Repair of DNA Loops Involves DNA-Mismatch and Nucleotide

letters to nature

Acknowledgements. We thank T. Tanaka for IL-4 and CT4S cells; S. Nagata for pEF-BOS vector and Jak2 Msh3 or Msh6 and the MutL homologues Mlh1 and Pms1 (refs 6, cDNA; J. Krolewski for Tyk2 cDNA; K. Yasukawa for recombinant IL-6 and sIL-6R; Y. Shima, H. Danno, K. Kunisada and H. Tagoh for technical assistance; H. Saito for discussion; and A. Nobuhara for secretarial 7). Rad1 acts in conjunction with Rad10 to cleave 5Ј to the assistance. This work was supported by a Grant-in-Aid from the Ministry of Education, Science and pyrimidine dimer during nucleotide-excision repair1. We used Culture, Japan. two different types of mutant his4 alleles: his4-lopd (a 26-base- Correspondence and requests for materials should be addressed to T.K. (e-mail: [email protected]. pair (bp) non-palindromic insertion in HIS4; ref. 5) and his4-AAG osaka-u. ac.jp). The sequence of SSI-1 has been deposited with Genbank, under accession number AB000710. (a base-pair change within the initiating ATG codon of HIS4; ref. 8). Diploid strains heterozygous for one of these alleles and homozygous for rad1, msh2 or both mutations were sporulated and tetrads were dissected. Spore colonies grown on rich medium were replica-plated to medium lacking histidine and scored as Hisþ,HisϪ and sectored Repair of DNA loops involves Hisþ/His Ϫ (representing PMS events) (Tables 1 and 2). In DNY27, the wild-type strain heterozygous for his4-lopd, only DNA-mismatch and 12% of the aberrant segregation events represent PMS events, indicating that the 26-base loop expected within the heteroduplex nucleotide-excision is efficiently recognized and repaired. Homozygous rad1 or msh2 mutations increase the frequency of PMS about threefold, demon- repair proteins strating that Rad1 and Msh2 are involved in the repair of large DNA David T. Kirkpatrick & Thomas D. Petes loops. Moreover, they function in the same repair pathway, as the rad1 msh2 mutant has the same PMS frequency as the rad1 and Department of Biology, Curriculum in Genetics and Molecular Biology, msh2 strains. Surprisingly, the rad1 mutation also significantly University of North Carolina at Chapel Hill, Chapel Hill, increases the PMS frequency for his4-lopd when heterozygous, North Carolina 27599-3280, USA suggesting that the RAD1 gene product may be rate-limiting for ...... repair in the RAD1 MSH2-dependent pathway. rad1 and msh2 A number of enzymes recognize and repair DNA lesions1. The mutations have no significant effect on the crossover frequency in DNA-mismatch repair system corrects base–base mismatches the HIS4 to LEU2 interval (Table 2). and small loops, whereas the nucleotide-excision repair In a wild-type strain heterozygous for the point mutation his4- system removes pyrimidine dimers and other helix-distorting AAG (PD83), we found that 14% of the aberrant segregants were lesions. DNA molecules with mismatches or loops can arise as PMS (Table 2). The rad1 mutation slightly (but significantly a consequence of heteroduplex formation during meiotic (P ¼ 0:03)) increased PMS frequency. In an msh2 derivative, PMS recombination2. In the yeast Saccharomyces cerevisiae, repair of constituted 87% of the total aberrant segregation events, as expected mismatches results in gene conversion or restoration, and failure from previous studies9. From these results, we suggest that Rad1 to repair the mismatch results in post-meiotic segregation (PMS) plays a smaller role in the meiotic repair of base–base mismatches (Fig. 1). The ratio of gene-conversion to PMS events reflects the than it does in the repair of loops. In previous studies10,11, no effects efficiency of DNA repair3,4. By examining the PMS patterns in of the rad1 mutation on PMS frequency of heterozygous mutations yeast strains heterozygous for a mutant allele with a 26-base-pair (presumably point mutations) were observed, although crossing- insertion, we find that the repair of 26-base loops involves Msh2 (a over was reduced in ultraviolet-irradiated rad1 diploids12. DNA-mismatch repair protein) and Rad1 (a protein required for The nucleotide-excision repair (NER) system in yeast requires the nucleotide-excision repair). concerted action of a large number of proteins. The NER endonu- We examined the effects of msh2 and rad1 mutations on the cleases, Rad1/Rad10 and Rad2, cleave 5Ј and 3Ј to the pyrimidine meiotic segregation patterns of heterozygous markers in the HIS4 dimer respectively, whereas Rad14 seems to be involved in the gene, a recombination hotspot5. Msh2 (a homologue of the MutS recognition of DNA damage1. To determine whether the effect on protein of Escherichia coli) is required for the repair of both base– the repair of large loops observed with Rad1 required other NER base subsitutions and small DNA loops, acting together with either proteins, we examined the effects of rad2 and rad14 mutations on

Table 1 Effects of rad1 and msh2 mutations on the meiotic segregation patterns of heterozygous markers at the HIS4 locus

Number of tetrads with various meiotic segregation patterns*

Diploid Mutant RAD1† MSH2† Other Ab Other Total his4 allele alleles 4 : 4 6 : 2 2 : 6 5 : 3 3 : 5 4:4 7:1 1:7 8:0 0:8 PMS tetrads DNY27 his4-lopd þ=þþ=þ – 252 54 38 11 1 0 0 1 1 1 0 359 TP1012 his4-lopd þ= Ϫ þ=þ – 201 37 25 10 9 1 0 0 1 0 0 284 TP1013 his4-lopd Ϫ = Ϫ þ=þ – 294 86 28 37 24 3 4 2 3 0 0 481 DTK223 his4-lopd þ=þ Ϫ = Ϫ – 832112174 3002 0 0 142 DTK230 his4-lopd Ϫ = ϪϪ=Ϫ –611210131202020 0112 DTK241 his4-lopd þ=þþ=þrad2 16422141 4 0000 0 0 205 DTK242 his4-lopd þ=þþ=þrad14 16239185 1 0002 1 0 228 ...... MW103‡ his4-Sal þ=þþ=þ – 202 31 50 0 0 0 0 0 2 6 0 291 DTK257 his4-Sal Ϫ = Ϫ þ=þ – 249 66 70 0 0 0 0 0 4 3 0 392 ...... PD83§ his4-AAG þ=þþ=þ – 142 54 58 11 10 0 1 2 13 23 1 315 DTK224 his4-AAG Ϫ = Ϫ þ=þ – 80 41 48 13 11 4 2 1 10 10 0 220 No. 5k his4-AAG þ=þ Ϫ = Ϫ –4063162580100 12111 DTK232 his4-AAG Ϫ = ϪϪ=Ϫ – 4066221810110 1 11116 ...... * For all segregation patterns, the ﬁrst number represents the wild-type allele and the second represents the mutant allele. The segregation patterns include: 4 : 4 (normal mendelian segregation), 6 : 2 and 2 : 6 (gene conversion), 5 : 3 and 3 : 5 (tetrads with a single PMS event), Ab 4 : 4 (aberrant 4 : 4; one wild-type, one mutant, and two sectored colonies), 7: 1 and 1 : 7 (tetrads yielding three spore colonies of one genotype and one sectored colony), 8 : 0 and 0 : 8 (tetrads yielding four spores of a single genotype). The ‘other PMS’ class includes aberrant 6 : 2 and 2 : 6 tetrads as well as tetrads with three PMS events. † Notation: þ=þ, homozygous wild-type at the designated locus; þ= Ϫ , heterozygosity for the mutant allele; Ϫ = Ϫ , homozygosity for the mutant allele. ‡ Ref. 5. § Ref. 8. k Aberrant segregation patterns from this strain sporulated at 30 ЊC were examined previously9. As here all strains were sporulated at 18 ЊC, the earlier results are not included.

Nature © Macmillan Publishers Ltd 1997 NATURE | VOL 387 | 26 JUNE 1997 929 letters to nature the segregation of his4-lopd (DTK241 and DTK242; Table 1). of restoration repair would lead to an increase in 5 : 3 events without Neither mutation increased the PMS frequency, indicating that a change in 6 : 2 events. Rad1 may be used for conversion repair in the complete NER complex is not required for the repair of large heteroduplexes in which the mutant strand is donated (Fig. 1b), DNA loops. because loss of conversion repair would lead to both a gain of 3 : 5 As the his4-AAG and his4-lopd mutations are located about events and a loss of 2 : 6 events. Thus, Rad1 may be involved in repair 470 bp apart in HIS4, an alternative explanation of our results is in which the wild-type strand is excised from the heteroduplex, that Rad1 affects DNA-mismatch repair as a function of the position because such events are restorations when the wild-type strand is of the mutation within the gene. Consequently, we examined the donated and conversions when the mutant strand is donated (Fig. effects of the rad1 mutation on the aberrant segregation pattern of 1). Alternatively, as the sequences at the mismatch depend on which his4-Sal, a 4-bp insertion located at the same position as his4-lopd. strand is donated, Rad1 may have a sequence-specific role in repair. We found that none of the aberrant segregants was a PMS tetrad Note that there is an increase in PMS without a compensating (Table 1). As we previously found similar segregation patterns for decrease in either the 6 : 2 or 2 : 6 classes for the msh2 and rad1 msh2 his4-Sal in a RAD1 strain5, we conclude that Rad1 specifically affects strains containing the his4-lopd mutation (Table 2). Either the the repair of large DNA loops. number of tetrads analysed for these strains was too small to There are several other interesting points relating to Tables 1 and detect an altered ratio of the 6 : 2 and 2 : 6 classes, or the roles of 2. First, for strains with the his4-lopd mutation, we find that rad1 Rad1 and Msh2 in repairing loops are slightly different. increases the frequency of both 5 : 3 and 3 : 5 tetrads and reduces the Second, in the msh2 strain heterozygous for his4-AAG, the frequency of 2 : 6 tetrads, but has no effect on the frequency of 6 : 2 increase in PMS is balanced by an overall decrease in conversion tetrads (compare DNY27 and TP1013). This result suggests that (Table 2). This indicates that Msh2 is required primarily for Rad1 is used for restoration repair for loops in which the wild-type conversion-type repair for his4-AAG and restoration-type repair strand is donated in forming the heteroduplex (Fig. 1a), because loss for his4-lopd. This difference probably reflects the nature of the

Figure 1 Heteroduplex formation and DNA-mismatch repair during meiosis. duplexes contain DNA loops, representing sequences present in the mutant Paired meiotic chromosomes are shown, with each chromosome being double- gene but absent in the wild-type gene. Repair of the mismatches can occur either stranded. Centromeres are indicated by black and white ovals; rectangles depict by excision of the loop (followed by resynthesis using the wild-type strand as a genes with mutant insertions shown in black. a, Tetrads in which a wild-type template), or by excision of the sequences opposite the loop (followed by strand is non-reciprocally donated to a mutant gene, or b, in which a mutant DNA resynthesis using the mutant strand as a template). Failure to repair the loop strand is non-reciprocally donated to a wild-type gene4. The resulting hetero- leads to 5 : 3 or 3 : 5 postmeiotic segregation (PMS).

Table 2 Effects of rad1 and msh2 mutations on the frequencies of aberrant segregation and crossover frequencies in diploid strains heterozygous for his4- lopd or his-AAG

Relevant diploid Ab. seg. Conversion PMS PMS/ab. seg. HIS4–LEU2 genotype (strain) (%) (%) (%) (%) distance (cM) his4-lopd strains Wild-type (DNY27) 30 (1.0 ϫ ) 26 (1.0 ϫ ) 4 (1.0 ϫ ) 12 (1.0 ϫ ) 39 (1.0 ϫ ) rad1/RAD1 (TP1012) 29 (0.9 ϫ ) 22 (0.8 ϫ ) 7 (1.8 ϫ ) 24 (2.0 ϫ )* 33 (0.8 ϫ ) rad1/rad1 (TP1013) 39 (1.3 ϫ ) 25 (0.9 ϫ ) 14 (3.5 ϫ )** 36 (3.0 ϫ )** 36 (0.9 ϫ ) msh2/msh2 (DTK223) 42 (1.4 ϫ )* 25 (0.9 ϫ ) 17 (4.3 ϫ )** 41 (3.4 ϫ )** 38 (1.0 ϫ ) rad1/rad1 msh2/msh2 (DTK230) 46 (1.5 ϫ )** 22 (0.8 ϫ ) 23 (5.8 ϫ )** 51 (4.3 ϫ )** 34 (0.9 ϫ ) ...... his4-AAG strains Wild-type (PD83) 55 (1.0 ϫ ) 47 (1.0 ϫ ) 7 (1.0 ϫ ) 14 (1.0 ϫ ) 28 (1.0 ϫ ) rad1/rad1 (DTK224) 64 (1.2 ϫ )* 50 (1.1 ϫ ) 13 (1.8 ϫ )* 21 (1.5 ϫ ) 31 (1.1 ϫ ) msh2/msh2 (no. 5) 64 (1.2 ϫ ) 9 (0.2 ϫ )** 55 (7.9 ϫ )** 87 (6.2 ϫ )** 26 (0.9 ϫ ) rad1/rad1 msh2/msh2 (DTK232) 66 (1.2 ϫ ) 12 (0.3 ϫ )** 53 (7.6 ϫ )** 82 (5.9 ϫ )** 46 (1.6 ϫ ) ...... The data are derived from Table 1. Numbers in parentheses represent the increase relative to wild-type. A single asterisk indicates a statistically signiﬁcant difference compared to wild type at a P value between 0.01 and 0.05; double asterisks indicate a statistically signiﬁcant difference at a P value of less than 0.01. Ab. seg., aberrant segregation.

DNA mismatch (loop as compared to base–base mismatch), SspI-treated pKM55 DNA (from R. Schiestl). rad14 derivatives of AS4 because it was previously found that the msh2 mutation raised (DTK240) and DNY24 (DTK238) were constructed by one-step transplace- PMS and decreased gene conversion for base–base mismatches ment using XbaI/BglII-treated p14.4 DNA (provided by L. Prakash). The distributed throughout the HIS4 gene9. diploids (haploid strains crossed indicated in parentheses) were: DNY27 (ref. Our results demonstrate a previously undetected requirement for 5) (AS4 ϫ DNY24); PD83 (ref. 8) (AS4 ϫ PD73); TP1012 (AS4 ϫ TP1009); Msh2 for the repair of large loops formed during meiotic recombi- TP1013 (TP1011 ϫ TP1009); DTK223 (RKY1721 ϫ DTK222); DTK230 nation. Previous studies showed that msh2 increased PMS for (DTK229 ϫ DTK228); DTK224 (TP1011 ϫ TP1010); no. 5 (RKY1721 ϫ heterozygous base substitutions and for insertions of 4 bp9,13. RKY1452); DTK232 (DTK229 ϫ DTK231); DTK257 (TP1011 ϫ DTK256); Mutations in the MutL homologue MLH1 lead to an increase in DTK241 (DTK239 ϫ DTK237); DTK242 (DTK240 ϫ DTK239). PMS for the MAT loci (which differ by a large region of heterology) Meiotic analysis. Strains were sporulated at 18 ЊC and dissected using (N. Hunter and R. Borts, personal communication), whereas protocols described in ref. 5 (wild-type and rad strains) and ref. 9 (msh2 mutations in PMS1 (another MutL homologue) do not increase strains). PMS frequency for a 38-bp heterology3. Msh2 can bind base–base mismatches and DNA loops in vitro in the absence of other Received 12 December 1996; accepted 22 April 1997. mismatch-repair proteins14, but it is unclear whether other MutS 1. Friedberg, E. C., Walker, G. C. & Siede, W. DNA Repair and Mutagenesis 1–698 (ASM, Washington DC, 1995). and MutL homologues or an entirely different repair pathway are 2. Modrich, P. & Lahue, R. Mismatch repair in replication fidelity, genetic recombination, and cancer involved in Msh2-dependent loop repair in meiosis. In mitotic cells, biology. Annu. Rev. Biochem. 65, 101–133 (1996). 3. Fogel, S., Mortimer, R. K. & Lusnak, K. in The Molecular and Cellular Biology of the Yeast msh2 mutants seem to be incapable of efficiently correcting loops of Saccharomyces: Life Cycle and Inheritance (eds Strathern, J. N., Jones, E. W. & Broach, J. R.) 289– 15,16 20 bases or larger which result from DNA-polymerase slippage , 339 (Cold Spring Harbor Laboratory Press, New York, 1981). but these studies would fail to detect small effects. 4. Petes, T. D., Malone, R. E. & Symington, L. S. in The Molecular and Cellular Biology of the Yeast Saccharomyces: Genome Dynamics, Protein Synthesis, and Energetics (eds Broach, J. R., Pringle, J. R. & Other connections between the mismatch-repair and nucleotide- Jones, E. W.) 407–521 (Cold Spring Harbor Laboratory Press, New York, 1991). excision-repair systems have been reported. For example, the 5. Nag, D. K., White, M. A. & Petes, T. D. Palindromic sequences in heteroduplex DNA inhibit mismatch a repair in yeast. Nature 340, 318–320 (1989). human MutS mismatch-repair activity can recognize bulky 6. Marsischky, G. T., Filosi, N., Kane, M. F. & Kolodner, R. Redundancy of Saccharomyces cerevisiae 17,18 adducts normally repaired by NER proteins , although repair of MSH3 and MSH6 in MSH2-dependent mismatch repair. Genes Dev. 10, 407–420 (1996). these adducts by the mismatch-repair system has not been demon- 7. Johnson, R. E., Kovvali, G. K., Prakash, L. & Prakash, S. Requirement of the yeast MSH3 and MSH6 genes for MSH2-dependent genomic stability. J. Biol. Chem. 271, 7285–7288 (1996). strated. Conversely, the NER system can repair base–base mis- 8. Detloff, P., Sieber, J. & Petes, T. D. Repair of specific base pair mismatches formed during meiotic matches, although much less efficiently than pyrimidine dimers19. recombination in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol. 11, 737–745 (1991). 9. Alani, E., Reenan, R. A. G. & Kolodner, R. D. Interaction between mismatch repair and genetic With certain direct repeat substrates, msh2 has been shown to lower recombination in Saccharomyces cerevisiae. Genetics 137, 19–39 (1994). mitotic recombination to the same extent as rad1, and the msh2 10. DiCaprio, L. & Hastings, P. J. Post-meiotic segregation in strains of Saccharomyces cerevisiae unable to rad1 mutations have about the same effect as the single mutations, excise pyrimidine dimers. Mutat. Res. 37, 137–140 (1976). 20 11. Dowling, E. L., Maloney, D. H. & Fogel, S. Meiotic recombination and sporulation in repair-deficient indicating that Msh2 and Rad1 might function in a single pathway . strains of yeast. Genetics 109, 283–302 (1985). In addition, Msh2 and Msh3 are required in the Rad1/Rad10- 12. Resnick, M. A., Game, J. C. & Stasiewicz, S. Genetic effect of UV irradiation on excision-proficient and -deficient yeast during meiosis. Genetics 104, 603–618 (1983). dependent removal of non-homologous tails from double-strand 13. Reenan, R. A. G. & Kolodner, R. D. Characterization of insertion mutations in the Saccharomyces breaks, but have no subsequent role in mitotic DSB-induced cerevisiae MSH1 and MSH2 genes: evidence for separate mitochondrial and nuclear functions. recombination (N. Sugawara, F. Paques, M. Colaiacovo and Genetics 132, 975–985 (1992). 14. Alani, E., Chi, N.-W. & Kolodner, R. The Saccharomyces cerevisiae Msh2 protein specifically binds to J. Haber, personal communication). Finally, mutations in the duplex oligonucleotides containing mismatched DNA base pairs and insertions. Genes Dev. 9, 234– Drosophila homologue of RAD1, mei-9 (ref. 21), increase PMS22. 247 (1995). 15. Sia, E. A., Kokoska, R. J., Dominska, M., Greenwell, P. & Petes, T. D. Microsatellite instability in yeast: Our results indicate that both Rad1 and Msh2 are involved in the dependence on repeat unit size and DNA mismatch repair genes. Mol. Cell. Biol. 17, 2851–2858 meiotic repair of large DNA loops. As we still observe meiotic gene (1997). conversion in the rad1 msh2 his4-lopd strain, a RAD1 MSH2- 16. Tran, H. T., Gordenin, D. A. & Resnick, M. A. The prevention of repeat-associated deletions in Saccharomyces cerevisiae by mismatch repair depends on size and origin of deletions. Genetics 143, independent pathway for meiotic repair of DNA loops must exist. 1579–1587 (1996). For example, some gene conversion events may not involve hetero- 17. Duckett, D. R. et al. Human MutS␣ recognizes damaged DNA base pairs containing O6-methylgua- nine, O4-methylthymine, or the cisplatin-d(GpG) adduct. Proc. Natl Acad. Sci. USA 93, 6443–6447 duplex intermediates. Whatever the mechanism, our results provide (1996). evidence for multiple pathways for meiotic gene conversion. Ⅺ 18. Mello, J. A., Acharya, S., Fishel, R. & Essigman, J. M. The mismatch repair protein hMSH2 binds ...... selectively to DNA adducts of the anticancer drug cisplatin. Chem. Biol. 3, 579–589 (1996). 19. Huang, J.-C., Hsu, D. S., Kazantsev, A. & Sancar, A. Substrate spectrum of human excinuclease: repair Methods of abasic sites, methylated bases, mismatches, and bulky adducts. Proc. Natl Acad. Sci. USA 91, 12213– Strains and plasmids. All strains are derived from the haploid strains AS13 12217 (1994). (a leu2 ura3 ade6 rme1) or AS4 (␣ and trp1 arg4 tyr7 ade6 ura3 MAL2)23. Strain 20. Saparbaev, M., Prakash, L. & Prakash, S. Requirement of mismatch repair genes MSH2 and MSH3 in the RAD1–RAD10 pathway of mitotic recombination in Saccharomyces cerevisiae. Genetics 142, 727– 8 DNY24 is AS13 with the his4-lopd (ref. 5) allele, PD73 is AS13 with his4-AAG , 736 (1996). and MW1 is AS13 with his4-Sa1 (ref. 5). Strains RKY1452 (AS4 with msh2) and 21. Sekelsky, J. J., McKim, K. S., chin, G. M. & Hawley, R. S. The Drosophila meiotic recombination gene 9 mei-9 encodes a homologue of the yeast excision repair protein Rad1. Genetics 141, 619–627 (1995). RKY1721 (PD73 with msh2) were obtained from E. Alani . rad1 derivatives 22. Carpenter, A. T. C. Mismatch repair, gene conversion, and crossing-over in two recombination- of haploid strains AS4 (TP1011), PD73 (TP1010), DNY24 (TP1009) and defective mutants of Drosophila. Proc. Natl Acad. Sci. USA 79, 5961–5965 (1982). ␤ MW1 (DTK256) were constructed by transforming strains with BamHI- 23. Stapleton, A. & Petes, T. D. The Tn3- -lactamase gene acts as a hotspot for meiotic recombiantion in Ϫ yeast. Genetics 127, 39–51 (1991). treated pDG18 (rad1::URA3; provided by R. Schiestl). Ura derivatives of 24. Boeke, J. D., Lacroute, F. & Fink, G. R. A positive selection for mutants lacking orotidine-5Ј-phosphate these strains were selected on 5-fluoroorotic acid24, creating strains DTK225, decarboxylase activity in yeast: 5-fluoro-orotic acid resistance. Mol. Gen. Genet. 197, 345–346 (1984). DTK226 and DTK221, respectively. msh2 derivatives of haploid strains Acknowledgements. This work was supported by the NIH. We thank R. Borts and J. Haber for DNY24 (DTK222), DTK225 (DTK229), DTK226 (DTK231) and DTK221 communicating unpublished information, and M. Dominska, M. Mears, R. Pukkila-Worley and Q.-Q. (DTK228) were constructed by one-step integration of SpeI–BglII-digested Fan for help with tetrad analysis. pII-Tn10LUK msh2 (ref. 13) (from E. Alani). rad2 derivatives of AS4 (DTK239) Correspondence and requests for materials should be addressed to T.D.P. (e-mail: [email protected]. and DNY24 (DTK237) were constructed by one-step transplacement using edu).