Saccharomyces cerevisiae MMS2 encodes a ubiquitin-conjugating enzyme variant, belongs to the error-free branch of the RAD6 postreplication repair (PRR) pathway, and is parallel to the REV3-mediated mutagenesis branch. A mutation in genes of either the MMS2 or the REV3 branch does not result in extreme sensitivity to DNA-damaging agents; however, deletion of both subpathways of PRR results in a synergistic phenotype. Nevertheless, the double mutant is not as sensitive to DNA-damaging agents as a rad6 or rad18 mutant defective in the entire PRR pathway, suggesting the presence of an additional subpathway within PRR. A synthetic lethal screen was employed in the presence of a sublethal dose of a DNA-damaging agent to identify novel genes involved in PRR, which resulted in the isolation of RAD9 as a candidate PRR gene. Epistatic analysis showed that rad9 is synergistic to both mms2 and rev3 with respect to killing by methyl methanesulfonate (MMS), and the triple mutant is nearly as sensitive as the rad18 single mutant. In addition, rad9 rad18 is no more sensitive to MMS than the rad18 single mutant, suggesting that rad9 plays a role within the PRR pathway. Moreover, deletion of RAD9 reduces damage-induced mutagenesis and the mms2 spontaneous and induced mutagenesis is partially dependent on the RAD9 gene. We further demonstrated that the observed synergistic interactions apply to any two members between different branches of PRR and G1/S and G2/M checkpoint genes. These results suggest that a damage checkpoint is essential for tolerance mediated by both the error-free and error-prone branches of PRR.
EUKARYOTIC cells are constantly challenged by environmental stresses and endogenous cellular processes that can cause DNA damage and compromise the integrity of the genome. Organisms have evolved surveillance mechanisms that sense and respond to DNA damage. These surveillance mechanisms, known as DNA damage checkpoints, were initially identified when inactivation of genes resulted in defects in cell-cycle arrest in response to genotoxic treatments.
The DNA damage checkpoint was initially discovered by Weinert and Hartwell (1988) when analyzing the rad9 mutant of Saccharomyces cerevisiae. In addition to RAD9, several genes have been found to be involved in the DNA damage response pathway. The G1, G2, and intra-S DNA damage checkpoints encompass several groups of proteins that function in combination with a central signal transduction cascade. Rad17, Mec3, and Ddc1 form a heterotrimeric complex with a structural similarity to proliferating cell nuclear antigen (PCNA) (Kondo et al. 1999; Thelen et al. 1999; Melo et al. 2001). Rad24 is related to replication factor C (RFC), a protein complex responsible for loading PCNA onto DNA during replication (Waga and Stillman 1998). The interaction of Rad24 with the four smaller RFC subunits acts to load the Rad17–Mec3–Ddc1 complex to sites of DNA damage (Kondo et al. 2001; Melo et al. 2001; Majka and Burgers 2003). Once the damage has been detected, the checkpoint pathway transmits the signal through a kinase cascade.
The RAD9 gene functions predominantly in the G1/S and G2/M transitions of the DNA damage checkpoints (Siede et al. 1993). Rad9 is phosphorylated during normal cell-cycle progression (Vialard et al. 1998) and hyperphosphorylated after DNA damage in a Mec1- and Tel1-dependent manner (Emili 1998; Vialard et al. 1998). It is proposed that Rad9 recruits and catalyzes the activation of Rad53 by acting as a scaffold that brings Rad53 molecules into close proximity, facilitating autophosphorylation of Rad53 (Toh and Lowndes 2003). Activated Rad53 is involved in cell-cycle arrest, transcriptional induction of repair genes, inhibition of late replication origin firing, and stabilization of stalled replication forks (Santocanale and Diffley 1998; Santocanale et al. 1999; de la Torre Ruiz and Lowndes 2000; Lopes et al. 2001; Tercero and Diffley 2001).
Mec1 is a member of the evolutionarily conserved phosphatidylinositol 3-kinase (PI3-kinase)-related kinase family (Elledge 1996) and, like Rad53, plays a central role in the DNA damage checkpoint at all cell-cycle stages (Longhese et al. 1998). Mec1 functions in a partially redundant manner with Tel1, another member of the PI3-kinase family (Lowndes and Murguia 2000). Several species have orthologs of MEC1 and TEL1, including humans (ATR and ATM, respectively) and the fission yeast Schizosaccharomyces pombe (rad3) (Abraham 2001). Several cellular proteins become rapidly phosphorylated in a Mec1/Tel1-dependent manner in response to DNA damage, including Rad53 and Chk1 (Lowndes and Murguia 2000).
Like the damage checkpoint pathway, the DNA postreplication repair (PRR) pathway also does not remove DNA lesions but instead attempts to bypass the lesion encountered. It is thought that PRR serves to resume DNA replication blocked by lesions, but its molecular mechanism is largely unknown (Broomfield et al. 2001; Barbour and Xiao 2003). To help define the PRR pathway, yeast cells were screened for mutants displaying synergistic sensitivity in the absence of MMS2. By taking advantage of the synergism between the error-free PRR pathway and the error-prone mutagenesis pathway (Broomfield et al. 1998), we used a novel screening protocol in the presence of extremely low doses of methyl methanesulfonate (MMS, 0.005%) that will not affect the growth of the single mutant, but will effectively kill the double mutants. One of the mutations identified in this screen revealed a role for checkpoint proteins in the PRR pathway.
We report here the synergistic interaction of rad9 with PRR mutations. Our results show the partial requirement of RAD9 in the mutagenesis observed in the mms2Δ mutants. These findings suggest a role for the checkpoint pathway in PRR, potentially for delaying the cell cycle and allowing time for the cells to tolerate the damage via the PRR pathway.
MATERIALS AND METHODS
Yeast strains and cell culture:
The yeast strains used in this study are listed in Table 1. All the strains are isogenic derivatives of DBY747, originally obtained from D. Botstein (Stanford University), HK578-10A and HK578-10D, obtained from H. Klein (New York University), or BY4741. BY4741 and its gene deletion derivatives were created by the Saccharomyces Genome Deletion Project consortium and purchased from Research Genetics (Invitrogen, Carlsbad, CA).
Yeast cells were cultured at 30° in either a rich YPD medium or a synthetic glucose (SD) medium supplemented with various nutrients as instructed (Sherman et al. 1983). Intact yeast cells were transformed by a modified lithium acetate method (Ito et al. 1983). For one-step targeted gene disruption (Rothstein 1983), plasmid DNA containing the desired disruption cassette was cleaved with restriction enzymes prior to yeast transformation. Source and use of rad9Δ∷hisG-URA3-hisG (Schiestl et al. 1989), rad17Δ∷HIS3 (Zhu and Xiao 1998), rad18Δ∷TRP1 (Xiao et al. 1996), rad18Δ∷LEU2 (Xiao et al. 2000), rev3Δ∷LEU2 (Xiao et al. 1996), and rev3Δ∷ hisG-URA3-hisG (Roche et al. 1994) are as described. The mms2Δ∷HIS3 and mms2Δ∷TRP1 disruption cassettes were produced using a PCR method. The plasmid pJJ215 (HIS3) or plasmid pJJ248 (TRP1) (Jones and Prakash 1990) was amplified using the primers YGL87 (5′-TTCTTATTCTGTATATGCAACGTAGAAGAAGCAGCCGTTTACACAAACAGCTATGACCATG-3′) and YGL88 (5′-GTGGCTTGGAATGCTGCAAATACTGTTTAGGAAAAAGTAGATAACGTTTTCCCAGTCACGAC-3′). The PCR products contained the 5′ and 3′ flanking sequences of the MMS2 (underlined) open reading frame disrupted by HIS3 or TRP1, respectively.
Synthetic sensitivity screen:
Yeast cells (WXY1228) harboring the pSLS-based plasmids (Barbour et al. 2000) were grown overnight in 10 ml of SD medium lacking uracil. The cells were harvested by centrifugation, resuspended in the same volume of YPD medium, and incubated for another 4 hr. The cells were collected, washed twice in 50 mm potassium phosphate buffer (pH 7.5), and resuspended in 10 ml of the same buffer. Ethyl methanesulfonate was added to a final concentration of 3% and the culture was incubated at 30° for 30 min. Ten-percent filter-sterilized sodium thiosulfate was added to stop the reaction. The cells were washed twice, diluted, plated onto YPD or YPGal medium, and incubated for 4 days at 30°. Individual nonsectoring colonies were picked and further characterized by two steps. First, they were streaked onto the same medium to monitor color segregation. Cells from the nonsectoring colonies were then used to inoculate 2 ml of liquid YPD. After an overnight incubation, cells were diluted and plated onto YPD to record colony-color segregation. For a synthetic sensitivity screen, cells were plated onto YPD containing 0.005% MMS.
Cell killing by DNA-damaging agents:
MMS- and UV-induced liquid killing was performed as previously described (Xiao et al. 1996). Briefly, overnight yeast cultures were used to inoculate fresh YPD at ∼5 × 106 cells/ml and allowed to grow until the culture contained ∼2 × 107 cells/ml. For MMS treatment, MMS was added to the culture at a final concentration as specified and aliquots were taken at given intervals. For UV treatment, cells were plated in duplicate at different dilutions and then exposed to 254 nm UV light in a UV crosslinker (Fisher Science model FB-UVXL-1000 at ∼2400 mW/cm2) at given doses in the dark. For γ-irradiation, cells were collected by centrifugation, resuspended in sterile water, exposed to γ-rays from a 60Co γ-ray source at a dose rate of 22 rad/sec, and plated.
The gradient plate assay was performed as a semiquantitative measurement of relative MMS sensitivity. A total of 30 ml of molten YPD agar was mixed with the appropriate concentration of MMS to form the bottom layer. The gradient was created by pouring the media into tilted square petri dishes. After brief solidification for 1 hr, the petri dishes were returned flat and 30 ml of the same molten agar without MMS was poured to form the top layer. A 0.1-ml sample was taken from an overnight culture, mixed with 0.4 ml sterile water and 0.5 ml of molten YPD agar, and then immediately imprinted onto freshly made gradient plates via a sterile microscope slide. Gradient plates were incubated at 30° for the indicated time.
MMS sensitivity was also determined by a serial dilution assay. Yeast cells were inoculated in 3 ml of YPD medium (or selective medium if required) overnight and subcultured into 3 ml of fresh medium. Cells were incubated at 30° until a midlogarithmic phase was reached. The cell density was adjusted to 2 × 106 cells/ml, as determined by a hemocytometer, and further diluted serially 10-fold with ddH2O. The relative MMS sensitivity was determined using freshly made YPD plates containing the indicated amount of MMS. Five-microliter aliquots of each dilution were applied onto YPD and YPD + MMS plates. The plates were incubated at 30° for 2 days and photographed.
Spontaneous Trp+ reversion rates of DBY747 derivatives were measured by a modified Luria and Delbruck fluctuation test as described (Von Borstel 1978). Trp+ reversions were measured using the trp1-289 amber allele. An overnight culture was used to inoculate five tubes, each containing 10 ml of fresh YPD, to a final titer of 20 cells/ml. Incubation was continued until the cell titer reached 2 × 107 cells/ml. Cells were collected, washed, resuspended, and plated. Each set of experiments contained five independent cultures of each strain, and each culture was plated onto YPD in duplicate to score total survivors and onto SD–Trp plates to score Trp+ revertants. Spontaneous mutation rates (number of revertants per cell per generation) were calculated as previously described (Williamson et al. 1985).
For MMS-induced mutagenesis, single-cell clones were used to inoculate fresh YPD liquid medium and log-phase cells were treated with 0.05% MMS for 30 min, washed twice with ddH2O, concentrated 10-fold, and plated onto two selective plates at ∼5 × 107 cells/plate to score for Trp+ revertants. The same sample was diluted and plated on YPD plates to score for total viable cells. The plates were incubated at 30° for 3 days. The percentage of survival of each strain after MMS treatment was determined by comparing with the same strain without MMS treatment.
Synthetic genetic array and confirmation tests:
Y8389 (rev1Δ) and Y8426 (ubc13Δ) were created by a nonessential gene-switching method and crossed with the deletion mutant array as described (Tong et al. 2001). The relevant diploids from the array were received from C. Boone (University of Toronto) and the random spore analysis was carried out as described (Tong et al. 2001). Briefly, spores obtained from the synthetic genetic array analysis were resuspended in 500 μl of ddH2O and an aliquot was plated onto selective media with or without 0.004% MMS as follows: 20 μl onto SD–His/Arg/Lys + canavanine/thialysine, 40 μl onto SD–His/Arg/Lys + canavanine/thialysine/G418, 40 μl onto SD–His/Arg/Lys + canavanine/thialysine/clonNAT, and 80 μl onto SD–His/Arg/Lys + canavanine/thialysine/G418/clonNAT. The plates were incubated at 30° for 3 days and scored for synergistic interactions.
Identification of synthetic sensitive mutations with mms2:
The synthetic lethal screen is a method of isolating novel mutants whose survival is dependent on the presence of the gene of interest. A colony-color-based assay has been developed in the budding yeast (Hieter et al. 1985; Koshland et al. 1985), and we recently reported an improved method to enhance the screening efficiency (Barbour et al. 2000). Here we wanted to apply this method to the concept of a synthetic sensitivity screen. It is well established that MMS2/UBC13 and REV1/3/7 represent two alternative branches for tolerance to many kinds of DNA damage (Broomfield et al. 1998; Xiao et al. 1999). Figure 1A demonstrates that 0.005% MMS in the culture medium does not result in any notable lethal effect to the mms2 or rev3 single mutant; however, it completely inhibits the growth of the mms2 rev3 double mutant. Hence, mms2 and rev3 exhibit a strong synergistic interaction and the double mutant is inviable in the presence of 0.005% MMS.
On the basis of the above assertion and the hypothesis that novel genes whose mutation causes synergistic interactions with mms2/ubc13 and/or rev1/3/7 may be identified, we adapted a synthetic lethal screen protocol to isolate mutations that are synergistic with mms2. This resulted in the isolation of mutations known to be synergistic with mms2 with respect to killing by MMS, such as rev1 and rev3, as well as novel mutations that do not belong to the above genes. One such mutant, SLM-11, displays a synergistic sensitivity on plates containing 0.005% MMS as compared to the mms2 single mutant (Figure 1B). To identify the gene whose mutation is responsible for the synergistic effect in SLM-11, a single- and a multi-copy yeast genomic library were screened for functional complementation of the MMS-sensitive phenotype found in SLM-11, which resulted in the isolation of RAD9. As shown by a reconstructed experiment (Figure 1B), SLM-11 transformed with a RAD9-containing plasmid restored the MMS resistance to a level comparable to that of the mms2 single mutant, whereas SLM-11 transformed with both RAD9 and MMS2 completely restored the MMS resistance to the wild-type level. This result indicates that a rad9 mutation is synergistic with mms2 to killing by MMS. Note that SLM-11 transformed with MMS2 alone is expected to restore the MMS resistance to the rad9 single-mutant level, which can be distinguished from that of wild type only at a higher MMS concentration (data not shown).
rad9 is synergistic with both mms2 and rev3:
To ask whether RAD9 is the gene mutated in SLM-11 or an extragenic suppressor, we deleted RAD9 from an mms2Δ strain and examined the sensitivity of this double mutant relative to the corresponding single mutants. As shown in Figure 2A, while the rad9Δ mutant showed little sensitivity and the mms2Δ single mutants showed slight sensitivity to the DNA-damaging agent, the rad9 mms2 double mutant was extremely sensitive to MMS. The effect of the two mutations was clearly synergistic, as judged by a quantitative liquid killing assay (Figure 2C). To further prove that SLM-11 contains a rad9 mutation, we crossed SLM-11 with an mms2Δ rad9Δ mutant and performed random spore analysis. All 50 haploid spores analyzed displayed extreme sensitivity—for example, SLM-11 or the mms2 rad9 double mutant—confirming that the unknown mutation in SLM-11 and RAD9 occupies the same chromosomal locus.
The synergistic interaction between mms2 and rad9 prompted us to look at the genetic interaction between RAD9 and another branch within the PRR pathway, namely the mutagenesis pathway mediated by REV1/3/7. We combined rad9 with rev3 and determined the MMS-induced killing of the rad9 rev3 double mutant. Both single mutants alone showed no sensitivity on plates containing 0.005% MMS. To our surprise, the rad9 rev3 double mutant did not grow at all on the same MMS plate (Figure 2B). Quantitative analysis (Figure 2D) clearly indicates that the interaction is synergistic. These results suggest that the RAD9 gene genetically interacts with both the error-free and error-prone branches of PRR and may constitute a third branch of the PRR pathway.
RAD9 functions within the PRR pathway:
The synergistic interactions among rad9, mms2, and rev3 prompted us to ask whether the triple mutation is synergistic compared to each of the double mutations. The liquid killing experiment showed that the triple mutant is significantly more sensitive than the most sensitive double mutant mms2 rev3 (Figure 3A), suggesting a three-branch relationship represented by MMS2, REV3, and RAD9.
We have previously demonstrated that the mms2 rev3 double mutant is significantly less sensitive than the rad18 single mutant (Broomfield et al. 1998) and that the combination of any known mutations within the PRR branches did not reach the level of rad18 (Xiao et al. 2000; data not shown). The level of MMS sensitivity of the mms2 rad9 rev3 triple mutant relative to the mms2 rev3 double mutant (Figure 3A) is reminiscent of the rad18 single mutant, which led us to speculate that RAD9 may function in the PRR pathway. To further determine the genetic involvement of RAD9 in the PRR pathway we combined the rad9 mutation with rad18 and compared the double mutant with other relevant mutants. The rad9 rad18 double mutant appears to be more sensitive than the rad18 single mutant in a liquid killing experiment (Figure 3A). We suspected that the increased MMS resistance of the rad18 single mutant was due to a high number of revertants as previously observed (Xiao et al. 1996), possibly due to the acquisition of srs2 mutations capable of rescuing rad18 cells from killing by DNA-damaging agents (Aboussekhra et al. 1989; Schiestl et al. 1990). To partially overcome the above problem, we took freshly isolated individual colonies and performed liquid killing experiments, including early time points (5 and 10 min). Indeed, within the first 10 min, the killing curves of rad18 and rad9 rad18 mutants were indistinguishable (Figure 3A). We further analyzed individual colonies that survived after a 40-min MMS treatment and found that >86% (50 of 58) of rad18 clones gained resistance to MMS, whereas none of the rad9 rad18 survivors was resistant to MMS. We also cultured cells without MMS treatment and plated ∼107 cells on plates containing various concentrations of MMS. On 0.001% MMS plates, rad18 cells formed about eightfold more colonies than rad9 rad18 cells did. Upon restreaking these colonies, 43 of 48 rad18 colonies inherited resistance, while none of 50 rad9 rad18 colonies displayed MMS resistance. This would explain most, if not all, observed differences between rad18 and rad9 rad18 cells in response to MMS treatment (Figure 3A). We also performed a gradient plate assay, which is able to distinguish individual revertants, and confirmed that indeed the rad9 rad18 double mutant is no more sensitive than the rad18 single mutant (Figure 3B, lanes 6 and 7), and that the rad9 mms2 rev3 triple mutant (lane 5) is only slightly less sensitive than the above two strains, while none of the above mutant strains displayed a growth defect in the absence of MMS (Figure 3B). Taken together, although we cannot formally rule out the possibility that differences in liquid killing (acute treatment) and plate-based assays (chronic treatment) also contribute to the above discrepancy, the above results generally support our hypothesis that RAD9 functions within the PRR pathway as the third branch independent of MMS2 and REV3 and that RAD18 is epistatic to all of the above three branches.
RAD9 is partially required for mutagenesis:
It has been previously determined that PRR is composed of an error-free branch and an error-prone branch with characteristic phenotypes of spontaneous and damage-induced mutagenesis (Broomfield et al. 1998; Xiao et al. 1999). The effects of RAD9 deletion on mutagenesis were determined and compared with other relevant mutants using the trp1-289 reversion assay, which detects primarily base-pair substitutions. As expected, the mms2 single mutant had a spontaneous mutation rate increased by ∼22-fold compared to wild-type cells, whereas the spontaneous mutagenesis was abolished in the mms2 rev3 double mutant. The spontaneous mutation rate in the rad9 single mutant is not significantly higher than in wild-type cells; however, the rad9Δ mutation reduced spontaneous mutagenesis seen in mms2Δ cells from 22- to 12-fold, and the mms2 rad9 rev3 triple mutant displayed a spontaneous mutation rate indistinguishable from that of mms2 rev3 cells (Table 2 ). These results suggest that the RAD9 gene is partially responsible for the spontaneous mutagenesis seen in the absence of MMS2, in a REV3 dependent manner.
We also determined the involvement of RAD9 in MMS-induced mutagenesis. The MMS-induced reversion of the trp1-289 allele was also elevated in the mms2 mutant, which appears to be partially suppressed by RAD9 deletion. To our surprise, while deletion of RAD9 alone only slightly reduced MMS-induced mutagenesis, we were unable to score a single revertant from the rad9 rev3 strain during the entire experiment (Table 3 ). These results suggest that, like REV3, the RAD9 gene is also involved in MMS-induced mutagenesis, which is consistent with a role for RAD9 in the PRR pathway.
The DNA damage checkpoint is required for efficient PRR:
Like RAD9, several other genes, such as RAD17, RAD24, DDC1, and MEC3, are also classified as G1/S and G2/M damage checkpoint genes (Siede 1995; Longhese et al. 1998). However, it was reported that RAD9 may act in a branch alternative to the other four genes (de la Torre-Ruiz et al. 1998). To determine whether the above observed rad9 mutant phenotypes are unique or common to other G1/S and G2/M checkpoints, we combined a rev3 mutation with additional checkpoint mutations, such as rad17 and rad24, and compared the double mutants with corresponding single mutants. None of the rev3, rad17, or rad24 single mutants displayed a sensitivity to 0.005% MMS on a plate assay; however, we observed a strong synergistic phenotype in the rad17 rev3 and rad24 rev3 double mutants in the same experiment (Figure 4, A and B). Furthermore, the rad18 rad24 double mutant is no more sensitive to MMS than the rad18Δ single mutant as judged by a gradient plate assay (data not shown). These results suggest that RAD17 and RAD24 checkpoint genes participate in the PRR pathway, most likely in a manner similar to RAD9.
The RAD53 and CHK1 genes define two parallel pathways that regulate multiple cell-cycle transitions (Foiani et al. 2000), and Rad9 is required for activation of both Rad53 and Chk1 (Navas et al. 1996; Sanchez et al. 1999). To determine if the involvement of checkpoints in the PRR response encompassed the entire DNA damage checkpoint pathway, rad53 rev3 and chk1 rev3 double mutants were created and tested for their sensitivity to MMS on a plate assay. The rev3 mutant showed no obvious sensitivity to 0.01% MMS, and the rad53 mutant was slightly sensitive on the MMS plate. However, the rev3 rad53 double mutant was inviable on this concentration of MMS, which unmistakably shows a synergistic interaction between the two genes (Figure 4C). In contrast, under the same experimental conditions, the rev3 chk1 double mutant displayed a phenotype no more sensitive than that of the single mutants (Figure 4D). These results suggest that the RAD53 branch, and not the CHK1 branch, of the checkpoint pathway is responsible for facilitating the tolerance of the DNA lesion by the PRR pathway.
The involvement of checkpoint genes and other genes in PRR was further assessed by a genomewide synthetic genetic array (SGA) approach (Tong et al. 2001; Tong and Boone 2006). In this case, we utilized two alternative PRR mutants, ubc13 and rev1, instead of mms2 and rev3, and screened for lethal phenotypes with the entire yeast mutant array on both synthetic minimal medium and the minimal medium containing 0.004% MMS. As shown in supplemental Figure S1 at http://www.genetics.org/supplemental/ and in similar data not shown, both ubc13 and rev1 are synergistic with rad9, rad17, rad24, or ddc1 (the mec3 single mutant did not grow on the tester plate) on a minimal medium containing 0.004% MMS. In contrast, the corresponding double mutants with genes involved in other cell-cycle checkpoint pathways, such as chk1, pds1, mrc1, tof1, and csm3, did not display synthetic sensitive phenotypes.
RAD9 involvement in PRR is lesion specific:
DNA-damaging agents produce specific types of lesions that are detrimental to the cell at specific stages of the cell cycle. For example, UV damage causes predominantly (T<>T) dimers and (6-4) photoproducts that block replication, whereas ionizing radiation such as X rays and γ-rays produce strand breaks (Friedberg et al. 2006). These agents induce mainly G1/S and G2/M arrest while MMS causes S-phase delay (Friedberg et al. 2006). To determine if the synergistic effect seen in the mms2 rad9 and rev3 rad9 double mutants is specific to MMS-induced damage and thus restricted to the intra S-phase of the cell cycle, the sensitivity of these mutants to UV and γ-irradiation was determined (Figure 5). When treated with UV at the given dose range, rad9, mms2, and rev3 single mutants all display sensitivity. When combining the rad9 mutation with either the mms2 or the rev3 mutation, the double mutants display an additive phenotype to UV irradiation (Figure 5, A and B), suggesting that RAD9 does not have a strong genetic interaction with either MMS2 or REV3 in response to UV irradiation.
The rad9 mutant shows a >10-fold elevated level of sensitivity to γ-irradiation compared to the mms2 or rev3 single mutant. When the rad9 mutation is combined with rev3, the double mutant shows a weak additive phenotype to γ-irradiation (Figure 5D). In contrast, combining rad9 with the mms2 mutation resulted in a slight rescuing effect of the rad9 phenotype (Figure 5C). The plate-based assay showed similar results (supplemental Figure S2, A and B, at http://www.genetics.org/supplemental/). We also determined genetic interactions when replication is stalled due to lack of substrates and found that the mms2 rad9 and rev3 rad9 double mutants hardly displayed any additional sensitivity to hydroxyurea (HU) compared to the corresponding single mutants (supplemental Figure S2, C and D, at http://www.genetics.org/supplemental/). Overall, these results suggest a lack of synergistic interactions between rad9 and mms2 or rev3 with regard to UV, γ-irradiation, and HU-induced killing.
All living organisms have developed sophisticated networks to deal with spontaneous and induced DNA damage, and this is particularly true for eukaryotic cells from yeast to human (Friedberg et al. 2006). PRR is one such pathway dealing with damage that escapes repair but impedes replication. Unlike most other repair pathways, PRR does not remove the replication-blocking lesion, but coordinates different bypass mechanisms to avoid cell death (Barbour and Xiao 2003). To date, two subpathways have been well defined: one utilizes nonessential DNA polymerases for translesion DNA synthesis (Nelson et al. 1996a,b), which is considered a mutagenic or error-prone pathway, and another is mediated by Rad5–Ubc13–Mms2 (Ulrich and Jentsch 2000; Xiao et al. 2000), an E2–E3 ubiquitination complex capable of poly-ubiquitinating PCNA via a noncanonical Lys63 chain (Hofmann and Pickart 1999; Hoege et al. 2002). While it is conceivable that such poly-ubiquitinated PCNA promotes error-free bypass of replication-blocking lesions, the underlying mechanism of such bypass is yet unclear (Pastushok and Xiao 2004). A recent report (Zhang and Lawrence 2005) suggests that the error-free pathway employs recombination between partially replicated sister strands.
Analysis of previously reported genetic screens (Lemontt 1971, 1972; Prakash and Prakash 1977) led us to believe that there are genes involved in each of the error-free and error-prone subpathways yet to be discovered and, most importantly, there may exist an additional subpathway, since cells with simultaneous inactivation of both error-free and error-prone subpathways are significantly less sensitive to DNA-damaging agents than a rad6 or a rad18 mutant thought to be defective in the entire PRR pathway (Broomfield et al. 1998; Xiao et al. 1999). In this study we adapted a synthetic lethal screen approach in an attempt to identify novel genes involved in PRR. To achieve this goal, a genetic approach was developed by supplementing the screening medium with a small dose of MMS that would not interfere with single mutant growth, but is sufficient to eliminate the double-mutant growth if the two mutations are synergistic. This method allowed us to isolate genes, such as REV1 and REV3, whose mutations are known to be synergistic with mms2, as well as RAD9, which is novel and surprising. This study provides several lines of evidence suggesting the involvement of RAD9 in the PRR pathway. First, quantitative analysis shows that rad9 is synergistic to both mms2 and rev3 with respect to killing by MMS. The synergistic relationship with both PRR subpathways indicates that RAD9 most likely functions as a third branch within the same pathway and is less likely to be independent of the PRR pathway. Second, RAD9 is partially required for the increased spontaneous mutagenesis as well as MMS-induced mutagenesis seen in the mms2 cells. This phenotype is reminiscent of REV3, which is also required for spontaneous and induced mutagenesis (Lawrence 1982). Furthermore, rad9 and rev3 mutations appear to have an additive effect on MMS-induced mutagenesis, suggesting that these genes confer nonoverlapping functions in mutagenesis. Finally, we demonstrated that while rad9 is synergistic with both mms2 and rev3, it is hypostatic to rad18. Most interestingly, the level of MMS sensitivity of the mms2 rad9 rev3 triple mutant is much more sensitive than any of the corresponding double mutants and is approaching that of the rad18 single mutant. Taken together, our data are consistent with a model that PRR consists of three relatively independent branches represented by MMS2, REV3, and RAD9. We further demonstrated that each branch contains several genes as defined by previous studies and that the combination of any two mutations from genes belonging to different branches (Figure 6) generates a characteristic synergistic interaction. It is of particular interest to note that RAD9 was originally placed within the RAD6 epistasis group on the basis of its phenotypes and genetic interactions (Lawrence and Christensen 1976; McKee and Lawrence 1980; Friedberg 1988). We favor an argument that it is the damage checkpoint function that is involved in PRR because the observed genetic interaction is true for all G1/S and G2/M damage checkpoint genes, regardless of their biochemical activities. Our data support a notion that when one or both error-free and error-prone branches are inactivated, a cell-cycle delay becomes increasingly important for efficient PRR; however, in the absence of RAD18 function this delay has no effect on cell survival. It is interesting to note that neither PRR nor checkpoints actually remove DNA lesions; instead, they provide tolerance mechanisms to ensure cell survival in the presence of DNA damage. Data presented in this report suggest that, under certain conditions, the above two responses are coordinately regulated.
An unexpected result during this study was the differential response of yeast mutants to different DNA-damaging agents. Various reports indicate that rad6 and rad18 mutants are extremely sensitive to a broad range of DNA-damaging agents regardless of lesions produced (Broomfield et al. 2001). It appears that the common end product handled by the PRR pathway is the replication block containing a single-stranded DNA region (Broomfield et al. 2001). Within the PRR pathway, error-free and error-prone branches appear to deal with the same set of lesions (Xiao et al. 1999), whereas the checkpoint response to different lesions may vary. For example, compared with mms2, the rad9 mutant appears to be less sensitive to MMS, more sensitive to UV, and much more sensitive to γ-irradiation. We do not know the exact reason for the quantitative and qualitative differences in terms of genetic interactions; however, it can be speculated that in the case of UV treatment, nucleotide excision repair can efficiently remove UV-induced lesions and is probably required for the initiation of the RAD9-mediated G1 checkpoint (Siede et al. 1994). During S-phase, RAD30-encoded Polη can efficiently bypass T<>T dimers (Johnson et al. 1999), undermining the requirement for REV3. In the case of γ-irradiation, which induces largely strand breaks, the error-free and error-prone branches of PRR may play much less critical roles than the damage checkpoint. This argument appears to be consistent with MMS-induced mutagenesis, where RAD9 favors a mutagenic bypass pathway instead of template-based error-free bypass. Alternatively, the differential contribution of RAD9, or the entire damage checkpoint pathway, is simply dependent on the nature of agents that induce delay or arrest at different cell-cycle stages.
It has been previously reported that RAD9 functions to prevent genomic instability (Schiestl et al. 1989; Weinert and Hartwell 1990; Barrington et al. 1999); however, most reports were based on assays detecting gross genome rearrangements instead of point mutations. In this study, we utilized a base-substitution assay and report that inactivation of Rad9 partially reduces spontaneous and MMS-induced mutagenesis. This observation is reminiscent of a report on roles of RAD9 in UV-induced mutagenesis (Paulovich et al. 1998) and the 9-1-1 checkpoint clamp (composed of Ddc1, Rad17, and Mec3) in spontaneous mutagenesis (Sabbioneda et al. 2005). It would be interesting to determine whether rad9 enhances gross genomic rearrangement under the same conditions.
Finally, while it is generally agreed that mono-ubiquitinated PCNA promotes translesion synthesis (Stelter and Ulrich 2003) whereas poly-ubiquitinated PCNA is required for error-free PRR (Hoege et al. 2002), it is not clear what signal within PRR controls checkpoint involvement. One can argue that since rad18 is epistatic to rad9 with respect to MMS-induced killing, probably the RAD6-RAD18 function is required for checkpoint signaling. In this regard, the Rad6–Rad18 complex may serve as a sensor for MMS-induced DNA damage and elicit an SOS type of response, including a DNA damage checkpoint response. Whether the sensor is via PCNA covalent modification is unclear. Nevertheless, recent reports (Papouli et al. 2005; Pfander et al. 2005) that small ubiquitin-related modifier-modified PCNA has a higher affinity for Srs2 than unmodified PCNA and recruits Srs2 to the damage site to prevent recombination may shed light on the possible roles of the damage checkpoint in PRR, since Srs2 itself is considered a component of the damage checkpoint (Liberi et al. 2000) and interacts genetically with G1/S and G2/M checkpoint genes (Liberi et al. 2000; Vaze et al. 2002). Another recent report (Sabbioneda et al. 2005) of physical interaction between the PCNA clamp-like structure and the Rev7 subunit of Polζ provides an alternative link between damage checkpoint and error-prone PRR. Further investigation is required to determine whether this interaction is responsible for the observed genetic interactions reported in this study.
The authors thank C. Boone and Renee Brost for technical assistance with SGA experiments and yeast strains, others for yeast strains and plasmids, and Michelle Hanna for proofreading the manuscript. This work was supported by a Canadian Institutes of Health Research operating grant (MOP-38104) to W.X. L. Barbour is a recipient of College of Medicine and Arthur Smyth Memorial Scholarships.
↵1 Present address: NIDDK, National Institutes of Health, Bethesda, MD 20892-0840.
Communicating editor: M. Lichten
- Received May 27, 2006.
- Accepted October 10, 2006.
- Copyright © 2006 by the Genetics Society of America