Genetics, Vol. 160, 49-62, January 2002, Copyright © 2002
Differential Suppression of DNA Repair Deficiencies of Yeast rad50, mre11 and xrs2 Mutants by EXO1 and TLC1 (the RNA Component of Telomerase)
L. Kevin Lewis1,a,
G. Karthikeyana,
James W. Westmorelanda, and
Michael A. Resnicka
a Laboratory of Molecular Genetics, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709
Corresponding author:
Michael A. Resnick, National Institutes of Health, National Institute of Environmental Health Sciences, 111 Alexander Dr., Research Triangle Park, NC 27709., resnick{at}niehs.nih.gov (E-mail)
Communicating editor: A. NICOLAS
 | ABSTRACT |
|---|
Rad50, Mre11, and Xrs2 form a nuclease complex that functions in both nonhomologous end-joining (NHEJ) and recombinational repair of DNA double-strand breaks (DSBs). A search for highly expressed cDNAs that suppress the DNA repair deficiency of rad50 mutants yielded multiple isolates of two genes: EXO1 and TLC1. Overexpression of EXO1 or TLC1 increased the resistance of rad50, mre11, and xrs2 mutants to ionizing radiation and MMS, but did not increase resistance in strains defective in recombination (rad51, rad52, rad54, rad59) or NHEJ only (yku70, sir4). Increased Exo1 or TLC1 RNA did not alter checkpoint responses or restore NHEJ proficiency, but DNA repair defects of yku70 and rad27 (fen) mutants were differentially suppressed by the two genes. Overexpression of Exo1, but not mutant proteins containing substitutions in the conserved nuclease domain, increased recombination and suppressed HO and EcoRI endonuclease-induced killing of rad50 strains. exo1 rad50 mutants lacking both nuclease activities exhibited a high proportion of enlarged, G2-arrested cells and displayed a synergistic decrease in DSB-induced plasmid:chromosome recombination. These results support a model in which the nuclease activity of the Rad50/Mre11/Xrs2 complex is required for recombinational repair, but not NHEJ. We suggest that the 5'–3' exo activity of Exo1 is able to substitute for Rad50/Mre11/Xrs2 in rescission of specific classes of DSB end structures. Gene-specific suppression by TLC1, which encodes the RNA subunit of the yeast telomerase complex, demonstrates that components of telomerase can also impact on DSB repair pathways.
DNA double-strand breaks (DSBs) in eukaryotic chromosomes are repaired primarily by two pathways, homologous recombination and nonhomologous end-joining (NHEJ). Additional, nonconservative mechanisms of DSB repair have been described for selected DNA substrates, e.g., within segments of DNA containing multiple repeat sequences (PETES et al. 1991 
; KLEIN 1995 ). End-joining is the predominant pathway of DSB repair in higher eukaryotes, although some classes of homologous recombination events occur at high frequency in cultured mammalian cells (e.g., SARGENT et al. 1997 ; LIANG et al. 1998 ; TAKATA et al. 1998 ).
Recombinational repair mechanisms are highly efficient in the yeast Saccharomyces cerevisiae and DSB ends containing modified bases or sugars, such as those produced by ionizing radiation, are almost exclusively repaired by this process. Genes specifically associated with the homologous recombination pathway in mitotic cells include RAD51, RAD52, RAD54, RAD55, RAD57, RAD59, and RPA (GAME 1993 ; BAI and SYMINGTON 1996 ; LEWIS et al. 1999 ; PAQUES and HABER 1999 ). Protein:protein interactions and strand exchange reactions mediated by purified Rad51, Rad52, Rad54, Rad55, Rad57, and Rpa have been described (CLEVER et al. 1997 ; BENSON et al. 1998 ; GASIOR et al. 1998 ; NEW et al. 1998 ; SHINOHARA and OGAWA 1998 ). Mutations in three other genes, RAD50, MRE11, or XRS2, cause only modest defects in spontaneous intrachromosomal recombination in vegetative cells (described below).
NHEJ in yeast can be precise or error prone and requires the functions of several genes. These include YKU70 and YKU80 (encoding subunits of the Ku heterodimer); DNL4 (LIG4) and LIF1 (yeast homologs of human DNA ligase IV and XRCC4, respectively); SIR2, SIR3, and SIR4 (mediating transcriptional silencing at yeast mating type loci and telomeres); as well as RAD50, MRE11, and XRS2 (TSUKAMOTO et al. 1997 ; CRITCHLOW and JACKSON 1998 ). Cells lacking any of these genes have a reduced ability to recircularize linear, cohesive-ended plasmid DNA after cellular transformation (e.g., see BOULTON and JACKSON 1998 ; LEWIS and RESNICK 2000 ). In addition, yku70, yku80, rad50, mre11, xrs2, dnl4, sir2, sir3, and sir4 mutants are hypersensitive to the expression of EcoRI endonuclease in vivo (BARNES and RIO 1997 ; LEWIS et al. 1998 , LEWIS et al. 1999 ; MILLS et al. 1999 ). Important roles for YKU70, RAD50, MRE11, and XRS2 in the repair of DSBs produced by HO endonuclease have also been described (ELIAS-ARNANZ et al. 1996 ; MOORE and HABER 1996 ; LEE et al. 1998 ; MARTIN et al. 1999 ). The available data suggest that Yku70, Yku80, Dnl4, Lif1, Rad50, Mre11, and Xrs2 participate in the processing and/or joining of broken DNA ends (PANG et al. 1997 ; BAUMANN and WEST 1998 ; CRITCHLOW and JACKSON 1998 ; HABER 1998 ; LEWIS and RESNICK 2000 ), but the functions of Sir2, Sir3, and Sir4 in NHEJ remain unclear (ASTROM et al. 1999 ; IMAI et al. 2000 ; LANDRY et al. 2000 ).
rad50, mre11, and xrs2 strains are defective in both recombination and NHEJ. These mutants are highly sensitive to damage induced by ionizing radiation, methyl methanesulfonate (MMS), or EcoRI endonuclease, exhibit elevated levels of spontaneous interchromosomal recombination during mitotic growth, and are deficient in the formation and processing of DSBs in meiosis (HAYNES and KUNZ 1981 ; JOHZUKA and OGAWA 1995 ; HABER 1998 ; LEWIS et al. 1999 ). rad50, mre11, and xrs2 cells also display reduced rates of spontaneous intrachromosomal recombination (RATTRAY and SYMINGTON 1995 ; TRAN et al. 1995 ; KOUPRINA et al. 1999 ) and feature delayed kinetics of mating type switching, an intrachromosomal gene conversion event (HABER 1998 ). In addition, gross chromosomal rearrangements and chromosome arm loss events occur more frequently in such strains than in other DSB repair mutants (CHEN and KOLODNER 1999 ; KOUPRINA et al. 1999 ).
The Rad50, Mre11, and Xrs2 proteins associate in vivo and in vitro and constitute a nuclease complex with manganese-dependent 3'–5' exonuclease and single-stranded DNA endonuclease activities (BRESSAN et al. 1998 ; FURUSE et al. 1998 ; USUI et al. 1998 ; MOREAU et al. 1999 ; CHAMANKHAH et al. 2000 ). A similar nuclease complex exists in human cells and consists of the products of the hRAD50, hMRE11, and hNBS1 genes (PAULL and GELLERT 1999 ; PETRINI 1999 ). Mutations in hNBS1 cause the autosomal recessive disorder Nijmegen breakage syndrome. Cells obtained from individuals with this disorder are hypersensitive to ionizing radiation and exhibit chromosome instability. In addition, mutations in hMRE11 have been linked to an ataxia telangiectasia-like disorder that is phenotypically similar to Nijmegen breakage syndrome (STEWART et al. 1999 ).
Stable maintenance of the lengths of yeast telomeres involves multiple genes. These include components of the telomerase complex such as EST1, EST2, EST3, TLC1 (encoding the RNA template subunit of telomerase), and CDC13, as well as many genes within the NHEJ pathway of DSB repair (YKU70, YKU80, RAD50, MRE11, XRS2, SIR2, SIR3, and SIR4, but not DNL4 or LIF1; PORTER et al. 1996 ; KIRONMAI and MUNIYAPPA 1997 ; LOWELL and PILLUS 1998 ; NUGENT et al. 1998 ; LEWIS and RESNICK 2000 ). Mutations in these genes result in telomere shortening in haploid cells. Genetic analysis has suggested that at least three epistasis groups are involved in maintaining telomere stability in yeast, including (i) the genes encoding the Ku heterodimer; (ii) CDC13; and (iii) EST1, EST2, EST3, TLC1, RAD50, MRE11, and XRS2 (NUGENT et al. 1998 ). Telomere stability may also be influenced by genes associated with recombination (LE et al. 1999 ; TENG and ZAKIAN 1999 ) and general DNA replication (ADAMS and HOLM 1996 ; PARENTEAU and WELLINGER 1999 ).
The precise functions of the Rad50/Mre11/Xrs2 complex (and its human cell counterpart) in recombination, NHEJ, and telomere maintenance are unknown. In an attempt to clarify its roles in DSB repair we have identified genes that, when expressed at elevated levels, suppress the severe DSB repair defects of rad50, mre11, and xrs2 mutants. Clones identified in the screen contained cDNAs encoding Exo1 (a 5'–3' exonuclease implicated in DNA replication, mismatch repair, and homologous recombination) and TLC1, the RNA subunit of the yeast telomerase complex. Results obtained in subsequent analyses were consistent with a model in which elevated levels of the 5'–3' exo activity of Exo1 substitute for the Rad50/Mre11/Xrs2 nuclease complex in recombinational DNA repair, but not in NHEJ or in maintenance of telomere ends. Several lines of evidence suggested that the mechanism of suppression by TLC1 RNA is different from that of EXO1 and revealed additional connections between DSB repair and telomere length maintenance pathways.
| MATERIALS AND METHODS |
|---|
Yeast strains and media:
S. cerevisiae strains used in the study are displayed in Table 1. VL6 and derivatives are descended from the S288c strain YPH857 (LARIONOV et al. 1994 ). The reg1-501 host strain T334 (described below) was employed for cell transformations requiring continuous transcription of the galactose-inducible GAL1 promoter. YPD and synthetic yeast growth media were prepared as described (LEWIS et al. 1999 ). Methyl methanesulfonate was purchased from Fluka (Buchs, Switzerland). G418 (Life Technologies, Rockville, MD) and 5-fluoroorotic acid (5-FOA; American Biorganics, Niagara Falls, NY) were used as described (LEWIS et al. 1999 ). Hygromycin B (Boehringer Mannheim, Indianapolis) was employed as recommended in GOLDSTEIN and MCCUSKER (1999). Yeast cell transformation was accomplished using a LiAc-based protocol (GIETZ et al. 1995 ). Photographic analysis of yeast colony growth and morphology was accomplished using a FOTO/Analyst Archiver Eclipse electronic documentation system from Fotodyne (New Berlin, WI). All wild-type and repair-deficient cells were grown at 30° except 
yku70 mutants, which were propagated at 25°. Treatment of cells with ionizing radiation was accomplished using a 137Cesium source emitting at a dose rate of 2.75 krad/min.
Strains containing rad50::hisG, rad51::hisG, rad52::hisG, and rad54::hisG deletion disruptions were created as previously described (LEWIS et al. 1998 , LEWIS et al. 1999 ). Strains containing rad50::G418, mre11::G418, rad59::G418, and xrs2::G418 deletions were created using PCR fragment-mediated gene disruption in conjunction with the G418-resistance plasmid pFA6MX4 (WACH et al. 1994 ). Similarly, creation of mre11::HygB involved amplification of the gene for Hygromycin B resistance within pAG32 (GOLDSTEIN and MCCUSKER 1999 ) and yku70::TRP1 alleles were created after PCR amplification of pRS304 (SIKORSKI and HIETER 1989 ). Primers used for each deletion contained 20–25 bases homologous to plasmid sequences and 45–55 bases of sequence from the 5' or 3' ends of the coding regions of genes to be deleted. Oligonucleotide sequences were as follows:
grad50C, GCATGAGCGCTATCTATAAATTATCTATTCAGGGCATACGGTCTTCGTACGCTGCAGGTCGAC; grad50D, CGCAGTCTTATAGGAGAGCTCCGTTTCTTCCAGGACATCATTATAATCGATGAATTCGAGCTCG; gmre11A, GGACTATCCTGATCCAGACACAATAAGGATTTTAATTACTACAGACGTACGCTGCAGGTCGAC; gmre11B, GGAAGGAATCTAGCCCATTACCATTGAATGCGAAATTTGTCTCATATCGATGAATTCGAGCTCG; GxrsE, GGTGATAACTATAAATTTATGTGGGTAGTACGATACCAGAATACATTGGAAGATGCGTACGCTGCAGGTCGAC; GxrsF, AAAAGAGCCACGTGATCTACTAGTTGCGCTTCTACTGTGTCTACTTTGCTTGGAAATCGATGAATTCGAGCTCG; grad59A, ATGACGATACAAGCGAAGCCCAGTTCGAGCATATCGTATGATTCGCGTACGCTGCAGGTCGAC; grad59B, TTATTTGATATGCGTGCCTTTAGCATCCTCCAATTTGATAAAAGTATCGATGAATTCGAGCTCG; prsku70a, GGCCAGTCACTAATGCATTTGGCAATAGTGGAGAACTTAACGATCAAGTGGATCAGAGCAGATTGTACTGAGAGTGCACC; prsku70b2, AGATTACTGTCGTGCATAAATATCTTGCTAATAGTTGTACAGTACAACGTCGCATCTGTGCGGTATTTCACACCGC; 5'rad27-G418, AAAATTAACAACAAGAACATTATTATTCGATAGGAATGGGCGTACGCTGCAGGTCGAC; 3'rad27-G418, ACCTTCGAACATATATATACACCACTTAGAAAATTGATGGATCGATGAATTCGAGCTCG; DHS1-KanMX-5', AGGTATGAAGGAGAAGTGTTAGCCATTGATGGCTATGCATCGTACGCTGCAGGTCGAC;DHS1-KanMX-3', TTGGCTTGACTTAGTAGTTTCGATGTCCCTTTTCTTACTTATCGATGAATTCGAGCTCG.
cDNA library screening and DNA sequence analysis:
Strain YLKL440 (rad50::G418 reg1-501) was transformed with plasmid DNA prepared from an amplified stock of a library containing haploid mitotic cell cDNAs under the control of the GAL1 promoter (LIU et al. 1992 ). The presence of the reg1-501 allele in this host permits modulated GAL promoter induction by galactose (Gal) while cells continue growth in glucose (Glu) media (HOVLAND et al. 1989 ; LEWIS et al. 1998 , LEWIS et al. 1999 ). Transformants were plated onto synthetic media containing 2% Glu + 1% Gal without uracil and colonies were subsequently replica plated to the same media supplemented with 0.3 mM MMS. Analysis of 
70,000 initial transformants resulted in identification of >200 apparent MMS-resistant colonies after 4–5 days growth at 30°. Twelve of the clones remained resistant upon subsequent testing and were designated pCDNA50.1–pCDNA50.2 (pGAL1::EXO1 clones) and pCDNA50.3–pCDNA 50.12 (pGAL1::TLC1). Approximately 500 bp at the 5' and 3' ends of each plasmid insert were sequenced using the M13 Universal primer and primer Gal740 (CAACATTTTCGGTTTGTATTACTTC). The PRISM cycle sequencing system was employed in conjunction with an Applied Biosystems (Foster City, CA) model 373A DNA sequencer. Insert sequences were manipulated using DNA Strider 1.2 and homology searches were performed using the BLAST program (ALTSCHUL et al. 1995 ) in conjunction with the Saccharomyces Genome Database and the Munich Information Center for Protein Sequences (MIPS) yeast database. For analysis of the effects of EXO1 and TLC1 overexpression on cell growth, cycling, and survival in VL6 (REG1)-derived strains, GAL promoter activity was induced by growth in synthetic media containing 2% galactose without glucose. For assessment of sensitivity to MMS by dilution pronging, equal numbers of cells were serially diluted either fourfold (Fig 1) or fivefold (all other figures). Survival of cells containing pCDNA50.1–pCDNA50.12 was analyzed on 2% Gal-Ura plates containing the indicated concentration of mutagen.
View larger version (81K):
In this window
In a new window
Download PPT slide
|
Figure 1.
Repair of alkylation-induced DNA damage is enhanced in rad50 cells containing plasmids overexpressing EXO1 (pCDNA50.1 and pCDNA50.2) or TLC1 (pCDNA50.3–pCDNA50.7). Cultures of the reg1-501, rad50 strain YLKL440 containing control or library plasmids were diluted serially and pronged to plates containing glucose and galactose (2% Glu + 1% Gal - Uracil) with and without 0.3 mM (0.003%) MMS. Colonies containing GAL1::EXO1 plasmids were photographed after growth at 30° for 3 days and cells containing GAL1::TLC1 were photographed after 4 days.
|
|
Nonhomologous end-joining and homologous recombination assays:
T334-derived cells (reg1-501) containing pRS316Gal (LEWIS et al. 1998 ), pCDNA50.1 (pGAL::EXO1), or pCDNA50.3 (pGAL::TLC1) were grown in synthetic 2% Glu + 1% Gal media minus uracil prior to transformation. Two hundred nanograms of supercoiled or BamHI-cut pRS314 DNA was used to transform wild-type, rad50, and mre11 hosts (T334, YLKL499, and YLKL555, respectively) as described (GIETZ et al. 1995 ; BOULTON and JACKSON 1998 ). Continuous galactose-induced synthesis of EXO1 and TLC1 was maintained by selecting for transformants on 2% Glu + 1% Gal media without tryptophan. Each plasmid:strain combination was assayed three times and the results were averaged.
DNA damage-induced changes in cell cycling after treatment with MMS were analyzed as previously described (LEWIS et al. 1998 , LEWIS et al. 1999 ). Spontaneous rates of intrachromosomal recombination were assayed using a derivative of the strain VL6-48-12* (KOUPRINA et al. 1999 ) containing either pRS313 (CEN/ARS, HIS3) or pLKL53Y (CEN/ARS, HIS3, GAL1::EXO1). pLKL53Y was created by cloning the ApaI/NotI GAL::EXO1 fragment of pCDNA50.1 into ApaI/NotI-digested pRS313. Colonies grown on 2% Gal-His plates were harvested into water, diluted, and spread onto 2% Glu-His plates and 2% Glu-His-Ade-Lys plates supplemented with 5-FOA. The Ura- His+ Ade+ Lys+ colonies formed on the latter plates were subsequently replicated to Glu-Trp plates to identify the fraction of cells that were Ura- Trp- His+ Ade+ Lys+. Rates correspond to the number of events per generation per plasmid-containing cell.
Plasmid:chromosome recombination assays were performed by transforming wild-type, exo1, rad50, and exo1 rad50 strains with pLKL37Y (HIS3 URA3) that was cut at the NcoI site within URA3 and selecting for Ura+ integrant colonies. pLKL37Y was created by cloning the HindIII-HindIII URA3 gene fragment from YEP24 into the integrating vector pRS303. Replica plating and PCR analysis of wild-type transformants demonstrated that all Ura+ recombinants were also His+ with the plasmid integrated at the ura3-52 locus. The host strain, VL6
, contains a deletion of the HIS3 gene (his3-200). The chromosomal ura3-52 allele has a Ty insertion 121 nucleotides (nt) downstream of the URA3 start codon and the NcoI breakage site is 210 nt downstream, resulting in 89 bp of homology on one side of the DSB and >600 bp homology on the other. The same competent cell preparations were transformed separately with pRS313 (CEN/ARS, HIS3) and recombination frequencies were normalized for transformation efficiency. Each linear or supercoiled DNA transformation was performed three times and results were averaged.
The effects of EXO1 on HO endonuclease-induced growth inhibition were assessed after transformation of YLKL391 (rad52::LEU2 his3::[GAL::HO, TRP1] reg1-501) and YLKL574 (rad50::hisG his3::[GAL::HO, TRP1] reg1-501) with either the control plasmid pRS425 (2µ LEU2) or pRDK480 (2µ LEU2 EXO1; TISHKOFF et al. 1997 ). Growth of the rad50 and rad52 strains (but not Rad+ cells) containing the control vector was strongly inhibited after induction of HO endonuclease on plates containing 2% glucose plus galactose at concentrations ranging from 0.02 to 2.0%. Restoration of growth of rad50 strains by EXO1 was observed at reduced kinetics of HO induction (0.02–0.05% galactose), but not at higher levels of the inducer (0.5–2.0% galactose). Galactose-induced cell killing in reg1 rad50 and reg1 mre11 mutants containing an integrated GAL1::EcoRI cassette was also alleviated by constitutive expression of EXO1 (using the previously described strains YLKL372 and YLKL407 containing pRDK480; LEWIS et al. 1999 ). Assessment of the ability of Exo1 proteins containing substitutions within the nuclease domain to suppress EcoRI- and HO-induced cell killing was accomplished using plasmids pEAM69 (2µ LEU2 exo1-D171A) and pEAM71 (2µ LEU2 exo1-D173A; SOKOLSKY and ALANI 2000 ).
| RESULTS |
|---|
Repair of DNA damage in rad50 mutants is enhanced by overexpression of EXO1 and TLC1:
The functions of the Rad50/Mre11/Xrs2 nuclease complex in DSB repair are complex. To gain insight into the repair function(s) that is absent in such strains a screen was performed to identify genes that suppress the repair defects of rad50 mutants when expressed at high levels. Plasmid DNA prepared from a library containing yeast mitotic cDNAs under the control of the galactose-inducible GAL1 promoter (LIU et al. 1992 ) was transformed into the rad50 strain YLKL440. This derivative of T334 contains a mutation (reg1-501) that permits modulated induction of GAL promoter activity while cells continue to grow in glucose (HOVLAND et al. 1989 ; LEWIS et al. 1998 , LEWIS et al. 1999 ). Analysis of 70,000 transformants resulted in the identification of 12 clones that could form colonies on plates containing 2% glucose + 1% galactose + 0.3 mM (0.003%) MMS. Two clones (pCDNA50.1 and pCDNA50.2) provided much stronger resistance to MMS than the others and also improved the plating efficiency of rad50 strains on synthetic galactose media without mutagen (Fig 1). The other 10 clones (pCDNA50.3–pCDNA50.12) provided weaker resistance to MMS and did not significantly affect plating efficiency.
Sequence analysis revealed that the two clones that provided the strongest resistance to MMS contained the complete coding sequence of the EXO1 gene. This gene encodes a 5'–3' exonuclease previously implicated in multiple pathways including recombination, mismatch repair, and processing of Okazaki fragments during DNA replication (FIORENTINI et al. 1997 ; TISHKOFF et al. 1997 ; TRAN et al. 1999 ). The 10 remaining clones contained GAL::TLC1 promoter fusions. TLC1 RNA serves as template for DNA synthesis at the ends of yeast chromosomes during S phase and may have additional functions in telomere silencing and length maintenance (PRESCOTT and BLACKBURN 1997 ; SINGER et al. 1998 ; LUE 1999 ). One clone, pCDNA50.6, contained the entire 1301 bases predicted to be present in TLC1 RNA plus additional nucleotides upstream and downstream of the gene, but the other inserts contained slightly shorter sequences at their 5' ends (corresponding to total lengths ranging from 1270 to 1284 nucleotides). Gene lengths for pCDNA50.3–pCDNA50.12 were 1281, 1273, 1278, 1301, 1281, 1270, 1272, 1284, 1284, and 1282 nt, respectively. All inserts retained the template portion required for RNA:DNA annealing and DNA synthesis. The presence of TLC1 cDNAs in the library is explained by past observations that a fraction of TLC1 transcripts is polyadenylated (CHAPON et al. 1997 ). No library clones containing the RAD50 gene were recovered in the screen. This may be explained by the fact that RAD50 is a large gene (4000 bp) and the cDNA library was unlikely to contain full-length RAD50 clones.
Cells containing deletions of RAD52 epistasis group genes (RAD50-RAD59, MRE11, or XRS2) are hypersensitive to ionizing radiation, but mutants specifically deficient in end-joining (i.e., yku70, yku80, sir2, sir3, sir4, dnl4) are resistant (LEWIS and RESNICK 2000 ). Thus, radiation-induced DSBs are repaired almost exclusively by homologous recombination in this organism. The extreme sensitivity of rad50 mutants suggests that they are defective in some aspect of recombinational repair of radiation-induced DSBs (e.g., processing of the damaged, broken ends to create recombination substrates containing single-stranded 3' overhangs) (RESNICK 1976 ; OSMAN and SUBRAMANI 1998 ). Ionizing radiation survival curves for wild-type cells and for rad50 mutants expressing elevated levels of Exo1 (pCDNA50.1) or TLC1 RNA (pCDNA50.3) are presented in Fig 2. Survival at 20 krad was increased 100-fold by EXO1, suggesting a role for the exonuclease activity in recombinational repair of the damaged DSB ends. Interestingly, TLC1 overexpression produced a modest increase (9-fold) in resistance (Fig 2).
EXO1- and TLC1-mediated suppression is specific for rad50, mre11, and xrs2 strains:
In accord with past observations of phenotypic similarities between strains defective in RAD50, MRE11, and XRS2, we observed that EXO1 and TLC1 overexpression also suppressed the DNA repair defects of mre11 and xrs2 mutants (Fig 3). Complementation by EXO1 was similar in each of the three mutants, but a three- to fivefold greater MMS resistance was consistently observed in mre11 strains when pGAL:: TLC1 was induced (Fig 3). The enhanced repair of MMS-induced lesions did not correlate with a change in checkpoint responses in the mutant strains. As shown in Fig 4, rad50 and mre11 cells rapidly arrest cycling at the G2/M checkpoint in the presence of DNA damage produced by low levels of MMS (scored as distended, large-budded cells; LEWIS et al. 1998 , LEWIS et al. 1999 ). Overexpression of EXO1 or TLC1 did not significantly alter the kinetics or magnitude of this response.
View larger version (110K):
In this window
In a new window
Download PPT slide
|
Figure 3.
Galactose-induced overexpression of EXO1 suppresses MMS-induced lethality in rad50 (YLKL499), mre11 (YLKL503), and xrs2 (YLKL545) mutants more efficiently than TLC1 RNA overexpression. Strains were grown on 2% Gal - Uracil plates with and without 0.3 mM MMS. All dilutions were fivefold. Plasmids were pCDNA50.1 (pGAL::EXO1) and pCDNA50.3 (pGAL::TLC1).
|
|
MMS and radiation sensitivities of rad51, rad52, rad54, and rad59 mutants, which are specifically deficient in the recombinational pathway of DSB repair, were not rescued by EXO1 or TLC1 overexpression. Overexpression of EXO1 (but not TLC1) in rad51 and rad54 mutants reduced plating efficiency (5- to 25-fold) in the absence of damage and modestly impaired cell recovery after exposure to low doses of ultraviolet light and ionizing radiation (Fig 5A). These effects were greater in rad51 and rad54 mutants than in rad52 strains. We also observed that recovery of cells after exposure to DNA damage induced by MMS and ultraviolet light (but not ionizing radiation) was impaired in rad59 mutants synthesizing elevated levels of Exo1 (Fig 5B). One interpretation of these results follows from past observations that, unlike rad52 strains, rad51, rad54, and rad59 mutants are only partially defective in spontaneous or DSB-induced recombination. Inappropriate processing of the ends of damaged DNA by elevated levels of Exo1 may reduce the recombinational capability of the partial function mutants closer to that of rad52 cells. In addition, we observed that killing of two MMS-sensitive, NHEJ-deficient yku70 and sir4 strains (GA911 and GA429; MARTIN et al. 1999 ) was also not rescued by overexpression of EXO1 or TLC1 (data not shown).
View larger version (58K):
In this window
In a new window
Download PPT slide
|
Figure 5.
Effects of EXO1 overexpression on growth and survival of recombinational repair-deficient mutants after exposure to DNA-damaging agents. (A) Galactose-induced expression of EXO1 reduces plating efficiency in rad51, rad52, and rad54 mutants and recovery from DNA damage is significantly impaired in rad51 and rad54 cells. (B) Elevated levels of Exo1 impede the repair of MMS- and ultraviolet light-induced lesions in rad59 mutants. Cells were grown in 2% Gal - Uracil with or without exposure to UV (30 J/m2), ionizing radiation (5 krad), or MMS (0.3 mM). Strains YLKL276, YLKL514, YLKL531, and YLKL532 were employed for the assays. See Fig 2 for plasmid descriptions.
|
|
Disparate impacts of TLC1 and EXO1 on DNA repair and telomere instability defects of yku70 and rad27 mutants:
Mutations in 8 of the 10 genes required for NHEJ exhibit telomere shortening at normal growth temperatures. In ku mutants telomere destabilization is exacerbated at elevated temperatures and cells eventually stop dividing at 37° (FELDMANN and WINNACKER 1993 ; BARNES and RIO 1997 ; BOULTON and JACKSON 1998 ). A previous report indicated that overexpression of EST1, EST2, or TLC1 RNA (using high-copy 2µ plasmids) suppressed the 37° growth defect of Ku mutants, presumably through stabilization of telomeric repeat sequences (NUGENT et al. 1998 ). Each of the GAL::TLC1 constructs isolated in this study could rescue lethality in yku70 strains on Gal-Ura plates at 37° (Fig 6A). In contrast, overexpression of EXO1 had no effect on growth of yku70 cells. Interestingly, although full-length TLC1 rescued both telomere instability and DNA repair defects, overexpression of a truncated version of TLC1 RNA lacking the 16-nt template region did not rescue growth of yku70 mutants, but did partially rescue MMS resistance in rad50 and mre11 strains (Fig 6A and data not shown).
View larger version (75K):
In this window
In a new window
Download PPT slide
|
Figure 6.
DNA metabolic defects of yku70 and rad27 mutants are differentially suppressed by EXO1 and TLC1. (A) Telomere shortening-associated lethality of yku70 cells (YLKL494) at 37° is rescued by overexpression of TLC1 RNA, but not EXO1. (B) DNA repair defects of rad27 mutants (YLKL505) are rescued by EXO1, but not by TLC1. Growth arrest of rad27 cells at 37° (also causing telomere instability) was not suppressed by either EXO1 or TLC1 (see text). See Fig 2 for plasmid descriptions.
|
|
In previous studies, EXO1 overexpression was found to suppress the spontaneous mutator and temperature sensitivity phenotypes of rad27 strains (TISHKOFF et al. 1997 ; PARENTEAU and WELLINGER 1999 ). Cells containing a deletion of RAD27 (the yeast homolog of the human FEN1 gene) exhibit increased telomere instability and aberrant processing of DSB termini in plasmid NHEJ assays (PARENTEAU and WELLINGER 1999 ; WU et al. 1999 ). Telomeres in such mutants retain long 3' single-stranded DNA extensions and are unusually heterogeneous in length at 37°, but do not exhibit the rapid shortening observed in yku70 mutants. The GAL::EXO1 clones isolated in the current study did not suppress the 37° growth defect in the rad27 strains employed here, but did efficiently rescue MMS sensitivity (Fig 6B). In contrast, overexpression of TLC1 RNA did not suppress the MMS- or temperature-sensitive phenotypes of rad27 strains (Fig 6B and data not shown). Thus, elevated levels of TLC1 RNA rescued growth of yku70 mutants at elevated temperatures, but did not suppress the telomere instability of rad27 mutants at 37°.
EXO1 stimulates repair by the recombination pathway:
Additional experiments were performed to assess whether increased levels of Exo1 protein influenced recombinational repair, NHEJ, or both pathways. The effects of EXO1 on recombination proficiency were analyzed directly using an intrachromosomal recombination assay (KOUPRINA et al. 1999 ). Strains used for this assay contain a yeast artificial chromosome (YAC) consisting primarily of human chromosomal DNA with several common yeast markers flanking direct repeats of a pBR plasmid sequence (Fig 7A). Overexpression of EXO1 was accomplished using new constructs containing the GAL1::EXO1 cassette cloned into the centromeric vector pRS313 (see MATERIALS AND METHODS). Spontaneous recombination between the direct repeats was moderately reduced (four- to sixfold) in rad50, mre11, and xrs2 strains (Fig 7B). Overexpression of EXO1 elevated recombination rates of the mutant cells to near wild-type levels and also stimulated recombination in RAD cells (sixfold). These results suggest that EXO1 increases chromosomal recombination generally in cells that retain the homologous pairing and strand exchange activities of Rad51, Rad52, etc. (e.g., SHINOHARA and OGAWA 1998 ).
View larger version (19K):
In this window
In a new window
Download PPT slide
|
Figure 7.
Spontaneous recombination rates are elevated in cells overexpressing EXO1. (A and B) Recombination events between distant, direct repeat sequences (scored as Lys+, Ade+, Ura-, Trp- cells) were scored in wild-type mutant cells with and without overexpression of EXO1. (C) Effects of EXO1 or TLC1 RNA overexpression on NHEJ repair of linear plasmid DNA transformed into yeast cells (see MATERIALS AND METHODS for details). pCDNA50.3 (pGAL::TLC1) and an alternative GAL::EXO1 construct, pLKL53Y (see MATERIALS AND METHODS), were used in the assays.
|
|
All NHEJ-deficient mutants have a reduced ability to recircularize linear plasmids containing complementary overhangs after transformation into haploid cells. The impact of EXO1 overexpression on the end-joining pathway of DSB repair was examined by performing plasmid end-joining assays (see MATERIALS AND METHODS). Linearized pRS314 DNA yielded transformants as efficiently as supercoiled DNA in wild-type cells (152,400 ± 31,700 and 192,300 ± 34,200 transformants per microgram of DNA for supercoiled and linear plasmid DNA, respectively). However, recovery of transformants using broken plasmids was >50-fold lower than with supercoiled plasmids in rad50 and mre11 mutants (Fig 7C). These deficiencies were not rescued by overexpression of either EXO1 or TLC1 RNA, suggesting that the increased DNA repair proficiency is not due to direct or indirect effects on NHEJ.
rad50, mre11, and xrs2 mutants are able to accomplish HO endonuclease-induced mating type switching, an intrachromosomal gene conversion event, but do so with retarded kinetics (IVANOV et al. 1994 ). Expression of HO (using an integrated GAL10::HO cassette) inhibits growth of such strains. Growth inhibition of the rad50 strain YLKL574 (his3::GALHO reg1-501) on plates containing 2% glucose supplemented with small amounts of galactose (0.02 or 0.05%) was alleviated by the presence of a high-copy 2µ plasmid containing EXO1 (pRDK480; Fig 8C). This rescue of cell growth was not observed at high levels of HO expression (e.g., with 2% galactose) and did not occur in an isogenic rad52 strain at any concentration of galactose (data not shown). These results suggest that the exonuclease can process the ends of the DSB at MAT to facilitate mating type switching in cells lacking the Rad50/Mre11/Xrs2 nuclease complex. The dependence on low intracellular levels of HO endonuclease activity suggests that processing by Exo1 is relatively inefficient. Overexpression of EXO1 also rescued lethality in rad50 and mre11 strains expressing EcoRI endonuclease (Fig 8B), which generates multiple chromosomal DSBs containing 5' single-stranded DNA overhangs (LEWIS et al. 1999 ).
View larger version (31K):
In this window
In a new window
Download PPT slide
|
Figure 8.
Expression of mutant proteins containing substitutions within the conserved nuclease domain of Exo1 does not alleviate lethality induced by high-level expression of HO or EcoRI in rad50 mutants. (A) Conserved aspartic acid residues within the nuclease domain of Exo1 and several homologous enzymes (MCCREADY et al. 2000 ). (B and C) rad50 cells constitutively overexpressing Exo1 proteins were transferred to galactose plates to induce expression of EcoRI or HO and colony formation was scored. See MATERIALS AND METHODS for descriptions of strain genotypes.
|
|
Role of the 5'–3' exo activity in cells overexpressing EXO1 and in cells that lack the Rad50/Mre11/Xrs2 nuclease complex:
Formally, the DSB repair-enhancing effects of Exo1 might be a result of increased 5'–3' exonuclease processing of DSB ends or might be achieved by an alternative mechanism such as titration of Exo1-associated proteins (e.g., Msh2 or other DNA end-binding proteins; SOKOLSKY and ALANI 2000 ). Exo1 is a member of a conserved family of nucleases that includes the yeast and human equivalents of FEN-1 and XPG and various DNA polymerases (MCCREADY et al. 2000 ; Fig 8A). The requirement for the enzymatic function of the protein was assessed upon overexpression of mutant proteins (Exo1-D171A and -D173A) containing substitutions within the highly conserved nuclease domain of the protein (SOKOLSKY and ALANI 2000 ; Fig 8A). As shown in Fig 8B and Fig C, in contrast to the wild-type enzyme, neither of the mutant proteins was able to enhance repair of HO or EcoRI-induced DSBs in rad50 cells. This result supports the idea that Exo1-induced recombinational repair involves processing of broken DNA ends via the 5'–3' exo activity of the protein.
The possibility that the basal levels of Exo1 protein present in cells lacking the Rad50/Mre11/Xrs2 complex might provide a "backup" end-processing system for repair of spontaneous or induced DSBs was investigated. As shown in Fig 9, exo1 mutants were not more sensitive to ionizing radiation than wild-type cells and exo1 rad50 double mutants consistently exhibited only slightly more killing than rad50 mutants or severely Rec- rad52 strains. These data suggest that the "dirty-ended" DSBs produced by radiation, which frequently contain damaged or missing bases and sugars, have a strong dependence on processing by the Rad50/Mre11/Xrs2 complex and basal levels of Exo1 contribute very little to repair of such lesions.
View larger version (17K):
In this window
In a new window
Download PPT slide
|
Figure 9.
Cells lacking both nucleases (Exo1 and Rad50/Mre11/Xrs2) exhibit only modestly increased sensitivity to radiation-induced DNA damage (A), but cell cultures contain a large fraction of distended, large-budded cells (arrested at G2/M) (B), suggesting a high level of unrepaired lesions under normal growth conditions. Strains employed for the assays were VL6, YLKL276, YLKL499, YLKL546, and YLKL619. ( ) Wild type, ( ) exo1, ( ) rad50, (•) rad52, ( ) exo1 rad50.
|
|
Interestingly, the exo1 rad50 double mutants exhibited reduced plating efficiency, slower growth rates, increased cell volume, and an elevated fraction of distended, large-budded cells (arrested at G2/M) when propagated on synthetic or rich media (Fig 9B and data not shown). The latter phenotypes suggested that exo1 rad50 strains have elevated levels of unrepaired DNA damage and that Exo1 does serve a backup function for repair of some types of lesions. We investigated the possibility that recombinational repair of DSBs with undamaged ends (unlike radiation-induced DSBs) might exhibit a greater dependence on basal levels of Exo1 in rad50 mutants. An integrating plasmid (pLKL37Y; HIS3, URA3) was cleaved at a unique NcoI site within URA3 and transformed into wild-type and mutant cells along with a control, centromeric vector for normalization of transformation efficiencies (pRS313). The linearized plasmid containing complementary DSB ends efficiently recombined with the ura3-52 heteroallele on chromosome V to produce HIS3+ URA3+ integrants in wild-type cells (Fig 10). Integration into ura3-52 was reduced 4-fold in rad50 strains, but was reduced 50-fold in exo1 rad50 double mutants (similar to Rec- rad51 cells), demonstrating that basal levels of Exo1 can substitute for the Rad50/Mre11/Xrs2 complex for processing of complementary-ended DSBs.
View larger version (20K):
In this window
In a new window
Download PPT slide
|
Figure 10.
DSB-induced plasmid:chromosome recombination is synergistically decreased in cells lacking both the Exo1 and Rad50/Mre11/Xrs2 nuclease complexes. (A) Ura+ His+ recombinants were generated by recombination between the integrating vector pLKL37Y (linearized at the NcoI site within URA3) and a chromosomal ura3-52 allele. (B) Normalized recombination frequencies in wild-type, single-, and double-mutant cells (±SD). Strains VL6, YLKL499, YLKL532, YLKL546, and YLKL619 were employed for the assays.
|
|
| DISCUSSION |
|---|
EXO1 overexpression elevates recombination rates, but does not rescue telomere instability or NHEJ repair defects:
Overexpression of EXO1 from a GAL1 promoter enhanced resistance to MMS and ionizing radiation in rad50, mre11, and xrs2 strains, but did not mitigate the damage-induced cell cycle arrest response of the cells. In addition, we observed that EXO1 overexpression restored growth to rad50 strains after cleavage of chromosomal DNA by HO endonuclease. This result suggests an increased efficiency of gene conversion (mating type switching) due to the actions of the exonuclease. Overexpression of the enzyme also rescued EcoRI-induced cell killing in rad50 and mre11 mutants and increased spontaneous recombination rates in wild-type cells, as well as in rad50, mre11, and xrs2 strains. In contrast, overexpression of the exonuclease did not suppress the severe defect in NHEJ in these mutants and did not suppress the MMS sensitivities of NHEJ-deficient yku70 or sir4 strains. These latter results are consistent with recent observations that EXO1 overexpression does not affect telomere shortening, thought to be an NHEJ-associated defect, in rad50, mre11, and xrs2 mutants (CHAMANKHAH et al. 2000 ; TSUBOUCHI and OGAWA 2000 ).
Spontaneous intrachromosomal recombination rates are moderately decreased in exo1 mutants (FIORENTINI et al. 1997 ). In addition, survival and cell cycle arrest responses of exo1 mutants after induction of EcoRI endonuclease expression in vivo are similar to those of several recombination-defective strains (i.e., rad51, rad52, etc.), but are distinct from the responses of NHEJ mutants (LEWIS et al. 1999 ). These data, in conjunction with other recent reports of EXO1-mediated suppression of DNA repair and replication pathways (PARENTEAU and WELLINGER 1999 ; CHAMANKHAH et al. 2000 ; SOKOLSKY and ALANI 2000 ; TSUBOUCHI and OGAWA 2000 ), place the function(s) of Exo1 in the recombination pathway of DSB repair, but not in NHEJ. The primary function(s) of the enzyme in DNA metabolism is unclear, however. For example, Exo1 protein forms a complex with the mismatch repair protein Msh2 inside cells, deletion of the gene produces a mutator phenotype in haploid cells and is synthetically lethal when combined with mutations in the DNA replication gene RAD27, and EXO1 transcription is induced during meiosis (CHAMANKHAH et al. 2000 ; SOKOLSKY and ALANI 2000 ; TSUBOUCHI and OGAWA 2000 ).
Interestingly, Exo1 did not provide increased resistance to MMS or radiation in rad51, rad52, rad54, or rad59 strains, which are primarily defective in the recombinational repair pathway. Instead, recovery of rad51, rad54, and rad59 cells exposed to DNA damage was impaired by EXO1 overexpression. Although increased Exo1 activity did inhibit recovery in rad52 cells, the effects were not as pronounced as in the other mutants. Previous studies have demonstrated that rad51, rad54, and rad59 mutants are not as deficient as rad52 strains in several classes of mitotic recombination events (e.g., AGUILERA 1995 ; RATTRAY and SYMINGTON 1995 ; SUGAWARA et al. 1995 ; ZOU and ROTHSTEIN 1997 ; KOUPRINA et al. 1999 ; PARK et al. 1999 ). We suggest that the elevated exonuclease activity leads to inappropriate processing of DNA ends in the rad51, rad54, and rad59 mutants and interferes with the substantial residual recombination capabilities of these strains. This interference would therefore effectively reduce the repair capacity of the partially recombination-defective cells toward that of rad52 mutants.
exo1 rad50 double mutants were only slightly more sensitive to ionizing radiation than rad50 or rad52 strains, but grew slowly, displayed reduced viability, and had an increased fraction of distended, G2-arrested cells (60% of total cells). These data suggest that cells lacking both nuclease enzymes have elevated levels of unrepaired DNA damage, which produce chronic activation of the damage-responsive checkpoint system and extended time in G2. In agreement with the idea that Exo1 is functionally redundant with the Rad50/Mre11/Xrs2 complex for repair of some types of lesions, we found that targeted recombination between cohesive-ended, linear plasmid DNA and yeast chromosomal DNA was synergistically decreased in exo1 rad50 double mutants relative to either single mutant.
The Rad50/Mre11/Xrs2 complex of yeast (and its human cell counterpart consisting of hRAD50, hMRE11, and hNBS1) exhibits ssDNA endonuclease and 3'–5' exonuclease activities in vitro (FURUSE et al. 1998 ; USUI et al. 1998 ; MOREAU et al. 1999 ; PETRINI 1999 ; CHAMANKHAH et al. 2000 ). Purified Mre11 protein retains the individual nuclease functions and the Rad50 subunit is an ATP-binding protein (HOPFNER et al. 2000 ). The finding that the yeast complex is a nuclease suggests that it is required in recombinational repair to process the broken ends of DNA for subsequent homology search and strand invasion reactions (although a possible role in resolution of recombination intermediates cannot be ruled out yet). This idea is consistent with several models of recombination that posit the processing of DSB ends to generate a single-stranded 3' DNA tail as an early step (RESNICK 1976 ; SZOSTAK et al. 1983 ; THALER and STAHL 1988 ; OSMAN and SUBRAMANI 1998 ). The Rad50/Mre11/Xrs2 complex (and purified Mre11 protein) is a 3'–5' exonuclease, which is inconsistent with recombination models since this activity should result in formation of a 5' single-stranded tail rather than a 3' tail. Recently, PAULL and GELLERT 1999 demonstrated that the trimeric complex (but not the individual proteins) contains a weak DNA unwinding activity. Thus, it is formally possible that sequential DNA unwinding and ssDNA endonucleolytic cleavage of the 5'-ended strand could function to generate an extended 3' tail for strand invasion. Alternatively, a separate DNA helicase enzyme may participate in the early steps of the reaction, but further experiments are needed to assess this possibility.
Recent analyses of mre11 mutants (containing substitutions within conserved sequence motifs thought to be important for nuclease function) have implied that the nuclease activities are not essential for many repair functions of the complex in vivo (BRESSAN et al. 1998 , BRESSAN et al. 1999 ; USUI et al. 1998 ; MOREAU et al. 1999 ). However, the observation that overexpression of EXO1 specifically elevates recombinational repair in rad50, mre11, and xrs2 mutants strongly suggests that one or both nuclease activities of the complex participate in this pathway of repair. We suggest that the 5'–3' exonuclease activity of Exo1 partially substitutes for missing activities of the Rad50/Mre11/Xrs2 nuclease complex at an early stage in recombinational repair of induced DSBs (though a role for the complex later, during the resolution step, cannot yet be ruled out). In this model the 5'–3' exonuclease activity of Exo1 would associate with either one or both ends of the DNA and degrade the strand containing a 5' end to produce a 3' tail. This structure would then serve as substrate for the subsequent actions of the recombination apparatus consisting of Rad51, Rad52, Rad54, Rad55, Rad57, etc. The observation that EXO1 expression was unable to alleviate the sensitivity of rad51, rad52, rad54, or rad59 mutants to MMS or radiation is consistent with this scheme as these proteins are required for downstream events. Furthermore, we note that EXO1 overexpression did not rescue MMS-induced killing of Rec+ Nhej- yku70 or sir4 mutants, which also suggests a specificity for recombinational repair. Other possible consequences of EXO1 overexpression, e.g., involving mismatch repair or the processing of Okazaki fragments (TISHKOFF et al. 1997 ; SOKOLSKY and ALANI 2000 ), appear less likely to lead to enhancement of recombinational repair.
TLC1 RNA overexpression selectively suppresses defects in DSB repair and telomere maintenance:
Ten independent galactose-regulated TLC1 cDNAs were isolated as suppressors of MMS-induced cell killing in rad50 mutants. Only one of the clones contained the complete 1301 bp of the gene, though the remaining clones were only slightly less than full length. All isolates retained the centrally located 16-nucleotide region postulated to act as template for extension of the 3' ends of chromosomal DNA during S phase (SINGER and GOTTSCHLING 1994 ). As with EXO1, overexpression of TLC1 RNA increased resistance to MMS in rad50, mre11, and xrs2 mutants, but not in several Rec- mutants (rad51, rad52, rad54, or rad59) or Nhej- mutants (yku70 or sir4). TLC1 RNA did not alter the DNA damage-induced cell cycle arrest response of rad50 or mre11 strains and did not rescue the severe defect in NHEJ in these mutants. Overexpression of TLC1 RNA increased resistance to ionizing radiation (thought to be primarily dependent on recombinational repair processes) to a lesser extent than to MMS, which produces lesions that are repaired by several pathways (XIAO et al. 1996 ). In contrast to results obtained with EXO1, the small impact of TLC1 on radiation resistance suggests that potential effects on recombinational repair are quite modest (see below).
Like Ku-deficient strains, rad50, mre11, and xrs2 mutants display telomere shortening and senescence at normal growth temperatures (LEWIS and RESNICK 2000 ), although these mutants do not exhibit the exacerbation of this defect and concomitant growth impediment of yku70 or yku80 cells at 37° (a slight temperature sensitivity has been reported in RAD50-deficient cells in one strain background; BOULTON and JACKSON 1998 ). Each of the new GAL::TLC1 clones (but not GAL::EXO1) was able to rescue the 37° growth impairment of yku70 mutants. Growth defects of yku70 strains were also suppressed by EST2 expressed from a 2µ plasmid, as previously observed by NUGENT et al. 1998 (data not shown). The results are consistent with the idea that increased expression of specific components of telomerase results in suppression of the rapid telomere shortening defects of ku mutants. These data also provide additional evidence that the effects of EXO1 and TLC1 overexpression are mechanistically distinct.
Interestingly, TLC1 RNA overexpression did not rescue the 37° growth impairment of rad27 mutants. Cells lacking Rad27 (the yeast homolog of the human flap endonuclease FEN1) are not deficient in standard assays for NHEJ (e.g., recircularization of linear plasmid DNA, EcoRI sensitivity), but do feature aberrant processing of noncomplementary DSB ends in plasmid NHEJ repair (WU et al. 1999 ). Such mutants also exhibit increased spontaneous mutation and recombination and defective processing of Okazaki fragments during DNA replication and may function in base excision repair (KUNKEL et al. 1997 ; GARY et al. 1999 ). In addition, telomeres retain long 3' single-stranded DNA extensions and become highly unstable at 37° in rad27 mutants (PARENTEAU and WELLINGER 1999 ). At this elevated temperature cells arrest growth and the lengths of telomeres become heterogeneous, but the telomeres do not exhibit the progressive shortening observed in yku70 and yku80 mutants. In the current study, EXO1 expression complemented the DNA repair defect (MMS sensitivity) of rad27 mutants, but had no effect on growth at 37° (see above). Although overexpression of TLC1 RNA (and EST2) rescued growth of yku70 strains at the restrictive temperature, it had no effect on MMS sensitivity or the telomere instability-linked 37° growth defects of rad27 mutants. These results suggest that increased levels of telomerase components can alleviate deficits involving progressive loss of telomeric repeats (as in ku strains), but not the types of telomere instability observed in rad27 mutants.
Formation of new telomeres at sites of chromosomal DSBs has been observed previously in yeast cells (KRAMER and HABER 1993 ; SCHULZ and ZAKIAN 1994 ; CHEN and KOLODNER 1999 ). However, several results suggest that the enhanced DNA repair contributed by TLC1 overexpression in rad50, mre11, and xrs2 mutants is not due to a general increase in capping of DSB ends. First, elevated levels of TLC1 RNA did not increase MMS resistance in wild-type cells or in rad51, rad52, rad54, rad59, yku70, or sir4 strains, arguing against a general capping mechanism. Second, 10 independent library plasmids containing GAL::TLC1 were isolated in the suppressor screen, but other genes associated with telomerase activity (e.g., EST1, EST2, EST3, or CDC13) were not found. Third, overexpression of TLC1 RNA has been shown to cause a modest reduction in average telomere lengths in yeast cells (SINGER and GOTTSCHLING 1994 ). Finally, we have recently observed that overexpression of a truncated form of TLC1 RNA lacking the 16-nt template region partially rescues MMS resistance in rad50 and mre11 strains (unpublished results). These data imply that complementation by TLC1 is not due to a general chromosome "healing" process whereby a subset of MMS- or radiation-induced DSBs are capped by de novo telomere formation. Increased telomere stabilization (possibly due to increased telomerase activity) appears to be involved in suppression of ku phenotypes by TLC1 (and by EST2), but not for suppression of rad50, mre11, and xrs2 defects. It is possible that suppression involves a competition between different enzyme complexes (e.g., consisting of telomerase, NHEJ, or recombinational repair proteins) for access to the ends of induced DSBs. For example, increased numbers of TLC1 RNA molecules might titrate away components of telomerase (or associated proteins) and alter protein: DNA interactions and/or processing of broken DNA ends produced by MMS or radiation.
DNA at the ends of yeast chromosomes forms complex local and higher order structures within the nucleus. For example, telomeres on different chromosomes appear to cluster in specific foci and form T-loops, whereby the DNA at the end of a chromosome loops back and anneals to internal repeat sequences (LOWELL and PILLUS 1998 ; GRIFFITH et al. 1999 ). In addition, telomeric and subtelomeric DNA regions are subject to transcriptional silencing, a phenomenon mediated through epigenetic changes in chromatin structure (LOWELL and PILLUS 1998 ; LUSTIG 1998 ). TLC1 was originally identified as one of several genes that reduced silencing at telomeres (but not at internal mating type loci) when overexpressed (SINGER and GOTTSCHLING 1994 ; SINGER et al. 1998 ). Reduced telomeric silencing has also been observed in several mutants, including sir3 and sir4 strains, where changes in telomere structure and localization within the nucleus were also noted (PALLADINO et al. 1993 ; GOTTA et al. 1996 ). Interestingly, a recent report revealed that three nuclear pore-associated proteins (Mlp1, Mlp2, and Nup145) are required for proper localization of telomeres and Yku70 protein within the nucleus and for efficient repair of DNA damage induced by bleomycin (GALY et al. 2000 ). The possibility that TLC1 RNA might also influence telomere clustering and/or localization at the nuclear periphery remains to be investigated.
| FOOTNOTES |
|---|
1 Present address: Department of Chemistry and Biochemistry, Southwest Texas State University, San Marcos, TX 78666. 
| ACKNOWLEDGMENTS |
|---|
We are grateful to Eric Alani, Alan Goldstein, Jakob Kirchner, Victoria Lundblad, and John McCusker for plasmids used in the study. The authors also thank Kirill Lobachev, Jim Mason, and Doug Thrower for providing comments on the manuscript.
Manuscript received April 18, 2001; Accepted for publication October 17, 2001.
| LITERATURE CITED |
|---|
TOP
ABSTRACT
MATERIALS AND METHODS
) Rad+. rad50: (
) pGAL::EXO1, (
) pGAL::TLC1, (
) pRS316.