A Genomics-Based Screen for Yeast Mutants With an Altered Recombination/End-Joining Repair Ratio

Corresponding author: Thomas E. Wilson, University of Michigan Medical School, 1301 Catherine Rd., M4214 Med Sci I, Box 0602, Ann Arbor, MI 48109-0602., wilsonte{at}umich.edu (E-mail)

Communicating editor: L. S. SYMINGTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We recently described a yeast assay suitable for genetic screening in which simple religation nonhomologous end-joining (NHEJ) and single-strand annealing (SSA) compete for repair of an I-SceI-created double-strand break. Here, the required allele has been introduced into an array of 4781 MATa deletion mutants and each strain screened individually. Two mutants (rad52 and srs2) showed a clear increase in the NHEJ/SSA ratio due to preferential impairment of SSA, but no mutant increased the absolute frequency of NHEJ significantly above the wild-type level. Seven mutants showed a decreased NHEJ/SSA ratio due to frank loss of NHEJ, which corresponded to all known structural/catalytic NHEJ components (yku70, yku80, dnl4, lif1, rad50, mre11, and xrs2); no new mutants in this category were identified. A clearly separable and surprisingly large set of 16 other mutants showed partial defects in NHEJ. Further examination of these revealed that NEJ1 can entirely account for the mating-type regulation of NHEJ, but that this regulatory role was distinct from the postdiauxic/stationary-phase induction of NHEJ that was deficient in other mutants (especially doa1, fyv6, and mck1). These results are discussed in the context of the minimal set of required proteins and regulatory inputs for NHEJ.

EUKARYOTIC cells possess two enzymatically distinct pathways for double-strand break repair (DSBR; reviewed in PAQUES and HABER 1999 ; JACKSON 2001 ). Repair by homologous recombination involves the concerted action of the RAD52 epistasis group of genes (RAD50–52, 54–55, –57, –59, MRE11, and XRS2). The products of these genes, along with other cellular factors, execute a resection of the 5' ends at a DSB to create 3' nucleoprotein filaments, followed by strand exchange with a homologous donor duplex, synthesis from the broken 3' termini, and ultimately resolution of the extended D-loop. In the absence of a donor duplex, but where a sequence is repeated in tandem on either side of the break, Rad52 and Rad59 can also catalyze the direct annealing of the two resected 3' ends independently of the other epistasis group members to create a deletion in a pathway known as single-strand annealing (SSA; IVANOV et al. 1996 ; SHINOHARA et al. 1998 ; SUGAWARA et al. 2000 ). In contrast, repair by nonhomologous end-joining (NHEJ) entails engagement and likely end-to-end bridging by the Ku heterodimer (Yku70/Yku80 in budding yeast; JONES et al. 2001 ) and ultimately ligation by the DNA ligase IV/XRCC4 complex (Dnl4/Lif1 in yeast; WILSON et al. 1997 ; HERRMANN et al. 1998 ). Interestingly, the Mre11-Rad50-Xrs2 complex is also required for NHEJ, suggesting an early role for this complex in DSBR (PETRINI 1999 ). Other genes are only variably required for NHEJ. POL4 is required for yeast NHEJ only when the termini are incompatible and require processing prior to religation (WILSON and LIEBER 1999 ). Artemis and DNA polymerase µ likely serve similar roles during mammalian NHEJ (MA et al. 2002 ; MAHAJAN et al. 2002 ). Mutants of SIR2-4 are NHEJ deficient, but only because they are functionally of the a/ mating type, which represses NHEJ by a mechanism now known to involve NEJ1 (LEE et al. 1999 ; FRANK-VAILLANT and MARCAND 2001 ; KEGEL et al. 2001 ; VALENCIA et al. 2001 ). Finally, higher eukaryotic cells depend on DNA-PKcs for efficient NHEJ, but this gene is not conserved in budding yeast (SMITH and JACKSON 1999 ).

Many questions remain regarding the mechanism of each DSBR pathway, as well as how these seemingly competitive and redundant processes are coordinated to optimize the likelihood of genome restoration. Genetic screens for DSBR-deficient mutants have played an important role in the discovery process, but with limitations. In mammalian systems, studies of radiosensitive Chinese hamster cell lines (THOMPSON et al. 1980 ) and immunodeficient mice and human patients (GENNERY et al. 2000 ) have revealed some components of both homologous and nonhomologous repair mechanisms, but the laboriousness of the genetic manipulations and diploid nature of the cells have left the picture incomplete. In yeast, early screens for radiosensitive mutants revealed the RAD52 epistasis group genes (RESNICK 1969 ; GAME and MORTIMER 1974 ), but uniformly failed to detect NHEJ components due to the relatively greater importance of recombinational repair. Indeed, the power of the yeast genetic system for screening DSBR mutations has not been fully realized due to this and other technical limitations.

In this study, I capitalized on three recent technological developments to perform a comprehensive yeast genetic screen that had the ability to find not only those mutants deficient in the SSA and NHEJ repair pathways, but also those that changed the relative NHEJ/SSA repair ratio. These developments were, first, the availability of array sets of deletion mutants of nearly all genes of Saccharomyces cerevisiae (WINZELER et al. 1999 ); second, the description of an assay, termed suicide deletion, that not only can detect both SSA and NHEJ repair events in a simple plating format, but also can distinguish them and reveal their ratio by a simple color readout (KARATHANASIS and WILSON 2002 ); and third, the development of techniques to rapidly introduce the critical test allele into the mutant array (TONG et al. 2001 ; VANCE and WILSON 2002 ). The screen revealed all known, but no novel, genes required for catalysis of NHEJ, as well as several novel genes that proved to serve two separable regulatory roles promoting NHEJ in the haploid and postdiauxic/stationary growth stages.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains, manipulation, and media:
All strains were isogenic derivatives of BY4741 (BRACHMANN et al. 1998 ). The subarray of DNA damage response mutants will be described elsewhere (VANCE and WILSON 2002 ). The MATa array was generated by the Saccharomyces Genome Deletion Project (WINZELER et al. 1999 ) and obtained from Research Genetics (Birmingham, AL). Media were as described (KARATHANASIS and WILSON 2002 ). Mutant arrays were manipulated using ethanol-sterilized manual replicators.

Plasmid construction:
All PCR for the following constructions was performed using the Advantage HF high-fidelity PCR kit (CLONTECH, Palo Alto, CA). pMATa was constructed by ligating a PCR fragment corresponding to the MATa allele of BY4741 into the BamHI and SalI polylinker sites of the CEN/HIS3 vector pRS413. pNEJ1 plasmid was made as follows. First, we constructed pTW367, a derivative of the CEN/LEU2 plasmid pTW300 (WILSON and LIEBER 1999 ) in which a single SmaI site was created downstream of the ADH1 promoter and start codon. SmaI-digested pTW367 was then cotransformed into a nej1 strain with a PCR fragment corresponding to the NEJ1 coding sequence that bore 5' tails to direct gap repair with pTW367 (5'-ACCATGGCGTCCGAGCAAAAGCTCATTTCTGAAGAGGACTTGCGC and 5'-TTTATGTAACGTTATAGATATGAAGGATTTCATTCGTCTGTCGAC for the amino- and carboxyl-terminal sides, respectively). The pNEJ1 plasmid was recovered from an isolate in which the nej1 mutation had been complemented prior to reintroduction into other strains.

Allele and strain construction:
Construction of the ade2::SD2+::URA3 and ade2::SD0+::URA3 alleles has been described previously (KARATHANASIS and WILSON 2002 ). The ade2::SD2+::URA3 allele was further modified to facilitate high-throughput mating by adding to it the -mating-type-specific marker gene STE3-MET15. This was achieved by creating PCR fragments corresponding to the STE3 promoter region and MET15 coding sequence and then fusing them in a second round of PCR by virtue of overlaps in the primers at the STE3-MET15 junction. The product was then recombined into the ADE2-RGA1 intergenic region by virtue of tails on the outside primers. The function of both ADE2 and RGA1 was preserved because no genomic sequence was deleted and the insertion point was in the 3' untranslated region of both genes. The resulting strain, YW798 (MAT ade2::SD2+::URA3::STE3-MET15 his31 leu20 met150 ura30), was identical to BY4741 except at the MAT and ADE2 loci.

The strategy used to introduce the suicide deletion allele into the arrays will be described in detail elsewhere (VANCE and WILSON 2002 ). Briefly, YW798 was mated to an array in liquid culture in microtiter dishes and then sporulated on plates. Spore cultures were transferred into microtiter dishes, diluted, and then spotted back to glucose germination plates lacking methionine and uracil and containing 200 µg/ml G418. Red colonies on these plates must be haploid MAT ade2::SD2+::URA3::STE3-MET15 mutx::kanMX4 (where MUTx refers to any of the genes deleted in the different array strains) because the STE3 promoter is active only in cells of the mating type. These were picked and passaged twice in the same liquid medium prior to spotting. All screen positives were repurified prior to quantitative testing, and the identities of the doa1, fyv6, and mck1 mutants were all verified by allele-specific PCR.

The nej1 mutx double-mutant strains were constructed by first isolating the nej1::kanMX4 strain from the MATa array and changing its marker from kanMX4 to LEU2 by PCR-mediated replacement. The resulting nej1::LEU2 allele was then introduced into a mini-array of the desired MAT ade2::SD2+::URA3::STE3-MET15 mutx::kanMX4 strains by the above method except that leucine was also omitted from the media.

Suicide deletion screening assay:
Three assay plates were routinely spotted (all additionally lacked methionine and contained 200 µg/ml G418): glucose complete (as a growth control), galactose complete (to detect mutants with an increased NHEJ/SSA ratio), and galactose lacking adenine (to detect NHEJ-deficient mutants; see Fig 1 and Fig 2 for phenotype scoring). In retests, a fourth glucose plate lacking adenine was also spotted to rule out the presence of rare contaminating diploid cells that had somehow managed to become Met⁺; when detected, these mutants were purified prior to further testing. Scoring of plates was performed by scanning them into computer image files. These were magnified and examined using a Microsoft Access database where all spots for an individual strain could be readily compared. Numerous parameters were recorded for each strain that were subsequently used to assign one of the following screen outcomes: failure (insufficient growth to score), negative (no significant difference from the typical spot appearance), or positive (spot growth or color differed on galactose but not on glucose plates). All primary data from the screen can be viewed at http://tewlab.path.med.umich.edu/SDScreen/frames.html.

View larger version (54K): In this window In a new window Download PPT slide   Figure 1. Sensitivity of the suicide deletion screen in spotting format. (A) Diagram of the ade2::SD2+ suicide deletion allele used in these experiments. After induction with galactose, I-SceI is expressed and cuts its gene from the genome, along with the URA3 marker, via two flanking cleavage sites. The resulting chromosome ends can be fused by SSA via the 28-bp direct repeat (red box, DR) or by simple religation NHEJ, which for this allele give ade2/red and ADE2/white yeast, respectively. (B) A test mini-array was spotted to a glucose complete control plate (top) as well as galactose indicator plates with and without adenine (middle and bottom, respectively). Positions marked "-" all contained pure cultures of wild-type yeast bearing the ade2::SD2+ allele shown in A. Positions marked with numbers contained the indicated percentages of the following strains: rad52 and yku70 yeast bearing the same ade2::SD2+ allele and wild-type yeast bearing the ade2::SD0+ allele (not drawn) for which SSA events are ADE2/white and NHEJ events are ade2/red. This last strain served as a sensitivity indicator for ade2::SD2+ mutants in which the NHEJ/SSA (i.e., color) ratio was reversed (see text). The remaining fraction of the mixtures was made of wild-type yeast bearing the ade2::SD2+ allele.

View larger version (77K): In this window In a new window Download PPT slide   Figure 2. Validation of the suicide deletion spotting screen using a "subarray" of DNA damage response mutants. The ade2::SD2+ allele was introduced into an array of 96 mutants by the method diagrammed in A and described in MATERIALS AND METHODS. The final strains then were spotted to the control and indicator plates (B). Grid positions that contain mutants known to be defective in SSA or NHEJ are labeled. The identities of the remaining mutants are omitted here for clarity; the complete searchable data set may be viewed at http://tewlab.path.med.umich.edu/SDScreen/frames.html. Two strains could not be picked in this experiment and so their grid positions are empty (grid positions D3 and G4, corresponding to rad6 and mre11, respectively). Grid position D1 corresponds to msh1; this strain illustrates the phenotype typical of a petite/very slow-growing mutant and does not reflect a DSBR deficiency of this strain.

Suicide deletion quantitative assay:
Two different methods, which differ by whether DSB induction occurred on plates or in liquid medium, were used. For each, source cultures were routinely grown in synthetic complete glucose medium lacking uracil and containing 40 µg/ml adenine (further lacking histidine and/or leucine as required for plasmid maintenance) for 48 hr to ensure that even slow-growing mutants had achieved early stationary phase. When indicated, exponential-phase source cultures were instead grown from high dilution in the same medium overnight so that the OD₆₀₀ was <1.0 in the morning. In method 1, 10-fold serial dilutions of the glucose source cultures were made in water, and appropriate volumes plated to synthetic defined medium. The absolute frequencies of SSA and NHEJ were determined by the percentage survival on galactose plates relative to parallel platings to glucose, where completely red colonies on galactose complete were counted as SSA events, and all colonies on galactose lacking adenine were counted as NHEJ events. The NHEJ/SSA repair ratio was calculated from these. In method 2, 5 x 10⁶ cells from the glucose source cultures were used to inoculate 2.5 ml of synthetic complete galactose medium containing 40 µg/ml adenine (further lacking histidine and/or leucine as required for plasmid maintenance), followed by shaking at 30° for 48 hr. Tenfold serial dilutions of these cultures were made and appropriate volumes plated to synthetic defined glucose plates with and without adenine. The NHEJ/SSA ratio was calculated by dividing the corrected colony count from the plate lacking adenine by the corrected red colony count from the plate containing adenine.

Viability and thermotolerance in stationary phase:
Exponential-phase cultures were diluted to a calculated starting OD₆₀₀ of 0.005 in 20 ml YPAD and allowed to grow with shaking in 50-ml tubes for 13 days. After most cultures had reached an OD₆₀₀ 5 (empirically determined to correspond to the diauxic shift), the colony-forming units (cfu) per milliliter of the cultures were determined at various time points by plating appropriate dilutions to YPAD and counting colonies formed after 3 days at 30°. Because the fyv6 strain never achieved as high a density as the other mutants tested, the cfu per milliter values obtained for each strain were normalized to the average of the values obtained for the first three time points (i.e., before the onset of stationary phase) to facilitate comparison of strains. At the last three time points, a 500-µl aliquot was heat-shocked in a 55° water bath for 5 min, followed by cooling on ice for 5 min, prior to diluting and plating to determine the thermotolerant cfu per milliliter.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Sensitivity of the suicide deletion screen:
The DSBR assay used throughout this study is based on the ade2::SD2+ suicide deletion allele diagrammed in Fig 1A and described in detail in KARATHANASIS and WILSON 2002 . Briefly, the I-SceI endonuclease creates two DSBs when its expression is induced with galactose, allowing repair in which the two chromosome ends are either ligated by NHEJ or joined by SSA via 28-bp terminal direct repeats. These DSBs result in the excision of the GAL1-I-SceI gene cassette from the chromosome that terminates I-SceI expression and largely prevents the recleavage of the newly created I-SceI site that would otherwise prevent outgrowth of the efficient simple religation NHEJ event. SSA and NHEJ events are distinguished by ADE2 status, allowing for both selection for the Ade⁺ NHEJ events as well as color readout in both screening (i.e., spotting) and quantitative plating formats. NHEJ repairs 3% of the broken chromosomes in this system whether or not the direct repeats are present, with SSA being 10-fold more efficient (KARATHANASIS and WILSON 2002 ).

To estimate the sensitivity of the ade2::SD2+ suicide deletion assay in screening format, I spotted the wild-type strain as well as strain mixtures in which varying percentages had been replaced with either rad52 or yku70 mutant cells (Fig 1B). When spotted to galactose-complete plates, the mixture containing 90% rad52 cells could reliably be detected as SSA-deficient as evidenced by increased whiteness of the spot. When spotted to galactose plates lacking adenine, the mixture containing 80% yku70 cells could reliably be detected as NHEJ deficient as evidenced by decreased spot density. These correspond to detection limits of an 10-fold decrease in SSA and a 5-fold decrease in NHEJ. The screen was also designed to detect putative regulatory mutants in which the total repair rate was preserved but the NHEJ/SSA repair ratio was increased. Since no such mutant was known, this sensitivity was estimated by mixing the wild-type ade2::SD2+ strain with a wild-type strain bearing the ade2::SD0+ allele (KARATHANASIS and WILSON 2002 ), for which SSA events are ADE2/white and NHEJ events are ade2/red, thus mimicking a partial to complete reversal of the color ratio. The screen was in fact most sensitive to this mutant class, being able to consistently detect a mixture containing only 40% ade2::SD0+ cells, corresponding to a <2-fold change in the color ratio.

Validation of the suicide deletion screen with a panel of DNA damage response mutants:
Prior to embarking on large-scale screens, the approach was further validated using a single-plate "subarray" of deletion mutants known to be deficient in the spectrum of damage response functions. To perform this test screen, as well as the full screen described below, it was of course necessary to introduce the ade2::SD2+ allele into the mutant strains. This was accomplished by a high-throughput mating strategy as described in MATERIALS AND METHODS and Fig 2A. The final subarray contained three mutants known to be deficient in SSA and six deficient in NHEJ. All were readily scored as positive except rad59 (Fig 2B). This mutant also proved negative on further screening, which likely reflects the atypical nature of the suicide deletion SSA event (see DISCUSSION). No other mutants showed a defect, demonstrating the power of examining strains individually and the specificity of the phenotypes for the predicted changes in DSBR efficiency.

Suicide deletion screen of 4781 haploid deletion mutants:
The ade2::SD2+ suicide deletion allele was next introduced into an array of nearly all viable haploid yeast deletion mutants (WINZELER et al. 1999 ), and the resulting strains were scored for the phenotypes described above. The screen progress is diagrammed in Fig 3A. A total of 136 strains were ultimately scored as positive. This number is large because the inclusion threshold was deliberately kept low to maintain high sensitivity. Quantitative analysis provided a facile and precise way of subsequently eliminating the false positives. This was initially performed using method 1, in which cells were plated to galactose so that both absolute and relative repair frequencies could be assessed (see MATERIALS AND METHODS). The NHEJ/SSA repair ratio was examined first. This parameter provides the greatest power of the suicide deletion assay, since all counted events must have induced GAL genes and grown in the presence of galactose, which in turn means that they must have all broken and repaired chromosome XV. This ratio is therefore largely independent of potential growth biases. As shown in Fig 3B, most screen-positive mutants showed no alteration in the NHEJ/SSA ratio. This was expected and represents the fact that growth defects in many strains had caused them to be scored as falsely positive in the screen. These strains indeed provided essential controls demonstrating the reproducibility of the quantitative assay, its insensitivity to general growth defects, and the significance of those mutants that did deviate from the wild-type range. Among the DSBR-deficient strains, a nearly 4-log range of NHEJ/SSA ratios was observed (Fig 3B). Considering absolute repair frequencies (i.e., percentage survival), the wild-type strain showed NHEJ and SSA frequencies of 4.5 and 58%, respectively (Table 2), similar to previous results (KARATHANASIS and WILSON 2002 ). Mutant strains with altered NHEJ/SSA ratios nearly always showed a decrease in only one of these frequencies with the other falling in the normal range and with hpr5/srs2 being a notable exception (Table 2; see DISCUSSION). The combined data allowed these mutants to be categorized into four classes of DSBR defect (Table 1; see DISCUSSION).

View larger version (35K): In this window In a new window Download PPT slide   Figure 3. Results of the screen of the complete MATa array. (A) The chart shows the number of mutants removed from consideration at various stages of the screening process. Untestable mutants and screen failures are described in the text; the majority of these were petite. The 736 retested strains were picked as new segregants from fresh mating and sporulation reactions. (B) The 136 mutants positive by spotting were subjected to quantitative analysis by method 1. The data are plotted here as the average mutant NHEJ/SSA ratio expressed relative to the average wild-type NHEJ/SSA ratio. The cluster of open circles to the left in the chart represents replicates of the wild-type strain as a demonstration of the inherent variability in the assay (the average of these points is 1.0). The cluster of solid circles to the right of this corresponds to the majority of mutants, tested only once since they showed no defect. The remaining points are the mean ± standard deviation for the wild-type strain (open circle) and those mutants tested in replicate (solid circles), ordered from lowest to highest NHEJ/SSA ratio. The horizontal dashed lines indicate the approximate NHEJ/SSA ratios that corresponded to the statistical significance thresholds of Table 1 and Table 2.

One potential bias of method 1 was that the NHEJ frequency of a mutant would be underestimated if it had a specific growth defect on medium lacking adenine, where there was already an inherent colony size heterogeneity due to limited I-SceI recleavage (KARATHANASIS and WILSON 2002 ). For this reason, many positives were repeated by quantitative method 2, in which galactose induction of all cells, and therefore DSB induction and repair, occurs in the same nonselective liquid culture. While this allows only estimation of the NHEJ/SSA ratio, it is largely free of growth biases. Method 2 generally agreed with method 1, but with a few exceptions. Specifically, the yaf9, ubi4, htz1, and ard1 mutants showed a lesser NHEJ deficiency with method 2, although yaf9, ubi4, and ard1 mutants still showed a significant and reproducible defect. Of these, ubi4 was surprising because this mutant grew at wild-type rates, so the basis for the method dependence is not clear.

NEJ1 mediates mating-type regulation of NHEJ:
At least two yeast cell-cycle states influence DSBR pathway utilization by enhancing NHEJ: haploid mating types (LEE et al. 1999 ) and the transition to postdiauxic/stationary phase (KARATHANASIS and WILSON 2002 ). Identification of genes that contribute to this regulation was a major goal of the screen. Indeed, among the class of mutants with deficient NHEJ, all novel genes showed only a partial deficiency as compared with the previously known mutants such as yku70, consistent with the notion that they may fulfill regulatory roles. At this point in my work, other groups reported that nej1 mutant strains were NHEJ deficient because they were unable to induce NHEJ in the haploid state (FRANK-VAILLANT and MARCAND 2001 ; KEGEL et al. 2001 ; OOI et al. 2001 ; VALENCIA et al. 2001 ). Indeed, the nej1 array mutant had been independently identified in my screen and verified as specifically deficient in NHEJ (Fig 3, Table 2). I noted a critical difference, however, in that the nej1 mutant was clearly only partially NHEJ defective in suicide deletion while it had showed a deficiency equivalent to Ku and DNA ligase IV mutants in others' transformed plasmid assays (KEGEL et al. 2001 ; OOI et al. 2001 ; VALENCIA et al. 2001 ). The residual ADE2 events in the nej1 mutant were NHEJ-dependent suicide deletion by all previously described criteria (KARATHANASIS and WILSON 2002 ) and by the observation that they were still dependent on RAD50 (not shown). This afforded me the opportunity to ask to what extent the mating-type and nej1 effects overlap and whether the stationary-phase effect is mediated through NEJ1.

As seen in Fig 4A, adding a plasmid-borne MATa gene to MAT strains caused the same partial (as compared to rad50) 10-fold decrease in NHEJ efficiency as did the nej1 mutation, with MATa/MAT nej1 strains being no more deficient, demonstrating an epistatic relationship of the MATa/MAT and nej1 genotypes with regard to NHEJ deficiency. Further, both the nej1 and MATa/MAT NHEJ defects were corrected by a plasmid expressing NEJ1 from the strong constitutive ADH1 promoter, confirming that regulated loss of NEJ1 expression is in fact responsible for the MATa/MAT effect (Fig 4A). It is thus apparent that NEJ1 expression can entirely account for the mating-type effect on NHEJ efficiency, but that Nej1 is not an obligatory participant in chromosomal NHEJ.

View larger version (28K): In this window In a new window Download PPT slide   Figure 4. NEJ1 expression can completely account for the mating type but not the growth-phase regulation of NHEJ efficiency. (A) Haploid yeast of the indicated genotypes were tested by suicide deletion quantitative method 2. All strains were chromosomal MAT, but where indicated also bore the MATa allele on a plasmid (a/, open bars). pNEJ1 expresses NEJ1 from the constitutive ADH1 promoter. (B) Similar experiment to A, except that the growth stage of the yeast used to inoculate the galactose medium was varied between active glucose exponential phase and early stationary phase.

Mating-type and growth-phase regulation of NHEJ are separable phenomena:
We have previously considered whether NEJ1 might be responsible for mediating the stationary phase as well as the mating-type effects on NHEJ (KARATHANASIS and WILSON 2002 ). Each of these effects, like the nej1 mutation, leads to only partial loss of chromosomal NHEJ, however, which makes it more likely that these effects are separable and that NHEJ-stimulating genes other than NEJ1 exist. Indeed, when tested in combination using wild-type yeast, the a/ mating type and exponential growth phase combined to drive chromosomal NHEJ to levels even lower than those of each parameter individually (Fig 4B). The combination showed an almost 100-fold decrease in the NHEJ/SSA ratio relative to MAT stationary-phase yeast, very nearly the same defect seen in rad50 yeast. While yeast bearing the nej1 mutation were again insensitive to the mating-type effect, both MATa nej1 and MATa/MAT nej1 yeast showed a significantly decreased NHEJ/SSA ratio when tested in exponential phase. It is thus clear that unlike the mating-type effect, NEJ1 is not required for postdiauxic/stationary-phase stimulation of NHEJ.

Identification of genes required for growth-phase regulation of NHEJ:
The identity of other partially NHEJ-deficient mutants suggested that they may in fact play a role in coordinating growth-phase-dependent induction of NHEJ (see DISCUSSION). To further explore this possibility I examined their phenotypes and relationships with nej1 in detail. Although many strains were tested in part, this discussion focuses on those strains that showed the largest consistent decrease in the NHEJ/SSA ratio, namely doa1, mck1, and fyv6. Each of these mutants showed a six- to sevenfold decrease in the NHEJ/SSA ratio by method 2 that was similar in magnitude to the decrease observed in wild-type exponential cells (Table 3). This similarity proved to be more than coincidental on the basis of several observations. First, the doa1, mck1, and fyv6 mutants were each largely insensitive to a further exponential-phase decrease in the NHEJ/SSA ratio (Table 3), parallel to the manner in which nej1 mutant cells were insensitive to the mating-type effect. Second, the doa1, mck1, and fyv6 mutations all showed synthetic decreases in the NHEJ/SSA ratio when combined with the nej1 allele (Table 3), indicating that, like the exponential-phase regulation of NHEJ, these genes act separately from NEJ1. It was thus not surprising that none of these mutants was corrected by the pNEJ1 plasmid (Table 3). In total, the data are fully consistent with the hypothesis that DOA1, MCK1, and FYV6 promote fully efficient NHEJ by a mechanism activated in postdiauxic/stationary phase that is consequently distinct and separable from the action of NEJ1.

Finally, I measured two parameters to determine whether the defect of the doa1, mck1, and fyv6 cells is in the induction of stationary phase per se (i.e., in the global response to nutritional deprivation) or in a downstream signaling of this response to the NHEJ apparatus. These parameters were maintenance of viability and induction of thermotolerance over many days in culture. As seen in Fig 5, three different patterns were observed. The doa1 mutant was deficient in stationary-phase induction in that it progressively lost viability from 4 days after the diauxic shift and never achieved wild-type levels of thermotolerance. In contrast, the mck1 mutant behaved as wild type in this analysis, showing both maintenance of viability and induction of thermotolerance, thus differentiating its growth-dependent NHEJ defect from stationary-phase induction. The fyv6 mutant was intermediate in that it did not lose viability but had poor induction of thermotolerance, further demonstrating that the various physiological changes that occur during prolonged nutrient deprivation are independent.

View larger version (15K): In this window In a new window Download PPT slide   Figure 5. Delineating the stationary-phase deficiency of doa1, fyv6, and mck1 mutants. The graph shows the relative cfu per milliliter in prolonged cultures of wild-type, doa1, fyv6, and mck1 strains. The zero time point corresponds to cultures at or just after the diauxic shift in YPAD. To the right are the percentages of cells showing tolerance to a 55° heat shock; values are the means ± standard deviations of measurements made at the last three time points.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Recent observations have suggested that unknown regulatory genes influence DSBR pathway utilization, in particular NHEJ, in S. cerevisiae (LEE et al. 1999 ; KARATHANASIS and WILSON 2002 ). Further, the lack of a comprehensive genetic screen capable of detecting NHEJ mutants has left the possibility that unknown genes participate structurally in this repair event. The screen described here was undertaken to search for these genes. As it was being completed, OOI et al. 2001 reported their screen, using an approach in which a pool of deletion mutants was transformed with linearized plasmids with microarray analysis used to identify those with NHEJ deficiencies. Because of the unique aspects of each approach (see below), results here complement and extend their findings. This discussion is organized according to the classes of DSBR-deficient mutants that were detected (Table 1).

Class I mutants, isolated SSA deficiency:
Although not a primary goal, the screen described here was able to detect mutants with a deficiency in SSA via short terminal direct repeats. The only mutant identified in this class was rad52 (but note srs2, below). This is consistent with previous findings that SSA, unlike true recombination, is independent of RAD52 epistasis group members required for strand invasion (IVANOV et al. 1996 ). As with NHEJ (see below), this does not mean that no other genes are involved, however. For example, SMITH and ROTHSTEIN 1999 found that specific mutations of the essential gene RFA1 act as suppressors of the rad52 SSA defect, indicating that RPA and Rad52 interact in the process of single-strand coating and annealing. It is noteworthy that rad59 was not among the class I mutants based on several spotting screens, since recent studies have demonstrated a role for Rad59 in SSA both in vitro and in vivo (PETUKHOVA et al. 1999 ; SUGAWARA et al. 2000 ; DAVIS and SYMINGTON 2001 ). I have not tested the rad59 mutant by quantitative analysis and so cannot rule out a minor SSA defect, and indeed the screen was least sensitive for this mutant class. Nonetheless, it seems that it is not strongly impaired in suicide deletion SSA, a conclusion supported by observations that this same mutant displays a recombination-defective phenotype when treated with various replication inhibitors (J. R. VANCE, A. IACCO and T. E. WILSON, unpublished results). This likely reflects the fact that the suicide deletion repeats are unusual in that they are short (28 bp) and at the termini, so that the enzymatic requirements for repair are relaxed due to a relative ease of homology searching (note that even the rad52 defect is comparatively modest). Finally, an important negative in all SSA screens to date is the failure to find a mutant deficient in 5' resection, which is required for SSA and recombination alike. This likely reflects an enzymatic redundancy in 5' resection (TSUBOUCHI and OGAWA 2000 ; MOREAU et al. 2001 ; LEWIS et al. 2002 ), and consequently more involved genetic analyses will be required to elucidate these mechanisms.

Class II mutants, combined NHEJ and SSA deficiency:
The array hpr5 mutant (more commonly known as srs2) was initially scored in the screen as SSA deficient. While a minor SSA deficiency has indeed been observed for srs2 by the Haber laboratory (SUGAWARA et al. 2000 ), the phenotype here is surprisingly severe, especially given the rad59 result above. Again, this may reflect the nature of the suicide deletion repeats, whose short length may increase the need for Srs2 to unwind nonproductive strand associations. On quantitative analysis the srs2 mutant proved to have an additional modest but reproducible twofold deficiency in NHEJ. This is entirely consistent with a previous report from the Klein laboratory (HEGDE and KLEIN 2000 ) and, importantly, verifies that the srs2 NHEJ defect is observed in the repair of chromosomal as well as plasmid DSBs. It remains unclear why the Srs2 helicase would be required for NHEJ via fully compatible 3' overhangs, however. Importantly, rad50, mre11, and xrs2 mutants are known to be deficient in both homologous recombination and NHEJ. The fact that these (and others?) were not detected as class II mutants reflects the fact that SSA is independent of the Rad50/Mre11/Xrs2 complex.

Class III mutants, increased NHEJ efficiency:
The screen described here was especially sensitive not only to DSBR pathway deficiencies, but also to mutants in which NHEJ efficiency was increased, which would be manifested as an increased ADE2/ade2 ratio. In particular, I anticipated that mutants showing altered pathway regulation or deficient 5' resection might favor NHEJ over SSA. The msn5 and especially bud31 mutants did show a reproducible and significant increase in the NHEJ/SSA ratio, but this was modest and not accompanied by an obvious increase in the absolute NHEJ frequency. Msn5 is an exportin required in various nuclear transport cycles (GORLICH and KUTAY 1999 ), while Bud31 is a protein of uncertain molecular function that might be involved in bud site selection (NI and SNYDER 2001 ). It is conceivable that either of these might affect DSBR pathway utilization, although it seems equally likely that these mutant phenotypes could be explained by secondary effects. Finally, it is noteworthy that rad5 was not scored as positive in the screen (and so was not tested further in the quantitative assay), as others have found Rad5 to function in avoidance of NHEJ of incompatible ends (AHNE et al. 1997 ). Although I again cannot rule out a minor effect below the sensitivity of the screen, my result is more consistent with data from HEGDE and KLEIN 2000 that simple religation NHEJ is neither increased nor decreased in rad5 mutants.

Class IVa mutants, structural/catalytic NHEJ deficiency:
At the time that this work was initiated, seven mutants were known to show NHEJ deficiencies consistent with structural or enzymatic participation in rejoining of compatible overhangs: yku70, yku80, dnl4, lif1, rad50, mre11, and xrs2. Each of these, but no novel genes in this class, was readily uncovered in the suicide deletion screen. This is in contrast to the results of OOI et al. 2001 in which mre11 and xrs2 mutants were not identified due to the fact that they gave insufficient signal in the pool due to growth deficiencies. This demonstrates a particularly powerful feature of the spot-screening approach used here and also recently by BENNETT et al. 2001 in a screen for radiosensitive mutants, namely that strains are examined individually and so the array is screened comprehensively. It is thus relevant to ask whether Ku, DNA ligase IV, and the Mre11/Rad50/Xrs2 complex in fact represent the complete set of genes required to execute a simple-religation NHEJ event. Answering this requires a consideration of the technical and genetic limitations of my approach.

To ultimately be scored as positive, any array strain needed to be mating proficient, Met⁺, Ura⁺, Ade⁺, Gal⁺, and not petite (most petite mutants did not grow sufficiently on galactose to be scored). In general, it is unlikely that mutants in these failure classes would be structurally required for NHEJ, although some may be deficient in mating type and nutritional regulation of NHEJ. For example, sterility prevented the recovery of sir2-4 mutants, and possibly others, that are not structurally required for NHEJ but nonetheless lead to impairment of NHEJ via loss of HMR and HML silencing (LEE et al. 1999 ). Beyond this, these technical classes of screen failure in fact proved to be quite helpful. In spotting analysis strains are identified only by well position, in contrast to the positive identification provided by "molecular bar codes" in pooled microarray analysis, and so it was possible that cross-contamination could lead to misidentification. In total, >350 mutants with predictable phenotypes behaved as expected during screening, which in conjunction with PCR verification of novel DSBR-deficient mutants demonstrated that the array had not decayed to a measurable degree during the handling steps.

Mutants of true NHEJ genes may also have been missed for genetic reasons if they were inviable (or otherwise absent from the array), highly redundant with other genes, or linked to either the ADE2 or MAT loci. Linkage was of surprisingly little concern, given that a great many asci could be spotted; only 10 open reading frames (ORFs) to either side of the ADE2 and MAT loci needed to be discounted from consideration. It is of course impossible to judge the likelihood of redundancy, but importantly the screen sensitivity would have allowed detection of partially redundant functions. Regarding inviability, it is clear that NHEJ is not an essential function in yeast. However, the finding that histone modifications are required for fully efficient NHEJ makes it clear that repair must of course occur in the context of chromatin, components of which frequently are essential or redundant (DOWNS et al. 2000 ). In total, the properties of my screen make it increasingly likely (although not completely certain) that Ku, DNA ligase IV, and the Mre11/Rad50/Xrs2 complex represent the complete set of proteins required specifically for catalysis of simple religation NHEJ, although these almost certainly interact with unidentified chromatin components during repair. This interpretation is supported by recent biochemical studies in which purified fractions of Ku, DNA ligase IV, and the Mre11/Rad50 complex appear to effectively reconstitute NHEJ (CHEN et al. 2001 ; HUANG and DYNAN 2002 ). Finally, I note that the present suicide deletion assay and this discussion address only simple religation NHEJ. Alternative approaches are being developed to identify genes that collaborate with POL4 and so participate in NHEJ only when ends are incompatible.

Class IVb mutants, partial/regulatory NHEJ deficiency:
The high sensitivity of the suicide deletion screen also allowed for detection of partially NHEJ-defective mutants that proved, as hypothesized, to serve regulatory roles. Somewhat surprisingly, some mutants had less severe defects than the initially estimated detection threshold of a fivefold decrease in NHEJ. This reflects the fact that the inclusion criteria were deliberately relaxed during the screening phase. A necessary and important corollary is that this class of mutants is almost certainly incomplete due to false negatives, in contrast to the discussion above regarding frank catalytic NHEJ deficiency. For example, checkpoint mutants previously reported to have minor deficiencies in NHEJ of transformed plasmids (DE LA TORRE-RUIZ and LOWNDES 2000 ) were consistently scored as negative. It is thus apparent that a large number of genes contribute to promoting maximal NHEJ efficiency. At the same time, it is difficult to be certain that the smallest defects are biologically meaningful. My attention thus focuses on those mutants with greater than fivefold NHEJ deficiencies or functions that suggest a role in NHEJ regulation.

NEJ1 mediates mating-type regulation of NHEJ:

NEJ1 has now been identified by several independent means as a regulator of NHEJ efficiency, including the functional screen described here (FRANK-VAILLANT and MARCAND 2001 ; KEGEL et al. 2001 ; OOI et al. 2001 ; VALENCIA et al. 2001 ). That this gene in fact mediates the mating-type regulation of NHEJ was suggested by the fact that it is expressed only in haploid cells (FRANK-VAILLANT and MARCAND 2001 ; KEGEL et al. 2001 ; VALENCIA et al. 2001 ). Indeed, expression of NEJ1 from a mating-type-independent promoter (in our case the ADH1 promoter) has in all hands proven to relieve the inhibition of NHEJ seen in MATa/MAT cells (Fig 4 and KEGEL et al. 2001 ; VALENCIA et al. 2001 ), thereby demonstrating that mating-type regulation is dependent on transcriptional regulation of NEJ1. In the case of the chromosomal suicide deletion assay used here, it was further evident that the NHEJ defects seen in nej1 and MATa/MAT cells were quantitatively equivalent and epistatic. This conclusion was made possible by the fact that these were only partial defects (10- to 20-fold) as reflected in both the absolute and relative NHEJ frequencies (Table 2). This first makes clear that Nej1 is not an obligatory participant in NHEJ catalysis in the same fashion as, for example, Ku (although this does not rule out that Nej1 might participate in the NHEJ structural complex). Further, it appears likely that NEJ1 is itself the sole mediator of mating-type regulation and that the other partially NHEJ-deficient mutants affect a different regulatory input. This conclusion is based on the facts that the nej1 and MATa/MAT effects are equivalent, that no mutant in the screen behaved epistatically to nej1, and that no mutant (except nej1 itself) was complemented by pNEJ1 (Fig 4 and Table 3). Indeed, previous observations strongly suggest that both the upstream regulation of NEJ1 and the downstream effect of Nej1 are mediated directly. Specifically, the NEJ1 promoter has sites for the Mata1-Mat2 repressor encoded by the MAT alleles of MATa/MAT cells (KEGEL et al. 2001 ; VALENCIA et al. 2001 ), and Nej1 itself interacts strongly with Lif1 (FRANK-VAILLANT and MARCAND 2001 ; KEGEL et al. 2001 ; OOI et al. 2001 ).

Multigenic NEJ1-independent growth-phase regulation of NHEJ:

The partial defect of nej1 in suicide deletion contrasts with its Ku- or DNA ligase IV-equivalent defect seen in plasmid transformation assays (KEGEL et al. 2001 ; OOI et al. 2001 ; VALENCIA et al. 2001 ). In addition to the chromosomal nature of the suicide deletion break and enhanced sensitivity resulting from limited recleavage of ligated I-SceI sites, this difference can be accounted for at least in part by the use of early stationary-phase cells in the standard suicide deletion methods. We have previously argued that there is a stimulation of NHEJ in the postdiauxic/stationary phase that is masked in plasmid assays using cells growing exponentially in rich glucose medium; simply transforming cells from late stage cultures causes a substantial increase in NHEJ event recovery (KARATHANASIS and WILSON 2002 ). Enhancement of NHEJ in postdiauxic/stationary cultures is also evident in the suicide deletion assay. Although we had initially considered that NEJ1 might mediate this effect, experiments here consistently demonstrated that this component of NHEJ regulation is still active in the absence of Nej1. In contrast, the screen did identify other mutants deficient in growth-phase regulation of NHEJ. Specifically, the doa1, fyv6, and mck1 mutants themselves not only were NHEJ deficient to approximately the same degree as exponential wild-type cells (approximately fivefold), but also were insensitive to a further significant decrease in NHEJ when assayed in the exponential phase. A difficulty in the initial description of the growth-phase regulation of NHEJ was that it might be an experimental artifact reflecting a growth-dependent change in suicide deletion dynamics independent of a true change in the DSBR efficiency (KARATHANASIS and WILSON 2002 ). The fact that this second input to increased NHEJ efficiency can also be abrogated by genetic alteration of equivalently grown early stationary-phase cells demonstrates that the phenomenon is not a growth artifact.

The nature of these genes provides the final evidence that they are specifically defective in a postdiauxic/stationary-phase induction of NHEJ. Doa1 (also known as Ufd3) is a protein required for the degradation of ubiquitin-tagged proteins (JOHNSON et al. 1995 ; GHISLAIN et al. 1996 ). Its precise function is unknown, but mutants have very low levels of free ubiquitin, and at least some mutant phenotypes are known to be corrected by increased expression of UBI4. UBI4, also identified here as a class IVb mutant, is the yeast polyubiquitin gene known to be induced in numerous stress responses including nutritional deprivation (OZKAYNAK et al. 1987 ). Although the mechanism is not established, ubi4 mutants do not establish stationary phase and die after 4 days in culture (PECK et al. 1997 ), a phenotype shared by doa1 mutants (Fig 5). The fact that the ubi4 NHEJ deficiency was significantly less than that of doa1 (at least by method 2), despite being profoundly deficient in stationary-phase induction, provides a first indication that these are overlapping but separable phenomena, however.

MCK1 encodes a dual-specificity protein kinase of the glycogen synthase kinase-3 (GSK-3) family that affects a surprisingly large number of cellular processes. Though too numerous to list here (see RAYNER et al. 2002 ), the collection has suggested a role for Mck1 in mediating nutritional/stress responsiveness, to which we add enhancement of NHEJ in postdiauxic/stationary phase. But again, this regulation is distinct from stationary-phase induction per se, because the mck1 mutant was able to maintain viability and achieve thermotolerance during prolonged culture.

The reserved gene name FYV6 stands for function required for yeast viability in response to K1 killer toxin, although no report has appeared to date. It is induced during stationary phase (PLANTA et al. 1999 ). This coupling might suggest a role in stress responsiveness. Strikingly, it presents yet a third pattern of stationary-phase deficiency, since it maintained viability but failed to achieve thermotolerance.

Summary:
The screen described here has provided a clear picture of the catalytic/structural requirements for NHEJ and suggests that Ku, DNA ligase IV, and the Rad50/Mre11 complex are likely to provide all of the obligatory functions for simple religation NHEJ. Moreover, the screen has revealed a dual regulatory input into the regulation of NHEJ. NEJ1 mediates an input dependent on mating type, with no other genes identified as cooperating with NEJ1 in this regard. In contrast, a series of genes, most notably DOA1, FYV6, and MCK1, mediate a separate input dependent on the nutritional status of the culture. This latter input is correlated with the passage into stationary phase, but is distinct from it. Although Doa1, Fyv6, or Mck1 might modify the function of the NHEJ machinery, their action need not be direct and in fact no such interactions have been identified by systematic screening (UETZ et al. 2000 ).

	ACKNOWLEDGMENTS

I thank the members of my laboratory for many helpful discussions during the development of this project and Sarah Sutter and Elissa Karathanasis for technical help. pTW367 was constructed by Jamuna Thimmarayappa. This work was supported in part by the Pew Scholars Program in the Biomedical Sciences of the Pew Charitable Trusts and Public Health Service grant CA-90911 (T.E.W.).

Manuscript received May 3, 2002; Accepted for publication July 19, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AHNE, F., B. JHA, and F. ECKARDT-SCHUPP, 1997  The RAD5 gene product is involved in the avoidance of non-homologous end-joining of DNA double strand breaks in the yeast Saccharomyces cerevisiae.. Nucleic Acids Res. 25:743-749.[Abstract/Free Full Text]

BENNETT, C. B., L. K. LEWIS, G. KARTHIKEYAN, K. S. LOBACHEV, and Y. H. JIN et al., 2001  Genes required for ionizing radiation resistance in yeast. Nat. Genet. 29:426-434.[Medline]

BRACHMANN, C. B., A. DAVIES, G. J. COST, E. CAPUTO, and J. LI et al., 1998  Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast 14:115-132.[Medline]

CHEN, L., K. TRUJILLO, W. RAMOS, P. SUNG, and A. E. TOMKINSON, 2001  Promotion of Dnl4-catalyzed DNA end-joining by the Rad50/Mre11/Xrs2 and Hdf1/Hdf2 complexes. Mol. Cell 8:1105-1115.[Medline]

DAVIS, A. P. and L. S. SYMINGTON, 2001  The yeast recombinational repair protein Rad59 interacts with Rad52 and stimulates single-strand annealing. Genetics 159:515-525.[Abstract/Free Full Text]

DE LA TORRE-RUIZ, M. and N. F. LOWNDES, 2000  The Saccharomyces cerevisiae DNA damage checkpoint is required for efficient repair of double strand breaks by non-homologous end joining. FEBS Lett. 467:311-315.[Medline]

DOWNS, J. A., N. F. LOWNDES, and S. P. JACKSON, 2000  A role for Saccharomyces cerevisiae histone H2A in DNA repair. Nature 408:1001-1004.[Medline]

FRANK-VAILLANT, M. and S. MARCAND, 2001  NHEJ regulation by mating type is exercised through a novel protein, Lif2p, essential to the ligase IV pathway. Genes Dev. 15:3005-3012.[Abstract/Free Full Text]

GAME, J. C. and R. K. MORTIMER, 1974  A genetic study of x-ray sensitive mutants in yeast. Mutat. Res. 24:281-292.[Medline]

GENNERY, A. R., A. J. CANT, and P. A. JEGGO, 2000  Immunodeficiency associated with DNA repair defects. Clin. Exp. Immunol. 121:1-7.[Medline]

GHISLAIN, M., R. J. DOHMEN, F. LEVY, and A. VARSHAVSKY, 1996  Cdc48p interacts with Ufd3p, a WD repeat protein required for ubiquitin-mediated proteolysis in Saccharomyces cerevisiae.. EMBO J. 15:4884-4899.[Medline]

GORLICH, D. and U. KUTAY, 1999  Transport between the cell nucleus and the cytoplasm. Annu. Rev. Cell Dev. Biol. 15:607-660.[Medline]

HEGDE, V. and H. KLEIN, 2000  Requirement for the SRS2 DNA helicase gene in non-homologous end joining in yeast. Nucleic Acids Res. 28:2779-2783.[Abstract/Free Full Text]

HERRMANN, G., T. LINDAHL, and P. SCHAR, 1998  Saccharomyces cerevisiae LIF1: a function involved in DNA double-strand break repair related to mammalian XRCC4. EMBO J. 17:4188-4198.[Medline]

HUANG, J. and W. S. DYNAN, 2002  Reconstitution of the mammalian DNA double-strand break end-joining reaction reveals a requirement for an Mre11/Rad50/NBS1-containing fraction. Nucleic Acids Res. 30:667-674.[Abstract/Free Full Text]

IVANOV, E. L., N. SUGAWARA, J. FISHMAN-LOBELL, and J. E. HABER, 1996  Genetic requirements for the single-strand annealing pathway of double-strand break repair in Saccharomyces cerevisiae.. Genetics 142:693-704.[Abstract]

JACKSON, S. P., 2001  Detecting, signalling and repairing DNA double-strand breaks. Biochem. Soc. Trans. 29:655-661.[Medline]

JOHNSON, E. S., P. C. MA, I. M. OTA, and A. VARSHAVSKY, 1995  A proteolytic pathway that recognizes ubiquitin as a degradation signal. J. Biol. Chem. 270:17442-17456.[Abstract/Free Full Text]

JONES, J. M., M. GELLERT, and W. YANG, 2001  A Ku bridge over broken DNA. Structure 9:881-884.[Medline]

KARATHANASIS, E. K. and T. E. WILSON, 2002  Enhancement of Saccharomyces cerevisiae end-joining efficiency by cell growth stage but not by impairment of recombination. Genetics 161:1015-1027.[Abstract/Free Full Text]

KEGEL, A., J. O. SJOSTRAND, and S. U. ASTROM, 2001  Nej1p, a cell type-specific regulator of nonhomologous end joining in yeast. Curr. Biol. 11:1611-1617.[Medline]

LEE, S. E., F. PAQUES, J. SYLVAN, and J. E. HABER, 1999  Role of yeast SIR genes and mating type in directing DNA double-strand breaks to homologous and non-homologous repair paths. Curr. Biol. 9:767-770.[Medline]

LEWIS, L. K., G. KARTHIKEYAN, J. W. WESTMORELAND, and M. A. RESNICK, 2002  Differential suppression of DNA repair deficiencies of yeast rad50, mre11 and xrs2 mutants by EXO1 and TLC1 (the RNA component of telomerase). Genetics 160:49-62.[Abstract/Free Full Text]

MA, Y., U. PANNICKE, K. SCHWARZ, and M. R. LIEBER, 2002  Hairpin opening and overhang processing by an Artemis/DNA-dependent protein kinase complex in nonhomologous end joining and V(D)J recombination. Cell 108:781-794.[Medline]

MAHAJAN, K. N., S. A. NICK MCELHINNY, B. S. MITCHELL, and D. A. RAMSDEN, 2002  Association of DNA Polymerase µ with Ku and ligase IV: role for pol µ in end-joining double-strand break repair. Mol. Cell. Biol. 22:5194-5202.[Abstract/Free Full Text]

MOREAU, S., E. A. MORGAN, and L. S. SYMINGTON, 2001  Overlapping functions of the Saccharomyces cerevisiae Mre11, Exo1 and Rad27 nucleases in DNA metabolism. Genetics 159:1423-1433.[Abstract/Free Full Text]

NI, L. and M. SNYDER, 2001  A genomic study of the bipolar bud site selection pattern in Saccharomyces cerevisiae.. Mol. Biol. Cell 12:2147-2170.[Abstract/Free Full Text]

OOI, S. L., D. D. SHOEMAKER, and J. D. BOEKE, 2001  A DNA microarray-based genetic screen for nonhomologous end-joining mutants in Saccharomyces cerevisiae.. Science 294:2552-2556.[Abstract/Free Full Text]

OZKAYNAK, E., D. FINLEY, M. J. SOLOMON, and A. VARSHAVSKY, 1987  The yeast ubiquitin genes: a family of natural gene fusions. EMBO J. 6:1429-1439.[Medline]

PAQUES, F. and J. E. HABER, 1999  Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae.. Microbiol. Mol. Biol. Rev. 63:349-404.[Abstract/Free Full Text]

PECK, V. M., E. K. FUGE, P. A. PADILLA, M. A. GOMEZ, and M. WERNER-WASHBURNE, 1997  Yeast bcy1 mutants with stationary phase-specific defects. Curr. Genet. 32:83-92.[Medline]

PETRINI, J. H., 1999  The mammalian Mre11-Rad50-nbs1 protein complex: integration of functions in the cellular DNA-damage response. Am. J. Hum. Genet. 64:1264-1269.[Medline]

PETUKHOVA, G., S. A. STRATTON, and P. SUNG, 1999  Single strand DNA binding and annealing activities in the yeast recombination factor Rad59. J. Biol. Chem. 274:33839-33842.[Abstract/Free Full Text]

PLANTA, R. J., A. J. BROWN, J. L. CADAHIA, M. E. CERDAN, and M. DE JONGE et al., 1999  Transcript analysis of 250 novel yeast genes from chromosome XIV. Yeast 15:329-350.[Medline]

RAYNER, T. F., J. V. GRAY, and J. W. THORNER, 2002  Direct and novel regulation of cyclic AMP-dependent protein kinase by Mck1p, a yeast GSK-3. J. Biol. Chem. 4:4.

RESNICK, M. A., 1969  Genetic control of radiation sensitivity in Saccharomyces cerevisiae.. Genetics 62:519-531.[Free Full Text]

SHINOHARA, A., M. SHINOHARA, T. OHTA, S. MATSUDA, and T. OGAWA, 1998  Rad52 forms ring structures and co-operates with RPA in single-strand DNA annealing. Genes Cells 3:145-156.[Abstract]

SMITH, G. C. and S. P. JACKSON, 1999  The DNA-dependent protein kinase. Genes Dev. 13:916-934.[Free Full Text]

SMITH, J. and R. ROTHSTEIN, 1999  An allele of RFA1 suppresses RAD52-dependent double-strand break repair in Saccharomyces cerevisiae.. Genetics 151:447-458.[Abstract/Free Full Text]

SUGAWARA, N., G. IRA, and J. E. HABER, 2000  DNA length dependence of the single-strand annealing pathway and the role of Saccharomyces cerevisiae RAD59 in double-strand break repair. Mol. Cell. Biol. 20:5300-5309.[Abstract/Free Full Text]

THOMPSON, L. H., J. S. RUBIN, J. E. CLEAVER, G. F. WHITMORE, and K. BROOKMAN, 1980  A screening method for isolating DNA repair-deficient mutants of CHO cells. Somatic Cell Genet. 6:391-405.[Medline]

TONG, A. H., M. EVANGELISTA, A. B. PARSONS, H. XU, and G. D. BADER et al., 2001  Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294:2364-2368.[Abstract/Free Full Text]

TSUBOUCHI, H. and H. OGAWA, 2000  Exo1 roles for repair of DNA double-strand breaks and meiotic crossing over in Saccharomyces cerevisiae.. Mol. Biol. Cell 11:2221-2233.[Abstract/Free Full Text]

UETZ, P., L. GIOT, G. CAGNEY, T. A. MANSFIELD, and R. S. JUDSON et al., 2000  A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae.. Na