Phosphorylation of Slx4 by Mec1 and Tel1 Regulates the Single-Strand Annealing Mode of DNA Repair in Budding Yeastᰔ Sonja Flott,1† Constance Alabert,2 Geraldine W

Phosphorylation of Slx4 by Mec1 and Tel1 Regulates the Single-Strand Annealing Mode of DNA Repair in Budding Yeastᰔ Sonja Flott,1† Constance Alabert,2 Geraldine W. Toh,1 Rachel Toth,1 Neal Sugawara,3 David G. Campbell,1 James E. Haber,3 Philippe Pasero,2 and John Rouse1* MRC Protein Phosphorylation Unit, James Black Centre, University of Dundee, Dundee DD1 5EH, United Kingdom1; Institute of Human Genetics, CNRS UPR 1142, 141 Rue de la Cardonille, 34396 Montpellier, France2; and Rosenstiel Basic Medical Sciences Research Centre, Waltham, Massachusetts3

Received 22 January 2007/Returned for modiﬁcation 1 June 2007/Accepted 28 June 2007

Budding yeast (Saccharomyces cerevisiae) Slx4 is essential for cell viability in the absence of the Sgs1 helicase and for recovery from DNA damage. Here we report that cells lacking Slx4 have difﬁculties in completing DNA synthesis during recovery from replisome stalling induced by the DNA alkylating agent methyl methanesulfonate (MMS). Although DNA synthesis restarts during recovery, cells are left with unreplicated gaps in the genome despite an increase in translesion synthesis. In this light, epistasis experiments show that SLX4 interacts with genes involved in error-free bypass of DNA lesions. Slx4 associates physically, in a mutually exclusive manner, with two structure-speciﬁc endonucleases, Rad1 and Slx1, but neither of these enzymes is required for Slx4 to promote resistance to MMS. However, Rad1-dependent DNA repair by single-strand annealing (SSA) requires Slx4. Strikingly, phosphorylation of Slx4 by the Mec1 and Tel1 kinases appears to be essential for SSA but not for cell viability in the absence of Sgs1 or for cellular resistance to MMS. These results indicate that Slx4 has multiple functions in responding to DNA damage and that a subset of these are regulated by Mec1/Tel1-dependent phosphorylation.

The RecQ helicases are important regulators of genome resulting structures may be converted to pseudo-dHJs, in a stability. Human cells have several RecQ family helicases, in- Rad51/Rad52-dependent manner, and cells lacking Sgs1 accu- cluding BLM and WRN, that are mutated in Werner’s and mulate these pseudo-dHJs presumably because it normally Bloom’s syndromes, respectively (19, 25). Yeasts have a single helps to dissolve them (27). There is experimental evidence to RecQ helicase: Sgs1 in Saccharomyces cerevisiae and Rqh1 in support template switching at blocked replisomes in yeast (53), Schizosaccharomyces pombe. Cells lacking Sgs1 or Rqh1 exhibit but the mechanisms involved are unclear, as are many of the high rates of chromosome loss and rearrangements, an ele- relevant genes. vated incidence of loss of heterozygosity, increased rates of Despite these important roles, Sgs1 is not essential for cell sister chromatid exchange, hypersensitivity to agents that dam- viability. However, a screen for genes required for viability in age DNA, and defects in meiosis (15, 25, 30, 48, 50). Sgs1 is the absence of SGS1 (or TOP3) identified six “SLX” genes. also important for preventing recombination between diver- The products of these genes form three heterodimeric com- gent sequences in the single-strand annealing (SSA) mode of plexes in cells: Slx2 (Mms4)/Slx3 (Mus81), Slx5/Slx8, and Slx1/ double-strand break (DSB) repair (32, 42, 43). Sgs1 interacts Slx4 (31). The precise cellular roles of Slx5 and Slx8 are un- with DNA topoisomerase III (Top3), and together Sgs1 and clear, although cells lacking either protein showed an increased Top3 can dissolve double Holliday junctions (dHJs) in vitro rate of gross chromosomal rearrangements (52). Mms4-Mus81 (49). It is thought that this activity promotes the resolution of is a structure-specific endonuclease that, at least in vitro, pref- recombination intermediates during restart of stalled or erentially cleaves branched DNA structures resembling struc- blocked replisomes (25). tures that arise during recombinational processing of stalled Replisomes blocked by DNA damage can bypass lesions by replication forks (2, 23). The synthetic lethality of sgs1 mus81 at least two different mechanisms: translesion synthesis (TLS) cells is suppressed when RAD52 is deleted (2, 10), suggesting across the damaged base by TLS polymerases and error-free Mus81 cleaves dHJs when they are not dissolved by Sgs1 dur- bypass, also known as “template switching” (34, 47). Template ing recombinational processing of stalled replisomes. switching is thought to involve unpairing of the nascent strands Slx1-Slx4 from budding yeast and fission yeast is also a struc- from the parental template and their subsequent annealing ture-specific endonuclease with preference in vitro for (20, 39). The nascent strand that had been blocked is then branched DNA substrates, especially simple-Y, 5Ј-flap, or rep- elongated using the nascent sister strand as template. The lication fork-like structures (7, 14). Slx1-Slx4 is likely to define a pathway distinct from Mms4-Mus81, because the synthetic * Corresponding author. Mailing address: MRC Protein Phosphor- lethality of sgs1⌬ slx1⌬ or sgs1⌬ slx4⌬ cells cannot be rescued ylation Unit, James Black Centre, University of Dundee, Dundee DD1 by deletion of RAD52 (2, 10). Slx4 has no obvious catalytic or 5EH, United Kingdom. Phone: 44-1382-385490. Fax: 44-1382-223778. structural motifs, apart from a cryptic SAP domain, but Slx1 E-mail: [email protected]. has a PHD-type zinc finger and is the founding member of a † Present address: Gurdon/CRUK Institute, University of Cam- bridge, Tennis Court Rd., Cambridge, United Kingdom. conserved family of nucleases defined by a UvrC-intron-endo- ᰔ Published ahead of print on 16 July 2007. nuclease (URI) domain (14).

6433 6434 FLOTT ET AL. MOL.CELL.BIOL.

While some of the cellular functions of Slx1 and Slx4 pro- genomic DNA of strain GA1701 (9). PEP4 disruption was found to curtail Slx4 teins are likely to overlap, given that these proteins interact degradation in native cell extracts. Strain SFY013 was constructed by replacing and are both required for viability in the absence of Sgs1, cells the KANMX cassette in strain SFY008 with a NAT resistance gene. To make plasmid pSLX4, the SLX4 open reading frame plus 1 kb of 5Ј sequence was lacking Slx4 are hypersensitive to DNA alkylation damage cloned into pRS413, and the SLX4 start ATG was mutated to an NcoI site. After whereas cells lacking Slx1 are not (5, 31). Furthermore, phos- digestion with NcoI, 13 copies of a MYC epitope tag were ligated into the NcoI phorylation of Esc4/Rtt107 is defective in cells lacking Slx4 but site. All mutations in SLX4 were introduced using the QuikChange site-directed not in cells lacking Slx1 (36). Moreover, Slx4 is required for mutagenesis kit (Stratagene). Mouse monoclonal antibodies against Myc (clone 9E10) and hemagglutinin (HA) were from Roche, and antibodies against Rad1 recovery from methyl methanesulfonate (MMS)-induced repli- were from Santa Cruz Biotechnology. Antibodies against Rad10 were a kind gift some stalling but Slx1 is not (36). Therefore, at least a subset from Errol Friedberg. of the cellular roles of Slx1 and Slx4 appears to be distinct. Analysis of chromosomes by PFGE. Cells were grown to early log phase Slx4 has been also shown to interact physically with proteins (optical density [OD600], 0.5) in yeast extract-peptone-dextrose (YPD) at 30°C ␣ ␮ other than Slx1. A genome-wide two-hybrid screen identified and arrested in G1 by addition of -factor (5 g/ml). When budded cells ac- counted for less than 5% of the population, cells were released from arrest by the Rad1 endonuclease as an Slx4 interactor (22). Rad1 cata- filtration and extensive washing and incubated in YPD for 10 min before addition lyzes DNA incision on the 5Ј side of UV-induced lesions and of MMS (0.05%). After 45 min in MMS, cells were filtered, washed extensively cleaves nonhomologous tails generated by DNA end resection with YPD containing 2.5% (wt/vol) sodium thiosulfate, and incubated in YPD at ϫ 8 during the SSA mode of DNA repair, responsible for repair of 30°C. At the times indicated, 1 10 cells were removed and fixed in 70% ethanol at 4°C overnight before preparation of chromosomes, exactly as de- double-strand breaks between repeated sequences (1, 11). scribed in the CHEF DRII instruction manual (Bio-Rad). Pulsed-field gel elec- However, it is not clear if the endogenous cellular form of Slx4 trophoresis (PFGE) was carried out using a Bio-Rad CHEF DRII apparatus at interacts with Rad1 or if this impacts on the function of either 14°C in a 1% agarose gel (pulsed-field certified; Bio-Rad) in 0.5ϫ Tris-borate- protein. Several groups reported that Slx4 interacts with Esc4/ EDTA for 24 h at 6 V/cm using a 120o included angle with a 6.8- to 158-s ␮ Rtt107 (6, 36, 51). Esc4 was originally identified in a screen for switch-time ramp. Gels were stained with 10 g/ml ethidium bromide for 30 min and washed for 2 min in water before DNA was visualized. genes that regulate retrotransposition in yeast (40) and in Measurement of spontaneous mutation frequency. The frequency of forward global genome screens for genes required for resistance to mutations at the CAN1 gene locus was determined by the frequency of appear- MMS (5, 17). Esc4 was subsequently shown to be required for ance of canavanine-resistant colonies that grew on selective minimal medium completion of chromosome replication after replisome stalling plates lacking Arg but containing canavanine (60 ␮g/ml) (for example, see reference 42a). Cultures were grown to stationary phase for 24 h in minimal (36, 37), but it is not yet known how it fulfills this task. Cells medium lacking Arg. The OD600 of cultures was measured, and from the same lacking Esc4 are hypersensitive to a wide range of agents that culture, in parallel, approximately 2 ϫ 107 cells were plated onto canavanine cause replisome stalling: camptothecin (CPT; causes S-phase- plates and 2 ϫ 102 cells were plated onto YPD plates. The colonies were counted specific DSBs and replisome collapse) and hydroxyurea (HU; after incubation at 30°C for 3 days. To calculate spontaneous mutation frequen- slows replication down by depleting deoxynucleoside triphos- cies, the number of canavanine-resistant colonies per ml of culture was divided by the number of CFU (on YPD) per ml of culture. Mutation frequencies phates), as well as MMS, whereas cells lacking Slx4 are not represented the average from three independent triplicate experiments, and the hypersensitive to CPT or HU, suggesting that Esc4 has cellular relative frequency was calculated from the increase or decrease in mutation roles not shared by Slx4. However, esc4⌬ and slx4⌬ cells show frequency in comparison to the wild-type strain. a comparable level of hypersensitivity to MMS and are epi- DNA combing. Wild-type and slx4⌬ cells containing the human nucleotide transporter hENT1 on a centromeric plasmid (pRS415) and seven copies of the static in this regard, suggesting that they share a similar role in herpes simplex thymidine kinase gene were synchronized in late G1 for 2.5 h with promoting resistance to MMS. 2 ␮g/ml ␣-factor (GenePep). Cells were released in S phase with 50 mg/ml Slx4 becomes phosphorylated in response to a wide range of Pronase (Calbiochem) in the presence of 30 ␮g/ml bromodeoxyuridine (BrdU) different types of DNA damage, at all cell cycle stages, and this and 0.05% MMS. After 60 min, MMS was quenched with sodium thiosulfate and ␮ requires both the Mec1 and Tel1 protein kinases (12). These cells were resuspended in fresh medium containing 30 g/ml BrdU. Genomic DNA was extracted in LMP agarose plugs (800 ng/plug) and was stained with kinases, the yeast orthologues of ATR and ATM in higher YOYO-1 (Molecular Probes). DNA was resuspended in 50 mM morpho- eukaryotes, respectively, belong to the PIKK family of kinases, lineethanesulfonic acid, pH 5.7, to a final concentration of 150 ng/ml. DNA fibers which also includes DNA-dependent protein kinase (DNA- were stretched on silanized coverslips as described elsewhere (29) and were PK) (45). Mec1 and Tel1 phosphorylate target proteins, in- denatured for 25 min in 1 N NaOH. BrdU was detected with a rat monoclonal cluding the Chk1 and Rad53 kinases, on Ser/Thr-Gln (S/T-Q) antibody (clone BU1/75; AbCys) and a secondary antibody coupled to Alexa 488 (Molecular Probes). DNA molecules were counterstained with an antiguanosine motifs and are critically important regulators of several aspects antibody (Argene) and an anti-mouse IgG coupled to Alexa 546 (Molecular of the cellular response to DNA damage and stalled repli- Probes). Images were recorded with a Leica DM6000B microscope coupled to a somes (45). The Slx4 residues phosphorylated by Mec1 and CoolSNAP HQ charge-coupled device camera (Roper Scientific) and were pro- Tel1 and the functional significance of Slx4 phosphorylation cessed as described previously (33). MetaMorph v6.2 (Universal Imaging Corp.) was used to measure BrdU signals and DNA fibers, and statistical analysis was are not yet known. Here we report that Slx4 has at least three performed with Prism 4 (GraphPad Software, Inc.). independent cellular functions that appear to require different Antibody production. All peptides were synthesized by Graham Bloomberg, Slx4-interacting proteins, and we show that one of these func- University of Bristol. Antibodies against phosphorylation sites in Slx4 were raised tions is regulated by Slx4 phosphorylation. at the Scottish Antibody Production Unit (Lanarkshire, Scotland) by immunizing sheep with the following peptides coupled to keyhole limpet hemocyanin: AQK SPMpTQETTKN (phospho-Thr-72), LDNQESpSQQRLWT (phospho-Ser- 289), and VNFLSLpSQVMDDK (phospho-Ser-329), where pS and pT are phos- MATERIALS AND METHODS pho-Ser and phospho-Thr, respectively. Antibodies were purified by affinity Yeast strains, plasmids, and antibodies. All strains used in this study are listed chromatography on CH-Sepharose to which the phosphopeptide immunogen in Table 1. Disruption of SLX4 was carried out by transforming the relevant had been covalently coupled. Phospho-specific antibodies were used at a final strains with the KAN-MX-disrupted SLX4 gene amplified from strain SFY008. concentration of 4 ␮g/ml in the presence of 50 ␮g/ml non-phospho peptide in a Gene disruption was tested where possible by screening for MMS hypersensitiv- Western blot analysis. ity and then verified by PCR. PEP4 disruption was achieved by transforming cells Miscellaneous methods. Western blotting of extracts prepared by the trichlo- with a PCR product corresponding to the LEU2-disrupted PEP4, amplified from roacetic acid lysis method (38), preparation of native cell extracts and Myc-Slx4 VOL. 27, 2007 PHOSPHORYLATION MODULATES Slx4 FUNCTION 6435

TABLE 1. Yeast strains used in this study

Strain Genotype Background Reference or source W303 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 rad5-535 43a BY4741 MATa his3⌬1 leu2⌬ met15⌬ ura3⌬ EUROSCARF SFY002 SLX4 (MYC13)::KANMX::slx4 BY4741 12 SFY006 SLX4 (MYC13)::KANMX::slx4 mec1⌬::TRP1 tel1⌬::HIS3 sml1-1 W303 12 SFY008 slx4⌬::KANMX BY4741 EUROSCARF SFY009 slx1⌬::KANMX BY4741 EUROSCARF SFY010 sgs1⌬::KANMX BY4741 EUROSCARF SFY012 rad52⌬::KANMX BY4742 EUROSCARF SFY013 slx4⌬::kanmx::NAT BY4741 This study SFY015 SLX4 (MYC13)::KANMX::slx4 SLX1(HA12)::URA3::slx1 pep4⌬::LEU2 BY4741 This study SFY016 mec1⌬::TRP1 tel1⌬::HIS3 pep4⌬::LEU2 SLX4(MYC13)::KANMX::slx4 sml1-1 W303 This study HP30 POL30 (HIS6)::LEU2 DF5 32a SFY018 POL30 (HIS6)::LEU2 slx4⌬::kanmx::NAT DF5 This study SFY019 POL30 (HIS6)::LEU rad18⌬::KANMX DF5 32a DF5 MATa ura3-52 his3-A 200 trp l-1 leu2-3,112 lys2-801 32a SFY020 mag1⌬::KANMX BY4741 EUROSCARF SFY021 mag1⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY022 apn1⌬::KANMX BY4741 EUROSCARF SFY023 apn1⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY024 rad14⌬::KANMX BY4741 EUROSCARF SFY025 rad14⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY026 rad6⌬::KANMX BY4741 EUROSCARF SFY027 rad6⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY028 rad18⌬::KANMX BY4741 EUROSCARF SFY029 rad18⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY030 mms2⌬::KANMX BY4741 EUROSCARF SFY031 mms2⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY034 rev3⌬::KANMX BY4741 EUROSCARF SFY035 rev3⌬::KANMX slx4⌬::kanmx::NAT BY4741 This study SFY036 pol30(k164r)::LEU DF5 32a SFY037 pol30(k164r)::LEU slx4⌬::kanmx::NAT DF5 This study SFY038 ubc13⌬::KANMX BY4741 EUROSCARF PP108 RAD5 TK W303 44a SFY063 slx4⌬::kanmx::NAT PP108 This study SFY060 rad1⌬::KANMX BY4741 EUROSCARF SFY062 SLX4 (MYC13)::KANMX::slx4 pep4::LEU2 BY4741 This study GTY001 rad1⌬::HIS3 slx1⌬::KANMX BY4741 This study GTY002 rad1⌬::HIS3 slx4⌬::kanmx::NAT slx1⌬::KANMX BY4741 This study GA1701 MATa rad53::TRP1 sml1::KanMX6 Sgs1-Myc13::HIS pep4::LEU2 ade2-1 trp1-1 9 his3-11,15 ura3-1 leu2-3,112 SFY047 CTG-0 YPH500 13 SFY048 CTG-0 slx4⌬::KANMX SFY047 This study SFY049 CTG-250 YPH500 13 SFY050 CTG-250 slx4⌬::KANMX SFY049 This study SFY066 CTG-250 mms2⌬::KANMX YPH500 This study SFY071 CTG-250 rad1⌬::HIS3 YPH500 13 NJY561 MAT␣ ade2-1 ade3::hisG ura3-1 his3-11,15 trp1-1 leu2-3,112 can1-100 W303 30 sgs1::loxP slx4::loxP ͓pJM500::SGS1 URA3 ADE3͔ EAY1141 mat::leu2::hisG hmr3 thr4 leu2 trp1 THR4-ura3-A(205bp)-HOcs-URA3-A 16 ade3::GAL10-HO::NAT GTY003 slx4⌬::KANMX EAY1141 This study

immunoprecipitation, measurement of SSA induced by a run of trinucleotide 1A). Cells were released from G1 arrest into S phase in the repeats placed between direct repeats (13), measurement of HO-induced SSA presence of MMS. After 45 min in MMS, all cells were still in (16) and ﬂuorescence-activated cell sorter (FACS) analysis (12) were all carried out as described previously. S phase as judged by FACS analysis (Fig. 1B), and since chromosomes were not completely replicated, they did not enter the gel (Fig. 1A, lanes 2, 7, 12, and 17) due to the presence of RESULTS forks and bubbles that impede chromosome migration (8, 18). Analysis of DNA replication during recovery from MMS- When cells were washed free of MMS, chromosome replica- induced replisome stalling in cells lacking Slx4. The major tion recovered and was almost complete after 5 hours in wild- DNA lesion induced by MMS, N3-methyl adenine, potently type cells (Fig. 1B), enabling chromosomes to enter the gel blocks replisome progression (3). Roberts et al. showed that (Fig. 1A). In contrast, there was a major defect in recovery of Slx4 is required for recovery of DNA replication after MMS- chromosome replication in slx4⌬ cells and in sgs1⌬ cells (46), induced replisome stalling (36). We obtained similar data (Fig. but not in cells lacking Slx1 (Fig. 1A). 6436 FLOTT ET AL. MOL.CELL.BIOL.

FIG. 1. Analysis of DNA replication during recovery from MMS-induced replisome stalling in slx4⌬ cells. (A) Strains BY4741 (wild type), ⌬ ⌬ ⌬ ␣ SFY008 (slx4 ), SFY009 (slx1 ), and SFY010 (sgs1 ) were grown to mid-log phase, arrested in G1 by the presence of -factor, and then released from G1 arrest into YPD. After 10 min, MMS (0.05%) was added for 45 min. Cells were then filtered, washed extensively, and incubated in YPD at 30°C for the times indicated. Samples for PFGE were taken before treatment with MMS (lanes 1, 6, 11, and 16), at 0 (lanes 2, 7, 12, and 17), 3 (lanes 3, 8, 13, and 18), 4 (lanes 4, 9, 14, and 19), and 5 (lanes 5, 10, 15, and 20) hours after cells were washed free of MMS. After PFGE, chromosomes were visualized by staining with ethidium bromide. (B) Same experiment as in panel A, except that samples were subjected to FACS ⌬ ␣ analysis after staining with propidium iodide. (C) Wild-type (PP108) and slx4 (SFY063) cells were arrested in G1 with -factor and were released into S phase in the presence of BrdU and MMS (0.05%). After 60 min, cells were washed and resuspended for the indicated times in fresh medium containing BrdU. Chromosomal DNA was stretched on silanized coverslips, and replicated tracks were detected with an anti-BrdU antibody (green). DNA fibers were counterstained with an antiguanosine antibody (red). Merged images (BrdU and DNA) and BrdU signals alone are shown for representative DNA fibers, with arrowheads pointing to unreplicated gaps. (D) Box plots of the length distribution of BrdU tracks. The Mann-Whitney rank-sum test was used to show that the differences between wild-type and slx4⌬ cells are significant (P Ͻ 0.001). (E) For each time point, DNA fibers and BrdU tracks were measured and the percentage of replication was determined for wild-type cells (filled circles) and slx4⌬ cells (open circles). Mean fiber length, which is indicative of the persistence of DSBs or fragile replication intermediates, was significantly shorter in slx4⌬ mutants. Mutant cells also show a 3.2-fold increase of DNA fibers with unreplicated gaps relative to wild type after 130 min of recovery. VOL. 27, 2007 PHOSPHORYLATION MODULATES Slx4 FUNCTION 6437

FIG. 1—Continued.

We wished to test if this lack of recovery from MMS-induced fibers isolated from slx4⌬ cells were significantly shorter than replisome stalling in the absence of Slx4 reflected an inability those isolated from wild-type cells (Fig. 1E), presumably be- to resume DNA synthesis or whether instead cells could re- cause they are more fragile and break during the combing sume DNA replication but could not complete it. Cells were process due to the persistence of unreplicated chromosomal released from ␣-factor arrest into S phase in the presence of gaps and stalled replisomes. Taken together, these data indi- MMS and then washed free of MMS. FACS analysis (Fig. 1B) cate that although cells lacking Slx4 restart DNA synthesis revealed that, like wild-type cells, slx4⌬ cells had an apparently when replisomes stall, DNA replication is slower than normal. 2C DNA content by 4 h postrecovery, indicating that the ma- Also, at the latest time point examined, chromosomes from jority of the chromosomes in these cells had been replicated. slx4⌬ cells are left with unreplicated gaps. This would account More than 90% of slx4⌬ cells were viable at this time point for the inability of chromosomes to enter pulsed-field gels (data not shown). DNA replication during recovery from (Fig. 1A). MMS-induced replisome stalling in slx4⌬ cells was investigated SLX4 interacts with genes involved in error-free DNA dam- further by DNA combing (28, 29). Cells were released from G1 age bypass genes. One possible explanation for the recovery into MMS-containing medium in the presence of BrdU. After defect described above could be that Slx4 promotes repair of 45 min, MMS was washed out and cells were allowed to re- DNA alkylation damage. Epistasis analysis with mutants de- cover for 90 min or 130 min in fresh medium in the continued fective in base excision repair (BER), the principal pathway for presence of BrdU. After combing, chromosome fibers were repair of DNA alkylation damage, was carried out. Cells lack- stained with anti-DNA antibodies (red) or anti-BrdU antibod- ing both SLX4 and methyladenine DNA glycosylase (MAG1), ies (green) (Fig. 1C). This analysis revealed striking defects a BER factor that cleaves methylated bases to leave an abasic associated with loss of Slx4. BrdU tracks were almost 50% site in DNA, are more sensitive to MMS than either of the shorter in slx4⌬ cells than in wild-type cells 90 and 130 min respective single mutants (Fig. 2A). Similar results were ob- after recovery (Fig. 1C and D), suggesting that DNA replica- tained with APN1 (Fig. 2A), which is also required for BER. tion during recovery from MMS is slower in slx4⌬ cells. Even We conclude that BER or nucleotide excision repair (NER) so, at 130 min after release of cells from MMS, more than 90% (data not shown) is unlikely to be the major function of SLX4. of the stalled replisomes had resumed DNA replication in the Alternatively, Slx4 could promote recovery from replisome absence of Slx4 (Fig. 1C and E). However, unreplicated gaps stalling by regulating bypass of fork-blocking lesions. Bypass were observed in around 26% of the fibers in slx4⌬ cells, requires Rad6/Rad18-dependent ubiquitination of Lys164 of compared with around 8% in wild-type cells (Fig. 1E). These proliferating cell nuclear antigen (PCNA) (34, 47). Conse- data are reminiscent of the defects seen in cells lacking the quently, cells in which Lys164 of PCNA (Pol30 in budding cullin Rtt101, which is also required for recovery from DNA yeast) is mutated to Arg are hypersensitive to MMS (21). alkylation damage (45). In rtt101⌬ cells, unreplicated gaps When SLX4 was disrupted in pol30 lys164arg mutant cells, the were detected in 22% of chromosome fibers, compared with double mutants were not more sensitive to MMS than the most only 2% in wild-type cells, and CldU tracks were 50% shorter sensitive single mutant (Fig. 2B, top panel). Consistent with as observed in slx4⌬ cells (45). We also noticed that DNA this, disruption of SLX4 did not further sensitize rad18⌬ or 6438 FLOTT ET AL. MOL.CELL.BIOL.

FIG. 2. SLX4 interacts with genes involved in DNA damage bypass. Strains BY4741 (wild type), SFY008 (slx4⌬), SFY022 (apn1⌬), SFY023 (apn1⌬ slx4⌬), SFY020 (mag1⌬), SFY021 (mag1⌬ slx4⌬), SFY036 (pol30k164r⌬), SFY037 (pol30k164r⌬ slx4⌬), SFY026 (rad6⌬), SFY027 (rad6⌬ slx4⌬), SFY028 (rad18⌬), SFY029 (rad18⌬ slx4⌬), SFY034 (rev3⌬), SFY035 (rev3⌬ slx4⌬), SFY030 (mms2⌬), and SFY031 (mms2⌬ slx4⌬) were grown to saturation in liquid culture, and 10-fold serial dilutions were spotted on YPD agar plates with or without the indicated amounts of MMS. Plates were incubated for 3 days at 30°C.

TABLE 2. Analysis of the frequency of spontaneous mutations ⌬ a rad6⌬ cells to MMS (Fig. 2B, lower panel). Therefore SLX4 in slx4 cells has an epistatic relationship with genes that regulate DNA Mutation Relative Strain damage bypass. frequency (106) frequency There are two major pathways in cells for bypassing DNA Wild type 2.0 Ϯ 0.4 1.0 lesions: TLS and error-free bypass (47). TLS occurs by tran- pol30(k164r) 3.1 Ϯ 0.3 1.5 sient recruitment to stalled replisomes of specialized TLS ubc13⌬ 17.2 Ϯ 2.9 8.5 ␦ rev3⌬ 1.1 Ϯ 0.1 0.5 DNA polymerases that, unlike the replicative polymerases ⌬ Ϯ ε slx4 10.2 4.7 5.1 and , can replicate across DNA lesions (47). Rad6/Rad18- slx4⌬ rev3⌬ 1.1 Ϯ 0.3 0.5 dependent mono-ubiquitination of PCNA at Lys164 is thought a Spontaneous forward mutation frequencies at the CAN1 gene locus were to recruit TLS polymerases to stalled replisomes. Error-free determined for strains BY4741, SFY008 (slx4⌬), SFY037 (pol30(k164r)), bypass requires the addition of further ubiquitin moieties to SFY034 (rev3⌬), SFY038 (ubc13⌬), and SFY035 (rev3⌬ slx4⌬) as described in mono-Ub PCNA, and this is catalyzed by the Rad5/Mms2/ Materials and Methods. Mutation frequencies represent the frequencies of canavanine-resistant colonies per CFU (presented as means with the standard errors Ubc13 complex (47). We sought to determine which of these of the means). The relative frequency shows the fold increase or decrease in bypass pathways is regulated by SLX4. The major error-prone mutation frequency compared to the wild-type strain. VOL. 27, 2007 PHOSPHORYLATION MODULATES Slx4 FUNCTION 6439

translesion polymerase in budding yeast is polymerase ␨ (Pol ␨), comprising Rev3, the catalytic subunit, and Rev7 (35, 47). The MMS sensitivity of slx4⌬ rev3⌬ cells was greater than that of the single mutants (Fig. 2C), suggesting that SLX4 is not involved in error-prone translesion synthesis and pointing instead to error-free bypass. Consistent with this, slx4⌬ mms2⌬ double mutants were not more sensitive to MMS (Fig. 2D) than the most sensitive single mutants. These data suggest that the inability of slx4⌬ cells to complete DNA synthesis after replisome stalling may be due to an inability to carry out error-free bypass. Translesion synthesis by Pol ␨ is responsible for 50 to 75% of spontaneous cell mutations, and so deletion of Rev3 decreases cell mutation frequencies (26). In contrast, mutations in error- free bypass factors increase spontaneous mutation frequency because of compensatory increases in TLS (4, 41). We reasoned that if Slx4 regulates error-free bypass, then Slx4 defi- ciency should increase the frequency of spontaneous mutation. Therefore, the frequency of forward mutation in the CAN1 gene, which gives rise to canavanine resistance, was measured. In the genetic background used in this study, deletion of REV3 decreased spontaneous mutation frequency, whereas disruption of UBC13 caused an approximately 8.5-fold increase in mutation frequency (Table 2), consistent with previous reports (4). Cells lacking Slx4 showed an approximately fivefold increase in the frequency of mutation, and this was abrogated by deletion of Rev3 (Table 2). Therefore, cells lacking Slx4 have a “mutator” phenotype that is caused by increased translesion synthesis, like cells lacking error-free bypass factors. This is consistent with, but does not prove, a role for Slx4 in error-free DNA damage avoidance. Slx4 interacts with Slx1 and with Rad1-Rad10 in a mutually exclusive manner. Slx4 does not have any catalytic motifs that could provide clues as to how it may participate in DNA processing reactions at blocked replisomes. However, Slx4 interacts with the Slx1 structure-specific endonuclease that can cleave branched structures, such as flaps, in vitro (14, 31). Genome-wide two-hybrid analysis revealed the Rad1 subunit of the Rad1-Rad10 endonuclease as a putative Slx4-interactor

left to recover for 30 min, respectively. Native lysates were prepared, and Myc-Slx4 immunoprecipitates or HA-Slx1 or anti-Rad1 precipitates were subjected to SDS-PAGE and Western blot analysis with the antibodies indicated. (B) Same experiment as described for Fig. 2, except that different (indicated) strains were used. (C) Top: a schematic diagram of the assay to measure SSA according to reference 13 is shown. Bottom: yeast strains SFY047 (CTG-0), SFY049 (CTG-250), SFY048 (slx4⌬ CTG-0), SFY050 (slx4⌬ CTG-250), SFY064 (rad1⌬ CTG-250), and SFY066 (mms2⌬ CTG-250) were grown to mid-log phase and plated on YPD agar and on plates containing FOA, according to the methods described in reference 13. Frequency of URA3 marker loss was determined as described in Materials and Methods. Average frequencies of FOA-resistant colonies per CFU and the standard deviations are shown. (D) HO endonuclease cleaves at its recognition site between 205-bp repeats of the ura3 sequence (boxes). The probe used for blotting is a HindIII-BamHI fragment as indicated. (E) FIG. 3. Slx4 interacts with Rad1-Rad10 and is required for SSA. The kinetics of SSA was monitored as described previously (43) at the

(A) Cells expressing Myc13-Slx4 and HA6-Slx4 (in which PEP4 was indicated times after HO induction by Southern blot hybridization of disrupted) or cells expressing Slx4 were grown to mid-exponential BglII-digested genomic DNA using the probe indicated in panel D. phase in liquid culture and left untreated or incubated with MMS The uncut SSA product and HO-cut bands are 8.3, 5.5, and 4.8 kb, (0.02%) for 30 min or exposed to ionizing radiation (IR; 180 Gy) and respectively. Two different SLX4-disrupted strains are shown. 6440 FLOTT ET AL. MOL.CELL.BIOL.

(22). We tested if cellular Slx4 could interact with Rad1 and if TABLE 3. Sequences surrounding residues in Slx4 phosphorylated Slx1 was found in the same complex. Endogenous Slx4 was by DNA-PK in vitro tagged at the C terminus with 13 copies of the Myc epitope in Peptide Sequence sitea Phosphorylation cells in which Slx1 was tagged at the N terminus with 12 copies of the HA epitope. Disruption of Sgs1 did not cause lethality in 1 SPM-TQ-ETTKNDTER Thr-72 2 NKDVDKSCNPVSTSHPDLGGSN Thr-113 this background and the cells were not more sensitive to MMS IEENIFIN-TQ-IQSR than wild-type cells, strongly suggesting that epitope tagging 3 GDSTS-SQ-EYGNGLEPQQPVG Ser-355 did not affect the function of Slx4 or Slx1 (data not shown). As NVVGEDIELAVGTR shown in Fig. 3B, Slx1, Rad1, and Rad10 were detected in 4 IRD-TQ-SAVNFLSL-SQ-VMDDK Thr-319, Ser-329 5 DNQES-SQ-QR Ser-289 anti-Myc (Slx4) immunoprecipitates before and after exposure of cells to MMS or ionizing radiation (IR). In contrast, Slx4 a Residues phosphorylated by DNA-PK are underlined. was detected in Slx1 immunoprecipitates, but Rad1 and Rad10 were not. Consistent with this observation, Slx4 and Rad10 were detected in anti-Rad1 immunoprecipitates but Slx1 was when TNRs (CTG-250) are located between the direct repeats, not. These data indicate that Slx4 interacts with Slx1 and with but not in the absence of TNRs (CTG-0). Disruption of Slx4, Rad1-Rad10 in a mutually exclusive manner. This is consistent however, like disruption of Rad1 (13), causes a severe reduc- with the previous observation that Slx1 (31), but not Rad1 (48), tion in SSA in this assay (Fig. 3C). However, disruption of is synthetically lethal with Sgs1. In addition, although both Slx4 MMS2 that regulates error-free bypass did not have a statisti- and Slx1 interact with Esc4 (36, 51), neither Rad1 nor Rad10 cally significant effect on marker loss (Fig. 3C, bottom), and was found in anti-Esc4 immunoprecipitates (Fig. 3A). Taken this demonstrates that marker loss in this assay is not due to together, these data demonstrate that Slx4 exists in cells in at sister strand slippage caused by template switching at the least two complexes: Slx1-Slx4-Esc4 and Slx4-Rad1-Rad10. TNRs. Furthermore, disruption of Slx1 had no effect on SSA It is easy to envisage how the catalytic activity of Rad1 and (Fig. 3C, bottom). This argues that Slx4 regulates Rad1-depen- Slx1 could be useful during error-free bypass. However, cells dent SSA independently of Slx1 and independently of its role lacking Slx1 (Fig. 1A) or Rad1 (data not shown) are not de- in error-free bypass. fective in recovery from MMS-induced replisome stalling and We tested the effect of disruption of Slx4 in a different assay are not hypersensitive to MMS (Fig. 3B). To rule out redun- for SSA that did not rely on TNRs for formation of DSBs. dancy between Slx1 and Rad1, a strain lacking both nucleases Instead, the HO endonuclease was induced to create a single was made. However, slx1⌬ rad1⌬ cells were not more hyper- DSB between 205-bp repeated segments of the URA3 gene sensitive to MMS than wild-type cells, and slx1⌬ rad1⌬ slx4⌬ (Fig. 3D) (43). Induction of a GAL::HO gene efficiently in- cells were not more hypersensitive to MMS than slx4⌬ cells duced SSA in wild-type cells and created deletion products (Fig. 3B). Therefore, the role of Slx4 in promoting cellular that could be monitored by Southern hybridization (Fig. 3E), resistance to MMS, which may involve regulating error-free and it is well established that this requires Rad1-Rad10 (11). bypass, appears to be independent of Slx1 and Rad1. Strikingly, disruption of Slx4 in this background also abol- Slx4 regulates single-strand annealing. The observation that ishes SSA. Taken together, the data in this section indicate Slx4 interacts physically with Rad1-Rad10 in cells prompted us that Slx4 interacts with Rad1-Rad10 and promotes repair of to test the role of Slx4 in Rad1-Rad10-dependent processes. DSBs by SSA. Rad1-Rad10 is involved in NER, but as slx4⌬ cells are not Phosphorylation of Slx4 by Mec1 and Tel1 regulates SSA. hypersensitive to UV (data not shown), Slx4 does not appear to DNA damage induces Mec1/Tel1-dependent phosphorylation be involved in NER. Rad1-Rad10 also plays an important role of Slx4 that is independent of Rad53 and Chk1 (12). Mec1/ during the SSA mode of DSB repair that can repair a DSB Tel1, ATR/ATM, and DNA-PK all have identical specificity in between repeated sequences (Fig. 3D). After resection of the vitro, and even though DNA-PK is not found in yeasts, it ends, complementary strands of the homologous regions flank- phosphorylates the same motifs in vitro in Slx4 as Mec1 and ing the DSB can anneal, producing an intermediate that has Tel1 (45). Since DNA-PK is readily available, we determined two nonhomologous 3Ј-ended tails. Rad1 cleaves the 3Ј single- the residues in Slx4 phosphorylated by DNA-PK in vitro as a strand tails, and the resulting nicks are sealed by ligation (1, starting point to ultimately find the residues phosphorylated in

11). Thus, SSA results in deletion of one copy of the repeat vivo by Mec1/Tel1. Incubation of recombinant His6-Slx4 with plus the sequence located between the repeats. DNA-PK resulted in phosphorylation of His6-Slx4 with a stoi- Zakian and colleagues reported a system for measuring SSA chiometry of approximately 6 moles of phosphate per mole of repair of DSBs induced by a hairpin placed between two direct Slx4 (data not shown). The sites phosphorylated in Slx4 were repeats (Fig. 3C, top) (13). When a run of 250 tandem CTG identified by a combination of mass spectrometry and Edman trinucleotide repeats (TNRs) and the URA3 gene are placed degradation (data not shown). These analyses revealed that between two direct repeats (Fig. 3C, top), the TNRs form a Slx4 is phosphorylated in vitro by DNA-PK on the following six hairpin that causes replisome stalling and DSB formation. residues: Thr72, Thr113, Ser289, Thr319, Ser329, and Ser355, These breaks are repaired by SSA in a manner dependent on and all of these conformed to the S/T-Q consensus sequence RAD1and RAD52, resulting in loss of the URA3 marker from (Table 3). between the direct repeats and, consequently, cells become To test whether Slx4 is phosphorylated on these residues in resistant to 5-fluoroorotic acid (5-FOA); no marker loss was vivo, phospho-specific antibodies were raised. Phosphopep- observed in the absence of CTG repeats (13). As shown in Fig. tides corresponding to Thr113, Ser355, and Thr319 failed to 3C (bottom), a high frequency of URA3 loss is observed in generate antibodies capable of recognizing the phosphopep- VOL. 27, 2007 PHOSPHORYLATION MODULATES Slx4 FUNCTION 6441

FIG. 4. Identification of Ser/Thr residues in Slx4 phosphorylated by Mec1 and Tel1 after DNA damage. (A) Schematic diagram of Slx4. Sites phosphorylated by DNA-PK are indicated. Asterisks denote all S/T-Q motifs. (B) Different amounts of phosphopeptide (“phospho”) corresponding to Slx4 Thr72, Ser289, and Ser329 and the corresponding non-phospho peptides (“nonphospho”) were spotted onto nitrocellulose, and Western blot analysis was carried out with the relevant phospho-specific antibody. (C to E) The yeast strains indicated were grown to mid-exponential phase in liquid culture and left untreated or incubated with MMS (0.02%) for 90 min (D and F) or with MMS (0.02%) or CPT (5 ␮g/ml) for 15 or 90 min, respectively (E). Native lysates were prepared, and Myc-Slx4 immunoprecipitates were subjected to SDS-PAGE and Western blot analysis with antibodies against Myc, phospho-Thr72, phospho-Ser289, or phospho-Ser329. Thr72 (high), long exposure time; Thr72 (low), short exposure time. tide immunogen (data not shown). In contrast, dot blot anal- which endogenous Slx4 was tagged at the C terminus with 13 ysis showed that affinity-purified phospho-specific antibodies copies of the Myc epitope and that lacked the Pep4 protease to raised against Thr72, Ser289, and Ser329 recognized the cor- curtail Slx4 degradation were prepared (12). Slx4-Myc was rect phosphopeptide immunogen but not the corresponding immunoprecipitated and subjected to sodium dodecyl sulfate- non-phospho peptide (Fig. 4B). Native extracts from cells in polyacrylamide gel electrophoresis (SDS-PAGE) followed by 6442 FLOTT ET AL. MOL.CELL.BIOL.

FIG. 5. Phosphorylation of Slx4 promotes SSA. (A) Wild-type cells (strain HP30) transformed with pRS413 (empty vector) or slx4⌬ cells (SFY018) transformed with empty vector or plasmids pSLX4 or pSLX4-MUT6 were grown to saturation in liquid culture. Tenfold serial dilutions, starting from an OD600 of 0.6, were spotted on YPD agar or onto YPD agar plates containing MMS. Cells were then incubated at 30°C for 3 days. (B) sgs1⌬ slx4⌬ [pSGS1-URA3] cells (strain NJY561) (30) were transformed with the following HIS3 plasmids: empty plasmid (pRS413), pSLX4, or pSLX4-MUT6. Cells were then restreaked on YPD agar or on agar containing 5-FOA. Plates were incubated for 3 days at 30°C. (C) Same experiment as in Fig. 3C except that strains SFY049 (CTG-250) transformed with pRS413 (empty vector) and SFY050 (slx4⌬ CTG-250) transformed with pRS413 (empty vector), pSLX4, pSLX4-MUT3, or pSLX4-MUT6 were used.

Western blotting with the Slx4 phospho-speciﬁc antibodies. in Fig. 4E, no phosphorylation of Thr72, Ser289, and Ser329 This showed clearly that Slx4 is phosphorylated on Thr72, was observed in mec1⌬ tel1⌬ cells. Phosphorylation of these Ser289, and Ser329 after exposure of cells to MMS (Fig. 4C). residues occurred at wild-type levels in cells lacking either Phosphorylation of Slx4 on Thr72 was evident 15 min after Mec1 or Tel1 and in cells lacking both Rad53 and Chk1 (data MMS treatment and was much stronger after 90 min (Fig. 4D). not shown). Thus, both Mec1 and Tel1 phosphorylate Slx4 on In addition, Thr72 became phosphorylated after exposure of several S/T-Q motifs after DNA damage in vivo. cells to CPT, which causes double-strand breaks during S To test the potential role of Mec1/Tel1-mediated phosphor- phase (Fig. 4D). Slx4-Myc was then immunoprecipitated from ylation in regulating Slx4, different combinations of point mu- cells lacking both Mec1 and Tel1, and precipitates were sub- tations were introduced into the SLX4 gene expressed under jected to SDS-PAGE followed by Western blotting. As shown the control of its own promoter from a low-copy-number plas- VOL. 27, 2007 PHOSPHORYLATION MODULATES Slx4 FUNCTION 6443 mid. Initially, the three residues in Slx4 that were shown to be bility in sgs1⌬ cells and is not required for cellular resistance to phosphorylated by Mec1 and Tel1 in vivo—Thr72, Ser289, and MMS. Ser329 (Fig. 4C)—were all mutated to alanines, resulting in The inability of slx4⌬ cells to recover from MMS-induced pSLX4-MUT3. In another mutant, these three sites plus the replisome stalling (Fig. 1A) (36) prompted us to investigate if three other sites phosphorylated in vitro by DNA-PK, Thr113, DNA replication could resume in these cells during recovery or Thr319, and Ser355, were mutated to Ala (pSLX4-MUT6). whether problems arose instead in completing S phase. We When these plasmids and wild-type pSLX4 were introduced found that although DNA replication can resume after repli- into slx4⌬ cells, the Slx4 mutants were expressed at similar some stalling in cells lacking Slx4 (Fig. 1B), DNA synthesis is levels to wild-type Slx4 (data not shown). slow and there is a major defect in the completion of replica- We tested if the Slx4 phosphorylation site mutants could tion, resulting in unreplicated gaps in the genome (Fig. 1C to rescue the MMS hypersensitivity of slx4⌬ cells. Cells expressing E). Precisely why slx4⌬ cells have problems in completing Slx4-Mut3 (data not shown) or Slx4-Mut6 (Fig. 5A) were no DNA replication is not yet clear. However, in this study we more sensitive to MMS than cells expressing wild-type Slx4. found that SLX4 interacts with genes involved in error-free However, it was possible that phosphorylation of Slx4 by Mec1/ DNA damage bypass, and several lines of evidence suggest that Tel1 at S/T-Q sites other than those mutated in pSLX4-MUT6 Slx4 might regulate this process. Firstly, Slx4 is epistatic with may have provided resistance to MMS. This is unlikely, since error-free bypass genes in terms of MMS hypersensitivity (Fig. mutation of 16 of the total 18 S/T-Q motifs in Slx4 had no 2). Secondly, cells lacking a known error-free DNA damage effect on the ability of Slx4 to rescue the MMS hypersensitivity bypass factor, Mms2, have a similar recovery defect to that of slx4⌬ cells (data not shown). Thus, it is highly unlikely that seen in slx4⌬ cells (data not shown). Thirdly, the spontaneous Mec1/Tel1-dependent phosphorylation of Slx4 is required for mutation frequency is elevated in slx4⌬ cells, caused by a com- its ability to promote resistance to MMS or, therefore, error- pensatory increase in TLS. This also occurs in known error- free bypass. However, we reasoned that phosphorylation of free bypass mutants (Table 2) and probably masks the real Slx4 may affect an aspect of its function other than cellular severity of the consequences of disrupting Slx4, especially given resistance to MMS, such as SSA or the ability to maintain cell the additive effect of disrupting REV3 in slx4⌬ cells (Fig. 2C). viability in the absence of Sgs1. Therefore, we investigated the Error-free bypass is thought to involve template switching role of phosphorylation in regulating these aspects of Slx4 initiated by unpairing and annealing of the blocked nascent function. strand with the intact nascent sister strand, but the mechanisms Cells lacking both sgs1⌬ and slx4⌬ are inviable but can be and gene products involved are not clear. After replication past kept alive by a low-copy-number [pSGS1-URA3-ADE] plas- the blockage, the nascent strands would unpair again and re- mid expressing SGS1. These cells are sensitive to killing by anneal with the parental strands (20, 39). There is experimen- FOA, which is converted to toxic intermediates by Ura3, since tal evidence to support template switching at blocked repli- cells cannot lose the SGS1 plasmid (30, 31). sgs1⌬ slx4⌬ ade2 somes in yeast (53), but since the mechanisms involved are ade3 [pSGS1-URA3-ADE] cells (strain NJY561) (5) were unclear, as are many of the relevant genes, it is difﬁcult to transformed with the following HIS3 plasmids: empty plasmid speculate about the potential role of Slx4. Template switching (pRS413), pSLX4, or pSLX4-MUT6. Cells were then re- should involve the action of helicases and nucleases, and it is streaked to YPD plates with or without 5-FOA. Cells trans- possible that Slx4 recruits one or more catalytic activities to formed with empty plasmid (pRS413) are FOA sensitive and replisomes blocked by DNA lesions. In this light, it is interest- red (since they cannot lose the ADE3 gene). Introduction of ing that Slx4 interacts with Slx1 that in vitro cleaves structures SLX4 on a low-copy-number plasmid allowed sgs1⌬ slx4⌬ that resemble those that may arise during error-free bypass. [pSGS1-URA3-ADE] cells to grow on FOA (Fig. 5B) and However, Slx1 is not required for cellular resistance to MMS. caused red-white sectoring on low-adenine plates (data not Genome-wide analysis revealed the Rad1 subunit of the Rad1- shown), since cells could now lose the SGS1-URA3 plasmid. Rad10 nuclease as a potential Slx4 interactor, and in this study Introduction of pSLX4-MUT6, like wild-type SLX4, resulted we demonstrated that cellular Slx4 interacts with Rad1-Rad10 in FOA resistance (Fig. 5B). This indicates Slx4 phosphoryla- (Fig. 3A). However, rad1⌬ cells recover normally from MMS tion is not essential for viability in sgs1⌬ cells. We next tested (data not shown) and are not MMS sensitive (Fig. 3B). It is the impact of Slx4 phosphorylation on SSA. Whereas wild-type unlikely, therefore, that Rad1 and Slx1 function redundantly in Slx4 could rescue the decrease in SSA observed in slx4⌬ cells, error-free bypass since slx1⌬ rad1⌬ cells are not hypersensitive the Slx4-Mut3 or Slx4-Mut6 phospho site mutants could not to MMS (Fig. 3B). (Fig. 5C). This is consistent with previous observations that Slx4 interacts with Esc4, which is also required for recovery cells lacking Mec1 show reduced SSA (24). Taken together, the from replisome stalling (36, 37), but unlike Slx4, disruption of data in this section indicate that Mec1/Tel1-mediated phos- Esc4 causes sickness but not lethality in sgs1⌬ cells (51). Cells phorylation of Slx4 regulates SSA but not cell viability in the lacking Esc4 are hypersensitive to CPT and HU (44), whereas absence of Sgs1 or cellular resistance to MMS. cells lacking Slx4 are not (data not shown), suggesting that Slx4 and Esc4 have distinct functions. However, esc4⌬ and slx4⌬ DISCUSSION cells show a comparable level of hypersensitivity to MMS and In this study we showed that Slx4 has at least three appar- are epistatic in this regard (6, 36). In addition, Slx4 is required ently independent cellular functions: protecting cells when for phosphorylation of Esc4 by Mec1 (36), indicating that the Sgs1 is not functional, promoting cellular resistance to MMS, Slx4-Esc4 interaction is functionally important. It remains to and SSA. We also demonstrated that phosphorylation of Slx4 be determined how the association with Esc4 impacts on Slx4 by Mec1 and Tel1 regulates SSA but is not required for via- function in promoting resistance to MMS. It will be of partic- 6444 FLOTT ET AL. MOL.CELL.BIOL.

in the absence of Sgs1 and in promoting resistance to MMS (Fig. 6). The Slx4-Rad1-Rad10 and Slx4-Slx1-Esc4 complexes appear to be functionally distinct, and phosphorylation of Slx4 by Mec1 and Tel1 appears to be required for the function of one of these complexes: Slx4-Rad1-Rad10, which mediates SSA. It will be vitally important to investigate the molecular basis for modulation of Slx4 function by phosphorylation. It is unlikely that phosphorylation promotes the interaction of Slx4 with Rad1-Rad10, since Slx4 interacts with these proteins even in the absence of DNA damage (Fig. 3A) (22, 31). However, it FIG. 6. Model for cellular roles of Slx4. Slx4 exists in at least two may be that phosphorylation modulates recruitment of Slx4- mutually exclusive, functionally distinct Slx4-containing complexes. Rad1-Rad10 to stalled replisomes or sites of DNA damage, Esc4-Slx4-Slx1 protects cells in the absence of Sgs1 but is not required for SSA. The Slx4-Rad1-Rad10 complex, on the other hand, promotes and this will be interesting to investigate. SSA. The two Slx4-containing complexes appear to be distinct in that Based on the data in this study, we postulate the existence of deletion of RAD1 is not synthetically lethal with deletion of SGS1 (4), at least two mutually exclusive, functionally distinct Slx4-con- and Rad1 is not required for resistance to, or for recovery from, MMS taining complexes (Fig. 6A). Esc4-Slx4-Slx1 protects cell via- (Fig. 3B and data not shown). Similarly, Slx1 is not required for recovery from MMS (Fig. 1A) or for SSA (Fig. 3C). It is clear that Slx4 bility in the absence of Sgs1 but is not required for SSA. The is required for recovery from MMS-induced replisome stalling, possi- Slx4-Rad1-Rad10 complex, on the other hand, promotes SSA. bly by regulating error-free bypass, but it is not yet clear which, if any, The two Slx4-containing complexes appear to be distinct, in of the known Slx4-interacting proteins is responsible. Since SLX4 is that deletion of RAD1 is not synthetically lethal with deletion epistatic with ESC4 in terms of MMS hypersensitivity, it is likely that of SGS1 (48) but deletion of SLX1 is (31). In addition, Rad1- this subset of the Esc4-Slx4-Slx1 complex is important. Alternatively, one or more as-yet-unidentified Slx4-interacting proteins may help Rad10 but not Slx1 is required for SSA (Fig. 3C and E). The promote resistance to MMS. Mec1 and Tel1 together phosphorylate ability of Slx4 to promote cellular resistance to MMS probably Slx4, and this is required for efficient SSA but not for viability in the involves the Esc4-Slx4-Slx1 complex: even though Slx1 is not absence of Sgs1, for cellular resistance to MMS, or for completion of required for recovery from MMS (Fig. 1A), Esc4 and Slx4 are DNA replication when replisomes stall. The precise molecular basis for the modulation of Slx4 function by phosphorylation is not yet epistatic with regard to MMS hypersensitivity (36). Alterna- known. tively, one or more as-yet-unidentified Slx4-interacting proteins may fulfill this task. It will be important to identify such proteins, to examine the dynamics of Slx4-containing com- ular importance to examine the ability of these proteins to plexes and how Slx4 is distributed among them, and to study associate with stalled replisomes and how this is regulated. how Slx4 phosphorylation influences Slx4 function at the mo- Rad1-Rad10 is essential for SSA (1, 11, 13), and because we lecular level. found that Slx4 interacts with this complex, we tested if Slx4 ACKNOWLEDGMENTS regulates SSA. Indeed, slx4⌬ cells showed a severe reduction in We thank Helle Ulrich, Michael Fasullo, Virginia Zakian, and Rad1-Rad10-dependent SSA in two different assays (Fig. 3). It EUROSCARF for yeast strains and Steve Brill for useful reagents and is not yet clear why deletion of Slx4 has such a profound effect advice. We are grateful to Susan Lees-Miller for the kind gift of on SSA but could reflect a role in Rad1-mediated removal of DNA-PK. We thank the Antibody Production Team and James Hastie nonhomologous tails, in annealing of the repeats, or in direct- and Hilary MacLauchlan in the Division of Signal Transduction Ther- apy, University of Dundee, for help with raising and purifying anti- ing new DNA synthesis prior to ligation of the ends. Since the bodies and the DNA sequencing group for technical assistance. We are Slx4-Slx1 complex also cleaves flaps and branched structures, it grateful to Tony Carr, Anton Gartner, and members of the Rouse is tempting to speculate that Slx4 recognizes and binds to these laboratory for useful discussions and to S. Gasser for technical advice. structures in a manner that somehow facilitates cleavage by S.F. was funded by a predoctoral fellowship from the Medical Re- associated enzymes. However, Slx1 does not appear to be in- search Council (MRC) UK, and work in the Rouse lab is funded by the MRC, the Association for International Cancer Research, and an volved in SSA, and the Slx4-Slx1 and Slx4-Rad1-Rad10 com- EMBO Young Investigator Award. plexes appear to be distinct. It is not, therefore, clear how the different complexes recognize the appropriate lesions. Bio- REFERENCES chemical analysis of these complexes and the identification of 1. Bardwell, L., A. J. Cooper, and E. C. Friedberg. 1992. Stable and specific association between the yeast recombination and DNA repair proteins separation-of-function mutants may help to elucidate this RAD1 and RAD10 in vitro. Mol. Cell Biol. 12:3041–3049. problem. 2. Bastin-Shanower, S. A., W. M. Fricke, J. R. Mullen, and S. J. Brill. 2003. The mechanism of Mus81-Mms4 cleavage site selection distinguishes it from the In this study we identified six residues in Slx4 phosphory- homologous endonuclease Rad1-Rad10. Mol. Cell. Biol. 23:3487–3496. lated by DNA-PK in vitro. Phospho-specific antibodies dem- 3. Beranek, D. T. 1990. Distribution of methyl and ethyl adducts following onstrated clearly that at least three of these sites—Thr72, alkylation with monofunctional alkylating agents. Mutat. Res. 231:11–30. 4. Brusky, J., Y. Zhu, and W. Xiao. 2000. UBC13, a DNA-damage-inducible Ser289, and Ser329—become phosphorylated in vivo in re- gene, is a member of the error-free postreplication repair pathway in Sac- sponse to DNA damage but not in cells lacking both Mec1 and charomyces cerevisiae. Curr. Genet. 37:168–174. Tel1 kinases. Mutation of these three sites in Slx4 phosphory- 5. Chang, M., M. Bellaoui, C. Boone, and G. W. Brown. 2002. A genome-wide screen for methyl methanesulfonate-sensitive mutants reveals genes required lated in vivo after DNA damage does not affect cell viability in for S phase progression in the presence of DNA damage. Proc. Natl. Acad. the presence of MMS or viability in cells lacking Sgs1 but does Sci. USA 99:16934–16939. 6. Chin, J. K., V. I. Bashkirov, W. D. Heyer, and F. E. Romesberg. 2006. inhibit Rad1-dependent SSA. This suggests that the role of Esc4/Rtt107 and the control of recombination during replication. DNA Slx4 in SSA is distinct from its role in promoting cell viability Repair (Amsterdam) 5:618–628. VOL. 27, 2007 PHOSPHORYLATION MODULATES Slx4 FUNCTION 6445

7. Coulon, S., P. H. Gaillard, C. Chahwan, W. H. McDonald, J. R. Yates III, 32a.Papouli, E., S. Chen, A. A. Davies, D. Huttner, L. Krejei, P. Sung, and H. D. and P. Russell. 2004. Slx1-Slx4 are subunits of a structure-specific endonu- Ulrich. 2005. Crosstalk between SUMO and ubiquitin on PCNA is mediated clease that maintains ribosomal DNA in fission yeast. Mol. Biol. Cell 15: by recruitment of the helicase Srs2p. Mol. Cell 19:123–133. 71–80. 33. Pasero, P., A. Bensimon, and E. Schwob. 2002. Single-molecule analysis 8. Desany, B. A., A. A. Alcasabas, J. B. Bachant, and S. J. Elledge. 1998. reveals clustering and epigenetic regulation of replication origins at the yeast Recovery from DNA replicational stress is the essential function of the rDNA locus. Genes Dev. 16:2479–2484. S-phase checkpoint pathway. Genes Dev. 12:2956–2970. 34. Prakash, L. 1981. Characterization of postreplication repair in Saccharomy- 9. Duncker, B. P., K. Shimada, M. Tsai-Pflugfelder, P. Pasero, and S. M. ces cerevisiae and effects of rad6, rad18, rev3 and rad52 mutations. Mol. Gasser. 2002. An N-terminal domain of Dbf4p mediates interaction with Gen. Genet. 184:471–478. both origin recognition complex (ORC) and Rad53p and can deregulate late 35. Prakash, S., R. E. Johnson, and L. Prakash. 2005. Eukaryotic translesion origin firing. Proc. Natl. Acad. Sci. USA 99:16087–16092. synthesis DNA polymerases: specificity of structure and function. Annu. Rev. 10. Fabre, F., A. Chan, W. D. Heyer, and S. Gangloff. 2002. Alternate pathways Biochem. 74:317–353. involving Sgs1/Top3, Mus81/ Mms4, and Srs2 prevent formation of toxic 36. Roberts, T. M., M. S. Kobor, S. A. Bastin-Shanower, M. Ii, S. A. Horte, J. W. recombination intermediates from single-stranded gaps created by DNA Gin, A. Emili, J. Rine, S. J. Brill, and G. W. Brown. 2006. Slx4 regulates replication. Proc. Natl. Acad. Sci. USA 99:16887–16892. DNA damage checkpoint-dependent phosphorylation of the BRCT domain 11. Fishman-Lobell, J., and J. E. Haber. 1992. Removal of nonhomologous protein Rtt107/Esc4. Mol. Biol. Cell 17:539–548. DNA ends in double-strand break recombination: the role of the yeast 37. Rouse, J. 2004. Esc4p, a new target of Mec1p (ATR), promotes resumption ultraviolet repair gene RAD1. Science 258:480–484. of DNA synthesis after DNA damage. EMBO J 23:1188–1197. 12. Flott, S., and J. Rouse. 2005. Slx4 becomes phosphorylated after DNA 38. Rouse, J., and S. P. Jackson. 2000. LCD1: an essential gene involved in damage in a Mec1/Tel1-dependent manner and is required for repair of checkpoint control and regulation of the MEC1 signalling pathway in Sac- DNA alkylation damage. Biochem. J. 391:325–333. charomyces cerevisiae. EMBO J. 19:5801–5812. 13. Freudenreich, C. H., S. M. Kantrow, and V. A. Zakian. 1998. Expansion and 39. Rupp, W. D., C. E. Wilde III, D. L. Reno, and P. Howard-Flanders. 1971. length-dependent fragility of CTG repeats in yeast. Science 279:853–856. Exchanges between DNA strands in ultraviolet-irradiated Escherichia coli. J. 14. Fricke, W. M., and S. J. Brill. 2003. Slx1-Slx4 is a second structure-specific Mol. Biol. 61:25–44. endonuclease functionally redundant with Sgs1-Top3. Genes Dev. 17:1768– 40. Scholes, D. T., M. Banerjee, B. Bowen, and M. J. Curcio. 2001. Multiple 1778. regulators of Ty1 transposition in Saccharomyces cerevisiae have conserved 15. Gangloff, S., J. P. McDonald, C. Bendixen, L. Arthur, and R. Rothstein. roles in genome maintenance. Genetics 159:1449–1465. 1994. The yeast type I topoisomerase Top3 interacts with Sgs1, a DNA 41. Schurer, K. A., C. Rudolph, H. D. Ulrich, and W. Kramer. 2004. Yeast helicase homolog: a potential eukaryotic reverse gyrase. Mol. Cell. Biol. MPH1 gene functions in an error-free DNA damage bypass pathway that 14:8391–8398. requires genes from Homologous recombination, but not from postreplica- 16. Goldfarb, T., and E. Alani. 2005. Distinct roles for the Saccharomyces cer- tive repair. Genetics 166:1673–1686. evisiae mismatch repair proteins in heteroduplex rejection, mismatch repair 42. Spell, R. M., and S. Jinks-Robertson. 2004. Examination of the roles of Sgs1 and nonhomologous tail removal. Genetics 169:563–574. and Srs2 helicases in the enforcement of recombination fidelity in Saccha- 17. Hanway, D., J. K. Chin, G. Xia, G. Oshiro, E. A. Winzeler, and F. E. romyces cerevisiae. Genetics 168:1855–1865. Romesberg. 2002. Previously uncharacterized genes in the UV- and MMS- 42a.Stelter, P., and H. D. Ulrich. 2003. Control of spontaneous and damage- induced DNA damage response in yeast. Proc. Natl. Acad. Sci. USA 99: induced mutagenesis by SUMO and ubiquitin conjugation. Nature 425:188– 10605–10610. 191. 18. Hennessy, K. M., A. Lee, E. Chen, and D. Botstein. 1991. A group of 43. Sugawara, N., T. Goldfarb, B. Studamire, E. Alani, and J. E. Haber. 2004. interacting yeast DNA replication genes. Genes Dev. 5:958–969. Heteroduplex rejection during single-strand annealing requires Sgs1 helicase 19. Hickson, I. D. 2003. RecQ helicases: caretakers of the genome. Nat. Rev. and mismatch repair proteins Msh2 and Msh6 but not Pms1. Proc. Natl. Cancer 3:169–178. Acad. Sci. USA 101:9315–9320. 20. Higgins, N. P., K. Kato, and B. Strauss. 1976. A model for replication repair 43a.Thomas, B. J., and R. Rothstein. 1989. The genetic control of direct-repeat in mammalian cells. J. Mol. Biol. 101:417–425. recombination in Saccharomyces: the effect of rad52 and rad1 on mitotic 21. Hoege, C., B. Pfander, G. L. Moldovan, G. Pyrowolakis, and S. Jentsch. 2002. recombination at GAL10, a transcriptionally regulated gene. Genetics 123: RAD6-dependent DNA repair is linked to modification of PCNA by ubiq- 725–738. uitin and SUMO. Nature 419:135–141. Tong, A. H., M. Evangelista, A. B. Parsons, H. Xu, G. D. Bader, N. Page, M. 22. Ito, T., T. Chiba, R. Ozawa, M. Yoshida, M. Hattori, and Y. Sakaki. 2001. A 44. Robinson, S. Raghibizadeh, C. W. Hogue, H. Bussey, B. Andrews, M. Tyers, comprehensive two-hybrid analysis to explore the yeast protein interactome. and C. Boone. Proc. Natl. Acad. Sci. USA 98:4569–4574. 2001. Systematic genetic analysis with ordered arrays of yeast 294: 23. Kaliraman, V., J. R. Mullen, W. M. Fricke, S. A. Bastin-Shanower, and S. J. deletion mutants. Science 2364–2368. Brill. 2001. Functional overlap between Sgs1-Top3 and the Mms4-Mus81 44a.Tourriere, H., G. Versini, V. Cordon-Preciado, C. Alabert, and P. Pasero. endonuclease. Genes Dev. 15:2730–2740. 2005. Mre1 and Tof1 promote replication fork progression and recovery 24. Kaye, J. A., J. A. Melo, S. K. Cheung, M. B. Vaze, J. E. Haber, and D. P. independently of Rad53. Mol. Cell. 19:699–706. Toczyski. 2004. DNA breaks promote genomic instability by impeding 45. Traven, A., and J. Heierhorst. 2005. SQ/TQ cluster domains: concentrated proper chromosome segregation. Curr. Biol. 14:2096–2106. ATM/ATR kinase phosphorylation site regions in DNA-damage-response 25. Khakhar, R. R., J. A. Cobb, L. Bjergbaek, I. D. Hickson, and S. M. Gasser. proteins. Bioessays 27:397–407. 2003. RecQ helicases: multiple roles in genome maintenance. Trends Cell 46. Ui, A., M. Seki, H. Ogiwara, R. Onodera, S. Fukushige, F. Onoda, and T. Biol. 13:493–501. Enomoto. 2005. The ability of Sgs1 to interact with DNA topoisomerase III 26. Lawrence, C. W., and V. M. Maher. 2001. Mutagenesis in eukaryotes depen- is essential for damage-induced recombination. DNA Repair (Amsterdam) dent on DNA polymerase zeta and Rev1p. Philos. Trans. R. Soc. London B 4:191–201. 356:41–46. 47. Ulrich, H. D. 2005. The RAD6 pathway: control of DNA damage bypass and 27. Liberi, G., G. Maffioletti, C. Lucca, I. Chiolo, A. Baryshnikova, C. Cotta- mutagenesis by ubiquitin and SUMO. ChemBiochem 6:1735–1743. Ramusino, M. Lopes, A. Pellicioli, J. E. Haber, and M. Foiani. 2005. Rad51- 48. Watt, P. M., I. D. Hickson, R. H. Borts, and E. J. Louis. 1996. SGS1, a dependent DNA structures accumulate at damaged replication forks in sgs1 homologue of the Bloom’s and Werner’s syndrome genes, is required for mutants defective in the yeast ortholog of BLM RecQ helicase. Genes Dev. maintenance of genome stability in Saccharomyces cerevisiae. Genetics 144: 19:339–350. 935–945. 28. Luke, B., G. Versini, M. Jaquenoud, I. W. Zaidi, T. Kurz, L. Pintard, P. 49. Wu, L., and I. D. Hickson. 2003. The Bloom’s syndrome helicase suppresses Pasero, and M. Peter. 2006. The cullin Rtt101p promotes replication fork crossing over during homologous recombination. Nature 426:870–874. progression through damaged DNA and natural pause sites. Curr. Biol. 50. Yamagata, K., J. Kato, A. Shimamoto, M. Goto, Y. Furuichi, and H. Ikeda. 16:786–792. 1998. Bloom’s and Werner’s syndrome genes suppress hyperrecombination 29. Michalet, X., R. Ekong, F. Fougerousse, S. Rousseaux, C. Schurra, N. in yeast sgs1 mutant: implication for genomic instability in human diseases. Hornigold, M. van Slegtenhorst, J. Wolfe, S. Povey, J. S. Beckmann, and A. Proc. Natl. Acad. Sci. USA 95:8733–8738. Bensimon. 1997. Dynamic molecular combing: stretching the whole human 51. Zappulla, D. C., A. S. Maharaj, J. J. Connelly, R. A. Jockusch, and R. genome for high-resolution studies. Science 277:1518–1523. Sternglanz. 2006. Rtt107/Esc4 binds silent chromatin and DNA repair pro- 30. Mullen, J. R., V. Kaliraman, and S. J. Brill. 2000. Bipartite structure of the teins using different BRCT motifs. BMC Mol. Biol. 7:40. SGS1 DNA helicase in Saccharomyces cerevisiae. Genetics 154:1101–1114. 52. Zhang, C., T. M. Roberts, J. Yang, R. Desai, and G. W. Brown. 2006. 31. Mullen, J. R., V. Kaliraman, S. S. Ibrahim, and S. J. Brill. 2001. Require- Suppression of genomic instability by SLX5 and SLX8 in Saccharomyces ment for three novel protein complexes in the absence of the Sgs1 DNA cerevisiae. DNA Repair (Amsterdam) 5:336–346. helicase in Saccharomyces cerevisiae. Genetics 157:103–118. 53. Zhang, H., and C. W. Lawrence. 2005. The error-free component of the 32. Myung, K., A. Datta, C. Chen, and R. D. Kolodner. 2001. SGS1, the Sac- RAD6/RAD18 DNA damage tolerance pathway of budding yeast em- charomyces cerevisiae homologue of BLM and WRN, suppresses genome ploys sister-strand recombination. Proc. Natl. Acad. Sci. USA 102:15954– instability and homeologous recombination. Nat. Genet. 27:113–116. 15959.