. LI, N. IYER,L. D. MCDANIEL,M. S. REAGAN et al., from the DHFR gene of CHO cells is much more efficient than 1995 The Cockayne Syndrome Group a gene encodes a WD in the genome overall. Cell 40: 359-369. repeat protein that interacts with CSB protein and a subunit of BOYER,J. C., A. Uw,J. I. RISINGER,J. R. LIPFORD,M. KANE et al., RNA polymerase I1 TFIIH. Cell 82: 555-564. 1995 Microsatellite instability, mismatch repair deficiency and IVANOV,E. L., and J. E.HABER, 1995 Mland RADIO, but not genetic defects in human cancercell lines. Cancer Res. 55 6063- other excision repair genes, arerequired fordouble-strand 6070. break-induced recombinationin Saccharomycescereuisiae. Mol. BRANCH,P., G. AQUILINA,M. BIGNAMIand P. KARRAN, 1993 Defec- Cell. Biol. 15 2245-2251. tive mismatch binding and a mutator phenotype in cells tolerant JONES,M., and R. WAGNER,1981 NMethyl-N-nitro-Nnitrosoguani- to DNA damage. Nature 362 652-654. dine sensitivity of E. coli mutants deficient in DNA methylation BRANCH,P., R. HAMPSONand P. KARRAN, 1995 DNA mismatch bind- and mismatch repair. Mol. Gen. Genet. 184: 562-563. ing defects, DNA damage tolerance, and mutator phenotypes in KAT, A,, W. G. THILLY,W. H. FANG,M. J. LONGLEY,G. M. LI et al., 1993 human colorectal carcinoma cell lines. Cancer Res. 55: 2304- An alkylation-tolerant, mutator human cell line is deficient in 2309. strand-specific mismatch repair. Proc. Natl. Acad.Sci. USA 90: BRONNER,C. E., S. M.BAKER, P. T. MORRISON,G. WARREN, L. G. 6424-6428. SMITHet al., 1994 Mutation in the DNA mismatch repair gene KOLODNER,R., 1995 Mismatch repair: mechanisms and relationship homologue hMMl is associated with hereditary non-polyposis to cancer susceptibility. Trends Biochem. Sci. 20: 397-401. colon cancer. Nature 368: 258-261. KRAMER,W., B. KRAMER,M. S. WILLIAMSONand S. FOGEI.,1989 Clon- DATTA, A.,A. ADJIRI, L. NEW,G. F. CROUSEand S. JINKS-ROBERTSON, ing and nucleotide sequence of DNA mismatch repairgene 1996 Mitotic crossovers between diverged sequences are regu- PMSl from Saccharomyces cerevisim: homology ofPMSl to procary- lated by mismatch repair proteinsin Saccharomyces cereuisiae. Mol. otic MutL and HexB. J. Bacteriol. 171: 5339-5346. Cell. Biol. 16: 1085-1093. LEACH,F. S., N. C. NICOWDES,N. PAPADOPOULOS,B. LIU, J. JEN et al., DONAHUE,B. A,, S. YIN,J. S. TAYLOR,D. REINES and P. C. HANAWALT, 1993 Mutations of a mutS homolog in hereditary nonpolyposis 1994 Transcript cleavage byRNA polymerase I1 arrested by a colorectal cancer. Cell 75 1215-1225. cyclobutane pyrimidine dimer in the DNA template. Proc. Natl. MCGRAW,B. R., and M. G. MARINUS, 1980 Isolation and character- Acad. Sci. USA 91: 8502-8506. ization of Damt revertane andsuppressor mutations that modify DRAPKIN,R., J. T. REARDON, A. ANSARI, J. C. HUANG,L. ZAWEI. et secondary phenotypes of dam-3 strains of Escherichiacoli K-12. al., 1994 Dual role of TFIIH in DNA excision repair and in Mol. Gen. Genet. 178: 309-315. transcription by RNA polymerase 11. Nature 368: 769-772. MELLON,I., and G. N. CHAMPE,1996 Products of DNA mismatch E xcision Repair in Mismatch MutantsMismatch in Repair Excision 1135

repair genes mutS and mutL are required for transcription-cou- mutant deficient in “mutation frequency decline” lacks strand- pled nucleotideexcision repair of the lactose operon in Esche- specific repair: in vitro complementation with purified coupling richia coli. Proc. Natl. Acad. Sci. USA 93: 1292-1297. factor. Proc. Natl. Acad. Sci. USA 88: 11574-11578. MELLON,I., G. SPIVAKand P.C. HANAWALT,1987 Selective removal SHERMAN,F., G. FINKand J. HICKS, 1986 Laboratory Course Manual oftranscription-blocking DNA damagefrom the transcribed fwMethods inYeast Genetics. Cold SpringHarbor Laboratory Press, strand of the mammalian DHFR gene. Cell 51: 241-249. Cold Spring Harbor, W. MELLON,I., D. K RAJFXL, M. KOI, C. R. BOIANDand G. N. CHAMPE, SKANDALIS,A,, B. N. FORDand B. W. GLICKMAN,1994 Strand bias in 1996 Transcription-coupled repair deficiency and mutations in mutation involving 5-methylcytosine deamination in the human human mismatch repair genes. Science 272: 557-560. hprt gene. Mutat. Res. 314 21-26. MIRET,J. J., M. G. MILLAand R. S. LAHUE, 1993 Characterization STRAND,M., T.A. PROLLA, R. M.LrsmvandT. D. PETES,1993 Destabi- of a DNA mismatch-binding activity in yeast extracts. J. Biol. lization of tracts of simple repetitive DNA in yeast by mutations Chem. 268 3507-3513. affecting DNA mismatch repair. Nature 365: 274-276. MODRICH, P.,1991 Mechanisms and biological effects of mismatch STRAND,M., M. C. EARLEY, G. F. CROUSEand T. D. PETES,1995 repair. Annu Rev Genet 25 229-253. Mutations in the MSH? gene preferentially leadto deletions within tracts of simple repetitive DNA in Saccharomyces cereuisiae. NICOMDES,N. C., N. PAPADOPOULOS,B. LIU, Y. F. WEI, K. C. CARTER Proc. Natl. Acad. Sci. USA 92: 10418-10421. et al., 1994Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 371: 75-80. SWEDER,K. S., 1994 Nucleotide excision repair inyeast. Curr. Genet. 27: 1-16. PALOMBO,F., P. GALLINARI,I. IACCARINO,T. LETTIERI, M. HUGHESet SWEDER,K. S., and P. C. HANAWALT,1992 Preferentialrepair of al., 1995 GTBP, a 160-kilodalton protein essential for mis cyclobutane pyrimidine dimers in the transcribed strand of a match-binding activity in human cells. Science 268: 1912-1914. gene in yeast chromosomes and plasmids is dependent on tran- PAPADOPOULOS,N., N. C. NICOLAIDES,Y. F. WEI, S. M. RUBEN,K. C. scription. Proc. Natl. Acad. Sci. USA 89: 10696-10700. CARTER et al., 1994 Mutation of a mutL homolog in hereditary SWEDER,K. S., and P. C. HANAWALT,1994 The COOH terminus of colon cancer. Science 263 1559-1560. suppressor of stem loop (SSL2/Rad25) in yeast is essential for PAPADOPOULOS,N., N. C. NICOMDES,B. LIU, R. PARSONS,C. LEN- overall genomic excision repair andtranscription-coupled tepair. GAUER et al., 1995 Mutations of GTBP in genetically unstable J. Biol. Chem. 269: 1852-1857. cells. Science 268: 1915-1917. VAN GOOL,A. J., R. VERHAGE,S. M. SWAGEMAKERS, P. VAN DE PUITE, PARKER,B. O., and M. G. MARINUS,1992 Repair of DNA hetero- J. BROUWERet al., 1994 RAD26, the functional S. cerevisiae ho- duplexes containing small heterologous sequences in Escherichia molog of the Cockayne syndrome B gene ERCC6. EMBO J. 13: coli. Proc. Natl. Acad. Sci. USA 89: 1730-1734. 5361 -5369. PROLLA,T. A,, Q. PANG, E.ALANI, R. D. KOLODNERand R. M. LISKAY, VAN HOFFEN,A,, A. T. NATARAJAN,L. V. MAWE, A. A. VAN ZEELAND, 1994a MLH1, PMS1, and MSH2 interactions during the initia- L. H. F. MULLENDERSet al., 1993 Deficient repair of the tran- tion of DNA mismatch repair in yeast. Science 265: 1091-1093. scribed strand of active genes in Cockayne’s syndrome cells. Nu- PROIU, T. A, D. M. CHRISTIEand R M. LISKAY 1994b Dual require- cleic Acids Res. 21: 5890-5895. ment in yeast DNA mismatch repair for MLHl and PMSI, two VENEMA,J., L. H. F. MULLENDERS,A. T. NATARAJAN, A. VAN A. ZEEIAND homologs of the bacterial mutL gene. Mol. Cell. Biol. 14 407-415. and L.V. MAWE, 1990 The geneticdefect in Cockayne syn- REENAN,R.A. G., and R.D. KOLODNER,1992 Characterization of drome is associated with a defect in repair of UV-induced DNA insertion mutations in the Saccharomycescereuisiae MSHl and damage in transcriptionally activeDNA. Proc. Natl. Acad. Sci. MSH2 genes: evidence for separate mitochondrial and nuclear USA 87: 4707-471 1 functions. Genetics 132: 975-985. VERHAGE,R., A. M. ZEEMAN,N. DE GROOT,F. GLEIG,D. D. BANGet ROTHSTEIN,R. J., 1983 One-step gene disruption in yeast. Methods al., 1994 The RAD7 and RAD16 genes, which are essential for Enzymol. 101: 202-211. pyrimidine dimer removal from the silent mating type loci, are SAMBROOK,J., E. F. FRITSCH and T. MANIATIS,1989 Molecular Clon- also required for repairof the nontranscribedstrand of an active ing: A Labmatory Manual. Cold Spring Harbor Laboratory Press, gene in Saccharomycescereuisiae. Mol. Cell. Biol. 14: 6135-6142. Cold Spring Harbor, Ny. VERHAGE,R., A. J. VAN COOL, N. DE GROOT,J. H. J. HOEIJMAKERS, SAUERBIER,W., and HERCULES,K, 1978 Gene and transcription unit P. VAN DE PU?TE et al., 1996 Double mutants of Saccharomyces mapping by radiation effects, pp. 329-363 in AnnualReviews Genet- cerevisiae with alterations in global genome and transcription- ics, edited by H. L. ROMAN. AnnualReviews Inc., Palo Alto, CA. coupled repair. Mol. Cell. Biol. 16: 496-502. SCHAEFFER, L., R. ROY,S. HUMBERT,V. MONCOLLIN,W. VERMEULEN WANG, Z., J. Q. SVEJSTRUP,W. J. FEAVER,X. WtJ, R. D. KORNBERGet et al., 1993 DNA repair helicase: a component of BTF2 (TFIIH) al., 1994 Transcriptionfactor b (TFIIH) is requiredduring basic transcription factor. Science 260: 58-63. nucleotide excision repair in yeast. Nature 368: 74-76. SCHAEFFER,L., V. MONCOLLIN,R. ROY, A. STAUB,M. MEZZINAet al., WILLIAMSON,M. S., J. C. GAMEand S. FOGEI.,1985 Meiotic gene 1994 The ERCC2/DNA repairprotein is associated with the conversion mutants in Saccharomyces rermisiae. I. Isolation and class I1 BTF2/TFIIH transcription factor. EMBO J. 13: 2388- characterization of pmsl-1 and pmsl-2. Genetics 110: 609-646. 2392. X1.40, W., L. RATHGEBER,T. FONTANIEand S. BAWA,1995 DNA SELBY,C. P., and A. 1990 Transcription preferentially inhib mismatch repair mutants do not increase Nmethyl-N-nitro-N- SANCAR, nitrosoguanidinetolerance in ab-methylguanine DNA methyl- its nucleotide excision repair of the template DNA strand in vitro. transferasedeficient yeast cells. Carcinogenesis 16: 1933-1939. J. Biol. Chem. 265: 21330-21336. ZHOU, W., and P. W. DOETSCH,1993 Effects of abasic sites and DNA SELBY, C. P., and A.SANCAR, 1995 Structure and function of tran- single-strand breaks on prokaryotic RNA polymerases. Proc. Natl. scription-repair coupling factor. 11. Catalytic properties. J. Biol. Acad. Sci. USA 90: 6601-6605. Chem. 270: 4890-4895. SELBY, C. P., E. M.WITKIN and A. SANCAR, 1991 Escha’chia colimfd Communicating editor: G. R. SMITH