Jul 30, 1990 - KUSHNER 1978) and in a X inversion assay (ENNIS,. AMUNDSEN and SMITH 1987). ..... YCp50 library (ROSE et al. 1987). One transformant.
Copyright 0 1991 by the Genetics Society of America
The Hyper-Gene Conversionhpr5-1 Mutation of Saccharomyces cerervisiae Is an Allele of the SRS2IRADH Gene Lei Rong, Francesca Palladino, And& Aguilera' and Hannah L. Klein Department of Biochemistry and Kaplan Cancer Center,New York University Medical Center,New York, New York 10016 Manuscript receivedJuly 30, 1990 Accepted for publication September28, 1990 ABSTRACT The HPRS gene has been definedby the mutation hpr5-1 that results in an increased rate of gene UV sensitive phenotype of rad18 mutations in hpr5-1 rad18 conversion. This mutation suppresses the double mutants by channeling the aborted repair events into a recombination repair pathway. The HPRS gene has been cloned and is shown to be allelic to the SRSPIRADH gene, a putative DNA helicase. The HPRS gene, which is nonessential, is tightly linkedto the ARG3 locus on chromosomeX . The hpr5-1 allele containsa missense mutation in the putative ATP binding domain.A comparison of the recombination properties of the hpr5-1 allele and the null allele suggests that recombination by several mechanisms. We propose that the HPR5 events in hbr5 defective strains can be .generated , gene functionsin the RAD6 repair pathway.
S
TUDIES of mutations that affect DNA metabolism have shown that DNA replication, repairand recombination have functional overlap in the utilization of common enzymes (KUNZand HAYNES 1981; SMITH 1988) which may reflect the use of similar substrates. Thus mutants that are defective in genes required for DNA replication often show increased sensitivity to agents that damage DNA, presumably the result of unrepaired lesions in the DNA. Under someconditionssome of thesemutants also show decreasedrecombinationrates, for example poZA (DNA polymerase I) and Zig (DNA ligase) mutants of Escherichia coli in Hfr crosses (ZEIG, MAPLESand KUSHNER1978) and in a X inversion assay (ENNIS, AMUNDSENand SMITH 1987). However, with other assays thesemutantshaveincreasedrecombination rates, hyperrecombination a phenotype (KONRAD 1977). Using nontandem lacZ duplication (KONRAD 1977) or Hfr-mediated conjugation assays (FEINSTEIN and Low 1986) hyperrecombination mutants of E. coli have been isolated. Most of these mutants were defective in the known DNA synthesis and/or repair genes, PoZA, lig, uvrD(helicase 11),dam (DNA adenine methylase) and dut (deoxyuridine triphosphatase),and mutH, mutLand mutS (mismatch repair). This may not indicate a direct role for these enzymes in recombination; rather, the mutantsmay accumulate increased nicks or gaps in the DNA which are recombinogenic substrates,thusleading to the hyperrecombination phenotype (FEINSTEIN and Low 1986). In the yeast Saccharomyces cerevisiae several genes have been identified in which mutations were subse-
' Current address: Departamento de Genetica, Facultad de Biologia, Universidad de Sevilla, 41080 Sevilla, Spain. Genetics 127: 75-85 (January, 1991)
quently shown to result in a mitotic hyperrecombination phenotype. These may be loosely grouped into three classes and similar to E. coli, many of these genes encode functions required in DNA metabolism. The first group are from the cdc (cell division cycle) mutants. Hyperrecombinant mutants have been found among the cdc mutants, especially those that arrest in the S or G P phase of the cellcycle. These include mutations in the CDC2 (DNA polymerase 11), CDC6, CDC9 (DNA ligase), CDCl? and CDCl7 (DNA polymerase I) genes (GAME,JOHNSTONand VON BORTSEL 1979; HARTWELL and SMITH 1985). cdc5 and c d c l l mutants, which arrest cellsin G2, also have been reported to be hyperrecombinant, although the observed increasein recombination is not as high asthat observed in mutants defective in DNA polymerase or ligase (HARTWELLand SMITH 1985). The second group comes from the rad (radiation sensitive) collection of mutants. T h e hyperrecombinant rem1 mutants, isolated on the basis of a mitotic mutator phenotype, have been found to be allelic to the RAD3 gene (MONTELONE,HOEKSTRAand MALONE1988) which encodesa DNA helicase (SUNG et aZ. 1987). Other rad mutantsthat are hyperrecombinant are rad6 (KOWALSKI and LASKOWSKI 1975; KERN and ZIMMERMAN1978), rad18 (BORAMand ROMAN1976), and rad50 (MALONEand ESPOSITO 1981). Other genes involved in DNA repair include the PMSl gene. The pmsl mutation, isolated as a mitotic mutator (WILLIAMSON, GAMEand FOCEL1985), also shows a hyperrecombinant phenotype. The PMSl gene showssequence homology to mutL of E. coli and hexB of Streptococcus pneumoniae and has been shown to be required for mismatch repair (BISHOPet aZ. 1987).
76
L. Rong et al.
The third group of mutants are the DNA topoisomerases. DNA topoisomerases have been shown to have a role insuppressingrecombinationbetween repeated sequences such as the rDNA repeats (TOPI a n d TOP2) (CHRISTMAN,DIETRICHa n d FINK 1988) a n d 6 sequences (TOP3) (WALLISet al. 1989). Finally, additional hyperrecombinant mutants include the rec mutants (ESPOSITOet al. 1984). The nature of these products has not yet been established. In yeast, the rad mutations define three epistasis groups of genes involved in DNArepair (for a review see HAYNES andKUNZ 1 9 8 1; FRIEDBERG1988). T h e RAD? epistasis group is required for excision repair andamongothersincludesthe R A D l ,R A D 3 a n d RAD4 genes. The RAD6 epistasis group is involved in error-prone or post-replicativerepair. T h e RAD6, RAD9 and RAD18 genes function in this repair pathway. The third epistasis group is required for recombination repair and includes the RAD50 a n d RAD52 genes. Mutations in genes in this group are defective in repair of double strand breaks andare sensitive to X-ray-induced lesions. We have isolated mitotic intrachromosomal hyperrecombinant (hpr) mutants in an attempt to understand the separate but mechanistically linked events of gene conversion and reciprocal exchange in Saccharomyces cerevisiae. Among these mutants were alleles of the CDC2 (hpr6) a n d CDCl7 (hpr?) DNA polymerase genes (AGUILERA and KLEIN 1988). The experiments reported here concern mutations in a third gene hpr5. T h e hpr5-1 mutation is slightly UV sensitive and results in a tenfold increase in gene conversion. This has been observed using an assay which distinguishes gene conversion events between nontanof a sedemrepeatsfrompop-outrecombination quence located between the repeats. The gene conversion seen in the hpr5-1 is highly biased, with o n e member of the duplicated genepair acting asa donor inapproximately 89% of the conversioneventsas compared to50% in wild-type strains. We have cloned the HPRS gene and determined that hpr5-1 is a missense mutation in a putative active siteof the protein. We also show that HPRS is allelic to RADH which encodes a protein that is highly homologous to the Rep and UvrD DNA helicases of E. coli (ABOUSSEKHRA et al. 1989). By comparingthephenotypes of the hpr5-I missense mutation and a hpr5 null allele we postulate that the HPRS gene functions normally in the RAD6 error-prone repair pathway and propose a model for the action ofHPRS in repair and recombination. MATERIALS AND METHODS Strains: The yeast strains used are shown in Table 1. Media and growth conditions: Rich medium YEPD, synthetic mediumSD, synthetic complete medium SCwith bases and amino acids omitted as specified, and sporulation
medium were prepared as described (SHERMAN, FINK and HICKS1986). All strains were grown either on solid medium or in liquid medium with rotatory shaking at 30". Genetic analysis and determinationof recombination rates: Genetic analysis was performed according to published procedures (SHERMAN, FINKand HICKS1986). Recombination rates were calculated by the median method of LEAand COULSON948) (1 as described (AGUILERA and KLEIN 1988). Strains were grown on YEPD for 2-3days.Since not all of the strains were isogenic at least three different strains of each genotype were analyzed. For each genotype, three or more independent clones were analyzed for each fluctuation test. As the hpr5-1 and hpr5A56 alleles do not affect strain viability, cell number was determined using a hemocytometer. Leu+ recombinants were selected on complete synthetic medium (SC) lackingleucine. The proportion of Ura+ events among the Leu+ recombinants was calculated by streaking independent colonies from YEPD onto SC-Leu medium to isolate independent Leu+ recombinants and testing these for the Ura phenotype on SC-Ura medium. His+ recombinants were selected on SC-His medium. Cloning of the HPR5 gene: A hpr5-1radl8-1 strain F28-ID6 was transformed with a partial Sau3A genomic library in YCp50 using the lithium acetate method of ITO et al. (1983). Ura+ transformants were picked and screened for sensitivity to ultraviolet irradiation. In 6000 transformants one was found to confer UV sensitivity to the F28-1D6 strain. The plasmidDNA pHK201 recovered from this transformant carried a 8.8-kb insert covering the HPR5 gene. The 5.6-kb EcoRI-XhoI fragment of pHK201 was used to replace the EcoRI-Sal1 region of YCp50 to give pHK207. Cloning and sequencing of the hpr5-1 allele: Plasmid pHK207 was digested with Sac1 restriction enzyme which cuts at -681 bp and 282 bp of the HPR5 coding sequence to give a deletion of 963 bp, as deduced from the sequence et al. 1989). The gapped plasmid of RADH (ABOUSSEKHRA was used to transform yeast strain F28-1D6. Ura+ unstable transformants were selected and UV sensitivity determined by exposure to 300-nm UV light for 15 sec. The UVresistant transformants contained the cloned hpr5-1 mutation on the plasmid while the U V sensitive transformants carried the wild-type HPRS gene. The 190 bp PstI-Sac1 fragment from the cloned hpr5-1 gene was subcloned into M13mp19 (MESSING1983) and sequenced by the dideoxychain termination method (SANGER, NICKLENand COULSON and 1977) using T7 DNA polymerase (Sequenase) (TABOR RICHARDSON 1987) and 5'-([~y~'S]thio)-triphosphate (BIGGIN, GIBSON and HONG 1980). DNA manipulation: Plasmid DNA was isolated from E. coli by CsCl gradient centrifugation as described (CLEWELL and HELINSKI1970). Yeast genomic DNAwas prepared FINKand from 5-ml YEPD cultures according to SHERMAN, HICKS(1986) and was used for Southern blotanalysis (SOUTHERN 1975). 3'P-labeled DNA probes were prepared according to FEINBERG and VOCELSTEIN(1984). Hybridization was performed in 6 X SSC, 1 X Denhardt's solution, 0.25 M NaH2P04pH 6.0 and 30 mM NapP207 at 65" for 18 hr.Linear DNA fragments were recovered directly from agarose gels and used in DNA labeling experiments or in ligation reactions using T4 DNA ligase overnight at 14". Yeast chromosomal DNA was isolated and run on OFAGE gels according to CARLE and OLSON(1984) andblotted and hybridized as described (AGUILERA and KLEIN 1990). Constructionsusedtostudybiasedgeneconversion: Plasmid pHK134 containing a 2.16-kb SalI-Xhol DNA fragand HABER ment bearing the leu2-r allele (BORTS, LICHTEN 1986) of the LEU2 gene inserted into the Sal1 site of pBR322 and a 1.1-kb Hind111 fragment bearing the URA3 gene
RADHwith HPR5 of Allelism
77
and SRS2 TABLE 1 Yeast strains
Strain
Genotype
Source
F425 MATa his4ABC radl8-1 G . FINK F554 MATa ade2-1 qcl-9 leul-12 arg4-17 his5-2 srs2-2 rad6-1 rad18-2 C. LAWRENCE ACUILERA and KLEIN(1988) MATaadel-101 ura3-52 leu2-I 12::URA3::leu2-khis3-513::TRPI::his?-5?7 hpr5-1 27C14-4A ACUILERA and KLEIN(1988) MATa adel-101 ura3-52 trpl Eeu2-II2::URA3::leu2-k his3-513::TRP1::his3-5?7 344-109D ACUILERA and KLEIN(1 989a) MATaadel-I01 ura3-52 trpl leu2-I 12::URA3::leu2-k hpr5-1 AA15-5B ACUILERA and KLEIN(1 989a) MATa his4-38 ura3-52 trpl leu2-I I2::URA?::leu2-k AAl5-5C ACUILERA and KLEIN(1989b) MATa adel-101 ura3-52 trpl leu2A68 his3-k::LEU2-leu2-r::his?Ah-URA3 hprl-1 AL644-4 R. KEIL MATa ade2-1 his4-260 lys2ABXho ura?-52 trpl-h leu2-3,112 canl KM84 A 1F28MATa adel-I01 ura3-52 trpl leu2-I 12::URA3::leu2-k radl8-1 study This ABOUSSEKHRA et al. (1 989) FF 18244 MATa ade5 ura3-52 trpl-289 leu2-3,l 12 can1 radH This study MATa ade5 ura3-52 trpl-289 leu2-3,112 canl radH radl8::LEU2 FF18244::plg KM84-56 MATa ade2-1 hu4-260 lys2ABXho ura3-52 trpl-h leu2-3,I 12 canl hpr5A56::TRPl Thisstudy AA15-5B202 MATa adel-101 ura3-52 trpl leu2-I12::URA3::leu2-khpr5-1-TRPl This study AA15-5C202 MATa his4-38 ura3-52 trpl leu2-I 12::URA?::leu2-k HPR5-TRPI study This F31-5D MATa adel-I01 ura3-52 trpl leu2-112::URA3::leu2-k hpr5-1 radl8-1 study This F28-1 D MATa adel-101 ura3-52 trpl leu2-112::URA3::leu2-k hpr5-1 radl8-1 This study F28-1D6 MATa adel-101 ura3-52 trpl leu2-I IP::TRPI::leu2-k hpr5-1 rad18-1 This study 386-5C MATa ura3-52 trpl leu2-I 12::URA3::leu2-k HPR5-TRPI radl8-1 study This 397-3C ura3-52 MATa trpl srs2-2 (rad62) rad18 study This F74-1B ura3-52 MATa trpl leu2-?,112 hpr5A56 study This F74-1C MATa adel-I01 ura3-52 trpl leu2-3,112 his3-513::TRPI::his?-537 hpr5A56 This study radlB::LEU2 trpl F65-2A ura2 MATa Eeu2-3,112 his3-11,15 study This F83-1B MATa arg3 ura3-52 trpl study This 1-2728 27 MATa his4-38 ura3-52 trpl leu2-101,I 12 rad6-1 study This 388-3B MATa adel-I01 ura3-52 trpl leu2-I 12::URA3::leu2-k hpr5-I-TRPI study This F103-2A MATa adel-I01 ade2-1 ura3-52 trpl-h lys2ALiXho leu2-112::URA3::leu2-k This study F103-2A56 MATa adel-I01 ade2-I ura3-52 trpl-h lys2ABXho leu2-I 12::URA3::leu2-k This study hpr5A56::TRPI 347-6A32 ura3-52 MATa trpl leu2-r::URA3::leu2-k study This 347” adel-101 MATa ura3-52 trpl leu2-r::URA3::leu2-khpr5-1 study This ura3-52 314R14-4B ade2-I01 MATa leu2-3 study This RL5 MATa ade2-101 ura3-52 leu2-r::URA3::leu2-3 study This RL5-SA MATa ura?-52 leuZ-r::URA3::leu2-3hpr5-1 study This OUl-ID MATa ura3-52 t r p l leu2-k::URA3::leu2-I12 This study OUl-IC ura3-52 MATa trpl leu2-k::URA3::leu2-I12hpr5-1 study This OU2-1A56 MATa lys2ABXho ura3-52 trpl-h leu2-k::URA?::leu2-112 hpr5A56::TRPI This study 34%-6A adel-101 MATa ura3-52 trpl leu2-k::URAjr::Eeu2-r study This 349s-9c ura3-52 MATa trpl leu2-k::URA3::leu2-rhpr5-1 study This 392-8B MATa adel-101 ura3-52 trpl leu2A68 his3-k::LEU2-leu2-r::his3Ah-URA3 study This 392-4B MATa ura?-52 trpl leu2A68 his?-k::LEU2-Eeu2-r::his?Ah-URA3hprl-I hpr5-1-TRPl This study R393-6D MATa ura3-52 trpl leu2A68 his3-k::LEU2-leu2-r::his3Ah-URA3 hpr5-I-TRPI This study F91-1C MATa ura3-52 trpl leu2A68 his3-k::LEU2-leu2-r::his3Ah-URA3 hpr5A56::TRPl This study F9-2B 1 MATa ura3-52 trpl leu2A68 his3-k::LEU2-leu2-r::his3Ah-URA3 hpr5A56::TRPJ This study and CHRISTENSEN 1979) as indeed are radH T h e srs2 alleles are dominant in suppression of rad6 and rad18 UV sensitivity (LAWRENCE (ABOUSSEKHRA et al. 1989) and hpr5 (Figure 2) mutations. However, since the mutations are fully recessive inRAD+ a background, the wildtype allele i s written in uppercase letters and the mutant alleles in lowercase letters.
inserted into the Hind111 site of pBR322 was used to conleuZ-r::URA3::leu2-3 duplication. The plasstruct the midwasdigested with BstEII and used totransform strain314R14-4B(seeTable1) to Ura+. The construction was confirmed by Southern analysis. The other duplications leu2-r::URA3::leu2-k, leuZ-k::URA3::leu2-r and leuP-k::URA3::leu2-112 have been described (KLEIN 1988). RESULTS
Epistatic interaction of the hfir5-I mutation with rad mutations: T h e h p 5 - l allele leads to an increase in mitotic intrachromosomalgene
conversion and
gene conversion between homologous chromosomes (AGUILERA and KLEIN 1988). We examined the effect of h p 5 - I rad double mutants on cellviability and recombination, using several rad mutations ( r a d l , rad3, rad4, rad6, rad9, rad18 and rad50), including those reported to result in a hyper-gene conversion phenotype. No effect on viability or recombination was seen when the 4 7 - 5 - I allele was combined with a radl,rad3,rad4 or rad9 mutation.However, we observed thatthe 47-5-I rad5OA doublemutant showed a strong decreasein growth rate as compared to the single mutant strains. We also observed that
78
L. Rong et al. , 1 kb,
HPRS
m B
(RADH) sc
C
TRPl (hpr5A 5 6 ) FIGURE1.-Restriction map of the yeast DNA insert in pHK2O 1 that complements the hpr5-I mutation. The thick solidline indicates the HPRliIRADH coding region as determined by ABOUSSEKHRA et al. (1989). Abbreviations of restriction sites: B, BamHI; Bg, BglII; C, Chi; H, HindIII; Hp; HPaI; P, PUuII; Ps, PstI; S, SalI; Sc, Sacl; X , XhoI; Xb, Xbal. The hpr5A56 deletion was constructed by replacement of theinternal 2.2-kbXbaI-BglII fragment from pHK201 with the 0.8-kb EcoRI-BglII TRPl fragment from plasmid YRp7 using XhoI linkers on the X6aI and EcoRI ends of the DNA fragments.
the hpr5-1 mutation suppressed the UV sensitivity of rad6 (data notshown) and rad18 mutations (see Figure 2). Intrachromosomal gene conversion was reduced 2-3-fold in the hpr5-1 radl8-1double mutant as compared to the hpr5-1 single mutant (data not shown). Thus the hpr5-1 mutation suppresses rad18 and rad6 mutants with only a slight effect on thehpr5-1 recombination phenotype. Cloning of the HPR5 gene: We used the suppression of the radl8-1 UV sensitivity by the hpr5-1 mutation to clone the wild-type HPRS gene.Strain F28-1D6 carrying the radl8-1 and hpr5-1 mutations and the ura3-52 mutation was transformed with the YCp50 library (ROSEet al. 1987). One transformant was identified by the ability to confer UV sensitivity to the UV resistant hpr5-1 radl8-1 strain (see MATERIALS AND METHODS). Plasmid DNA (pHK201)recovered from this transformant was used to retransform strain F28-1D6. All retransformants had a UV sensitive phenotype that cosegregated with the plasmid marker.Subcloningexperiments localized the complementing region to 6.0 kb from theinsertion in the pBR322vector tothe XhoI site of the insert (Figure l), scored by conferring UV sensitivity to the radl8-1 hpr5-1 strain F28-1D. To confirm that the 6.0-kb insert carried the HPR5 gene, the ClaI-ClaI 3.2-kb fragment (see Figure 1) was used to integrate a selectable marker, T R P l , intostrainAAl5-5C (Table 1) togive strain AA15-5C202. This strain also carries an intrachromosomal recombination assay for scoring hyperrecombination. Strain AAl5-56202 was crossed to an hpr5-1 strain carrying the same recombination system (strain AA15-5B) and the resulting diploid was analyzed by tetrad analysis. The cloned fragment was linked to the HPRS allele as shown by 38/38 tetrads segregating two Trp+ spores
(the marked HPR5 locus) to two hyperrecombinant (hpr5-1) spores. In the course of these studies we learnedthat ABOUSSEKHRA et al. (1989)had isolated amutant which suppressed the UV sensitivity of rad18 mutations. T h e amino acid sequence of the suppressor gene, RADH, shows a high degree of homology to the E. coli DNA helicases UvrD and Rep (ABOUSSEKHRA et al. 1989) andcontains all of the aminoacid sequence motifs defined by HODGMAN (1 988) characteristic as of DNA helicases. T h e RADH gene was found to be linked to a gene called SRSZ (LAWRENCE and CHRISTENSEN 1979) that was isolated as a suppressor of the UV sensitivity of rad6 and rad18 mutations. Three lines of evidence suggest that HPR5, RADH and SRSZ are the same gene. First, comparison of the restriction maps of RADH (ABOUSSEKHRA et al. 1989) and the insert in pHK201 (Figure 1) shows that the regions are identical, indicatingthat HPRS and RADH arethe same gene.Second, the HPRS containing plasmid pHK201 suppresses the UV resistance of a radH radl8A strain (strain FF18244::p18, Table 1) and a srs2-2rad18 strain (strain 397-3C, Table 1). Third, HPRS and RADH and HPRS and SRS2 are linked by tetrad analysis. Diploid radHIHPR5-TRPl radl8Alradl8-1 gave 99 of 101 tetrads parental ditype and diploid srs2-2/HPR5-TRPl radl8-l/radl8-l gave 93 of 96 tetrads parentalditype. The remaining tetradsaretetratypetetradsthat arise as a consequence of unequal crossing over in the duplicated 3.2kb ClaI-ClaI fragment used tointegratethe TRPl marker adjacent to the HPRS allele. These results combined with the observation that all attempts to isolate novel extragenic suppressors of RAD6 and RAD18 have identified only one complementation group (LAWRENCE and CHRISTENSEN 1979; ABOUSSEKHRA et al. 1989; SCHIESTL, PRAKASH and PRAKASH 1990; F. PALLADINO, unpublished observations) lead us to conclude that HPR5, RADH and SRSZ are the same locus. Genetic mapping of the HPRS locus: T h e 3.2-kb CZaI-ClaI fragment of pHK201 (Figure 1)was used to probe an OFAGE blot of whole yeast chromosomes. HPRS was located on chromosome X. Further mapping studies have placed HPRS tightly linked to the A R C 3 locus (58 of 58 tetrads were parental ditype). However HPR5 is not ARG3 as a null allele of HPRS (see below) has no arginineauxotrophyphenotype and the restriction map of HPRS shows no overlap with the restrictionmap of A R C 3 (CRABEEL et al. 1985). DNA sequence of the hpr5-1 allele: The hpr5-1 allele was cloned by gap repair as described in MATERIALS AND METHODS. T h e mutation was localized to the 0.96-kb Sad-Sac1 fragment at the 5’ end of the gene.Sequencing of the regionfrom PstI to Sac1
Allelism of HPRS with RADH and SRS2
-
-t-
.o 1
79
RAD18 HPR.5 RAD18 hpf5A56
rad18-1 hPr5A56
RAD18 hpr5-1 rad18.f hpr5.1
radl8-l HPRJ
FIGURE2.-UV survival curves of hpr5-1 and hpr5A56 single mutants compared to hpr5-i radl8-1 and hpr5A56radl8-1 double mutants. Strains used were: 344-109D (HPR5 RAD18); F74-1B (hpr5A56 RAD18); F74-IC (hpr5A56radl8-I); AA155 8 (hpr5-1 RADI8); F31-5D (hpr5-1 r a d l b l ) ; F28-1A (HPR5 radl8-I).
&"J 20
40
60
80
100
DOSAGE (Jim2 )
revealed thatthe hpr5-1 mutation is a single base transition of C to T, resulting in a change of the Pro-37residue to Leu-37. This mutation is in the conserved domain I , the ATP binding domain that is present in all DNA helicases (HODGMAN 1988). The affected proline residue is within the DNA helicase consensus sequence of (A/P)GXGK(S/T) (BURGESS, COUTOand GUTHRIE1990). T h e hpr5-1 transition mutation in the presumed ATP binding site of the HPR5 protein would introduce nonconserved a change (prolineto leucine) thatmay reduce or abolish helicase activity. The HPRS gene is not essential for mitotic growth: A null allele of HPR5, hpr5A56, was constructed by XbaI-BglII fragment with replacing the internal 2.2-kb the TRPI gene. This was used to replace the chromosomal copy of HPR5 in a haploid strain. T h e disrupted haploid strain showed normal viability, indicating that theHPRS gene does not encode a function essential for mitotic growth. U V sensitivity of hpr5 mutants: We examined the UV sensitivity of HPRS RADI8, hpr5-I RADI8, hpr5A56 RAD18, HPRS radl8-I, hpr5-1 radl8-1 and hpr5A56 radI8-I strains. Both haploid and diploid cultures in stationary phase were used. T h e survival curves show that thehpr5-1 RAD18 strain has a tenfold HPR5 RAD18 decrease in viability as compared to the strain(Figure 2; AGUILERA and KLEIN 1988). The HPR5 radI8-1 strain shows a greater than 103-fold decrease inviability at the samedose. T h e hpr5-1 radl8-I double mutant shows the same UV sensitivity as the hpr5-IRAD18 strain. T h e UV sensitivity of hpr5-I RAD18 and hpr5-1 rad18 is not apparentwhen patches of cells are replica-plated and thenexposed to UV irradiation, but can only be seen by UV survival curves. On replica plates hpr5-1 rad18 strains appear as UV resistant as wild-type strains. Similar results
were obtained using hpr5A56 RAD18 and hpr5A56 rudI8-1 strains. Both genotypesresulted in similar levels of UV sensitivity. T h e hpr5 deletion mutant is slightly more UV resistant than the hpr5-1 mutant, but otherwise is similar to the hpr5-l mutant in terms of the interaction with rad18 mutations. Experiments with diploid strains showed that hpr5-1 RAD18 and hpr5-I radl8-1 homozygous diploids aremore UV sensitive than the corresponding haploid strains (data not shown). Heterozygous HPRSlhpr5-1 and HPR5I hpr5A56 strains in a radl8/radl8 background suppressed the radI8 UV sensitivity to give the UVsensitive phenotype of the hpr5 allele, indicating that the suppression of rad18 UV sensitivity by hpr5 mutation is dominant. Mitoticintrachromosomalgeneconversion in hpr5-1 and hpr5A56 strains: T h e hpr5-1 mutation has been shown to result in biased gene conversion in a leu2-I I2::URA3::leu2-kduplication (duplication I of Figure 3; ACUILERAand KLEIN 1988), seen as an excess of conversion of the leu2-k allele to wild type compared to conversion of leu2-I I2 to wild type (expressed as the R/L ratio in Figure 3). The biased pattern of conversion in the hpr5-1 mutant could be due to (1) the nature of the two mutations used in the duplication, (2) the position of the allele within the gene (5' vs. 3' end of the gene) or (3) the location of the allele within the duplication (centromere distal or proximal). T o further investigate the biased gene conversion in hpr5-1 strains a series of leu2 duplications that contain different leu2 alleles was constructed. T h e four leu2 alleles used in these duplications were: (1) Zeu2-I 12, an ICR170 induced putative +I frameshift mutation that is 3' to the EcoRI site in the LEU2 coding region (M. CULBERTSON, personal communication); (2) leu2-k, a 7-bp deletion at the KpnI
80
L. Rong et al. Duplication
L kU.?-112
R HPRS 27 hprS-1 hprSA56
URA3
1 Ieu2-r
URA3
HPRS
5.1 58.0 ( x l l ) 16.0 ( ~ 3 )
36 6 7
2.6 28.8 1)(x1
21
46
URA~
leu.?-3
HPRS ltprs-1
1.6 11.928(x7)
Ieu2-k
ORA3
ku2-112
HPRS hprS-1 hprSA56
2.9 39 73 12.5 (x4) 92 1.9 ( ~ 0 . 7 ) 19
llPR5 hpr5-1
2.1 11.3 (x5)
4 IeuZ-k
ORA3
leu.?-r
35
24 21
12.9.
1.7
8 6.048
bul-r
3
1.3
48 8.0 31 5.29.4.0
6.70.
69 1.5 95 7.46.3.4
49 19
0.5 0.5 1.0
0.01 2.20
8 0.3 5 0.030.2
FIGURE 3.-Analysis of mitotic gene conversion events in LEU2 duplications. Strains used were: 344-1091) ( H P R 5 ) , 27C14-4A (hpr5-1) and F103-2A56 (hpr5A56) for duplication 1; 347-6A32 ( H P R 5 ) and 347s-6D (hpr5-1) for duplication 2; RL5 ( H P R 5 ) and RL5-9A (hpr5-1) for duplication 3; OUI-ID (HPRS), OUI-IC (hpr5-1) and OU2-1A56 (hpr5A56) for duplication 4; 349s-6A ( H P R 5 ) and 349s-9C (hpr5-1) for duplication 5. Additional related strains were also used in fluctuation tests so that the recombination rates shown for each genotype represents the average of the median values of three to four strains. The Leu+ Ura+ rates were determined from the percentage of Leu' Ura+ conversion events of the total Leu+ events and the Leu+ rates. From 50 to 100 independent Leu+ segregants from each strain were used to determine the Ura phenotype among the Leu+ segregants. In duplications 1, 2 and 3 the proportion of Ura+ among the Leu+ segregants is 50% in HPR5 strains and 80% in hpr5-l and hpr5A56 strains. In duplications 4 and 5 over 99% of the Leu+ segregants were Ura+ regardless of the HPR5 genotype. Some of these were found to be triplications upon Southern analysis and these were subtracted from the Leu+ Ura+ events to give Leu+ Ura+ conversion events. The numbers in parentheses adjacent to the rates are the fold increase in hpr5 mutant strains compared to the wild type HPR5 strain with the same duplication. L and R represent the left and right leu2 allelesas represented above in the duplications with the right allele being proximal to CENIII. The molecular analysis of the conversion events was done by Southern analysis of genomic DNA isolated from independent Leu+ Ura+ recombinants digested with KpnI or EcoRI restriction enzyme as required. Contingency x* tests were done on the distribution of conversion events at the L and R alleles comparing HPRS and hpr5 strains. Values greater than 3.84 were regarded as statistically significant ( P < 0.05) and are marked by asterisks. The data for the HPR5 and hpr5-1 strains in duplication 1 are taken from ACUILERA and KLEIN (1988).
site ofthe LEU2 coding region (H. KLEIN,unpublished results); (3) leu2-3, a +1-bp insertion 3 bp 5' from the KpnIincisionsite (M. CULBERTSON, personalcommunication); and (4) leu2-r, a +4-bp insertion from a Klenow fill-in reaction at the EcoRI site of the LEU2 coding region (BORTS,LICHTENand HABER1986). The duplicationswereanalyzed for recombination rates and biased gene conversion among the Leu+ Ura+ segregants. Duplication 1 leu2-112::URA3::1eu2-k results in an 1I-fold increase in Leu+ Ura+ segregants in a hpr5-I strain and gives biased gene conversion among these segregants that is significantly different from wild type (Figure 3, ACUILERA and KLEIN 1988). T o investigate whether the biased conversion seen in duplication 1 was due to the nature of the mutations used or the allelic position of the mutations, duplication 2 leu2-r::URA3::leu2-k (Figure 3) was used. In this duplication the leu2-112 allele of duplication 1 has been substituted with the leu2-r allele, changing a +1-bp frameshift mutation to a +4-bp insertion at a site 3' to theKpnI site inthe LEU2 coding region. Leu+ Ura+
segregants were analyzed by Southern blots to determinewhichallele was converted to wild type. The data show that, first, the hyperrecombinant phenotype is still observed and, second, that the leu2-k allele is preferentially converted to wild type in hpr5-I strains but not in HPRS strains (Figure 3). This indicates that the nature of the mutation is not important for the biased gene conversion. T o determine whether the nature of the allele in the centromere proximal (rightmost) gene is related to the gene conversion biasa third duplication, duplication 3 of Figure 3, was constructed. In this duplication the Zeu2-k allele of duplication 2 hasbeen replaced by the leu2-3 allele. The site of the leu2-3 +1 bp insertion overlaps the 7-bp deletion of leu2-k. As can be seen in Figure 3, this duplication still gives the hyperrecombinant phenotype and biased gene conversion of the right gene member of the duplication, in this case leu2-3, in hpr5-1 strains. These results suggest that it is not the nature, but the position of the mutation in the gene, whether it is located in the centromere proximal or centromere distal copyof the
81
Allelism of HPR5 with RADH and SRS2 His' Independent His' events Genotype recombination LeuTotal %Leu' rate ( ~ 1 0 ~ ) Leu'
I e u lf2 -r
---t--1.16kb
his3A
HPRS
6.1
100
84
184
44
hpr5-1
20.0
268
72
340
21
X'
33.0*
his3-k
10.0 45 26 71 37 1.35 hprSd56 FIGURE4.-Length of conversion tracts in HPR5, hpr5-1 and hpr5A56 strains. The inverted repeat construction used for these studies has been described by ACUILERAand KLEIN (1989b). Strains used are: 392-8B (HPR5); 392-4B and R393-6D (hpr5-Z); F91-1C and F91-2B (hpr5A56). The average of the median values of each strain is presented. The percentage of Leu- recombinants among the His+ segregants was determined by isolating independent His+ recombinants from SC-His medium and testing the Leu phenotype on SC-Leu medium. Contingency x* values were determined for the distribution of Leu' and Leu- segregants among the His+ recombinants in HPRZ and hpr5 strains. A value over 3.84 was considered statistically significant (P < 0.05) and is marked by an asterisk.
LEU2 duplication, is important for the gene conversion bias. To further investigate the importance of the allele position in gene conversion bias we examined whether the location of thegene within the duplication, whether it is centromere distal (left) or proximal (right) is critical. For this duplications 4 and 5 were used (Figure 3). Duplications 4 and 5 are identical to duplications1 and2, respectively, except thatthe order of the genes within the duplication is reversed. T h e rate of Leu+ Ura+ recombinantsin HPRS strains is reduced 2-3-fold using duplications4 and 5 as compared to duplications 1 and 2, respectively. The hpr5-1 mutants showed a 4-5-fold increase in the Leu+ Ura+ rate in duplications4 and 5ascomparedto HPR5 strains (Figure 3). This indicates that the hyperrecombinationphenotype of hpr5-1 is still seen with duplications 4 and 5 . When the Leu+ Ura+ segregants were analyzed, no evidence of biased gene conversion was seen in either duplication (Figure 3). These results suggest that both the position of the allele within the gene and theposition of gene within the duplication are two importantfactors in gene conversion bias in hpr5-Z strains. Since the alleles that show preferential conversion are those that areclosest to the pBR322vectorsequences thatseparatethe duplicatedgenes, the pBR322 vector mayin some way contribute to preferential initiation of recombination in hpr5 strains. Whatever the source of the biased conversion, it is specific to mutations in the HPRS gene. We compared the effect of the hpr5-1 missense mutation on recombination tothe recombination properties of the hpr5 null mutation. Studies similar to those using hpr5-1 strainswereperformed with hpr5A56 strains with the most relevant duplications, duplications 1 and 4. With the first duplication only a %fold increase in recombination was observed as compared to the 11-fold increase in hpr5-I strains (Figure 3). However, a bias in gene conversion similar to that
seen in hpr5-1 strains was observed with the hpr5A56 strain. When duplication 4was used, which has the reverse order of alleles to duplication 1, the 3-fold hyperrecombination phenotypeof hpr5A56 was not observed. As in hpr5-I strains, gene conversion bias is also absent. Therefore hpr5A56 null mutants have a similar effect on gene conversion bias to the hpr5-1 mutants. However the hyperrecombinationphenotype is approximately 4-fold lower with respect to the hpr5-I strains for each duplication tested. Gene conversion tract length in hpr5 strains: We have used the recombination system (shown in Figure 4) described by AGUILERA and KLEIN (1 989b) to examine the length of intrachromosomal gene conversion events in hpr5-1 and hpr5A56 strains. His+ prototrophs, resulting from geneconversion between the two his3 genes of the inverted repeat (Figure 4) are selected and thenanalyzed for the leucine phenotype. His+ Leu+ events result from gene conversion events that areless than 1.16 kb andmay be as short as a few nucleotides (AGUILERA and KLEIN 1989b). His+ Leuevents are coconversion events that extend at least 1.16 kb. Figure 4 shows that among the His+ segregants in HPR5 strains, 44% are Leu- (coconversion events),a value similar to that previously reported (AGUILERA and KLEIN 1989b). In the hpr5-Z mutant 2 1 % of the His+ segregants are also Leu-, a statistically significant reduction from thevalue observed with the HPR5 strain (x2 = 33.0, P < 0.05). This suggests that most of thegene conversion events seen in the hpr5-1 mutant have short tract alength. The hpr5A56 mutation does not significantly change the distribution of His+ events relative to wild type (Figure 4). We have also examined the association of crossing over with the His+ gene conversion events. We find no difference in crossover association in HPR5 and hpr5-1 strains, for either short conversion His+ Leu+ events (29% of events are associated with crossover in
82
L. Rong et al.
HPRS strains and 33% in hpr5-1 strains) or long conversion His+ Leu- events (53% in HPRS strains and 52% in hpr5-1 strains). DISCUSSION
We have demonstrated in this study that the HPR5 gene, defined by the hyper-gene conversion mutation hpr5-1, is the same gene as RADH and SRS2. The hpr5-1 mutation, isolated by the hyperrecombination phenotype(AGUILERA and KLEIN 1988), also suppresses the strong UV-sensitive phenotype of rad18 strains. Previously two groups have defined mutations that were able to suppress the UV-sensitive phenotype of rad6 and rad18 cells. LAWRENCE and CHRISTENSEN (1979) definedSRS2 by mutations that suppressed the trimethoprim sensitivity phenotype of rad6 mutants and subsequently showed thatthesemutations also suppressed the DNA repair defect phenotypeof rad 18 cells.ABOUSSEKHRA et al. (1989) defined RADH by mutations that suppressedthe UV sensitivity of rad18 cells. We have cloned HPRS and have shown that (1) t.he restriction map of HPRS is identical to the reported map of RADH (ABOUSSEKHRA et al. 1989), (2) HPR5, RADHand SRS2 are tightly linked genetically, and (3) the HPRS clone is able to complement the radH and srs2 phenotype of suppression of rad18 UV sensitivity. This togetherwith the fact that all reported attempts to isolate extragenic suppressorsof rad6 and/ or rad18 UV sensitivity have defined a single locus (LAWRENCE and CHRISTENSENABOUSSEKHRA 1979; et al. 1989; SCHIESTL, PRAKASH and PRAKASH 1990;F. PALLADINO,unpublished observations) clearly indicates that SRS2, RADH and HPRS are the same gene. In this regard it should be noted that the other allele of HPRS reported in this study, hprSA56, suppresses the rad18 UV sensitivity, although this was not the basis for the isolation of this allele. LAWRENCE and CHRISTENSEN ( I 979) andABOUSSEKHRA et al. (1989) have reported that srs2/radH mutants are dominant in suppressing rad18 UV sensitivity in diploid strains. We have also observed dominance of the hpr5-I mutation in suppressing rad18 UVsensitivity and similar tothe observations of ABOUSSEKHRA et al. (1989) conclude that the dominance appears to related to strain ploidy. A hpr5-l rad18 haploid bearing a wild-type HPRS gene on a centromere plasmid isUV sensitive while HPR5/ hpr5-1 radl8/radl8 diploids are UV resistant. These diploids are as U V resistant as hpr5-Ilhpr5-1 radl8/ observarad18 diploids (F. PALLADINO, unpublished tions). Thus the hpr5-I mutation is recessive in haploids with two copies of the HPRS gene, but dominant in diploids. We have also observed that the HPR51 hpr5A56 radl8/radl8 diploid is UV resistant (F. PALLADINO, unpublished observations), further suggest-
ing that the HPRS gene dosage is important in establishing dominance. The amino acid sequence of the RADH protein shows high homology to the bacterial DNA helicases Rep and UvrD (ABOUSSEKHRA et al. 1989). The location of the hpr5-l allele, a missense mutation in the putative ATP binding domain, suggests that this region is important for activity of the protein and further strengthens the hypothesis that HPR5/RADH/ SRS2 is a DNA helicase. A missense mutation in a similar region of the RAD3 gene,ademonstrated DNA helicase (SUNGet al. 1987), leads to loss of DNA helicase and ATPase activities (SUNGet al. 1988). Itis likely that the hpr5-1 allele results in a loss of helicase activity in the HPR5 protein althoughthis remains to be demonstrated. The mitotic role of the HPR5protein is not essential, based on our analysis of the deletion allele and on the reportedanalysis of the deletionallele of RADH (ABOUSSEKHRA et al. 1989). Double mutant strains containing the hpr5-I allele and various alleles of RAD3 which include rad3-2, reml-1, rad3-Arg48 (demonstrated helicase defective) (SUNGet al. 1988), rad3-20 and rud3-21 (putative helicase defective) (NAUMOVSKI and FRIEDBERG 1986) (data notshown) showed n o effect on cell viability, suggesting thatthere may beadditional DNA helicase activities inyeast. This is not surprising as eight DNA helicase activities in E. coli have been reported (WOODand MATSON 1987, 1989). The observed UV sensitivity of the hpr5-I and hpr5A56 alleles and the interaction with rad18 and rad6 suggest that the HPRS gene has a direct role on the error-prone repairsystem of yeast. In this context the HPRS gene would actbeforethe RAD18 and RAD6 steps, resulting in the epistasis of the hpr5 phenotype of slight UV sensitivity in hpr5 rad18 and hpr5 rad6 double mutants. Arrest at the HPRS mediated step would allow the DNA damage to bechanneled into alternatepathways for repair,thus resulting in only slight UV sensitivity. The hpr5 null mutant differs from the hpr5-1 missense mutant with respect to increased recombination rates. There is a fourfold difference in the levels of hyperrecombination in hpr5-1 mutants as compared to the hpr5A56 deletion mutants. This suggests that although absence of the HPR5 protein can stimulate recombination events,the presence of a HPR5 protein with defective helicase activity provides an additional mechanism by which recombination events are generated. However, both mutants show a similar biased pattern of gene conversion events in the LEU2 duplications. The biased gene conversion in hpr5 strains is not a property of the LEU2 duplication as a similar duplication at the HIS3 locus of HIS3 heteroalleles
+*+Allelism of HPRS with RADH and SRSP
HPRS
hpr5-1
FIGURE5,”Relative frequencies at which each allele is used as the recipient in gene conversion events in HPR5 and hpr5-1 strains. For each genotype, HPRS or hpr5-1, the percentage of conversion events at the L or R copy (see Figure 3) was multiplied by the rate of Leu+ Ura+ conversion events to give a rate of conversion for each allele. These rates were normalized for each genotype to the lowest rate from that genotype, setting this to equal one, resulting in the relative frequencies shown above.The data from duplications 1 , 2, 4 and 5 (Figure 3) were pooled so that the upper left number in the diagrams represents conversions of leu2-k in duplications 1 and 2, the lower left number is conversions of leu.2-k in duplications 4 and 5, the upper right number is conversions of leu2-112 and leu2-r in duplications 4 and 5, and the lower right is conversions of leu2-112 and leu2-r in duplications 1 and 2. Similar relative frequencies were obtained using the data from hpr5A56 strains.
also gives biased gene conversion in a hpr5-1 strain (L. RONG, unpublished observations). Both the position of the allele within a gene and the position of the allele within the LEU2 duplication (centromere proximal or distal) are important in establishing biased gene conversion in hpr5-1 strains. Although we have not fully examined biased gene conversion in hpr5A56 strains it is reasonable to think that this also is the case here. We interpret these results as an indication of the recombination events seen in hpr5 strains initiating at a few fixed sites. Figure 5shows the relative frequency at which each allele is a recipient in the conversion events in HPRS and hpr5-I strains as deduced from the data of Figure 3. We observe that the allele used more often as recipient is the one thatis closest to the LEU2 promoter region. Depending on whether the hpr5-I events initiate by a double strand break in the recipient copy (SZOSTAKet al. 1983) or a nicked or gapped donor copy (MESELSONand RADDING1975), the stronger initiation site would be upstream of the R copy or L copy, respectively. Structural features could make this DNA regionmore available to a defective hypothetical repair complex and may explain preferential initiation of recombination in hpr5 mutants. Amore likely alternative would involve transcription, where the DNA helix must be opened at the transcription start site. T h e effect of transcription in recombination in yeast has been shown to occur in both RNA polymerase I (VOELKEL-MEIMAN, KEIL and and ROEDER1987) and RNApolymerase I1 (THOMAS ROTHSTEIN 1989) transcribed genes. A related observation is that the site for initiation of meiotic recombination in the ARC4 locus cannot be separated from the promoterregion of this gene (NICOLAS et al. 1989). While we have not yet precisely determined the DNA sequences that are the source of the gene conversion bias in hpr5 strains, DNA sequences upstream of the
a3
LEU2 genes, either the pBR322 sequences adjacent to theR copy or theTyl-17 sequence adjacent to the L copy, could provide a site for enhanced initiation of recombination in hpr5 strains. It appears likely that the pBR322 sequences adjacent to the R copy could be critical, especially since gene conversion tracts are short in hpr5 strains. If preferential initiation of recombinationoccursnear the junction between the vector and the R copy and shorter conversion tracts, then those alleles closest to this region, the R copy alleles of duplications 1, 2 and 3, would be preferentially converted. Another difference between the hpr5-I and hpr5A56 mutations is the effect on coconversion events in the invertedrepeat system described by AGUILERAand KLEIN (1989b). The predominantly short conversion events observed in hpr5-1 mutants can be explained as a consequence of theHPR5 defective DNA-helicase blocking the advancement of the repair tract,whereas in hpr5A56 deletion mutants presumably no HPR5 protein participates in the recombination process itself. It is possible that those events seen in hpr5A56 strains proceed by a process that bypasses the HPR5-mediated event, maybe using other DNA helicases, and that this process has the wild-type distribution of coconversion events. T o summarize, we have compared the hpr5-1 and hpr5A56 mutations with regard to three parameters of recombination: (1) recombination rates, (2) biased gene conversion and (3)coconversion events. We have observed that both mutations result in biased gene conversion, butthatrecombinationratesalthough elevated, are not as high in the hpr5A56 mutant as compared to thehpr5-I mutant. Coconversion events are significantly reduced in the hpr5-1 strains, but not in hpr5A56 strains. We andothers have observedthat the HPR5I RADHISRSZ mediated suppression of rad6 and rad18 mutations and hyperrecombination requires the RAD52 repair pathway (AGUILERA and KLEIN 1988; ABOUSSEKHRA et al. 1989; SCHIESTL,PRAKASH and PRAKASH1990).Since the rad52-I mutation blocks mainly gene conversion events inmitosis (JACKSON and FINK 1981), this would be consistent with gene conversion being the major recombination event observed in hpr5-I strains. Alternatively since most of the events are short conversions these would not be associated with crossovers as alreadydemonstrated (AHN and LIVINGSTON 1986; AGUILERA and KLEIN 1989b). These data prompt us to propose a two-step role for the HPR5 proteinin the RAD6 DNA repair pathway as shown in Figure 6. In this model the first step is formation of a protein complex that binds to DNA. In the second step, the DNA helicase and ATPase domains, as indicated by the homology to the Rep and
+
L. Rong et al.
a4
RAD3 REPAIR PATHWAY
HPR5 COMPLEX
',
PRAKASH, R. MALONE, R. ROTHSTEIN and G. FINKfor providing strains and plasmids and F. FABRE for communicating results prior for helpful comments on to publication. We thank E. WOLINSKY the manuscript. This work was supported by National Institutes of Health grant GM30439.
RAD52 RECOMBINATON REPAIR PATHWAY
"-
"
\
..
4-
LITERATURE CITED
.
--
"
V P R 5
+RAD6 REPAIR PATHWAY
FIGURE6,"Model of the possible role of the HPR5 protein in DNA repair. The role of the HPR5 gene productin the RAD6 error prone repair pathway is presented. A two-step is shown to account for the phenotypes of the hpr5-1 and hpr5A56 mutations. Also shown are the steps where the RAD6 repair pathway intermediates can be channeled into the RAD52 recombination repair pathway, leading to the hyperrecombinant phenotype of hpr5 mutants. See DISCUSSION for a moredetailed explanation of the model.
UvrD proteins (ABOUSSEKHRA et al. 1989), would act on the bound DNA. If the HPR5 protein is required to initiate the error-prone repair by binding to DNA (i.e., a mismatch) as a complex with other proteins, then null alleles such as hpr5A56 would abort the RAD4 pathway at an early step. In the absence of the HPR5 protein the RAD3 repair pathway, for example, may eventually become the predominant repairpathway, bypassing the RAD52 pathway. This could explain the low hyperrecombinantphenotype of the deletion mutant. In contrast, the hpr5-1 missense allelewould permit the complex to form, but subsequent repair steps would be blocked by the defective helicase. If such lesions are specifically channeled into the RAD52 recombinationrepair pathway, thena hyperrecombinant phenotype will result. This effect is in fact observed in our experiments. T h e HPR5/ RADH/SRSS carboxyterminal domain seems to be a likely region for protein-protein or protein-DNA interactions (first step in the model of Figure 6) as this region has been implicated in these interactions in the case of the UvrD and Rep proteins in E. coli (CHAO and LOHMAN 1990). The molecular analysis of other hpr5 mutantsshouldprovide further evidencefor different functional domains of the HPR5 gene and its role in repair and recombination. Since the HPR5, RADH and SRS2 genes are identical, we propose to refer to this locus in the future as SRS2. The-hpr5-1 allele will be named srs2-101 and the hpr5A56allele will be named srs2A156. We thank P. MOSKOWITZfor her excellent technical assistance in cloning the HPR5 gene. Y. ZHANG performed the hpr5 reml-1 experiments. We thank F. FABRE, C. LAWRENCE, E. FRIEDBERG, L.
ABOUSSEKHRA, A., R. CHANET,Z. ZGAGA,C. GASSIER-CHAUVAT, M. HEUDE and F. FABRE, 1989RADH, a geneof Saccharomyces cerevisiae encoding a putative DNA helicase involved in DNA repair. Nucleic Acids Res. 17: 721 1-7219. AGUILERA, A., and H. L. KLEIN, 1988 Genetic control of intrachromosomal recombination in Saccharomyces cereuisiae. I. Isolation and genetic characterization of hyperrecombination mutations. Genetics 1 1 9 779-790. AGUILERA, A., andH. L. KLEIN,1989a Genetic and molecular analysis of recombination events in Saccharomyces cerevisiae occurring in the presence of the hyperrecombination mutation h p r l . Genetics 122: 503-517. AGUILERA, A,, and H. L. KLEIN,1989b Yeast intrachromosomal recombination: long gene conversion tracts are preferentially associated with reciprocal exchange and require the RAD1 and RAD3 gene products. Genetics 123: 683-694. AGUILERA,A,, and H. L. KLEIN,1990 H P R I , a novel yeast gene that prevents intrachromosomal excision recombination, shows carboxy-terminal homology to the Saccharomyces cerevisiae TOPI gene. Mol. Cell. Biol. 1 0 1439-1451. AHN,B.-Y., and D. M. LIVINGSTON, 1986 Mitotic gene conversion lengths, coconversion patterns, and the incidence of reciprocal recombination in a Saccharomyces cerevisiae plasmid system. Mol. Cell. Biol. 6 3685-3693. BIGGIN,M. D., J. GIBSON and G. F. HONG,1980 Buffer gradient gels and s5S label as an aid to rapid DNA sequence determination. Proc. Natl. Acad. Sci. USA 8 0 3963-3965. S . FOGELand R. D. KOLODNER, BISHOP, D.K., M. S. WILLIAMSON, 1987 The role of heteroduplex correction in gene conversion in Saccharomyces cerevisiae. Nature 328: 362-364. BORAM,W. R., and H. ROMAN,1976 Recombination in Saccharomyces cerevisiae: a DNA repair mutation associated with elevated mitotic gene conversion. Proc. Natl. Acad. Sci. USA 73: 2828-2832. BORTS,R, H., M. LICHTENand J. E. HABER,1986 Analysis of meiosis-defectivemutations in yeast by physical monitoring of recombination. Genetics 113: 551-567. BURGESS, S.,J. R. COUTOand C. GUTHRIE, 1990 A putative ATP binding protein influences the fidelity of branchpoint recognition in yeast splicing. Cell 60: 705-717. CARLE, G. F., and M. V. OLSON,1984 Separation of chromosomal DNA molecules from yeast by orthogonal-field-alternation gel electrophoresis. Nucleic Acids Res. 12: 5647-5664. CHAO,K., and T . M. LOHMAN, 1990 DNA and nucleotide-induced conformational changes in the Escherichia coli Rep and Helicase I1 (UvrD) proteins. J. Biol. Chem. 265: 1067-1076. CHRISTMAN, M. F., F. S . DIETRICH andG. R. FINK,1988 Mitotic recombination in the rDNA of S. cerevisiae by the combined action of DNA topoisomerases I and 11. Cell 55: 413-425. CLEWELL, D., and D. HELINSKI, 1970 Properties of a supercoiled deoxyribonucleic acid-protein relaxation complex and strand specificity ofthe relaxation event. Biochemistry 9 4428-4440. CRABEEI., M.,R. HYUGEN, K. VERSCHUEREN, F. MESSENGUY, K. 1985 General amino TINEL,R. CUNIN andN. GLANSDORFF, acid control and specific arginine repression in Saccharomyces cerevisiae: physical study of the bifunctional regulatory region of the ARG3 gene. Mol. Cell. Biol. 5: 3139-3148. ENNIS,D. G., S. K. AMUNDSENand G.R. SMITH,1987 Genetic functions promoting homologous recombination in Escherichia coli: a study of inversions in phage X. Genetics 115: 11-24.
Allelism of HPR5 with RADH and SRS2
85
netics 1 1 9 289-301. ESPOSITO,M. S., J. HOSADA, J. GOLIN,H. MOISE, K. BJORNSTADT NAUMOVSKI, L., and E. C. FRIEDBERG, 1986 Analysis of the essenand D. MALEAS,1984 Recombinationin Saccharomyces ceretial and excision repair functions of theRAD3 gene of Sacchavisiae: REC-genemutants and DNA-bindingproteins. Cold romyces- cerevisiae by mutagenesis. Mol. Cell. Biol. 6 1218Spring Harbor Symp. Quant. Biol. 4 9 41-48. FEINBERG,A. P., and B. VOGELSTEIN,1984 Atechniquefor 1227. NICOLAS,A., D. TRECO, N.P. SCHULTES and J. W. SZOSTAK,1989 radiolabeling DNA restriction endonuclease fragments tohigh Aninitiationsiteformeioticgeneconversion in the yeast specific activity. Anal. Biochem. 137: 266-267. Saccharomyces cerevisiae. Nature 338: 35-39. FEINSTEIN, S. O., and K. B. Low, 1986 Hyper-recombining recipROSE, M. D., P. NOVICK,J. H. THOMAS, D. BOTSTEINand G . R. ient strains in bacterial conjugation. Genetics 113: 13-33. FINK, 1987 A Saccharomyces cerevisiae genomic plasmid bank FRIEDBERG, E. C., 1988 Deoxyribonucleic acid repair in the yeast based on acentromere-containingshuttlevector.Gene 60: Saccharomyces cerevisiae. Microbiol. Rev. 52: 70-102. 237-243. GAME,J., L. JOHNSTON andR. VON BORSTEL,1979 Enhanced SANGER, F., S. NICKLEN and A. R. COULSON,1977 DNA sequencmitotic recombination in a ligase-defective mutant of the yeast ingwithchain-terminatinginhibitors.Proc.Natl.Acad. Sci. Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 76: 4589USA 7 4 5463-5467. 4592. SCHIESTL, R. H., S. PRAKASH and L.PRAKASH,1990 T h e SRSP HARTWELL, L. H., and D. SMITH,1985 Altered fidelity of mitotic suppressor of rad6 mutations of Saccharomyces cerevisiae acts by chromosome transmission in cell cycle mutants of S. cerevisiae. channeling DNA lesions into the RAD52 DNA repair pathway. Genetics 110 381-395. Genetics 124: 8 17-83 1 . HAYNES, R. H., andB. A. KUNZ,1981 DNA repair and mutagenSHERMAN, F., G. R. FINK and J. B. HICKS, 1986 Methods in Yeast esis in yeast, pp. 371-414 in The Molecular Biology of the Yeast Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, edited by J. N. Saccharomyces: Li&e Cycle andInheritance, N.Y. STRATHERN,E. W. JONES and J. R. BROACH.ColdSpring SMITH,G. R., 1988 Homologousrecombination in prokaryotes. Harbor Laboratory, Cold Spring Harbor, N.Y. Microbiol. Rev. 52: 1-28. HODGMAN, T. C., 1988 A new superfamily of replicative proteins. SOUTHERN,E. M., 1975 Detectionofspecificsequencesamong Nature 333: 22-23. DNA fragment separated by gel electrophoresis. J. Mol. Biol. ITO, H., Y. FUKUDA,K. MURATA and A. KIMURA, 1983 98: 503-517. SUNG,P., L. PRAKASH,S. W. MATSONand S. PRAKASH,1987 Transformation of intact cells treated with alkali cations. J. RAD3 protein of Saccharomyces cerevisiae is a DNA helicase. Bacteriol. 153: 163-168. Proc. Natl. Acad. Sci. USA 8 4 8951-8955. JACKSON, J. A., and G. R. FINK, 1981 Gene conversion between SUNG, P., D. HIGGINS, L. PRAKASH and S. PRAKASH,1988 duplicated genetic elements in yeast. Nature292: 306-3 1 1. Mutation of lysine-48 to arginine in the yeast RAD3 protein KERN, R., andF. K. ZIMMERMANN, 1978 T h e influence of defects abolishes its ATPase and DNA helicase activities but not the in excision and error prone repair on spontaneous and induced ability to bind ATP. EMBOJ. 7: 3263-3269. mitotic recombination and mutation Saccharomyces in cerevisiae. SZOSTAK, J. W., T . L. ORR-WEAVER, R. ROTHSTEIN J. and F. STAHL, Mol. Gen. Genet. 161: 81-88. 1983 T h e double-strand-break repair model for recombinaKLEIN,H . L., 1988 Different types of recombination events are tion. Cell 33: 25-35. controlled by the RAD1 and RAD52 genes of Saccharomyces TABOR,S., and C. C. RICHARDSON, 1987 DNA sequence analysis cerevisiae. Genetics 1 2 0 367-377. withamodifiedbacteriophage T7 DNApolymerase.Proc. KONRAD,E. B., 1977 Method for the isolation of Escherichia coli Natl. Acad. Sci. USA 8 4 4767-477 1 . mutants with enhanced recombination between chromosome THOMAS, B. J., and R. ROTHSTEIN,1989 Elevated recombination duplications. J. Bacteriol. 1 3 0 167-172. rates in transcriptionally active DNA. Cell5 6 619-630. KOWALSKI, S., and W. LASKOWSKI, 1975 T h e effect of three rad VOELKEL-MEIMAN, K., R. L. KEIL and G. S. ROEDER, 1987 genes on survival, inter- and intragenic mitotic recombination Recombination-stimulating sequences in yeast ribosomal DNA in Saccharomyces. Mol. Gen. Genet. 136: 75-86. correspond to sequences regulating transcription by RNA polymerase I. Cell 48: 1071-1079. KUNZ,B. A,, and R. H. HAYNES,1981 PhenomenologyandgeWALLIS,J. W., G. CHREBET,G. BRODSKY, M. ROLFEand R. ROTHnetic control of mitotic recombination in yeast. Annu. Rev. STEIN, 1989 Ahyperrecombinationmutationin S. cerevisiae Genet. 15: 57-89. identifies a novel eukaryotic topoisomerase.Cell 58: 409-4 19. LAWRENCE, C. W., and R. B. CHRISTENSEN, 1979 Metabolic supWILLIAMSON, M. S., J. C. GAME andS. FOGEL,1985 Meiotic gene pressors of trimethoprim and ultraviolet light sensitivities of conversion mutants in Saccharomyces cerevisiae. I. Isolation and Saccharomyces cermisiae rad6 mutants. J. Bacteriol. 1 3 9 866characterization of pmsl-1 and pmsl-2. Genetics 1 1 0 609876. 646. LEA, D. E.,andC.A. COULSON,1948 T h e distributionofthe WOOD,E. R., and S. W. MATSON,1987 Purification and characnumbersofmutantsinbacterialpopulations. J. Genet. 49: terizationofanewDNA-dependentATPase with helicase 264-284. activity from Escherichia coli. J. Biol. Chem. 262: 15269-1 5276. MALONE,R.E.,andR. E. ESPOSITO, 1981 Recombinationless WOOD, E. R., and S. W. MATSON,1989 T h e molecular cloning of meiosis in Saccharomyces cerevisiae. Mol. Cell. Biol. 1: 891-901. the gene encoding theEscherichia coli 75-kDa helicase and the MESELSON, M., and C.RADDING, 1975 A general model for gedetermination of itsnucleotidesequenceandgeneticmap netic recombination. Proc. Natl. Acad. Sci. USA 72: 358-361. position. J. Biol. Chem. 264: 8297-8303. MESSING,J.. 1983 New M13 vectorsforcloning.Methods EnZEIG,J., V. F. MAPLESand S. R. KUSHNER,1978 Recombination zymol. 101: 20-78. levels of Escherichia coli K12 mutants defective in variousr e p MONTELONE, B. A,, M. F. HOEKSTRA and R. E. MALONE,1988 lication, recombination,or repair genes.J. Bacteriol. 134: 958966. Spontaneous mitotic recombination in yeast: the hyperrecombinational rem1 mutations are alleles of the RAD3 gene. GeCommunicating editor:G . S. ROEDER
