Copyright 0 1993 by the Genetics Society of America

A Novel Mutation in DNA I of Yeast Causes DNA Damage and RAD9-Dependent Cell Cycle Arrest

Nikki A. Levin,*91Mary-Ann Bjornstit and Gerald R. Fink*" *Whitehead Institute for Biomedical Research, Cambridge, Massachusetts 02142, and Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, and +Department $Biochemistry and Molecular Biology, Thomas Jefferson University, Philadelphia, Pennsylvania, 19107-5541 Manuscript received October 9, 1992 Accepted for publication December 10, 1992

ABSTRACT DNA , that alter the superhelicity of DNA, have been implicated in such critical cellular functions as transcription, DNA replication, and recombination. In the yeast Saccha- romyces cerevisiae, a null mutation in the gene encoding topoisomeraseI (TOPl) causes elevated levels of mitotic recombination in the ribosomal DNA (rDNA), but has little effect on growth. We have isolated a missense mutation in TOPl that causes mitotic hyper-recombination notonly in the rDNA, but also at other loci,in addition to causing a number of other unexpectedphenotypes. This topoisomerase I mutation (topl-103) causes slow growth, constitutive expression of DNA damage- inducible genes, and inviability inthe absenceof the double-strand break repair system. Overexpres- sion of topl-103 causes RAD9-dependent cell cycle arrest in GP.We show that the Topl-103 nicks DNA in vitro, suggesting that it damages DNA directly. We propose that Topl-103 mimics the action of wild-type topoisomerase I in the presence of the anti-tumordrug, camptothecin.

YPE I DNA topoisomerases remove superhelical (CHRISTMAN,DIETRICH and FINK 1988). Deletion of T turns from DNA by making a transient single- TOPl in combination with a conditional mutation in stranded break in the DNA backbone and then acting the TOP2 gene, which encodes the yeast type I1 to- as a swivel point about which the strands may rotate poisomerase, results in recombinational excision of [Figure 1; reviewed in CHAMPOUX (1990)l. A major rDNA circles (KIM and WANG 1989).The association role of topoisomerases is believed to be reduction of between topoisomerase mutations and increased re- superhelicity that is generated by such processes as combination levels is also observed for a third yeast transcription and DNA replication (LIU and WANG topoisomerase gene, TOP3, which has been shown to 1987). In fact, a number of studies have implicated elevate mitotic recombination between delta ele- topoisomerases in these important cellular functions ments, which are the terminal repeats of the retro- [reviewed in YANACIDAand STERNCLANZ (1990)l. In transposon Ty (WALLISet al. 1989). the yeast Saccharomyces cereuisiae, topoisomerase mu- In order to explore the role of topoisomerase I in tants have been shown to build up supercoils on highly suppressing recombination, we have isolated new mu- transcribed (BRILLand STERNCLANZ 1988; tant alleles of the yeast TOPl gene.Among these GIAEVERand WANC1988), andtopoisomerase activity mutants, we identified one allele, designated topl-103, has been shown to be required for DNA replication that has a number of unexpected phenotypes. topl- and transcription (GOTOand WANG 1985). 103 causes an extremely high level of mitotic rDNA The yeast TOPl gene is a member of the highly recombination, giving an rDNA recombination rate conserved group of eukaryotic type I topoisomerases approximately 250-fold higher than wild type and 10- (GOTOand WANG1985; THRASHet al. 1985). Despite fold higher than that found in a topl deletion strain. the evidence for involvement of topoisomerase I in topl-103 also elevates recombination atother loci, essential functions such as transcription and DNA unlike previously studied mutations in TOPl or TOP2. replication, TOPl is notrequired for growth, and The topl-103 mutant grows slowly and expresses deletion of this gene results in only a modest growth DNA damage-induciblegenes constitutively. The defect. Deletion of TOPI does, however, cause a large topl-103 mutantrequires an intact double-strand increase in mitotic recombination in the tandemly break repair system for viability. In addition, overex- repeated ribosomal DNA arrays, a chromosomal do- pression of topl-103 causes cell cycle arrest in GS that main that is normally suppressed for recombination is dependent on the RAD9 gene product, a sensor of DNA damage. In the absence of RAD9, cells overex- ' Present address: Division of Molecular Cytometry, Department of Lab- pressing topl-103 traverse the cell cycle without re- oratory Medicine, MCEi-230, University of California, San Francisco, Cali- fornia94143. pairing DNA damage, resulting in rapid cell death. To whom correspondence should be addressed. We show here that Topl-103 enzyme nicks DNA

Genetics 133: 799-814 (April, 1993) 800 N. A. Levin, M.-A. Bjornsti and G. R. Fink

A FIGURE 1.-Model of eukaryotic topoisomerase I mechanism. (A) Topoisomerase I binds to super- coiledDNA with high affinity in a noncovalent interaction. The enzyme makes a single-stranded nick in DNA, simultaneously becoming covalently bound to the 3‘ end of the broken DNA strand. This topoisomerase-DNA intermediate is called the “cleavable complex.” The enzyme provides a Cleavable complex swivel about which the broken DNA strand may rotate to relieve superhelical tension. The enzyme then religates the broken DNA strands, releasing B itself from the DNA. After multiple rounds of nicking and resealing, the supercoiled DNA is con- verted to the completely relaxed form. (B) The anti-tumor drug camptothecin (CPT) inhibits the CPT religation step of the reaction, resulting in accu- + mulation of the cleavable complex. (After LW Cleavable complex 1989.) in vitro, suggesting that it can damage DNA directly. personal communication). pGAL::lacZ (pLR1 A23) is a 2-pm We propose that Topl-103mimics the actionof wild- containing URA3 and GALl-promoted lac2 (WEST, YOCUM and PTASHNE1984). pWE3 is a CEN-ARS plasmid type topoisomeraseI in the presence the anti-tumor of carrying URA3 and GALI-promoted TOPI(ENG et at. 1988). drug, camptothecin. This drugcauses topoisomerase pNKY3Ois an integrative plasmid for creating a I to become trapped during the relaxation reaction as rad5O::URA3 disruption (E. ALANIand N. KLECKNER,per- part of a covalent enzyme-DNA complex [Figure 1; sonal communication).pRR330 is an integrative plasmid for reviewed in Liu (1989,1990)l. The enzyme-bound creating a rad9::URA3 disruption (SCHIESTLet al. 1989). CB17 consists of the 3.8-kb HindIIl fragment of TOPl nicks generated camptothecin result in hyper-re- by filled in with Klenow and cloned into the SmaI site of YEp24 combination, induction of DNA damage responses, (M. F. CHRISTMAN,personal communication). CB32 and and cell cycle arrest, in close parallel to the effects of CB39 are integrative vectors carrying rDNA sequences topi-103. marked by URA3 and HIS3, respectively (CHRISTMAN,DIE- TRICH and FINK1988). CB45 carriesADE2 on a 6-kbBamH1 fragment cloned into the BamHI site of pBR322 and the MATERIALS AND METHODS 6.4-kb HindIII fragment of rDNA cloned into the HindIII site. CB51 and CB52 contain the 3.8-kb HindIIl fragment Media, strains and genetic techniques: Strains are de- of TOPI and of topl-727, respectively,cloned into the scribed Table l. YPD, SD, SC and 5-FOA media are as in HindIII site of YCp405. CB107 is the same as pWE3, but described in GUTHRIEand FINK(1 99 1). “Sectoring plates” with topl-727 replacing TOPl. are SD supplemented with 75 p~ adenine hemisulfate, 75 pNL2 contains the 3.2-kb HindIII-NruI fragment of p~ histidine, 400 p~ tryptophan and 1.7 mM leucine. The levelsof adenine and histidine are lower than those in TOPIcloned into the 7.0-kb HindIII-NruI fragment of standard media in order toachieve darker red coloration of YCp50. pNL2-102, pNL2-103 and pNL2-106 are pNL2 ade2 mutants. Yeast matings, tetrad dissection, lithium ace- containing the mutant alleles topl-102, topl-103 and topl- tate transformation and replica-plating are as in GUTHRIE 106, respectively, in place of TOPI. pNL8 is an integrative and FINK(1 99 1). vector consisting of the 4.3-kb HindIII-ApaI fragment of Plasmids: YCp50is a CEN-ARS vector marked with URA3 pNL2-103 cloned into the 3.5-kb HindIII-ApaI fragment (ROSE et al. 1987). YCp405 is a CEN-ARS vector marked of YIp5.pNL14 is a CEN-ARS plasmid consistingof the 2.6- with LYS2 (MA et al. 1987). YEp24 is a 2-pm vector marked kb MluI-NsiI fragment of topl-I03 from pNL2-103 cloned with URA3 (BOTSTEINet al. 1979). YIp5 is an integrative into the 12.8-kb MluI-NsiI fragment of CB51. pNL30 is a vector carrying URA3 (STRUHLet al. 1979). YIp357r is an CEN-ARS plasmid containing the 2.3-kb Bsu36I fragment integrative vector carrying the URA3, amp’ and lac2 genes of pNL14 cloned into the9.1-kbBsu36I fragment of pWE3 (MYERSet al. 1986). pRS303 and pRS305 are integrative to generate CALI-promoted topl-103. pNL46 is an integra- vectors carrying HIS3 and LEU2, respectively (SIKORSKY tive plasmid marked with LEU2 for fusing the GALl pro- and HIETER1989). moter to alleles of TOPI, consisting of the 2.0-kb EagI- pCT80ATOP is an integrative plasmid used for creating EcoRV fragment ofpNL3O cloned into the 5.5-kb EagI- a topl-7::LEU2 disruption (THRASHet al. 1985). pSR 16 and SmaI fragment of pRS305. pNL48 is an integrative plasmid pSR18 are integrative plasmids marked with URA3 for gen- for marking the rDNA withLYS2, consisting of a 2.8-kb erating lac2 fusions to DIN1 and DIN3, respectively (RUBY EcoRI fragment ofrDNA cloned into the EcoRI siteof and SZOSTAK1985). pR12Sclig2 is an ARS plasmid carrying pUC9 carrying LYS2. pNL5 1 is a CEN-ARS plasmid carrying URA3 and CDC9 (BARKER and JOHNSTON 1983). URA3 and topl-103,727 made by cloning the 865-bp Sphl YEpl3(RAD52) is a 2-pmplasmid carrying LEU2 and fragment of CB52 into the 9.3-kb SphI fragment of pNL2- RAD52 (SCHILDet al. 1983). pSM2 1 is an integrative plasmid 103. pNL52 is a CEN-ARS plasmid carrying URA3 and for generating the rad52::TRPI disruption, consisting of the GALI-promoted topl-l03,727 made by cloning the 865-bp 2-kb BamHIfragment of RAD52 cloned into theBamHI site SphI fragment of CB52 into the 10.6-kb SphI fragment of of pBR322 with an insertion of the BamHI-BglI1 fragment pNL3O. pNL53 is an integrative plasmid marked with HIS3 of TRPl cloned in at the BglII site of RAD52 (D. SCHILD, for fusing the GALl promoter to alleles of TOPI, consisting Topoisomerase I Mutant Causes DNA Damage 80 1

TABLE 1 Yeast strains

Strain Genotype Source FDY65 MATa leu2-3 ura3-52 CUPl::URA3 CHRISTMAN,DIETRICH and FINK(1988) topl-7::LEU2 (pCT80ATOP) CY 126 MATa leu2-3 ura3-52 hisl::URA3::hisl CHRISTMAN,DIETRICH and FINK (1988) toP1-7::LEU2 (pCT8OATOP) CY 184 MATa ade2-1 ura3-1 his3-11 trpl-1 leu2-3,112 canl-100 M. F. CHRISTMAN rDNA::ADE2 (CB45) CY185 CY 184 transformed to topl-7::LEU2 (pCT8OATOP) M.F. CHRISTMAN NLY 107 CY 185 carrying pNL2 This work NLY 108 CY 185 carrying YCp50 This work NLY143 CY185 carrying pNL2-103 This work NLY 163 CY 185 carrying pNL2-102 This work NLY 164 CY185 carrying pNL2-103 This work NLY 167 CY 185 carrying pNL2-106 This work NLY258 MATa leu2-3 ura3-52 lys2 rDNA::URA3 (CB32) topl-7::LEU2, This work carrying YCp405 NLY260 MATa leu2-3 ura3-52 lys2 rDNA::URA3 (CB32) topl-7::LEU2, This work carrying CB5 1 NLY336 MATa leu2-3 ura3-52 lys2 rDNA::URA3 (CB32) topI-7::LEU2, This work carrying pNL14 NLY360 MATa leu2-3 ura3-52 lys2 rDNA::URA3 (CB32), carrying This work pNL14 NLY363 CY 184 carrying pNL2-103 This work NLY373,374 CY 126 lys2 carrying YCp405 This work NLY375 CY 126 lys2 carrying CB5 1 This work NLY377.378 CY 126 lys2 carrying pNLl4 This work NLY379,380 FDY65 lys2 carrying YCp405 This work NLY381,382 FDY65 lys2 carrying CB51 This work NLY383 FDY65 lys2 carrying pNL14 This work NLY429 CY185 carrying pWE3 This work NLY 494 CY 185 carrying pGAL-lacZ This work NLY5 15 CY184 with TOP1 replaced by topl-103, using pNL8 This work NLY5 16 NLY515 switched to MATa This work NLY566 NLY5 15 carrying pNL2 This work NLY6 18 NLY515 carrying CB17 This work NLY620 CY 185 carrying pNL3O This work NLY638 NLY515 carrying YEp13(RAD52) This work NLY640 NLY515 carrying pR12Sclig2 This work NLY642 NLY515 transformed to rDNA::HIS3, using CB39 This work NLY650 MATa ade2-1 leu2-3,112 trpl-1lys2 his3 ura3 rad52::TRPI This work topl-7::LEU2 (pCT80ATOP) rDNA::ADE (CB45) NLY652 MATa ade2-1 trpl-1 leu2-3,112 his3 ura3lys2 top1-7::LEU2 This work (pCT8OATOP) rad52::TRPl (PSM21) rDNA::ADE2 (CB45) NLY678 CY184 transformed to pGAL1-TOP1 using pNL46 This work NLY681 NLY515 transformed to pGAL1-topl-103 using pNL46 This work NLY730,731 MATa ura3 ade2lys2 trpl his3 topl-103 rDNA::ADE2(CB45) This work rDNA::LYS2 (pNL48) rDNA::HIS3 (CB39) carry- ing YEp24 NLY732,733 Same as NLY730, 731, but carrying CB17 This work NLY737 NLY515 transformed to pGAL-topl-103 (HIS3)with This work pNL53 NLY755 MATa leu2 trplhis3 ade2 ura3 rad50::URA3 (pNKY3O) This work topl-7::LEU2 (pCT8OATOP) NLY814 CY 185 carrying pNL5 1 This work NLY816 CY185 carrying pNL52 This work NLY880 CY185 carrying CB107 This work NLY884 NLY681 transformed to rad9::URA3 with pRR330 This work NLY887 NLY681 transformed to DIN1::lacZ (URA3)with pSRl6 This work NLY89 1 MATa ura3 ade2 his3 trpl leu2DIN1::lacZ (URA3, pSRl6) This work pGAL-topl-103 (HIS3,pNL53) rDNA::ADE2 NLY894,895 NLY681 transformed to DIN3::lacZ (URA3) with pSRl8 This work NLY896,897 NLY737 transformed to DIN3::lacZ (URA3)with pSR18 This work 802 N. A. Levin, M.-A. BjornstiFink R. and G. of the 2.0-kb XhoI-Sac1 fragment of pNL46 cloned into the calculated by taking the average of sixindependent colonies. 4.4-kb Sad-XhoI fragment of pRS303. The frequencies of loss of the rDNA::URA3, rDNA::HIS3 Isolation of new alleles of TOPI: The TOPl gene on and rDNA::LYS2 markers were determined in an analogous plasmid pNL2 was mutagenized in vitro using hydroxyl- manner to rDNA::ADE2. amine HCI, by the method of NONETet al. (1987). The The rate of rDNA recombination was determined by mutagenized plasmid was transformed into , growing strains to be tested on SC - adenine medium to selecting ampicillin resistance, and approximately 1O6 colo- select for the rDNA::ADE2 insertion. Colonies were picked nies were pooled and grown to make plasmid DNA. An into sterile water, serially diluted, and plated for approxi- aliquot of the mutagenized plasmid was transformed into E. mately 200 colonies per plate on YPD (rich) medium. The coli strain DB6507, a pyrF mutant that is complemented by colonies on YPD plates were replica-plated to SC - adenine the wild-type yeast URA3 gene. By testing approximately to detect half-sectored colonies, which represent recombi- 300 colonies for the ability togrow without uracil, the nation events resulting in loss of ADEB from the rDNA in frequency of mutation of the URA3 gene was determined to the first cell division. The rate of rDNA recombination was be 2.2%, indicating that the mutagenesis procedure was calculated as the number of half-sectored colonies divided successful. by the total number of colonies, excluding totally red colo- Mutagenized pNL2 was transformed into the yeast topl- nies, which presumably had lost ADEB fromthe rDNA 7::LEU2 strain CY 185 (Table l),selecting for Ura+.Individ- before plating. The average rate for each strain was calcu- ual transformants were picked to 96-well microtiter dishes lated from measurements on six independent colonies. and grown overnight in synthetic complete medium lacking Frequency of recombination at HIS4::URA3::HIS4 and uracil at 30". The cells were collected by centrifugation, CUPl::URA3: Strains to be tested were streaked for single digested with Zymolyase (1 mg/ml final concentration), and colonies on medium lacking uracil to select for the URA3 lysed in the same microtiter dishes. Extracts were tested for insertion. Colonies were picked to liquid medium containing topoisomerase I activity by the method of THRASHet al. uracil (to allow URA3 to be lost during growth) in microtiter (1984, 1985), except that 4 pl of extract were added to 10 dishes. Strains were grown to saturation and serially diluted pl of assay buffer. Assays were at37" for 1 hr, using in sterile water. Dilutions were plated on medium containing approximately 200 ng perreaction of YIp357r asthe super- 5-fluoro-orotic acid (5-FOA) to determine the number of coiled substrate plasmid. Reactions were stopped with so- Ura- cells [5-FOA selects against Ura+ cells (BOEKEet al. dium dodecyl sulfate (SDS) and electrophoresed on an 0.7% 1987)] and on medium without 5-FOA to determine the agarose gel without ethidium bromide. The gel was stained total number of cells. The frequency of recombination for in ethidium bromide and photographed. All isolates that each strain was calculated by taking the average of six appeared defective in the initial screen were colony-purified independent colonies. and retested. Plasmid DNA was rescued from each mutant Determination of synthetic lethality of topl-103 with into E. coli (GUTHRIEand FINK 1991). The topl mutant DNA repair mutants: We assayed for a genetic interaction plasmids were restriction-mapped and retransformed into between topl-103 and a number of DNA repair mutants by CY185 to confirm the topoisomerase I activity defect. Of crossing a topl-103 haploid strain by a haploid strain of the 324transformants screened for lackof topoisomerase I opposite mating type carryinga marked deletion of the activity, three topl mutants were isolated. An additional 49 DNA repair gene to be tested, whichwe will refer to as mutants were isolated from 14,000 transformants screened RADX. The diploid generated by this cross was therefore by a colony-sectoring assay. heterozygous both at the TOPl locus and at RADX. When The topl-103 mutation was shown to reside in the plas- this diploid was sporulated, the alleles of TOPl and RADX, mid-borne TOPl gene, since transformation with plasmid which are unlinked genes in all cases tested, should have DNA (pNL2-103) isolated fromthe topl-103 mutant assorted independently,generating equal numbers ofall (NLY143) conferred on the parent strain all the same phe- four possible combinations of TOPl and RADX alleles. If notypes as the original mutant. In addition, the topl-103 topl-103 were lethal in combination with a deletion of allele was cloned onto an integrating plasmid (pNL8) and RADX, however, the double-mutant spore clones would not used to replace the wild-type TOPl gene in the chromosome. grow, and this combination ofalleles would not be re- This construct was confirmed by Southern hybridization. covered. We scored ascospore phenotypes by printing tetrad The topl-103 mutation in the chromosome conferred all the dissection plates to selective media in order to assay for the same phenotypes as the topl-103 allele on a plasmid. presence of the the marked deletion alleles. Molecular biology techniques: Plasmids were con- &Galactosidaseassays: The DlN1::lacZ (pSR16) and structed by standard methods (MANIATIS,FRITSCH and SAM- DIN3::lacZ (pSR18) fusions (see above) were integrated at BROOK 1982) and aredescribed above. Restriction enzymes DIN1 and DlN3, respectively, by transforming strains car- were purchased from New England Biolabs. Sequencing was rying a pCAL-topl-103 plasmid (pNL46), and selecting for performedon double-stranded DNA (plasmids pNL2, Ura+ transformants. Single colonies were grown overnight CB51, pNL14, pNL2-103)using the U.S. BiochemicalCorp. at 30"in 10 ml minimal synthetic raffinose medium to select Sequenase kit. Southern blotting was performed according for theDIN::lacZ fusion. The cultures were split in two, and to MANIATIS, FRITSCHand SAMBROOK(1982), and labeled galactose was added to one half to 2% final concentration, probes were prepared using the Prime Time random hex- in order to induce pGAL-topl-103. Glucose was added to amer kit of International Biotechnologies, Inc. theother half to 2% final concentration as a negative Assays of rDNA recombination rate and frequency: The control. Cells were grown for 24 hr, collected by centrifu- frequency of loss of the rDNA::ADE2 marker was measured gation, resuspended in 0.4 ml breaking buffer [0.1 M Tris, by picking colonies from medium lacking adenine (to select pH 8.0, 20% glycerol, 1 mM dithiothreitol (DTT), 1 mM for rDNA::ADE2) and streaking themon medium containing AEBSF protease inhibitor (CalBiochem)], and broken by adenine, to allow ADE2 to be lost from the rDNA during vortexing with 300 mg acid-washed glass beads for 2 min at growth. Colonies were scraped off the plates containing full speed. Cell debris and glass beads were removed by adenine, serially diluted in sterile water, and plated on centrifugation. Extracts were assayed as in MILLER(1972). medium containing adenine. The resulting colonies were @-Galactosidaseunits [nanomoles o-nitrophenyl @-D-galacto- replica-plated to medium lacking adenine to determine the pyranoside (ONPG) cleaved per min per mg protein] were frequency of Ade- cells. The frequency for each strain was calculated using the following formula: (A4eo)(assayvolume)/ Topoisomerase I DamageMutantDNA Causes 803 (0.0045)(protein concentration)(assay time)(extract vol- serum albumin (BSA). Reactions were performed in 34 pl ume), and protein concentration was determined using the total volume, containing 4 ng negatively supercoiled Bio-Rad dye reagent. Isogenic strains not carrying any lacZ pUC 1 18DNA and 6 pl diluted partially purified topoisom- fusion were assayed in parallel to determine the baseline erase I. Camptothecin (Sigma) in dimethylsulfoxide (DMSO) units for each strain, and these baseline values were sub- was added to reactions to a final concentration of 0.1 1 mg/ tracted from the values of strains with the DZN::lacZ fusion. ml or DMSO alone was added to 11% final concentration Quantitation of cell cycle stages: Strains were grown as a control. Reactions were at 37" for 10 min. Reactions overnight in synthetic raffinose medium lacking leucine were stopped by adding SDS to 0.5% final concentration, (NLY681 and NLY884) or uracil (NLY429, NLY494, followed by digestion with proteinase K (0.15 mg/ml final NLY620) to an OD600 of about 0.1. Galactose was added to concentration) for one hour at 37", except as noted. Reac- a final concentration of 2%, and cultures were grown for tion products were electrophoresed on 0.9% agarose gels 48 hr (NLY429, NLY494 and NLY620) or 29 hr (NLY681 containing 50 ng/ml ethidium bromide in order to separate and NLY884). An aliquot of each culture was sonicated to supercoiled, relaxed and nicked topoisomers. The gels were disperse aggregated cells, and3 PI of each culture was blotted to Biotrans nylon membranes (ICN) and visualized observed using a hemacytometer. At least 300 cells of each by hybridization to hexamer-labeled pUCll8, followed by culture were scored as unbudded, small-budded (bud di- autoradiography. Quantitation of radioactivity was per- ameter less than 1/2 that of mother), or large-budded (bud formed using a Fujix BAS-2000 phosphorimager. diameter greater than or equal to 1/2 that of mother). Purification of topoisomerase I from yeast: Topoisom- RESULTS erase I was purified from cells overexpressing TOPl (NLY429) or topl-103(NLY620) from the GAL1 promoter Isolation of new alleles of yeast TOPI: We ob- on a CEN, URA3 vector (pWE3 and pNL30, respectively), tained new topl mutations by randomly mutagenizing and a mock purification was done from the isogenic strain without any TOPl construct (NLYl08). The method of the cloned TOPI gene on a centromere-basedplasmid purification will be described in greater detail elsewhere (A. (pNL2) and transforming the library of mutagenized KNABand M.-A. BJORNSTI,unpublished results). Cells were plasmid DNA into a strain(CY 185) carryingthe topl- grown overnight at 30" in synthetic medium lacking uracil, 7::LEU2 deletion allele. The transformants were diluted 1: 100 into 300 ml synthetic raffinose medium lack- screenedfor a lackof topoisomerase activity by ing uracil, and grown to an OD600of approximately 1.0. I Galactose was added to 2%final concentration and the cells growingrandom transformants in liquid culture in were grown for approximately 6 hr. Cells were collected by microtiter dishes, making crude extracts fromeach of centrifugation, washed in sterile water, resuspended in 2 ml the cultures, and testing themfor activity in an in vitro of TEEG (Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA, 10% assay (THRASHet al. 1984, 1985). In this assay, the glycerol) per gram wet weight, supplemented with 0.3 M substrate, negatively supercoiled plasmid DNA, and KC1 and standard protease inhibitors, and stored frozen at -70". Cells were broken by vortexing withacid-washed theproduct, relaxed covalently closed circles, are glass beads for 15 min on a fixed-head vortexer. Cell break- easily distinguished because of their different mobili- age was monitored by determining the protein concentra- ties on an agarose gel. Extracts from a strain express- tion of the extracts. Cell debris and glass beads were re- ing wild-type TOPI (NLY 107) fully convert the super- moved by centrifugation. coiled substrate plasmid to the relaxed form, whereas Cell lysates were fractionated by ammonium sulfate pre- cipitations, collecting the material that was soluble at 30%, extracts from theisogenic topl-7::LEU2 deletion strain but insoluble at 70% saturation and resuspending it in (NLY108) fail to convert the substrate plasmid to the TEEG buffer with protease inhibitors but without salt. The relaxed form (Figure 2). Transformants receiving a conductivity was adjusted to that of TEEG + 0.2 M KCI. plasmid bearing a mutant allele of TOP2 would, like This material was loaded on a 2-ml bed volume phospho- topl-7::LEU2, beexpected to lack topoisomerase I cellulose (Whatman) column that had been equilibrated with TEEG + 0.2 M KC1 in a Bio-Rad 10-ml capacity Poly Prep activity. Topoisomerase I assays on several new topl column. The column was washed with five volumes TEEG mutants are shown in Figure 2. + 0.2 M KCI, and fractions were eluted intwo column We tested the new top1 mutants for their ribosomal volume steps using TEEG containing increasing concentra- DNA recombination levels. CHRISTMAN,DIETRICH tions of KC1 (0.4, 0.6, 0.8 and 1.O M KCI). Topoisomerase I and FINK(1 988) showed that yeast strains with a null activity eluted at 0.6 M KCI. Column fractions were stored at -20" in topoisomerase (deletion) allele of TOPl have elevated mitotic recom- I storage buffer (0.3 M KCI, Tris, pH 7.4, 1 mM EDTA, 1 bination in the rDNA arrays. The majority of new mM EGTA, 50% glycerol). The protein concentration of top1 mutants also had elevated rDNA recombination each fraction was determined by Bio-Radassay, and the levels that were indistinguishable from a strain carry- quantity of topoisomerase I enzyme in a given fraction was ing the topl-7::LEU2 deletion (NLY 108). One of the determined by electrophoresis on a polyacrylamide gel and visualization of the proteins by silver staining. The identity new topl mutants (NLY 164), however, had an unusu- of topoisomerase I, a 90-kD protein, was confirmed by ally highfrequency of rDNArecombination com- Western blotting of a gel run in parallel to the silver stained pared to the topl-7::LEU2 deletion strain, as judged gel, using polyclonal anti-yeast topoisomerase I antiserum by the colony sectoring assay (Figure 3). In this assay, (M.-A. Bjornsti, unpublished). the numberof red sectors in a colony reflects the level Nicking assays: The ability of wild-type andmutant topoisomerase I to nick a supercoiled plasmid substrate was of rDNA recombination. This unusual mutant, des- determined asin HSIANCet al. (1985), except thatthe ignated topl-103, was chosen for further study. reaction buffer was 40 mM Tris, pH 7.6, 50 mM KCI, 0.5 topi-I03 causes hyper-recombinationin the rDNA mM DTT, 15 mM EDTA, and 30 pg/ml acetylated bovine arrays and at other loci: Quantitation of the rDNA 804 N. A. Levin, M.-A. Bjornsti and G. R. Fink 103 is semidominant. topl-103 elevates rDNA recom- bination levels even in a strain expressing wild-type TOPI,although the frequency of rDNA recombina- tion is not as high as in a strain expressing topl-103 alone. topl-103 on a low-copy plasmid (pNL14) ele- vated rDNA recombination to 48-fold above the wild- type level when transformedinto a TOPI strain (NLY360),but to 338-fold above wild type when transformedinto a topl-7::LEU2 strain (NLY336, Table 2B). Increasing the dosage of TOPI in a topl- 103 mutant(NLY730 and 731) by transformation with a high-copy TOPI plasmid (CB17) resulted in a FIGURE2.-Topoisomerase I assays of extracts from wild-type greater reduction of rDNA recombination frequency and topl mutant strains. Extracts from isogenic wild-type and fop1 (28-40-fold) than when a single copy of TOPI was mutant strainswere incubated with a supercoiled plasmid substrate present (NLY732 and 733; Table 2C; also, compare (pUCI 18) for 1 hr at 37" in topoisomerase I assay buffer (THRASH Figure 3, C and D). In contrast to the reduction in et al. 1984, 1985), and the reaction products were separated by electrophoresis on an 0.7%agarose gel. The DNA was visualized rDNA recombination levels observed when the topl- by staining in ethidium bromide. See Table 1 for complete geno- 103 mutant was transformed withhigh copy TOPI types. (CB17), introducing the gene for DNA ligase (CDC9, NLY640) or the DNA repair gene RAD52 (NLY638) recombination effect using the rDNA::ADE2 marker on high-copy plasmids had no effect on the rDNA showed that a topl-7::LEU2 strain bearing topl-103 recombination level topl-I03 on a plasmid (NLY 164) had an rDNA recombination of a mutant (data not frequency 108-fold higher than the same strain car- shown). rying wild-type TOPI (NLY 107). This strain carrying Deletion of TOPI elevates mitotic recombination in the vector alone (NLYl08) had an rDNA recombi- the rDNA buthas little or noeffect on recombination nation frequency 19-fold above wild type (Table 2A). outside the rDNA arrays (CHRISTMAN,DIETRICH and Replacement of the wild-type TOPI gene (CY184) FINK 1988). We tested topl-103 for an effect on with topl-I03 (NLY642) by integrative transforma- recombination at two other loci, by measuring the tion also resulted in a large (164-fold) stimulation of frequency of loss of the URA3 gene from a tandem rDNArecombination frequency, in contrast to re- duplication of HIS4 or from the CUPl locus, a tandem placement of TOPI with topl-7::LEU2 (CY 185), which array of approximately 10 copies of the metallothi- elevatedrDNA recombination 23-fold (Table 2A). onein gene (FOGELand WELCH1982). Surprisingly, The stimulation of rDNA recombination by topl-103 the topl-103 allele increases the recombination fre- did not depend on using the rDNA::ADE2 marker, as quency at HIS4 approximately 24-fold (NLY377 and topl-I03 (NLY336) also increased the frequency of NLY378) over the isogenic wild-type strain loss of an rDNA::URA3 insertion to levels 338-fold (NLY375), whereas the topl-7::LEU2 deletion higher than wild-type (NLY260; Table 2B), and of (NLY373and NLY374) has only a twofold effect HIS3 and LYS2 insertions (NLY730, NLY731) in the (Table 4A). topl-103 also stimulates recombination at rDNA arrays (Table 2C). the CUPl locus (Table 4B). A topl-103 mutant Consistent with the elevated frequency of rDNA (NLY383) loses the CUPI::URA3 marker 11 times recombination in topl-103, the rate of rDNA recom- morefrequently than wild type (NLY381and bination was also significantly elevated in the topl-103 NLY382), whereas the topl-7::LEU2 deletion mutant. Whereas the frequency of rDNA recombi- (NLY379 and NLY380) has little or no effect on nation events is determined by measuring the fraction recombination at CUPl. of cells in a colony that have lost the rDNA marker, topl-103 strains have a growth defect:Although a the rate of rDNA recombination is expressed as loss deletion of TOPI causes only a slight growth defect, of the marker percell per generation (see MATERIALS the topl-I03 mutation causes a severe growth defect AND METHODS). topl-103 (NLY515) lost ADE2 from that is readily observed when topl-I03 is streaked for the rDNA ata rate 250 times greater than wild-type single colonies on solid medium (Figure 4). A TOPI (CY 184), whereas the topl-7::LEU2 strain (CY 185) strain (NLY 107) forms largecolonies on solid medium had a rate of loss of ADE2 that was 25 times greater and anisogenic topl-7::LEU2 mutant (NLY 108) forms than wild type (Table 3). As the effects of topl-103 on onlyslightly smaller colonies, whereas the isogenic the frequency and on the rate of recombination are topl-103 mutant (NLY164) grows very slowly, form- consistent, only the frequency was measured in sub- ing extremelysmall colonies. The slow growth of topl- sequent experiments. 103 is recessive to wild type. A single copy of TOPI The rDNA hyper-recombination phenotype of topl- on a centromere vector (NLY566) restores wild-type I

FIGURE3.-A colony sectoring assay for rDNA recombination. rDNA recombination levels in isogenic wild-type and fop1 mutant strains were measured using a colony sectoringassay. In this assay, the ADE2 gene is integrated into therDNA array of an nde2 mutant strain. Wild- type yeast cells form white colonies, whereas ade2 mutants form red colonies. Each red sector results from an rDNA recombination event in which the ADE2 insertion was lost. The number of red sectors in a colonyreflects the frequency of rDNA recombination events. The colonies shown here were grown on "sectoring medium" (see MATERIAIS AND METHODS).(A) NLY107 (TOPf).(B) NLY108 (fopI-7::LEU2).(C) NLY164 (fopl-103).(D) NLY618 [fopI-I03carrying CB17 (high copy TOPI)].See Table 1 for complete genotypes. growth to the topl-103 mutant as well as TOPI on a plasmid (pNL14) or the vector alone (YCp405). Al- high-copy plasmid (NLY618; Figure 4). though the vector control gave several hundred trans- topl-lo3 mutants require an intact double-strand formants, no topl-I03 transformants were recovered. break repair system: topl-I03 is inviable in combina- Likewise, transformation of a topl-103 strain tion with a deletion allele of the double-strand break (NLY515) with 1.2 pg of rad52::LEU2 disrupting repair gene RAD52. When a topl-103 RAD52 strain plasmid pSM2l was unsuccessful, whereas the same (NLY516) is crossed by a topl-7::LEU2 rad52::TRPI amount of this plasmid gave 14 Leu+ transformants strain (NLY650), a pattern of spore lethality is ob- of the isogenic wild-type strain (CY 184). served. Out of 29tetrads, no viable topl-103 topl-I03 is also lethal in combination with a deletion rad52::TRPI (Leu- Trp+) spores were recovered, allele of RAD50 (radSO::URA3), another gene in the whereas all other combinations of Leu and Trp phe- double-strand break repair pathway. Wecrossed a notypes were recovered. Since TOPI and RAD52 are topl-IO3 RAD50 strain (NLY515) by a topl-7::LEU2 unlinked, one would expect one quarterof the spores rad50::URA3 strain (NLY755)and sporulated thedip- to be Leu- Trp+. Further evidence for the synthetic loid. Out of 17 tetrads examined, noLeu- Ura+ (topl- lethality of topl-103 and rad52::TRPI comes from 103 rad50::URA3) spores were recovered. The other transformation experiments in which a rad52::TRPI combinations of markers were recovered at the ex- topl-7::LEU2 strain (NLY652) was transformed with pected frequencies. several micrograms (pg) of either a topl-103 bearing The synthetic lethality of topl-I03 with 806 N. A. Levin, M.-A. Bjornsti and G. R. Fink

TABLE 2 rDNA recombination frequenciesin isogenic wild-type, topl-7::LEUZ, and topl-IO3 strains

A. rDNA::ADE2

allele in allele Frequency of Frequency relative Strain TOPI TOPI loss of rDNA chromosome to on plasmid marker (%) TOPI+

NLY 107 topl-7::LEU2 TOPI 0.61 1 NLY 108 topl-7::LEU2 Vector 11.5 19 NLY 164 topl-7::LEU2 t0P1-I03 66.1 108 NLY363 TOPI topi-I03 27.2 45

CY184 TOPI - 0.51 1 CY 185 topl-7::LEUZ - 11.6 23 NLY642 topi-I03 - 83.6 164 B. rDNA::CJR43

allele in Frequency of Strain TOPI TOPI allele Frequency relative chromosome on plasmid loss of rDNA to marker (W) TOPI+ NLY260 TOPItopl-7::LEU2 0.24 1 NLY258 topl-7::LEU2 Vector 5.0 21 NLY336 top1-103topl-7::LEUZ 338 81.0 NLY360 TOPI 48 top1-103 11.6 C. rDNA::HIS3, rDNA::LYS2

Frequency of allele in allele Frequency of loss Strain TOPI TOPI loss of HIS3 chromosome of LYS2 (76) on plasmid (%) NLY730,731 topi-103 Vector (YEp24) 36.2 39.3 NLY732,733 top1-103 TOPl (CB17) 0.9 1 .o Ratio of vector to TOPI+: 40 28

The frequency of rDNA recombination was determined as described in MATERIALS AND METHODS. Each value is the average of six independent measurements. All strains are isogenic and their complete genotypes appear in Table 1.

TABLE 3 control of the GALl promoter. DINl and DIN3 are rDNA recombination rates in isogenic wild-type, topl-7::LEU2 transcriptionally induced by DNA damaging agents and topl-103 strains such as UV light or MMS, although the signal directly responsible for induction is not known (RUBY, SZOS- Rate of loss TOPI allele in of Rate relative TAK and MURRAY 1983; RUBY and SZOSTAK1985). Strain chromosome rDNA::ADE2 to TOPI+ We assayed the expression of the integratedDIN::lacZ CY184 TOPl 4.8 X 10-4 1 fusions by measuring P-galactosidase activity after in- CY185 topl-7::LEUZ 1.2 X25 lo-' ducing pGAL-topl-IO3 with galactose. We found that NLY515 topi-I03 1.2 x 10" 250 24 hrafter galactose inductionthe levelof DINl The rate of rDNA recombination was determined using the expression was elevated approximately 70-fold rela- rDNA::ADE2 insertion as described in MATERIALS AND METHODS. tive to the uninduced control (Table 5). The level of Each value is the average of six independent measurements. All strains are isogenic and their complete genotypes appear in Table DIN3::lacZ expression also increased in response to 1. induction of topl-103, reaching levels 5- to 12-fold higher than the uninduced control (Table 5), provid- rad5O::URA3 and rad52::LEU2 is a specific genetic ingadditional evidence that topl-103 causesDNA interaction, as the topl-103 mutation is viable in com- damage. In control experiments, we determined that bination with mutations in a number of other DNA galactose induction of wild-type TOPI under control repair genes, including RADI, RAD6, RAD9, RADIO, of the GALl promoter had no effect on DIN1::lacZ the DNA polymerase 6 subunit, CDC2, and the topo- expression. isomerase I1 gene, TOP2. topl-103 overexpression causes RADJdependent topl-103 causes inductionof DNA damage-induc- cell-cycle arrest:Overexpression of topl-I03 from the ible genes: The synthetic lethality of topl-103 with GAL1 promoter severely inhibits growth of wild-type rad50 or rad52 suggested that topl-103 causes DNA or topl-7::LEU2 mutant cells. A strain carrying topl- damage. We investigated this possibility by observing I03 under control of the GALl promoter on a cen- expression of two DNA damage-induciblegenes, tromere-based plasmid (pNL30, NLY620) or an in- DINl and DIN3, in strains carrying topl-103 under tegrated plasmid fails to form colonies on galactose Topoisomerase I Mutant Causes DNA Damage 807

TABLE 4 Recombination frequencies at the HIS4 and CUP1 loci in isogenic wild-type, topl-7::LEUZ and topl-IO3 strains

TOPI allele in TOPI allele Frequency of Frequency relative Strain chromosome on plasmid loss of uRA3 to TOPl A. HIS4::URA3::HIS4 NLY375 iopl-7::LEU2 TOP I 2.3 X 10-4 1 NLY373.374 top I-7::LEU2 Vector 4.9 x 10" 2.1 NLY377.378 topI-7::LEU2 top]-103 5.6 X 10" 24 B. CUPI:URA3 NLY38 1,382 tojII-7::LEU2 TOP I 3.8 x lo-4 1 NLY379,380 ioPI-7::LEU2 Vector 1.3 X 10" 0.34 NLY383 topl-7::LEU2 lop!-103 4.1 X 10-~ 11 The frequency of recombination was determined as described in MATERIALS AND METHODS. Each value is the average of six independent measurements. All strains are isogenic and their complete genotypes appear in Table 1.

TOPl fopl-7::LEU2 NLY816 (DGAL- tOpl-103,727)

NLY494 (pGAL- hCZ)

topi-103 NLY429 (pGAL- TOPl) [high copyTOP1] tOpl-103 NLY620 GAL- ropl-1o3 FIGURE 5.-Growth inhibition due to overexpression of topl- 103. Strain CY 185 (iof1-7::LEU2) was transformed with a series of [low copyTOPI] plasmids bearing the inducible GALl promoter fused to one of the tOpl-103 following genes: lncZ (NLY494). TOPI (NLY429). iopl-I03 (NLY620).iopl-727(NLY880).top1-103,727(NLY816).Thetrans- FIGURE 4."Growth defect of iopl-I03 mutant. Isogenic strains formants were streaked on synthetic galactose medium (lacking were streaked for single colonies on SD medium supplemented with uracil) in order to induce the CALI promoter, and incubated at tryptophan, histidine, and leucine. The relevant genotypes are as 30" for 4 days. follows: NLY107 (TOPI),NLY108 (toPI-7::LEUZ). NLYl64 (topi- 103). NLY566 [topl-103 carrying pNL2 (TOPI on a single-copy vector)], NLY618 [iopl-l03carrying CB17 (highcopy TOPI)].See 103 grows as well as the isogenic control strains on Table 1 for complete genotypes. media containing the carbon sources glucose or raffi- nose (data not shown), which do not induce transcrip- TABLE 5 tion of the GALl promoter. Expression of DIN1::lacZ and DIN3::lacZ fusions in strains When cells carrying topl-103 fused to the GALl expressing topl-103 from the GALl promoter promoter (NLY681) are grown in liquid raffinose medium and then induced with galactose, the number &Galactosidase units of viable cells rapidly stops increasing and remains Strain Fusion Galactose Glucose Ratio constant for greater than 24 hr (Figure 6). The num- NLY887 DIN1::lacZ 78.1 1.1 70.9 ber of viable cells in NLY68 1 begins to decrease after NLY891 DIN1::lacZ 45.5 0.68 66.9 prolonged incubation in galactose medium, eventually NLY894/895 DIN3::lacZ 4.98 0.86 5.8 reaching a level some 1000-fold lower than the unin- NLY896/897 DIN3::lacZ 4.0 0.34 11.8 duced control after 100 hr (data not shown). Control D1N::lacZ activity was determined 24 hr after induction of pCAL- strains carrying GALl-promoted lacZ (NLY494) or topl-103 with galactose, as described in MATERIALS AND METHODS. TOPl (NLY429) grow as well after galactose induction All values were normalized by subtracting &galactosidase units of an isogenic strain without any lacZ fusion. as in the uninduced cultures (data not shown). Cells overexpressing topl-103 arrest in the G:,phase medium (Figure 5). In contrast, isogenic control of the cell cycle. When strains bearing plasmids that strains carrying plasmids that overexpress lacZ overexpress topl-103, lacZ, or TOPl were induced (NLY494) or TOPl (NLY429) from the GALl pro- with galactose and grown to saturation, 78% of cells moter form normal colonies on galactose medium overexpressing topl-103 (NLY620) arrested in the G:! (Figure 5). The strain carrying GAL1-promoted topl- ("large budded") stage (Table 6). DAPI-staining of 808 N. A. Levin, "-A. Bjornsti and G. R. Fink

10' E A fopl-103 ++T7Bsu361 13420 Sal1 Y727 TAA "c NLY681 (RADS) I I CGCTTTA~AGCCG - -u - NLY884 (rad9) TTTAAAA U lYS B t s. cerevisiae 417 FALRAGGEKSEDE-ADTUGCCSL 438 lo4 I,~~I~~~I,~~I~~,~~~~I~~~I~~~I~,,I~~~~,,,I~,,I~,~~~~,I~~~Is. Pombe 462 FALRflGNEKGEDE-ADTUGCCSL 483 0 164 12 8 2028 24 H. sapiens 485 LALRAGNEKEEGETRDTUGCCSL 507 Time after induction (hours) A. fhaliana 680 LALRAGNEKDDDE-ADTUGCCTL 701 vaccinia virus 127 FFIRFGKIIKYLKE-NETUGLLTL 148 FIGURE6.-Viability of wild-type and rad9 mutant strains over- expressing topl-103. Isogenic wild-type (NLY681) andrad9 .at.* .t Ir ..Mr. Ir (NLY884) strains carrying GALl-promoted topl-103 were grown FIGURE7.-Sequence of the topl-lo3 mutation. (A) The coding to early log phase in synthetic raffinose medium lacking leucine. region of TOPl is represented by the open box. The topl-103 The GAL1 promoter was induced by addition of galactose to 2% mutation was localized to the 1.1-kb Bsu36ISalI fragment by re- final concentration. At each time point, an aliquot was removed, striction fragment swapping experiments. The wild-type and topl- serially diluted, and plated on SC - leucine (glucose) plates. The 103 mutant genes were sequenced in parallel, revealing a single number ofviable cells per miwas calculated. See Table 1 for base change (a G to A transition) at position 1758. This mutation complete genotypes. results in an arginine to lysine substitution at amino acid 420. The mutation also generates a new DraI site (TTTAAA) at position TABLE 6 1755, allowing confirmation of the sequence change. The active- Cell cycle arrest of RAD9 and rad9 strains overexpressingtopl- site tyrosine (Y727) is indicated. (B) The topl-103 mutation occurs I03 in a region of high sequence conservation. The amino acid sequence of topoisomerase I from S. cereuisiae,Schizosaccharomyces pombe, Homo sapiens, vaccinia virus (LYNNet al. 1989), and Arabidopsis Percent cells thaliano (KIEBER,TISSIER and SIGNER1992) is shown for a 23- Overexpression Small- Lar e amino acid region identified by MORHAMand SHUMAN(1990) as Strain Genotype plasmid Unbudded budded bud& being important for topoisomerase I activity. The numbers given for each species indicate the position of this region in the amino NLY429 RAD9 TOPl 91 6 3 acid sequence of each protein. Stars indicate residues that are NLY494 RAD9 lacZ 393 4 absolutely conserved in all five species, and dots indicate residues NLY620 RAD9 16 78 6 topi-103 that are identical in all speciesbut vaccinia virus, the least conserved NLY681 RAD9 topi-103 4.5 5.5 90 of the five. The arginine residue mutated in topl-103 is indicated NLY884 rad9 topi-103 55.5 17.6 26.9 by an arrow (After MORHAMand SHUMAN1990). The percentages of unbudded, small-budded and large-budded cells were determined for each strain as described in MATERIALS (Figure 6). Within eight hours of induction of topl- AND METHODS. At least 300 cells were counted for each strain. All strains are isogenic and their complete genotypes appear in Table 103, the number of viable cells has decreased almost 1. 70-fold. In contrast to the wild-type strain (NLY681) overexpressing topl-103, in which 90% of cells were these cells showed that they contained asingle nucleus arrested in the large budded stage (G2)after 29 hr of located at the neck of the bud (data not shown). In induction, rud9::URA3 (NLY884) cells were distrib- contrast,control strains overexpressing lacZ uted throughout the cell cycle, with only 27% in G2 (NLY494) or TOPl (NLY429), had only 3% cells in (Table 6). G2. These strains were overwhelmingly (93%and topl-103 has an arginine to lysine substitution at 91%, respectively) in GI(unbudded) atthis time point a conserved arginine residue:The topl-103 mutation (Table 6). was localized by exchanging restriction fragments be- The RAD9 gene product is required for G2 arrest tween the wild-type and mutant genes and determin- and maintenance of viability in cells overexpressing ing which fragments conferred the slow growth and topl-103. Whereas wild-type cellsoverexpressing topl- rDNA hyper-recombination phenotypes. The topl- 103 (NLY681)arrest in the G2 phase andremain 103 mutation was found to lie on a 1. l-kb Bsu36I- viable for at least 24 hr after induction, the isogenic Sal1 fragment (Figure 7A). Replacement of the wild- rud9::URA3 strain(NLY884) rapidly losesviability type TOPl sequence with this fragment from topl-103 Topoisomerase I Mutant Causes DNA Damage 809 resulted in a growth defect and quantitative rDNA 103 because of a difference in protein stability due to recombinationfrequency that was indistinguishable the topl-727 mutation. Purification of Topl-103 and from the originaltopl-103 mutant. Sequencing of this Top 1-727 from cells bearing identical overexpression fragment revealed only one base change, a G to A constructs (pNL3O and pNL52) yielded comparable transition at position 1758, resulting in an arginine amounts of the two mutant enzymes. No proteolysis (AGA) to lysine (AAA) change at amino acid 420. of eithermutant enzyme was observed. Since the This residue is absolutely conserved in all other eu- active-site tyrosine is requiredfor topl-I03 pheno- karyotic type I topoisomerases (LYNNet al. 1989; types, the Topl-103 mutant enzyme must at least be KIEBER, TISSIERand SIGNER1992; Figure 7B). able to nick DNA in order to exert its effects in the topl-lo3 phenotypes require the active-site tyro- cell. It is likely, therefore, that the mutant enzyme sine: Eukaryotic type I topoisomerases relax DNA by acts directly to damage DNA. making a single-stranded nick in the DNA backbone Topl-103 accumulates in the cleavable complex and becoming covalently attached to the3’ end of the in vitro: topl-I03 was isolated in a screen for mutants brokenstrand [Figure 1A; reviewed in CHAMPOUX defective in topoisomerase I activity. Whereas a wild- (1990)l. This covalent attachment occurs through an type cell extract completely relaxed 200 ng of nega- invariant tyrosine residue, referred to as the “active- tively supercoiled substrate DNA after 1 hr at 37’, site tyrosine.” Substitution of phenylalanine for the extract from atopl-103 mutant was unable to convert active-site tyrosine (tyrosine 727 in S. cerevisiae; see a detectable amount of the substrate to the relaxed Figure 7A) renders topoisomerase I catalytically in- form (Figure 2). We used a modified version of the active, as it cannotinitiate the relaxationreaction standard topoisomerase I assay to show that Topl- (LYNN and WANG 1989; ENG, PANDIT andSTERNG- 103 enzyme does relax supercoiled DNA, although it LANZ 1989). We examined whether the mutant Topl- is very inefficient and tends to accumulate in the 103 enzyme requires the active-site tyrosine in order nicked enzyme-DNA intermediate(the “cleavable to exert its effects in the cell. If Topl-103 must nick complex”; see Figure 1). In these experiments (Figure DNA in order to cause hyper-recombination and cell- 8), we used partially purified fractions of topoisom- cycle arrest, then tyrosine 727 would be required for erase I rather than crude extracts and we increased thesephenotypes, andthe double mutant topl- the enzyme to DNA ratio in order to make the assay 103,727 would notbe expected to display hyper- more sensitive. recombination or arrest. Alternatively, Topl-103 Experiments under anumber of different assay might not need to be catalytically active to exert its conditions and using several independent enzyme effects in the cell, if it acted, for example, as a domi- preparations led to the same conclusion, that Topl- nant inhibitory subunit of an essential protein com- 103 enzyme is defective in converting the cleavable plex. In this case, Topl-103 would not need to cleave complex to the relaxed covalently closed form. The DNA to cause hyper-recombination and cell-cycle ar- experiment shown in Figure 8A is typical ofthe results rest, and the topl-103,727 double mutant would be obtained. During a 10-min incubation at 37” in the expected to have the same phenotypes as the topl-103 absence of camptothecin, Topl-103 enzyme accumu- single mutant. lates in the cleavable complex toa much greater As measured by the colony-sectoring assay for degree than wild-type enzyme (compare lanes 6 and rDNA recombination, the topl-l03,727 double mu- 10). The amount of nicked DNA produced by wild- tant (NLY814) has much lower recombination levels type Topl (lane 6) is not significantly different from than the topl-I03 single mutant (NLY515; data not the amount produced when the supercoiled substrate shown), indicatingthat Topl-103must at least be able is incubated with enzyme storage buffer (lane 2) or a to initiate the topoisomerization reaction in order to column fraction of a mock purification from a Atopl cause hyper-recombination in therDNA. Quantita- strain (lane 4). In contrast, Topl-103 converts a large tion of rDNA recombination showed that strains car- fraction of thesupercoiled substrate to the nicked rying topl-727 and topl-l03,727 had comparable re- form(lane 10). It is unlikely thatthe nicked DNA combinationfrequencies. Inaddition, topl-103,727 produced in the Topl-103 reaction in the absence of overexpressedfrom the inducible GALl promoter camptothecin could have been generated by chemical permits colony formationon galactose plates means or by a contaminating copurifying , (NLY816), unlike the topl-103 mutationalone because no nicking activity is detected in the parallel (NLY620), which causes growth arrest when overex- column fractions from the Atopl strain or from the pressed from the GALl promoter (Figure 5). Over- strain expressing wild-type Topl. The same amount expressing topl-l03,727 fromthe GALl promoter of topoisomerase I was added to the wild-type and (NLY816) doescause some growth inhibition, butthis Topl-103 reactions, based on quantitation by silver effect is also seen with the topl-727 single mutant staining and Western blotting. (NLY880 in Figure5). It is unlikely thatthe topl- Consistent with a defect in converting the cleavable 103,727 phenotypes are less severe than thoseof topl- complex to relaxed covalently closed circles, Topl- 810 N. A. Levin, M.-A. Bjornstiand G. R. Fink A 103 enzyme produces less fully relaxed DNA (1 2.4% of total DNA in lane 10, based on phosphorimager analysis) than wild-type Topl (32.3% of total DNA in No lane 6) in parallel reactions. Experiments designed to Extract drop7 Topl 03Top1 -1 follow the conversion of supercoiled DNA tothe nn" relaxed form show that Topl-103 enzyme takes " " " ++ -- + + Camptothecin - + - + - + - + - + - + Proteinase K longer than 5 min to relax fully a supercoiled substrate DNA, whereas under the same conditions wild-tvpe Topl relaxes the same substrate in less than 20 sec, at least a 15-fold difference in efficiency (data not shown). These experiments also suggest that Topl- 103 is slower to make the initial nick in DNA than is wild-type Top 1 . The nicked DNA generated by Topl-103 enzyme is a genuinecovalent enzvme-bound species (the cleav- able complex), similar to the wild-type Topl enzyme- Nicked DNA complex in the presence of camptothecin. Un- likemost , topoisomerase I remains cova- lently bound to nicked DNA and does not dissociate Ir Supercoiled until it has religated the broken strand. This covalent 1 enzyme-DNA complex either remains in the wells of an agarose gel or migrates as a broad high molecular L e Relaxed weight smear (see Figure 8B, lanes 5, 6 and 7). Thus, the electrophoretic migration of this species asa single 1234 56 78 910 1112 band of free nicked DNA is dependent on proteinase B K digestion of the Topl reaction products(Figure 8B, lane 8). The intensitv of the band at the position Atop 7 Top1 TOP103 -1 of nicked DNA greatly increases following proteinase -CP T +CPT -CPT+CPT -CPT K treatment of Topl-103 reaction products (Figure 8A; compare lanes 9 and 10, lanes I 1 and 12), showing I- S P' I-S P' I- S H P' Treatment H H that Topl-103, like wild-type Topl in the presence Origin of camptothecin (Figure 8A; compare lanes 7 and 8). produces enzyme-bound nicked DNA. In contrast, the intensitv of the band at the position of free nicked DNA does not increase after proteinase K digestion of the reaction products of Topl in the absence of camptothecin (Figure 812; lanes 5 and 6). The Faint

band present at the "nicked " position in the Topl-

Nicked incubated for 10 minutes at 57". divided in half. and stopped by adding SDS to 0.5% final concentration. One half of each reaction was digested with 0. I5 rr~g//r~rlImteinase K (even-nur~~berrtI1;trlca) for 1 hr at 37". and the samples were electrophoresed on an 0.9% Supercoiled agarose gel with 50 ng/ml ethidium bromide. These electrophoresis 1 conditions resolve supercoiled, relaxed, and nicked topoisomers. The DNA was visualized by Southern blotting followed by hvbrid- izationwith a hexamer-labeled probe. (B) Reactions were per- ] Relaxed formed as in (A), except that camptothecin (0.1 I mg/ml) in DMSO was added to the reaction containing wild-type Topl enzyme (lanes 1234 567 8 9101112 5-8) whereas DMSO alone was added to the reactions containing FIGURE8.-Assays of nicking activity of the wild-type and 'ropl- Atop1 extract (lanes 1-4) or Topl-I05(lanes 9-12). The reactions 103 enzymes. (A) Diluted partially purified fractions of topoisom- were stopped by addition of SDS to 0.5% and divided into four erase I from strainsoverexpressing wild-type TOP1 (NLY429). fopl- parts that were treated as follows: reactions in lanes 1.5 and 9 ("-") 103 (NLY620) or no enzyme (AtopZ, NLY108) were added to a were untreated; NaCl ("S") was added to 0.5 M final concentration reaction mix containing 4 ng supercoiled pUCll8 DNA. Topoisom- to reactions in lanes 2.6 and 10: reactions in lanes 5.7 and 1 1 were erase storage buffer wasused as the "No Extract" control. The heated ("H") to 70" for 15 min; reactions in lanes 4.8 and 12 were reaction mixes contained either camptothecin in DMSO at a final digested with proteinase K ("P") at 0.15 mg/ml for one hour at concentration of 0.1 1 mg/ml (lanes 7, 8, 11 and 12) or DMSO 37". Thereactions were then electrophoresed, blotted. and hybrid- alone (all other lanes) at 1 1 % final concentration. Reactions were ized as in (A). Topoisomerase I Mutant Causes DNA Damage 81 1

103 reaction in the absence of proteinase K digestion wild-type strains with the anti-tumor drug, campto- (Figure 8A; lane 9)is attributable to contamination of thecin. the supercoiledsubstrate with a small amount of topl-103 mimics the in vivo effects of the anti- nicked DNA, as it is seen in lanes 1 to 6. tumor drug, camptothecin: Eukaryotic type I topoiso- We investigated the resistance of the Top1-103- merases relax DNA by a mechanism that involves a DNA complex to various treatments that would be covalent enzyme-DNA intermediate called the “cleav- expected to dissociate anoncovalent enzyme-DNA ablecomplex” [Figure 1A; reviewed in CHAMPOUX complex.After incubating the enzyme with super- (1990)l. Although the cleavable complex is ordinarily coiled DNA for 10 min at 37”, we stopped the reac- a transient intermediate, it can be trapped by treat- tions with SDS, divided them into aliquots,and treated ment with denaturing agents such as SDS, alkali, low them with either high salt (0.5 M NaCl for 1 hr), high pH, and by the topoisomerase I-specific drug, camp- temperature (70” for 15 min), or proteinase K (1 hr tothecin [Figure 1B; reviewed in LIU (1989, 1990)]. at 37”). A fourth aliquot was left untreated. Figure This drug does not alter the ability of topoisomerase 8B shows that only proteinase K digestion (lane 12) I to cleave the DNA substrate, but it impairs either fully removed Topl-103 from thenicked DNA, allow- the repositioning of thebroken DNA strandsfor ing it to migrate as a single band. High salt treatment ligation or the religation reaction itself (PORTERand (lane 10) produced no more free nicked DNA than in CHAMPOUX 1989; SVESTRUPet al. 199 1). the untreated control (lane 9), and high temperature The in vivo effects of camptothecin have been ex- treatment (lane 11) produced little more than in the tensively studied, as this drug and its analogues have control. Treating the reaction products of wild-type shown promise as anti-tumor agents[reviewed in POT- Topl in the presence of camptothecin with high salt MESIL and SILBER (1 990)]. Experimentsin mammalian (Figure 8B, lane 6), high temperature (lane 7),or tissue culture cells (HSIANGet al. 1985) have shown proteinase K (lane 8) gave results that paralleled the that camptothecin traps topoisomerase I in the cleav- results with Topl-103 in the absence of camptothecin. able complex in vivo as well as in vitro. Enzyme-bound Only proteinase K treatment fully released the nicked nicks generated by topoisomerase I in the presence of DNA to migrate at its characteristic position. camptothecin areexpanded double-strandedto Unlike wild-type Topl , Topl-103 is unaffected by breaks by replication fork passage during S phase camptothecin (Figure SA). Topl-103 accumulates in (HSIANG,LIHOU and LIU 1989; PORTER and CHAM- the cleavable complex in thepresence (lane 12) or POUX 1989;D’ARPA, BEARDMOREand LIU 1990; absence (lane 10) of camptothecin.Quantitation of ZHANG, D’ARPAand LIU 1990; DEL BINO, LASSOTA radioactivity shows that 10.6% of total DNA in the and DARZYNKIEWICZ1 ;199 RYAN et al. 199 1). Camp- Topl-103 reaction without camptothecin (lane 10) is tothecin treatment of mammalian cells induces recom- in the nicked form and 11.2% of the total DNA in bination, chromosomal aberrations (LIM et al. 1986), the Topl-103reaction with camptothecin (lane 12) is and fragmentation of DNA (HORWITZand HORWITZ in the nicked form.Camptothecin has no apparent 197 1). As in mammalian cells, treatment of yeast with effect on theconversion by Topl-103 of the cleavable sublethal levels of camptothecin causes hyper-recom- complex to relaxed covalently closed circles. bination, as well as induction of the theDNA damage- inducible gene, DIN3 (NITISS and WANG 1988; ENG DISCUSSION et al. 1988). Yeastcells lacking thedouble-strand break repair gene, RAD52, are much more sensitive We have identified a mutantallele of the yeast TOPI to killing by camptothecin (NITISS and WANC 1988). gene that has phenotypes that are strikingly different Camptothecin treatment also causes G2 arrest in a froma null allele. This mutation increases mitotic variety of mammalian cell lines, presumably as a result recombination 100-300-fold in the rDNA repeats and of the topoisomerase I-mediated DNA damage (LI, 10-25-fold at other loci, whereas a null mutation in FRASERand BHUYAN 1972;TOBEY 1972; DEL BINO, TOP1 results in a 20-fold elevation of mitotic rDNA SKIERSKIand DARZYNKIEWICZ 1990;TSAO, D’ARPA recombination and noeffect outside the rDNA.Other and LIU 1992). phenotypes of topl-103, including slow growth, invi- topl-103 has pleiotropic phenotypes that resemble ability in the absenceof the double-strand break repair the effects of camptothecintreatment ofyeast or system, and constitutive expression of DNA damage- mammalian cells. These phenotypes can be under- inducible genes, suggest that expression of the mutant stood in light of the model proposed for the in vivo enzyme leads to DNA damage. We have shown that effects of camptothecin. If the Topl-103enzyme has Topl-103 enzyme can generate nicked DNA in vitro, a tendency to become trapped in the cleavable com- and is therefore likely to be directly responsible for plex, replication fork passage in S phase would lead DNA damage. We propose that Topl-103 enzyme is to double-stranded breaks at sites where the enzyme trapped in the cleavable complex in vivo, leading to is bound. These breaks would generate a DNA dam- phenotypes similar to those seen upon treatment of age signal and cause cellcycle arrestat the G2-M 812 Levin,N. A. M.-A. Bjornstiand G. R. Fink boundary. Breaks occurring in S phase would not be fects suggest that preventing G2 arrest in camptothe- lethal to yeastcells with anintact recombinational cin-treated tumor cells might increase cell killing. A DNA repair system, since they could berepaired number of drugs have been identified that interfere during G2 by recombination with the intact sister with cell-cycle arrest in G2, including caffeine (SCHLE- chromatid. GEL and PARDEE 1986) andokadaic acid (YAMASHITA Consistent with this model, topl-103 causes mitotic et al. 1990; PATELand WHITAKER 1991).Simultane- hyper-recombination in the rDNA and at other loci, ous treatment of tumors with camptothecin and a just as camptothecin treatment elevates recombina- drug that blocked arrest might have a synergistic tion. In addition, topZ-103 is inviable in combination effect, reducing the required dose of camptothecin. with null mutations in either of the DNA-repair genes This synergy could be important for clinical use of RAD50 or RAD52, which are required for recombi- camptothecin, as the drug has toxic side effects (POT- national repair of double-strand breaks [reviewed in MESIL and SILBER 1990).The benefit of using drugs FRIEDBERG (1988)l.Expression of topl-IO3 from the that block G2 arrest in conjunction with camptothecin GALZ promoter causes cellcycle arrest in G2 and would, of course, depend on the relative toxicity of induction of the DNA damage-inducible genes DIN1 these drugs to normal and tumor cells. and DIN3, providing further evidence of topoisom- Topl-103 enzyme has altered activity in vitro: eraseI-mediated DNA damage. Additional support Topl-103 mutant enzyme differs from wild type in for the model that Topl-103 damages DNA by gen- three respects: it is much less efficient at relaxing erating enzyme-bound nicks comes from experiments DNA, it accumulates more covalent enzyme-DNA showing that topl-103 phenotypes are reversed by complex, and it is insensitive to camptothecin. Figure mutating the enzyme's active site tyrosine. 8 shows that wild-type enzyme converts more of the topl-103 overexpression leads to RADPdepend- supercoiled substrate to the relaxed form than does ent Gq arrest: Inboth yeast and mammalian cells, the same amount of Topl-103. Relative to wild type, DNA damage or incomplete DNA replication leads to the equilibrium between free Topl-103 enzyme and arrest in G2, preventing cells from dividing before the covalent enzyme-DNA complex is shifted toward their chromosomes have been completely and accu- the enzyme-DNA complex (see Figure 1B). More rately replicated [WEINERTand HARTWELL(1988) nicked DNA is produced in reactionscontaining and references therein]. In yeast cells, this arrest is Top 1- 103 enzyme than in those containing wild-type mediated by RAD9 (WEINERT and HARTWELL 1988). enzyme (Figure 8). We have demonstrated that most rad9 mutants fail to arrest properly in response to of the nicked DNA in the Topl-103 reactions is many types of DNA damage and are thus hypersensi- bound to protein, because digestion with proteinase tive to DNA-damaging agents. K prior to electrophoresis greatly increases the inten- Overexpression of top1 -103 causes cell-cycle arrest sity of the band at the position of free nicked DNA, in the G2 (large-budded) stage (Table 6), suggesting as compared with a sample that has not been digested that DNA damage in these cells leads to a cell division (Figure 8). block at the RAD9 checkpoint. Other mutants with a The amount of nicked DNA found in Topl-103 defect in DNA replication, including cdc9 (DNA li- reactions is unaffected by the presence of camptothe- gase), are known to arrest at this same point in the cin (Figure 8). Other camptothecin-resistant mutants cellcycle (SCHIESTLet al. 1989). Deletion of RAD9 have been isolated in mammalian cells (TANIZAWA prevents G2 arrest (Table 6) and maintenance of via- and POMMIER 1992) and yeast(M-A. BJORNSTI,un- bility (Figure 6) incells overexpressing topZ-103, published data). The insensitivity of Topl-103to showing that topoisomerase-induced DNA damage camptothecin could be explainedin several ways. One leads to cell-cycle arrest throughRAD9. Whereas wild- explanation is that the drug does not bind to Top1- type cells overexpressing top1-103 arrest uniformly 103 as well as it binds to the wild-type enzyme, and and remain viable for over 24 hr, the isogenic rad9 therefore cannot stabilize the enzyme-DNA complex. mutant continues to traverse the cell cycleand rapidly Anotherexplanation is thatcamptothecin stabilizes loses viability. the covalent intermediate by converting the religation Treatment of cultured mammalian cells with camp- reaction, which is normally a very fast step, into aslow tothecin results in G2 arrest (LI,FRASER and BHUYAN step. If the religation reaction is already a slow step 1972; TOBEY1972; DEL BINO,SKIERSKI and DARZYN- for Topl-103,which would explain the accumulation KIEWICZ 1990), analogousto the effect of overexpres- of covalent intermediate even in the absence of camp- sion of topl-lO3. The mechanism of this arrest is not tothecin, then the drug might have little effect. completely understood, butit has recently beenshown The in vitro activity of Topl-103 enzyme is similar to involve inhibition of ~34'~'~protein kinase activa- but not identical to that of wild-type enzyme in the tion (TSAO,D'ARPA and LIU 1992), andis presumably presence of camptothecin. Topl-103 enzyme is some- triggered by topoisomerase I-induced DNA damage. what defective in the entire reactionpathway, not just The parallels between topZ-ZO3 and camptothecin ef- the resealing step. However, the critical parameter in Topoisomerase IDamage Mutant DNA Causes 813 determining the effects of the mutant enzyme in the Fluoro-orotic acid as a selective agent in yeast molecular ge- cell is the relative level of the cleavable complex. netics. Methods Enzymol. 154 164-175. BOTSTEIN,D., S. C. FALCO,S. E. S~WART,M. BRENNAN,S. Topl-103 clearly accumulates more cleavable com- SCHERER,D. T. STINCHCOMB,K. STRUHLand R. W. DAVIS, plex than wild-type enzyme in vitro, and even a small 1979 Sterile host yeasts (SHY):a eukaryotic system of biolog- increase in cleavable complex in vivo might lead to ical containment for recombinant DNA experiments. Gene 8: severe consequences. 17-24. The topl-103 mutation falls in a critical catalytic BRILL,S. J.. and R. STERNGLANZ,1988 Transcriptiondependent DNA supercoiling in yeast DNA topoisomerase mutants. Cell domain of topoisomerase I: The topl-103 mutation 54: 403-4 11. consists of an arginine tolysine substitution at position CHAMPOUX,J. J., 1990 Mechanistic aspects of type I topoisomer- 420 of the yeast topoisomerase 1 enzyme (Figure 7). ases, pp. 217-242 in DNA Topology and its Biological Effects, Although this amino acid change is conservative, sub- edited by N. R. COZZARELLIand J. C. WANG.Cold Spring stituting one basic residuefor another, it altersa Harbor Laboratory, Cold Spring Harbor, N.Y. CHRISTMAN,M. F., F. S. DIETRICHand G. R. FINK,1988 Mitotic residue that is invariant in all cloned eukaryotic to- recombination in the rDNA of S. cerevisiue is suppressed by the poisomerase I genes (Figure7B). In addition, arginine combined action of DNA topoisomerases I and 11. Cell 55: 420 falls within a 23 amino acid stretch with very high 4 13-425. overall sequence identity (Figure 7B) that was shown D'ARPA,P., C. BEARDMOREand L. F. LIU, 1990 Involvement of by MORHAMand SHUMAN to be critical for nucleic acid synthesis in cell killing mechanisms of topoisom- (1990) erase poisons. Cancer Res. 506919-6924. enzyme function in the vaccinia virus topoisomerase DEL BINO,G., P. LASSOTAand Z. DARZYNKIEWICZ,1991 The S- I. Four out of the five mutants they isolated that phase cytotoxicity of camptothecin. EX~.Cell Res. 193: 27-35. abolished enzyme activity withoutaltering protein DELBINO, G.,J. S. SKIERSKIand Z. DARZYNKIEWICZ,1990 Diverse structure or stability fell in this region, suggesting an effects of camptothecin, an inhibitor of topoisomerase I, on the important role in catalysis or DNA binding. cell cycle of lymphocytic (L1210, MOLT-4) and myelogenous (HL-60, KGl) leukemic cells. Cancer Res. 50: 5746-5750. FRIESENand SADOWSKI(1 992)have recently studied ENG,W. K., S. D. PANDITand R. STERNGLANZ,1989 Mapping of the effect on catalysis of an arginine to lysine substi- the active site tyrosine of eukaryotic DNA topoisomerase I. J. tution in the yeast FLP protein, a site-specific recom- Biol. Chem. 264: 13373-13376. binase of the integrase family, whose members have a ENG, W. K., L. FAUCETTE,R. K. JOHNSON and R. STERNGLANZ, very similar reaction mechanism to topoisomerases. 1988 Evidence that DNA topoisomerase I is necessary for the cytotoxic effects of camptothecin. Mol. Pharmacol. 34 755- This enzyme class has an invariant tyrosine thatmakes 760. a covalent phosphotyrosine bond with the 3' end of FOGEL,S., and J. W. WELCH,1982 Tandem gene amplification the broken DNA strand, as well as invariant histidine mediates copper resistance in yeast. Proc. Natl. Acad. Sci. USA and arginine residues. Substitution of lysine for the 79 5342-5346. conserved arginine (RK191)results in a mutant en- FRIEDBERG,E. C., 1988 Deoxyribonucleic acid repair in the yeast Saccharomyces cerevisiae. Microbiol. Rev. 52 70-102. zyme that can initiate the strand exchange reaction, FRIESEN,H., and P. D.SADOWSKI, 1992 Mutagenesis of a con- but becomes trapped in the covalent reaction inter- served region of the gene encoding the FLP recombinase of mediate, a situationthat closely parallels that of Topl- Saccharomyces cerevisiae: a role for arginine 191 in binding and 103. ligation. J. Mol. Biol. 225: 313-326. Additional mutagenesis studies of yeast TOP1 are GIAEVER,G. N., and J. C. WANG,1988 Supercoiling of intracel- in progress. These studies should identify other resi- lular DNA can occur in eukaryotic cells. Cell 55 849-856. GOTO, T., and J. C. WANG,1985 Cloning of yeast TOPI: the gene dues that are important forcatalysis and forresistance encoding DNA topoisomerase I, and construction of mutants to camptothecin, and should provide insight into the defective in both DNA topoisomerase I and DNA topoisom- reaction mechanism and structural requirements of erase 11. Proc. Natl. Acad. Sci. USA 82: 7 178-7 182. topoisomerase I. GUTHRIE,C., and G.R. FINK, 1991 Guide to Yeast Genetics and Molecular Biology. Academic Press, San Diego. We thank STEPHANIERUBY, ROLF STERNGLANZ, JUN MA, ERIC HORWITZ,M. S., and S. B. HORWITZ,1971 Intracellular degra- ALANI,NANCY KLECKNER, LOUISE PRAKASH, MICHAEL CHRISTMAN, dation of HeLa and adenovirus type 2 DNA induced by camp DAVIDSCHILD and CRAIGBENNETT for providing plasmids and tothecin. Biochem. Biophys. Res. Commun. 45: 723-727. strains. We thank JAM= WANG,ROLF STERNGLANZ and LEROYLIU HSIANG,Y. H., M. G. LIHOUand L. F. L~u,1989 Arrest of for advice on nicking assays. We are grateful to DAVIDPELLMAN, replication forks by drug-stabilized topoisomerase I-DNA MICHAEL CHRISTMAN,JULIE BRILLand ALANSACHS for critical cleavable complexes as a mechanism of cell killing by campto- reading of the manuscript, and tomembers ofthe FINKlaboratory thecin. Cancer Res. 49 5077-5082. for helpful discussions. This work was supported by National Insti- HSIANG,Y. H., R. HERTZBERG,S. HECHT and L. F. LIU, tutes of Health Training Grant CA09541 and the Office of Naval 1985 Camptothecin induces protein-linked DNA breaks via Research Graduate Fellowship Program. mammalian DNA topoisomerase I. J. Biol. Chem. 260 14873- 14878. LITERATURECITED KIEBER,J. J., A. F. TISSIERand E. R. SIGNER,1992 Cloning and characterization ofan Arabidopsisthaliuna topoisomerase I BARKER,D. G., and L. H. JOHNSTON, 1983 Saccharomyces cereuisiae gene. Plant Physiol. 99 1493-1501. cdc9: a structural genefor yeast DNA ligase whichcomplements KIM, R. A., and J. C. WANG,1989 A subthreshold level of DNA Schizosaccharomyces pombe cdcI7. Eur. J. Biochem. 134 315- topoisomerases leads to the excision of yeast rDNA as extra- 319. chromosomal rings. Cell 57: 975-985. EKE, J. D., J. TRUEHART,G. NATSOULISand G. R. FINK,1987 5- LI, L. H., T. J. FRASERand B. K. BHUYAN,1972 Action of 814 N. A. Levin. M.-A. Bjornstiand G. R. Fink

camptothecin on mammalian cells in culture. Cancer Res. 32: 1991 Camptothecin cytotoxicity in mammalian cells isasso- 2643-2650. ciated with the induction of persistent double strand breaks in LIM, M.,L.F. LIU, K. D. JACOBSON and J. R. WILLIAMS, replicating DNA. Nucleic Acids Res. 19 3295-300. 1986 lnduction of sister chromatid exchanges by inhibitors SCHIESTL,R. H., P. REYNOLDS,S. PRAKASHand L. PRAKASH, of topoisomerases. Cell Biol. Toxicol. 2: 485-494. 1989 Cloning and sequence analysis of the Saccharomyces LIU, L. F., 1989 DNA topoisomerase poisons as antitumor drugs. cerevisiae RAD9 gene and further evidence that its product is Annu. Rev. Biochem. 58: 351-375. required for cell cycle arrest induced by DNA damage. Mol. LIU, L. H., 1990 Anticancer drugsthat convert DNA topoiso- Cell. Biol. 9 1882-96. merases into DNA damaging agents, pp. 371-389 in DNA SCHILD,D., B. KONFORTI, C. PEREZ,W. GISH and R. MORTIMER, Topology and Its Biological Effects, edited by N. R. COZZARELLI 1983 Isolation and characterization ofyeast DNA repair and J. C. WANG.Cold Spring Harbor Laboratory, Cold Spring genes. I. Cloning of the RAD52 gene. Curr. Genet. 7: 85-92. Harbor, N.Y. SCHLEGEL,R., and A. B. PARDEE, 1986 Caffeine-induced uncou- LIU, L. F., and J. C. WANG,1987 Supercoiling of the DNA pling of mitosis from the completion of DNA replication in template during transcription. Proc. Natl. Acad. Sci. USA 84: mammalian cells. Science 232: 1264-1 266. 7024-7027. SIKORSKI,R. S., and P. HIETER,1989 A system of shuttle vectors LYNN,R. M., and J. C. WANG, 1989 Peptide sequencing and site- and yeast host strains designed for efficient manipulation of directed mutagenesis identify tyrosine-3 19 as the active site DNA in Saccharomyces cerevisiae. Genetics 122: 19-27. tyrosine of Escherichia coli DNA topoisomerase I. Proteins 6: STRUHL,K., D. T. STINCHCOMB,S. SCHERERand R. W. DAVIS, 231-239. 1979 High frequency transformation of yeast: autonomous LYNN,R. M., M. A.BJORNSTI, P. R. CARONand J. C. WANG, replication of hybrid molecules. Proc. Natl. Acad. Sci. USA 76 1989 Peptide sequencing and site-directed mutagenesis iden- 1035-1039. tify tyrosine-727 as the active site tyrosine of Saccharomyces SVESTRUP,J. Q., K. CHRISTIANSEN,I. I. GROMOVA,A. H. ANDERSEN cerevisiae DNA topoisomerase I. Proc. Natl. Acad.Sci. USA and 0. WESTERGAARD,1991 New technique for uncoupling the cleavage and religation reactions of eukaryotic toposiom- 86: 3559-3563. erase 1. The mode of action of camptothecin ata specific MA, H., S. KUNES,P. J. SCHATZand D. BOTSTEIN,1987 Plasmid recognition site. J. Mol. Biol. 222 669-678. construction by homologous recombination in yeast. 58: 201- TANIZAWA,A., and Y. POMMIER,1992 Topoisomerase 1 alteration 216. in a carnptothecin-resistant cell line derived from Chinese ham- MANIATIS,T. E., E. F. FRITSCHand J. SAMBROOK, 1982Molecular ster DC3F cells in culture. Cancer Res. 52: 1848-1854. Cloning: A Laboratory Manual.Cold Spring Harbor Laboratory, THRASH,C., K. VOELKEL, S. DINARDOand R. STERNGLANZ, Cold Spring Harbor, N.Y. 1984 Identification of Saccharomyces cerevisiae mutants de- MILLER,J. L., 1972 Experiments in Molecular Genetics. Cold Spring ficient in DNA topoisomerase I activity. J. Biol. Chem. 259: Harbor Laboratory, Cold Spring Harbor, N.Y. 1375-1377. MORHAM,S. G., and S. SHUMAN,1990 Phenotypic selection and THRASH,C., A. T. BANKIER,B. G. BARRELLand R. STERNGLANZ, characterization of mutant alleles of a eukaryotic DNA topoi- 1985 Cloning, characterization, and sequence of the yeast somerase I. Genes Dev. 4: 515-524. DNA topoisomerase I gene. Proc. Natl. Acad.Sci. USA 82: MYERS,A. M., A. TZAGOLOFF,D.M. KINNEY and C. J. LUSTY, 4374-4378. 1986 Yeast shuttle and integrative vectors with multiple clon- TOBEY,R. A., 1972 Effectsof cytosine arabinoside, daumycin, ing sites suitable for construction of lacZ fusions. Gene 45: mithrarnycin, azacytidine, adramycin, and camptothecin on 299-3 10. mammalian cell-cycle traverse. Cancer Res. 32: 2720-2725. NITISS,J., and J. C. WANG, 1988 DNA topoisomerase-targeting TSAO,Y.-P., P. D'ARPAand L. F. LIU, 1992 The involvement of antitumor drugs can be studied in yeast. Proc. Natl. Acad. Sci. active DNA synthesis in camptothecin-induced G:! arrest: al- USA 85: 7501-7505. tered regulation of p34'd's/cyclin B. Cancer Res. 52: 1823- NONET,M., C. SCAFE,J. SEXTONand R. YOUNG, 1987 Eucaryotic 1829. RNA polymerase conditional mutant thatrapidly ceases mRNA WALLIS,J. W., G. CHREBET,G. BRODSKY,M. ROLFEand R. ROTH- synthesis. Mol. Cell. Biol. 7: 1602-1611. STEIN, 1989 A hyper-recombination mutation in S. cerevisiae PATEL,R., and M. WHITAKER,1991 Okadaic acid suppresses identifies a novel eukaryotic topoisomerase. Cell 58: 409-419. calcium regulation of mitosis onset in sea urchin embryos. Cell WEINERT,T. A,, and L. H. HARTWELL,1988 The RAD9 gene Regul. 2: 391-402. controls the cell cycle response to DNA damage in Saccharo- PORTER,S. E., and J. J. CHAMPOUX, 1989The basis for campto- myces cerevisiae. Science 241: 317-22. thecin enhancement of DNA breakage by eukaryotic topoisom- WEST, R. W., R. R. YOCUM and M. PTASHNE, 1984Saccharomyces erase I. Nucleic Acids Res. 17: 8521-8532. cerevisiae CALI-GALIO divergent promoter region: location POTMESIL,M., and R. SILBER,1990 DNA topoisomerases in clin- and function of the upstream activating sequence UASc. Mol. ical oncology, pp. 391-407 in DNA Topology and Its Biological Cell. Biol. 4: 2467-2478. Effects, edited byN. R. COZZARELLIand J. C. WANG.Cold YAMASHITA,K., H. YASUDA,J. PINES,K. YASUMOTO,H. NISHITANI, Spring Harbor Laboratory, Cold Spring Harbor, N.Y. M. OHTSUBO,T. HUNTER,T. SUGIMURAand T. NISHIMOTO, ROSE, M. D., P. NOVICK,J. H. THOMAS,D. BOTSTEIN and G. R. 1990 Okadaic acid, a potent inhibitor of type 1 and type 2A FINK,1987 A Saccharomyces cerevisiae genomic plasmid bank protein phosphatases, activates cdc2/H1 kinase and transiently based ona centromere-containing shuttle vector. Gene 60 induces a premature mitosis-like state in BHK2l cells. EMBO 237-243. J. 9 4331-4338. RUBY,S., and J. SZOSTAK,1985 Specific Saccharomyces cerevisiae YANAGIDA,M., and R. STERNGLANZ, 1990Genetics of DNA to- genes are expressed in response to DNA-damaging agents. poisomerases, pp. 299-320 in DNA Topology and Its Biological Mol. Cell. Biol. 5: 75-84. Effects, edited by N.R. COZZARELLIand J. C. WANG.Cold RUBY,S., J. SZOSTAKand A. W. MURRAY,1983 Cloning regulated Spring Harbor Laboratory, Cold Spring Harbor, N.Y. ZHANG,H., P. D'ARPAand L. F. LIU, 1990 A model for tumor yeast genes from a pool of lacZ fusions. Methods Enzymol. 101: cell killing by topoisomerase poisons. Cancer Cells 2: 23-27. 252-269. RYAN,A. J., S. SQUIRES,H. L. STRUTTand R. T. JOHNSON, Communicating editor: A. G. HINNEBUSCH