Nej1p, a cell type-specific regulator of nonhomologous end joining in yeast

Brief Communication 1611

Nej1p, a cell type-specific regulator of nonhomologous end joining in yeast Andreas Kegel, Jimmy O. O. Sjo¨strand and Stefan U. A˚stro¨m Mutant yeast strains lacking the silencing proteins Sir2p, Sir3p, or Sir4p have a defect in a DNA doublestrand break (DSB) repair pathway, called nonhomologous end joining (NHEJ) [1, 2]. Mutations in sir genes also lead to the simultaneous expression of a and ␣ mating type information, thus generating a nonmating haploid cell type [3] with many properties shared with a/␣ diploids. We addressed whether cell type or Sir proteins per se regulate NHEJ by investigating the role of a novel haploid-specific gene in NHEJ. This gene, NEJ1, was required for efficient NHEJ, and transcription of NEJ1 was completely repressed in a/␣ diploid [4] and sir haploid strains. The NEJ1 promoter contained a consensus binding site for the a1/␣2 repressor, explaining the cell type-specific expression. Expression of Nej1p from a constitutive promoter in a/␣ diploid and sir mutant strains completely rescued the defect in NHEJ, thus showing that Sir proteins per se were dispensable for NHEJ. Nej1p and Lif1P [5], the yeast XRCC4 homolog, interacted in two independent assays, and Nej1p localized to the nucleus, suggesting that Nej1p may have a direct role in NHEJ. Address: Umea˚ Center for Molecular Pathogenesis, Umea˚ University, SE-901 87 Umea˚, Sweden. Correspondence: Stefan U. A˚stro¨m E-mail: [email protected] Received: 16 June 2001 Revised: 2 August 2001 Accepted: 3 September 2001 Published: 16 October 2001 Current Biology 2001, 11:1611–1617 0960-9822/01/$ – see front matter  2001 Elsevier Science Ltd. All rights reserved.

Results and discussion At least two distinct repair pathways in eukaryotes, homologous recombination and NHEJ, can repair double-strand breaks (DSBs) [6]. In mammals, the NHEJ pathway appears to be the major pathway, whereas yeast mostly relies on homologous recombination. Nevertheless, yeast can perform NHEJ, and most of the protein components required for NHEJ appear to be evolutionary conserved. Previous studies suggested that a gene(s) repressed by the diploid-specific a1/␣2 heterodimer was important for

NHEJ, since haploid yeast strains performed NHEJ about ten times more efficiently than diploids did [7, 8]. There are seven genes known to be required for efficient NHEJ in yeast (YKU70, YKU80, LIF1, LIG4, RAD50, XRS2, MRE11), but none of these genes were transcriptionally repressed in diploids [4, 7], suggesting that a novel a/␣repressed gene was required for NHEJ. To identify this gene(s), we took advantage of a study in which a microarray experiment identified all haploid-specific genes in yeast [4]. One candidate was the YLR265c gene (named nonhomologous end-joining regulator 1, NEJ1), whose expression was very efficiently repressed in MATa/MAT␣ diploids [4]. We tested the NHEJ efficiency of a strain carrying a null allele of nej1 relative to an isogenic wildtype strain in a plasmid recircularization assay [1, 2]. We also tested a strain lacking the gene encoding DNA ligase IV (LIG4), a gene previously determined to be required for efficient NHEJ. By this assay, the NEJ1 gene was required for efficient NHEJ, since the nej1 strain showed an 8-fold reduction in recircularization efficiency (Figure 1a). Moreover, LIG4 and NEJ1 were in the same epistasis group, since the NHEJ efficiency in the lig4 nej1 doublemutant strain was not significantly different from either single-mutant strain. To determine if Nej1p was required for NHEJ, when DSBs were generated in chromatin, we expressed the HO endonuclease in nej1 and wild-type strains, thus generating a DSB in the MAT␣1 gene. Because the cryptic mating type loci, HML␣ and HMRa, were deleted in these strains, the sequence adjacent to the DSB lacked homology to other parts of the genome, thus ensuring that the DSBs could be repaired only by NHEJ [9]. Since a persistent DSB is lethal to cells, the survival of strains continuously expressing the HO endonuclease is a measure of the NHEJ efficiency. The growth of wild-type and nej1 strains on media causing continuous expression (galactose) or no expression (glucose) of the HO gene was compared. The nej1 strain performed NHEJ much less efficiently than the wild-type strain did (Figure 1b). This low level of repair was comparable to the effect of deleting YKU70 [8], a gene with a central role in NHEJ. Repair performed in wild-type strains during continuos HO expression results in small deletions or insertions in the HO recognition site. In fact, as much as 80% of the repair events in the wild-type strain is a 2-bp insertion (⫹CA), which destroys the HO recognition site and also renders the ␣1 protein nonfunctional [9] (Figure 1c). We determined the types of repair events that occurred in the nej1 mutant strain in 40 independent survivors. Many surviving colonies in the nej1 strain had large deletions around the HO-cutting site that extended into the MAT␣2 gene, since 38% of them had an a-like mating type, reflecting loss of both ␣1 and ␣2 function. Moreover, PCR

1612 Current Biology Vol 11 No 20

Figure 1

Nej1p is required for NHEJ. (a) Wild-type (W303␣), nej1 (SAY215), lig4 (W303␣ lig4), and lig4 nej1 (SAY301) strains were tested for end-joining efficiency by transforming super-coiled and HindIII-

linearized plasmid pRS316 into cells. The data represent the average of four independent experiments, and the standard deviations are indicated. (b) Survival of wild-type (JKM115) and nej1 (SAY264)

Brief Communication 1613

analysis using primers that flanked the HO recognition site revealed that the sterile isolates from the nej1 strain (32%) also had deletions of at least 100 bp (Figure 1c). The remaining isolates from the nej1 strain (30%) most likely had mutations in the HO gene, since they still mated as ␣-cells. Thus, most repair events that occurred in the nej1 mutant strain were large deletions similar to the types of repair events that occur in YKU70 mutant strains. NHEJ in nej1 strains was thus both inefficient and inaccurate. In yeast, haploid-specific genes are repressed by the a1/ ␣2 heterodimer, which is present in MATa/MAT␣ diploids and haploid sir mutant strains. To confirm that NEJ1 was a haploid-specific gene, we generated a NEJ1-lacZ reporter gene that was introduced into MATa or MAT␣ haploids, a MATa/MAT␣ diploid, and a MATa sir4 pseudodiploid strain. NEJ1 was indeed a haploid-specific gene, with expression being at background levels when the a1/ ␣2 repressor was present (Figure 1d). Furthermore, we suggest that the a1/␣2 heterodimer directly represses NEJ1 transcription, since a consensus operator site for a1/ ␣2 was present in the NEJ1 promoter region (Figure 1e). The absence of Nej1p was thus a possible explanation for the NHEJ defect observed in nonmating sir mutant strains [1, 2]. It was also possible that several haploidspecific genes were required for NHEJ and that NEJ1 was only one of them. To investigate if Nej1p expression was sufficient to correct the NHEJ defect of sir mutant strains, we expressed a functional GFP-Nej1p fusion protein from a glycerol phosphate dehydrogenase (GPD1) promoter. The GPD1 promoter directed GFP-NEJ1 expression independent of cell type, since both the haploid strains and the MATa/MAT␣ diploid showed localization of GFP fluorescence in the nucleus (data not shown). Expression of GFP-Nej1p completely rescued the NHEJ defect of sir2 and sir4 mutant strains and the MATa/MAT␣ diploid strain (Figure 2). Thus, by this assay, expression of Nej1p alone was sufficient to restore efficient NHEJ in strains expressing the a1/␣2 repressor. The HO endonuclease generates a DSB with a 4-bp 3⬘ overhang sequence in the MAT␣1 gene. The upper strand

strains during continuos HO expression. Ten-fold serial dilutions of the plasmid-containing strains are indicated on the left, spotted on synthetic complete medium lacking uracil, using either glucose (HO off) or galactose (HO on) as a carbon source. The right panel represents quantification of the survival on galactose medium from two independent experiments, and the standard deviations are indicated. (c) DNA repair performed in the nej1 strain was inaccurate. A schematic drawing of repair events from isolates subjected to continuous HO expression. Inactivation of the HO gene (top); small insertions or deletions at the HO recognition site (middle) or larger deletions extending into the ␣2 gene generate three different mating phenotypes. Percentages of mating phenotypes of 40 independent

Figure 2

Expression of Nej1p relieves the NHEJ defect of sir2, sir4, and MATa/ MAT␣ strains. The indicated strains containing either a plasmid expressing GFP-RAS2 (vector) or GFP-NEJ1 (pNEJ1) from a constitutive promoter (GPD1) were tested for end-joining efficiency by transforming super-coiled and HindIII-linearized plasmid pRS316 into cells using standard methods. The data represent the average of two independent experiments, and the standard deviations are indicated.

(as drawn, Figure 3a) is degraded in the 5⬘ to 3⬘ direction, producing a single-stranded 3⬘-protruding tail. Since mutations in genes required for NHEJ previously were shown to affect the rate of this degradation, we wanted to investigate if Nej1p also affected this process. Degradation of DNA can be observed after digestion of chromosomal DNA with StyI, followed by denaturing gel electrophoresis [10]. Since StyI cannot digest single-stranded DNA, longer restriction fragments are generated when the StyI sites are rendered single stranded. By measuring the rate by which a 700-bp restriction fragment generated from HO-StyI digestion is converted into larger fragments, it is possible to estimate the rate of degradation. In contrast to mutations in YKU70, which leads to a faster degradation rate [11], the absence of Nej1p slowed down the degradation rate (Figure 3b,c). The progression of degradation is reduced about 2-fold in both mre11 and rad50 mutants [12, 13], both of which are also required for efficient NHEJ [9, 14]. Mre11p has exonuclease activity, providing

isolates from JKM115 and SAY264. The asterisk indicates that PCR analysis showed that all sterile isolates from the nej1 strain had deletions of at least 100 bp. (d) NEJ1 is a haploid-specific gene. Yep357NEJ1-lacZ was introduced into the indicated strains, and ␤-galactosidase (␤-gal) units were determined. The background ␤-gal expression recorded from a strain containing Yep357 alone was 0.3 units. (e) The NEJ1 promoter contained an a1/␣2 operator site. A comparison between the consensus a1/␣2 operator site [3] and sequences spanning from 149 to 130 nt 5⬘ of the potential NEJ1 start codon. Matches to the consensus are indicated in bold. R ⫽ A or G, W ⫽ A or T, and N ⫽ A, C, G, or T.

1614 Current Biology Vol 11 No 20

Figure 3 Decreased 5⬘ to 3⬘ degradation rate in a nej1 mutant strain. (a) A schematic drawing of the MAT␣ locus with the StyI restriction sites, the HO recognition site, and the probe used indicated. (b) DNA blot analysis of StyIdigested chromosomal DNA prepared from strains JKM115 (wild-type) and SAY264 (nej1) containing a GAL1::HO plasmid. Cells grown in synthetic complete medium lacking uracil, containing glycerol as a carbon source, were induced to express HO by adding galactose (final concentration 2%) at t ⫽ 0. Molecular weight is indicated on the left. (c) Quantification of the data from three independent experiments. The ratio between the total radioactivity in the 700-bp fragment and a URA3-containing DNA fragment (inset) are plotted as a function of the time after HO induction. The amount of the HO-StyI 700-bp fragment at t ⫽ 1 hr was defined as 100%. Error bars indicate the standard deviations from three experiments.

a possible explanation for the decreased degradation rate. Since nej1 mutants also had decreased degradation rates, it was possible that Nej1p facilitated nuclease function. The GFP-Nej1p fusion protein was functional and localized to the nucleus, thus indicating that Nej1p exerted its function in the nucleus. We wanted to investigate the localization of Nej1p when it was expressed from the endogenous promoter, however, to make sure that the nuclear localization was not due to the overexpression of the fusion protein. This was particularly interesting since Nej1p contained two regions in the N-terminal part that were possible transmembrane helices. To investigate the localization of Nej1p expressed from the endogenous promoter, we generated an epitope-tagged NEJ1 allele in which 13 repeats of the Myc peptide were fused to the carboxyl terminus of Nej1p. The NEJ1-13xMyc allele encoded a functional protein, since this strain did not display increased sensitivity to the induction of the HO endonuclease (data not shown). We then performed indirect immunofluorescence experiments using a polyclonal antiserum raised against the Myc epitope. As a molecular marker for the nuclear periphery, we used a monoclonal antiserum specific for nuclear pore complex proteins (mAb 414). The results confirmed the previous results and showed that Nej1p indeed localized to the nucleus (Figure 4a–d). Nej1p, however, did not colocalize with the nuclear pores and did not show an obvious localization to the periphery of the nucleus. To investigate if Nej1p had a direct role in NHEJ, we tested for Nej1p interaction with the other known NHEJ components. The full-length NEJ1, LIF1, LIG4, YKU70,

YKU80, MRE11, XRS2, and RAD50 genes were cloned into the two-hybrid vectors pAS1 and pACT2 and were introduced pairwise into a yeast strain containing a GAL2ADE2 marker gene [15]. An interaction between these molecules could thus be scored on synthetic complete medium lacking adenine. In this assay, both Gal4BDNej1p and GAL4AD-Nej1p fusion proteins interacted with Lif1p, but not with any other fusion protein (Figure 4e, data not shown). Next, we wanted to confirm the interaction between Nej1p and Lif1p using an independent assay. Epitope-tagged Lif1p was immunoprecipitated from a yeast strain that expressed either full-length Nej1p or truncated Nej1p, lacking the first 20 amino acids (nej1⌬1-20). This truncated Nej1 protein partially lacked the first potential transmembrane domain. We investigated the immunoprecipitates from the respective strains for the presence of Nej1p (Figure 4f). Both full-length and truncated Nej1p could be coimmunoprecipitated with the Flag-Lif1 fusion protein. The truncated protein appeared to be more efficiently coprecipitated, possibly as a consequence of the fact that this protein was less likely to interact with membranes. Nej1p thus interacted with Lif1p using two independent assays. Cell type effects on DNA repair have been observed before in yeast [16], but the molecular explanation behind these effects has not been elucidated. In this study, we identified the first haploid-specific gene involved in DNA repair. We favor a model in which haploid cells in G1 must use the NHEJ pathway for DSB repair. Diploid cells, however, always have a homologous chromosome to perform homology-driven repair. Because homologydriven repair is more accurate, this pathway is preferred,

Brief Communication 1615

Figure 4

Nej1p localized to the nucleus. Indirect immunofluorescence of haploid strain SAY326 (NEJ1-13xMyc). (a–d) Cells were stained with DAPI (blue), goat fluorescein isothiocyanate (FITC)-conjugated IgG against rabbit anti-Myc (green), and donkey Cy3-conjugated IgG against mouse mAb 414 (red), as indicated. (d) A merger of the anti-Myc and mAb 414 images is also shown. Nej1p interacted with Lif1p. (e) Strain PJ69-4A (GAL2-ADE2) [15] containing plasmid pAS1-NEJ1 (Gal4BD-NEJ1) and plasmid pACT2 (Gal4AD), with gene fusions to the genes indicated on the left, were spotted as 10-fold serial dilutions on synthetic complete medium lacking adenine (⫺Ade)

and on rich medium (YEPD). (f) A Western blot using a rabbit antiNej1p antiserum of immunoprecipitates of Flag-Lif1p using a mouse anti-Flag antiserum. Protein extracts were prepared from strain SAY297 (nej1::KanMX lif1::GAL-8HIS-FLAG-LIF1::URA3) grown in galactose, expressing either full-length (NEJ1) or truncated (nej1⌬120) Nej1p. Input designates proteins present in the protein extracts, and IP designates proteins remaining after precipitation and washes. Addition (⫹) or no addition (⫺) of the anti-Flag antiserum to the IP lanes, as indicated.

1616 Current Biology Vol 11 No 20

and it thus seems logical that diploids would repress NHEJ. The induction of DSBs leads to a relocalization of Sir3p and Ku70p from telomeres to sites of DSB repair [17, 18], suggesting a possible role for Sir proteins in NHEJ. The current study did not exclude a minor role for Sir proteins in DNA repair, but showed that the NHEJ defect of sir mutants was largely attributable to the transcriptional repression of NEJ1. An exact mechanism by which Nej1p facilitates NHEJ was not revealed, and homology searches did not reveal genes highly homologous to NEJ1 in other organisms. Given the high degree of functional conservation of other NHEJ components between yeast and mammals, it would not be surprising to find a functional counterpart to Nej1p, perhaps highly diverged, in mammals. Because the rate of the 5⬘ to 3⬘ degradation close to a DSB was slower in cells lacking Nej1p and given that Nej1p localized to the nucleus and interacted with Lif1p, we suggest that Nej1p had a direct role in end joining. We cannot exclude more complex, indirect models, such as Nej1p regulating the transport or stability of some other protein involved in NHEJ. Nej1p, however, was required for efficient NHEJ even if a high level of Lif1p was present in the nucleus (see the Supplementary material available with this article online). Hence, the important role of Nej1p in NHEJ was not to regulate the stability or localization of Lif1p.

Materials and methods Strains and plasmids A wild-type (BY4741) and isogenic nej1::KanMX strain (5174) were purchased from Research Genetics. The nej1::KanMX allele was transferred into JKM115 [9] and W303 strain JRY2334 using a one-step gene disruption procedure [19], transforming the strains with a PCR fragment generated using primers flanking the nej1::KanMX allele in strain 5174. The disruptions were confirmed using DNA blots. A lig4 nej1 double-mutant strain (SAY301) was generated by crossing strain W303␣ lig4 (lig4::HIS3 [20]) with SAY215 (nej1::KanMX, this study). A NEJ1-13xMyc strain (SAY326) was generated using a PCR-based procedure [21] and integrated at the genomic NEJ1 locus in strain JKM115 using a one-step gene replacement. Immunoprecipitations were performed in strain SAY297 (nej1::KanMX lif1::GAL-8HIS-FLAGLIF1::URA3) transformed with plasmids containing full-length or truncated (nej1⌬ 1-20) NEJ1 in plasmid YcDE-2 (2␮ TRP1 ADH1 promoter). A NEJ1-lacZ fusion gene was generated by ligating a 252-bp XhoI-SphI PCR fragment corresponding to nucleotides ⫺225 to ⫹27 of the NEJ1 promoter into the SalI-SphI sites of Yep357 (2␮m URA3 lacZ). To generate a GFP-NEJ1 construct in which the fusion gene is driven by a glycerol phosphate dehydrogenase gene promoter, a 1.0-kb EcoRISalI PCR fragment containing the entire NEJ1 gene with the EcoRI site at the start codon was cloned into the EcoRI-SalI sites of vector pJW192 (2␮m TRP1 GPD-GFP-RAS2), generating p479. Two-hybrid vectors (pAS1 and pACT2) containing the NEJ1, YKU70, YKU80, LIG4, LIF1, XRS2, MRE11, and RAD50 ORFs were generated by PCR amplification from genomic DNA and standard cloning methods.

NHEJ measurements, immunofluorescence, immunoprecipitations, and measurement of degradation rate NHEJ assays [1, 2, 9] and estimations of degradation rates [11] were performed as described. Quantification of radioactivity on the blots was performed using a phoshorimager and the Imagequant software. The antiNej1p antiserum was raised in rabbits using a purified maltose binding protein-Nej1p fusion protein as an antigen, was affinity purified, and was used at a 1:500 dilution for Western blotting. For immunofluorescence,

the antisera used were diluted as follows: rabbit anti-Myc, 1:250 (Research Diagnostics); mAb 414, 1:2500 (BABCO); goat FITC-conjugated anti-rabbit, 1:1000 (Southern Biotechnology Associates); and donkey Cy3-conjugated anti-mouse, 1:500 (Jackson Immunoresearch Laboratories). For immunoprecipitations [22], 1 ␮g anti-Flag antiserum (M2, Sigma) was used for 1 mg protein extract.

Supplementary material Supplementary material including additional Materials and methods and three experiments that are shown in Figure S1 is available at http:// images.cellpress.com/supmat/supmatin.htm.

Acknowledgements We thank A. Bystro¨m, J. Haber, and S. Jackson for strains and plasmids. This work was supported by funds from the Swedish Natural Science Research Council (B-AA/BU 11279-306) and the Umea˚ Center for Molecular Pathogenesis.

References 1. Tsukamoto Y, Kato J, Ikeda H: Silencing factors participate in DNA repair and recombination in Saccharomyces cerevisiae. Nature 1997, 388:900-903. 2. Boulton SJ, Jackson SP: Components of the Ku-dependent nonhomologous end-joining pathway are involved in telomeric length maintenance and telomeric silencing. EMBO J 1998, 17:1819-1828. 3. Herskowitz I, Rine J, Strathern J: Mating-type determination and mating-type interconversion in Saccharomyces cerevisiae. In The Molecular Cellular Biology of the Yeast Saccharomyces: Gene Expression. Edited by Jones EW, Pringle JR, Broach JR. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1992:583656. 4. Galitski T, Saldanha AJ, Styles CA, Lander ES, Fink G: Ploidy regulation of gene expression. Science 1999, 285:251-254. 5. Herrmann G, Lindahl T, Scha¨r P: Saccharomyces cerevisiae LIF1: a function involved in DNA double-strand break repair related to mammalian XRCC4. EMBO J 1998, 17:4188-4198. 6. Paques F, Haber JE: Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae. Microbiol Mol Biol Rev 1999, 63:349-404. 7. A˚stro¨m SU, Okamura SM, Rine J: Yeast cell-type regulation of DNA repair. Nature 1999, 397:310. 8. Lee SE, Paques F, Sylvan J, Haber JE: Role of yeast SIR genes and mating type in directing DNA double-strand breaks to homologous and nonhomologous repair paths. Curr Biol 1999, 9:767-770. 9. Moore JK, Haber JE: Cell cycle and genetic requirements of two pathways of nonhomologous end-joining repair of doublestrand breaks in Saccharomyces cerevisiae. Mol Cell Biol 1996, 16:2164-2173. 10. White CI, Haber JE: Intermediates of recombination during mating type switching in Saccharomyces cerevisiae. EMBO J 1990, 9:663-673. 11. Lee SE, Moore JK, Holmes A, Umezu K, Kolodner RD, Haber JE: Saccharomyces Ku70, mre11/rad50 and RPA proteins regulate adaptation to G2/M arrest after DNA damage. Cell 1998, 94:399-409. 12. Tsubouchi H, Ogawa H: A novel mre11 mutation impairs processing of double-strand breaks of DNA during both mitosis and meiosis. Mol Cell Biol 1998, 18:260-268. 13. Sugawara N, Haber JE: Characterization of double-strand break-induced recombination: homology requirements and single-stranded DNA formation. Mol Cell Biol 1992, 12:563575. 14. Milne GT, Jin S, Shannon KB, Weaver DT: Mutations in two Ku homologs define a DNA end-joining repair pathway in Saccharomyces cerevisiae. Mol Cell Biol 1996, 16:4189-4198. 15. James P, Halladay J, Craig EA: Genomic libraries and a host strain designed for highly efficient two-hybrid selection in yeast. Genetics 1996, 144:1425-1436. 16. Heude M, Fabre F: a/alpha-control of DNA repair in the yeast Saccharomyces cerevisiae: genetic and physiological aspects. Genetics 1993, 133:489-498. 17. Martin SG, Laroche T, Suka N, Grunstein M, Gasser SM: Relocalization of telomeric Ku and SIR proteins in response to DNA strand breaks in yeast. Cell 1999, 97:621-633.

Brief Communication 1617

18. Mills KD, Sinclair DA, Guarente L: MEC1-dependent redistribution of the Sir3 silencing protein from telomeres to DNA double-strand breaks. Cell 1999, 97:609-620. 19. Rothstein RJ: One-step gene disruption in yeast. Methods Enzymol 1983, 101:202-211. 20. Teo S-H, Jackson SP: Lif1p targets the DNA ligase Lig4p to sites of DNA double-strand breaks. Curr Biol 2000, 10:165168. 21. Ba¨hler J, Wu J-Q, Longtine MS, Shah NG, MacKenzie A III, Steever AB, et al.: Heterologous modules for efficient and versatile PCRbased gene targeting in Scizosaccharomyces pombe. Yeast 1998, 14:943-951. 22. Feng Y, Song L-Y, Kincaid E, Mahanty SK, Elion EA: Functional binding between G␤ and the LIM domain of Ste5 is required to activate the MEKK Ste11. Curr Biol 1998, 8:267-278.