|Home | About | Journals | Submit | Contact Us | Français|
Most DNA alterations occur during DNA replication in the S phase of the cell cycle. However, the majority of eukaryotic cells exist in a nondividing, quiescent state. Little is known about the factors involved in preventing DNA instability within this stationary-phase cell population. Previously, we utilized a unique assay system to identify mutations that increased minisatellite alterations specifically in quiescent cells in Saccharomyces cerevisiae. Here we conducted a modified version of synthetic genetic array analysis to determine if checkpoint signaling components play a role in stabilizing minisatellites in stationary-phase yeast cells. Our results revealed that a subset of checkpoint components, specifically MRC1, CSM3, TOF1, DDC1, RAD17, MEC3, TEL1, MEC1, and RAD53, prevent stationary-phase minisatellite alterations within the quiescent cell subpopulation of stationary-phase cells. Pathway analysis revealed at least three pathways, with MRC1, CSM3, and TOF1 acting in a pathway independent of MEC1 and RAD53. Overall, our data indicate that some well-characterized checkpoint components maintain minisatellite stability in stationary-phase cells but are regulated differently in those cells than in actively growing cells. For the MRC1-dependent pathway, the checkpoint itself may not be the important element; rather, it may be loss of the checkpoint proteins' other functions that contributes to DNA instability.
Most eukaryotic cells spend the majority of their life span in a quiescent or nondividing state. It is thought that mutations arising within quiescent cells can allow them to escape quiescence, reentering the cell cycle and initiating tumorigenesis (1–3). Bypass of cellular growth controls results in unregulated cell divisions and the accumulation of genomic instability, two of the major hallmarks of cancer. At present, little is known about the initial mutagenic events that occur within quiescent cells that cause them to become cancerous. This is due to a lack of assay systems to assess genomic stability within this specific cell population. We previously developed an assay to monitor DNA stability in stationary-phase yeast cells. Since stationary-phase yeast cells are cells that have entered a cellular state called G0, which is equivalent to the mammalian quiescent state (4, 5), our assay allows us to analyze events in this critical population (6–8).
Checkpoints act as guardians of genomic stability, preventing the accumulation and replication of damaged DNA. In the yeast Saccharomyces cerevisiae, several checkpoints arrest the cell cycle when DNA damage, replication, or mitotic errors occur (Fig. 1). During S phase, the replication checkpoint is activated when replication forks stall (9, 10). This allows DNA repair to occur and prevents the formation of single-stranded DNA (ssDNA), which can be used as a template for chromosomal rearrangements. The protein kinase Mec1p is recruited to RPA-coated ssDNA at stalled replication forks, where it phosphorylates signal transducers, such as Mrc1p or Rad9p (11–13). Phosphorylation of Mrc1p or Rad9p results in the recruitment of Rad53p to the replication fork and activation of the replication checkpoint (14–16).
Intra-S-phase and DNA damage checkpoints utilize similar signal transduction pathways to block cell cycle progression (Fig. 1) (10, 17). The DNA damage signal is propagated by the kinases Mec1p and Tel1p, a protein functionally redundant with Mec1p (18–20). Mec1p and Rad24p (a component of the RFC-like complex which loads the PCNA-like clamp composed of Ddc1p, Rad17p, and Mec3p onto DNA [21, 22]) are independently recruited to the site of DNA damage (23, 24). This localization results in the activation of Mec1p and leads to the activation of Rad53p through the phosphorylation of Rad9p. Alternatively, Mec1p can also phosphorylate the kinase Chk1p to relay the signal and prevent cell cycle progression (25).
During mitosis, the spindle assembly checkpoint prevents the onset of anaphase until all chromosomes are properly attached to the mitotic spindle (26). In brief, Mps1p, Mad1p, Mad2p, Mad3p, Bub1p, and Bub3p localize to unattached kinetochores (27, 28). This binding prevents the activation of the anaphase-promoting complex (APC) through inhibition of its substrate Cdc20p (29). The inhibition of Cdc20p prevents the degradation of APC substrates, thereby delaying anaphase progression until all sister chromatids are properly aligned. These cell cycle surveillance mechanisms ensure the proper replication of DNA and segregation of chromosomes.
In yeast, stationary-phase cells recapitulate the phenotypes of quiescent cells found in other eukaryotic systems (4, 5). Recent studies in S. cerevisiae have shown that the checkpoint kinase Rad53p is phosphorylated in DNA repair mutants that are in stationary phase (30). Furthermore, the authors found a dramatic increase in the amount of spontaneous mutability within these cells. Due to the discovery of Rad53p activation in stationary-phase cells, we wanted to determine if other well-characterized checkpoint components play a role in maintaining genomic stability once cells have exited the cell cycle.
To allow us to study genomic rearrangement events in stationary-phase cells, we developed a novel color segregation assay that assesses the genomic stability of repetitive DNA tracts in stationary-phase yeast cells (6, 7). For this assay, we inserted a short minisatellite into the ADE2 gene (the ade2-min3 allele) of S. cerevisiae (Fig. 2a). Insertion of the minisatellite tract in the ADE2 gene pushes the gene out of frame, resulting in a red colony color and making the cells auxotrophic for adenine. Alterations occurring within the minisatellite are monitored by examining strains bearing the ade2-min3 allele for the presence of a color segregation phenotype. If the minisatellite tract loses one repeat or gains two repeats, the ADE2 gene is pushed back into frame, resulting in colonies that are now white in color and Ade+.
Previously, we performed a UV mutagenesis screen to find mutants that destabilized the minisatellite tract within the ade2-min3 allele (7). From this screen, we identified several mutations within the zinc homeostasis genes ZRT1 and ZAP1 that resulted in a novel color segregation phenotype called blebbing. Blebs are white Ade+ microcolonies that form on the surface of the red Ade− colony. Further characterization of the blebbing phenotype revealed that it resulted from the loss or gain of minisatellite repeats that occurred once cells entered stationary phase. Restoration of ADE2 function allowed the cell to escape the adenine nutrient limitation that had caused the cells to enter stationary phase and begin to grow again, forming the white microcolonies on the red colony. We subsequently demonstrated that the DNA alterations were occurring specifically within quiescent cells and not within a subpopulation of slowly growing G0 cells. We also demonstrated that the alterations were occurring independently of the ADE2 assay system, were not affected by genomic location, and were dependent upon an active recombination system (6, 7). Thus, our prior work establishes an assay system in which genomic instability can be readily studied within stationary-phase cells.
In this study, we used a modified version of synthetic genetic array (SGA) analysis (31, 32) to determine if well-characterized checkpoints play an active role in stabilizing minisatellites during stationary phase. In brief, we mated a query strain containing the ade2-min3 allele to mutant strains bearing either a complete deletion of a nonessential gene or a temperature-sensitive (ts) allele of an essential gene. We examined the resulting double mutants for the presence of a blebbing phenotype and then verified this phenotype in a separate strain background. Our results reveal that a specific subset of checkpoint components (namely, MRC1, CSM3, TOF1, DDC1, RAD17, MEC3, TEL1, MEC1, and RAD53) prevents minisatellite alterations in quiescent yeast cells in stationary phase.
Interestingly, CHK1, RAD9, and BUB3, despite their major roles in several cell cycle checkpoints, do not stabilize minisatellites in stationary-phase cells, implying that checkpoints in stationary-phase cells function in novel pathways and are regulated differently. This is supported by our pathway analysis in which we find that the key checkpoint components MEC1, RAD53, and the MRC1/CSM3/TOF1 complex function in independent genetic pathways during stationary phase, implying that the latter complex is not part of the Mec1p and Rad53p phosphorylation cascade. We also find that a RAD9-dependent pathway is active in stationary-phase cells and prevents minisatellite instability when Mrc1p function is compromised. Our work provides evidence that checkpoint proteins play an active role in maintaining genomic stability in stationary-phase cells and that these proteins interact differently in stationary-phase and actively growing cells.
Liquid and solid media used in this study were prepared as described previously (33). Solid yeast extract-peptone-dextrose (YPD) medium containing Geneticin (G418) was made by adding 200 mg/liter G418 sulfate (Cellgro) to YPD. Medium used for the SGA analysis was prepared as described elsewhere (31, 32). Presporulation solid medium consisted of 2% agar, 5% dextrose, 1% Difco yeast extract, and 3% Difco nutrient broth. Yeast mating, sporulation, and tetrad dissection were performed as stated previously (34).
All Saccharomyces cerevisiae strains used in this study are listed in Table 1. Each strain was derived from DTK264 (MATa his7-2 leu2::HisG trp1-289 ura3-52 ade2-min3) or DTK271 (MATα his7-2 leu2::HisG trp1-289 ura3-52 ade2-min3) (7). All strains with gene disruptions were constructed by isolating corresponding genomic DNA from the nonessential yeast deletion strain haploid set. In brief, PCR products containing the KANMX4 gene (G418 resistance) were generated from the genomic DNA using 5′ and 3′ flanking primers listed in Table 2. Parental strains were transformed with the KANMX PCR product and grown for 4 h at 30°C in liquid YPD. Transformants were selected on YPD plus G418 solid medium. All isolates were verified by PCR.
DTK1279, a strain bearing the pkc1-4 allele, was constructed by mating DTK271 with YKH25 (a gift from Lorraine Symington, Columbia University) (36). Spores were isolated by red color and lack of growth at 37°C. The resulting haploid was backcrossed twice to DTK271 to produce DTK1279.
The strain DTK1776, bearing a point mutation in MEC1, was constructed by mating DTK271 with the mec1-21 strain DKY256 (a gift from Deanna Koepp, University of Minnesota); this strain was originally from Y663 (37). The resulting diploid was sporulated and dissected, and a spore bearing the ade2-min3 allele, as well as the mec1-21 mutation, was isolated by colony color and sensitivity to 100 mM hydroxyurea (HU). This isolate was backcrossed twice to DTK271 to generate DTK1776. DTK1170, bearing a point mutation in RAD53, was constructed in a similar manner. The ade2-min3 strain DTK271 was mated with the rad53-1 strain DCY1793 (35). An ade2-min3 rad53-1 isolate was identified by colony color and sensitivity to HU. This spore isolate was backcrossed twice to the parental DTK271 strain as above to generate DTK1170, an ade2-min3 rad53-1 isolate.
DTK1630, bearing the ade2-min3 allele and a checkpoint-defective allele of MRC1 (mrc1AQ), was constructed by mating DTK271 with Y2298 (15). Spores were isolated by colony color and growth on medium lacking histidine. Progeny were backcrossed two times to DTK271. The strain DTK1683, bearing ade2-min3 and a replication-defective allele of MRC1 (mrc1-C14), was constructed by mating DTK271 with Y2544 (Y2298 and Y2544 are gifts from Steve Elledge, Harvard Medical School) (13). Spores were isolated by resistance to G418 and colony color. Progeny were backcrossed twice to DTK271. DTK1766, a strain bearing the C14 C-terminal truncation of Mrc1p combined with the AQ mutations, was constructed by transforming a KANMX PCR product into DTK1630. The KANMX PCR product, bearing MRC1 flanking sequence to delete residues 844 to 1096, was generated using pFAKANMX6 as the template and the primers 78415868 and 78415869. Transformants were selected by resistance to G418 and sequenced.
The strain DTK1979, bearing the rfa1-t11 allele, was constructed using the “pop-in/pop-out” method described elsewhere (38). Here, the rfa1-t11 open reading frame was excised from pKU1-t11 (a gift from Richard Kolodner, University of California San Diego) with SalI and HindIII (39). The resulting fragment was gel purified. The purified rfa1-t11 fragment was cloned into the multiple cloning site of pRS306, resulting in the plasmid pBMA105. Insertion of the rfa1-t11 open reading frame was confirmed by sequencing. pBMA105 was then linearized with MfeI. The digested plasmid was gel purified and transformed into DTK271. Transformants bearing an integrated rfa1-t11 allele were selected on synthetic dextrose without uracil (SD-Ura) solid medium. Ura+ transformants were patched onto YPD medium, grown at 30°C, and then replica plated to 5-fluoroorotic acid (5-FOA) medium. Single papillae were patched onto YPD, incubated at 30°C overnight, and then replica plated to SD-Ura solid medium. DTK1979 was isolated as a red Ura− transformant bearing the rfa1-t11 allele, as verified by sequencing and sensitivity to 0.02% methyl methanesulfonate (MMS).
The following haploid strains were generated by mating, sporulation, and tetrad dissection: DTK1171 (DTK878 × DCY1793), DTK1199 (DTK271 × DTK 1012), DTK1217 (DTK1170 × DTK1199), DTK1219 (DTK1088 × DTK1170), DTK1295 (DTK1170 × DTK1279), DTK1549 (DTK1392 × DTK1532), DTK1551 (DTK1392 × DTK1521), DTK1554 (DTK1528 × DTK1521), DTK1556 (DTK1539 × DTK878), DTK1558 (DTK1392 × DTK1170), DTK1559 (DTK1392 × DTK1199), DTK1587 (DTK1539 × DTK1360), DTK1588 (DTK1539 × DTK1361), DTK1603 (DTK1392 × DTK1279), DTK1693 (DTK1532 × DTK1361), DTK1694 (DTK1532 × DTK1360), DTK1695 (DTK1521 × DTK1360), DTK1700 (DTK1521 × DTK1361), DTK1762 (DTK1697 × DTK878), DTK1765 (DTK1697 × DTK1170), DTK1767 (DTK1697 × DTK1088), DTK1769 (DTK1697 × DTK1199), DTK1777 (DTK1697 × DTK1279), DTK1806 (DTK1539 × DTK1088), DTK1855 (DTK1776 × DTK1170), and DTK1871 (DTK1776 × DTK1603). For each strain, a spore containing the ade2-min3 allele as well as the marked gene deletion was isolated by color phenotype and growth on selective medium. Strains containing the mec1-21 or rad53-1 allele were also isolated by sensitivity to HU. Strains bearing the pkc1-4 allele were selected based on lack of growth at 37°C. Strains were verified by PCR. DTK1855 was verified by sequencing.
For SGA analysis, we constructed DTK893 (MATa his3-1 ura3-0 can1::MFA1pr-spHIS5 ade2-min3-URA3MX). A two-step PCR process was used to generate an ade2-min3 URA3MX-linked cassette. In brief, a URA3MX PCR product, flanked by a 5′ TEF promoter site and a 3′ TEF terminator site, was obtained from pDC369 (generous gift from Duncan Clarke, University of Minnesota), using the primers 14193004 and 14193005. An ade2-min3 PCR product was isolated from DTK271 genomic DNA using the primers 14193006 and 14193007. The URA3MX and ade2-min3 PCR products were combined using the primers 14193007 and 14193008. The complete ade2-min3 URA3MX-linked cassette was transformed into DCY2556, and Ura+ cells were selected on solid synthetic medium lacking uracil, yielding DTK893. DTK893 was crossed to DCY2557 (DCY2556 and DCY2557 are gifts from Duncan Clarke, University of Minnesota) to generate the diploid DTK1175. DTK1175 was sporulated and dissected, and red Ura+ spores were selected on synthetic medium lacking uracil. Isolates of DTK1175 were chosen as the query strains for the various SGA analyses: DTK1189 5a is a MATa derivate, and DTK1189 2b is a MATα derivate.
We conducted a modified version of synthetic genetic array analysis (31, 32). For this screen, we inoculated liquid YPD medium with a single colony of the DTK1189 5a query strain. After incubation at 30°C overnight with agitation, the culture was plated on solid YPD medium. Once these plates were dry, strains from the MATα nonessential yeast deletion strain haploid set (a gift from Robin Wright, University of Minnesota) (Invitrogen) was manually pinned onto the query strain in a 96-well format and incubated at 30°C overnight. Zygotes were pinned to SD-Ura plus G418 solid medium and incubated at 30°C overnight. To enhance the level of sporulation, we pinned the diploids to presporulation solid medium and incubated the diploids at 30°C overnight. Strains were sporulated at room temperature (RT) for 6 days. MATa progeny were selected on SD-His/Arg/Ura plus canavanine (Can) (US Biological) solid medium and incubated at 30°C overnight. This step was repeated. Double mutants containing the ade2-min3 allele were isolated by pinning the strains to SD-His/Arg/Ura plus Can plus G418 solid medium. This selection step was repeated. After selection, the haploids were left at RT on YPD for 5 days to assay for the blebbing phenotype. Each strain was pinned in duplicate. The screen was repeated three independent times. A MATα zrt1Δ haploid mutant, which has a strong degree of blebbing, was used as a positive control for our screen (7). All mutants producing a strong degree of blebbing (+++ or ++++ on a scale of + to ++++) were backcrossed three times to DTK271 to confirm the blebbing phenotype.
For this screen, we followed a protocol similar to the nonessential SGA protocol described above. The query strain used for the essential SGA was DTK1189 2b. Strains to be examined that were part of a MATa essential temperature-sensitive (ts) strain set containing 455 ts genes (a gift from Charles Boone, University of Toronto) were manually pinned onto the query strain in a 96-well format and incubated at RT for 2 days. Zygotes were pinned to SD-Ura plus G418 solid medium and incubated at RT overnight. Diploids were pinned to presporulation solid medium and incubated at RT overnight. Strains were then sporulated at RT for 6 days. MATa progeny were selected on SD-His/Arg/Ura plus Can solid medium and incubated at RT for 2 days. This step was repeated. Double mutants containing the ade2-min3 allele as well as a temperature-sensitive allele were isolated by pinning the strains to SD-His/Arg/Ura plus Can plus G418 solid medium and incubating at RT for 2 days. This selection step was repeated. After selection, the haploids were pinned to 5 separate YPD solid medium plates. Each plate was left at 26°C, 30°C, 32°C, 34°C, or 37°C for 5 days. Each strain was pinned in duplicate, and the screen was repeated three independent times. Blebbing for each plate was scored as stated above. We define a hit as follows: (i) a strain that produced a steadily increasing level of blebbing, reaching +++ or ++++ only when the critical temperature of the ts allele was reached for all 3 screens or (ii) a strain that produced a level of blebbing of +++ to ++++ only when the critical temperature of the ts allele was reached for all 3 screens. A MATa zrt1Δ haploid mutant was used as a positive control for this screen.
Strains were streaked on solid YPD medium and incubated at 30°C for 48 h to allow colonies to form. Five milliliters of liquid YPD medium was inoculated with single red colonies of each strain and incubated at 30°C for 4 h with agitation. Appropriate dilutions of each culture were plated on solid YPD medium and incubated at 30°C for 2 days. We then left the plates at RT for 6 days and counted the number of blebs on the surface of each colony. Blebs were counted on at least 100 colonies for each strain; we repeated this procedure for 3 independent experiments. Only colonies ranging from 1.26 to 1.32 mm in diameter were included. The average number of blebs per colony and the 95% confidence interval of the mean were calculated for each strain. Strains bearing the pkc1-4 mutation were grown at 35°C for 6 days, and then the number of blebs was counted.
We used the ade2-min3 reporter (6, 7) to determine the effect on minisatellite stability of mutations in checkpoint genes (Fig. 2a). The ade2-min3 construct consists of three 20-bp minisatellite repeat units plus a 5-bp linker inserted into the beginning of the ADE2 coding sequence, disrupting the reading frame and resulting in a red colony color. If one repeat unit is lost (or two repeats gained), the correct reading frame is restored and a white colony color results. All alterations observed involved increases or decreases of complete 20-bp repeat units.
We previously described the ade2-min3 reporter assay and its ability to distinguish between minisatellite alterations occurring during log-phase growth and those occurring during stationary phase (6–8). Minisatellite alterations occurring during growth of the colony result in a white sector integrated into the red colony, while minisatellite alterations that occur specifically during stationary phase result in a novel colony morphology called “blebbing,” in which white Ade+ microcolonies form on the surface of the red Ade− colony.
Approximately 100 genes have been annotated in the Saccharomyces Genome Database (www.yeastgenome.org) as having checkpoint-related functions (Table 3). To determine which of the checkpoint-related genes are important for maintaining minisatellite stability, we used a modified version of the synthetic genetic array (SGA) protocol (31, 32) to screen those genes in the yeast nonessential deletion haploid set and the essential temperature-sensitive haploid set (40). We mated a query strain containing the ade2-min3 allele to deletion set strains containing the KANMX4 allele in place of the wild-type (WT) open reading frame or to the essential strain set bearing temperature-sensitive alleles of genes. Isolated double mutant strains containing the ade2-min3 allele as well as the deletion or temperature-sensitive mutation were analyzed for a color segregation phenotype.
Our screening of the checkpoint genes identified those that are required for stationary-phase minisatellite stability (Table 3). Loss of the replication checkpoint gene MRC1, CSM3, or TOF1 gave a pronounced blebbing phenotype similar to that of our positive zrt1Δ control (7), indicating a role for these genes in maintaining minisatellite stability during stationary phase. Similarly, strains bearing temperature-sensitive alleles of the spindle checkpoint gene IPL1 (41, 42), the kinetochore gene CEP3 (43, 44), or the S-phase checkpoint gene RFC4 (45) produced a strong blebbing phenotype at their corresponding critical temperatures (Table 3). Additionally, strains containing deletions of ddc1, rad17, mec3, or TEL1 (components of the S-phase and DNA damage checkpoints) exhibited a blebbing phenotype but at a lower level than that for zrt1Δ.
To verify the results of our SGA analysis, we deleted several candidate nonessential genes with well-known cell cycle checkpoint roles (Fig. 1) in our standard strain background, which has been well-characterized in our previous work (6–8). We examined deletions of genes that exhibited stationary-phase minisatellite instability (MRC1, CSM3, TOF1, DDC1, RAD17, BUB3, MEC3, and TEL1) and genes that did not (RAD9 and CHK1).
We quantified the blebbing frequency of each of these strains (Fig. 2b and and3)3) as previously described, but with one modification: blebs were counted after colonies were incubated for 6 days at room temperature instead of 4 days (6, 7). This modification was made to accommodate the small size of the checkpoint strain blebs. Time course analysis monitoring the abundance of white cells in cultures of the parental strain, the mrc1Δ strain, or the zrt1Δ strain revealed that white cells accumulated at a later time point in stationary phase for the mrc1Δ strain than for the wild-type and zrt1Δ strains (data not shown). Our analysis suggests that minisatellite alterations in checkpoint mutants occur at a later time in stationary phase than for other strains.
Deletion of MRC1, CSM3, or TOF1 led to a strong blebbing phenotype (Fig. 2b). The mrc1Δ strain displayed an average of 23.9 blebs per colony (Fig. 3). This value is significantly higher than that of its parent, which had an average of 4.7 blebs/colony. The 95% confidence intervals were calculated for each strain; lack of overlap between the 95% confidence intervals indicated that the difference between the blebs quantified for each strain was statistically significant. Similarly, the csm3Δ strain and the tof1Δ strain displayed high levels of blebbing (26.2 and 24.4 blebs/colony, respectively). From these data, we conclude that components of the replication checkpoint play an important role in stabilizing minisatellites during stationary phase.
Loss of the DNA damage checkpoint gene DDC1, RAD17, MEC3, or TEL1 led to an average level of blebbing of 10.5, 12.5, 11.5, and 8.0 blebs/colony, respectively (Fig. 2b and and3).3). These values were significantly higher than that of the parental strain (4.7 blebs/colony). These results suggest that components of the S-phase and DNA damage checkpoints are also involved in minisatellite stability during stationary phase.
Of the mutants surveyed, the rad9Δ, chk1Δ, and bub3Δ strains did not exhibit a level of blebbing significantly greater than that of the parental strain, indicating no effect on minisatellite stability (Fig. 2b and and3).3). The deletion mutant strains had 3.2, 5.8, and 1.5 blebs/colony, respectively, compared to the 4.7 blebs/colony seen with the parental strain, indicating that RAD9, CHK1, and BUB3 do not influence the stability of minisatellites in stationary-phase cells.
For most of the genes examined in the two strain backgrounds (the lab strain and the SGA strain), deletion of each of the checkpoint genes in our strain background produced results similar to those obtained from the SGA analysis. For example, deletion of MRC1, CSM3, and TOF1 in both strain backgrounds produced strong blebbing phenotypes compared to results for other checkpoint mutants analyzed, in both the SGA analysis and blebbing quantification analysis (Table 3 and Fig. 3). Similarly, deletion of RAD9 and CHK1 did not produce significant levels of blebbing in either the SGA or blebbing quantification analysis. Interestingly, deletion of BUB3 produced a high level of blebbing in the SGA analysis, but blebbing was not significant in the lab strain background. This could be due to the presence of an enhancer or suppressor mutation that could affect a strain's blebbing phenotype.
To determine if the essential central checkpoint genes MEC1 and RAD53 were also involved in maintaining minisatellite stability, we constructed strains bearing a hypomorphic allele of MEC1 (mec1-21) or RAD53 (rad53-1) (35, 37). We then calculated the blebbing frequency for each strain. The mec1-21 strain had 14.2 blebs/colony, significantly greater than that of the WT strain. Similarly, the rad53-1 mutant had an average of 17.8 blebs/colony, significantly higher than the parental frequency of 4.7 blebs/colony (Fig. 3), demonstrating that loss of MEC1 or RAD53 results in a significant increase in minisatellite alterations during stationary phase. We were unable to accurately assess the average number of blebs per colony of a mec1-21 tel1Δ strain; in our strain background, this double mutant exhibited a severe defect in colony formation.
We observed a sectoring phenotype in both mec1-21 and rad53-1 strains (Fig. 2b), likely resulting from minisatellite alterations that occur during colony formation. This result is consistent with the previously described roles of Mec1p and Rad53p in regulating genome stability during cell cycle growth (46) and indicates that minisatellite alterations in the mec1-21 and rad53-1 strains occur in actively dividing cells as well as in cells that have entered stationary phase. No other blebbing strains produced a sectoring phenotype.
Previous work by the Werner-Washburne group demonstrated that yeast stationary-phase cultures consist of both true quiescent cells and nonquiescent cells that are still undergoing slow division (47, 48). Characterization of these subpopulations of stationary-phase cultures showed that each is differentially regulated by a specific set of genes (48). This genetic differentiation of stationary-phase cells allows us to distinguish which cell population gives rise to the blebbing phenotype. Loss of ETR1, which encodes a 2-enoyl thioester reductase, drastically reduces the ability of the quiescent cell population to reenter the cell cycle (48, 49). Similarly, loss of POR1, a gene that encodes mitochondrial porin, reduces the ability of nonquiescent cells to reenter the cell cycle (48, 50). If a cell cannot reenter the cell cycle, it cannot generate a blebbing colony.
To determine if minisatellite alterations in the checkpoint mutants occur in the quiescent or nonquiescent population of stationary-phase cells, we deleted ETR1 or POR1 in mrc1Δ, csm3Δ, tof1Δ, mec1-21, or rad53-1 strains and quantified the blebbing frequency of the resulting double mutants (Fig. 4). The mrc1Δ etr1Δ, csm3Δ etr1Δ, or tof1Δ etr1Δ mutant had 0.05, 0.01, or 0.01 bleb/colony, respectively. This is a significant decrease in blebbing from that of each of the single mutants. Similarly, deletion of ETR1 in a mec1-21 or rad53-1 background reduced the level of blebbing to 1.6 and 0.01 blebs/colony, respectively. In comparison, deletion of POR1 in a mrc1Δ, csm3Δ, or tof1Δ strain background did not result in a complete loss of blebbing. The mrc1Δ por1Δ, csm3Δ por1Δ, or tof1Δ por1Δ double mutant had a level of blebbing of 17.7, 18.1, or 13.9 blebs/colony, respectively. These results indicate that minisatellite alterations in a mrc1Δ, csm3Δ, or tof1Δ strain occur predominantly within the quiescent population rather than the nonquiescent population of stationary-phase cells. However, deletion of POR1 in a mec1-21 or rad53-1 background resulted in a decrease in blebbing from 14.2 blebs/colony (mec1-21) to 6.7 blebs/colony and from 17.8 blebs/colony (rad53-1) to 3.7 blebs/colony, respectively. We conclude that minisatellite alterations in a mec1-21 and rad53-1 background occur within both quiescent and nonquiescent cells populations, in keeping with phenotypes in actively growing cells.
To investigate how many independent pathways for stationary-phase minisatellite maintenance are represented by our blebbing mutants, we examined the blebbing phenotype of double mutant strains. The mrc1Δ csm3Δ, mrc1Δ tof1Δ and csm3Δ tof1Δ deletion strains produced equivalent blebbing frequencies (data not shown). We interpret this to mean that all three gene products function within the same pathway; therefore, for the remainder of the pathway analysis, we used only the mrc1Δ strain to construct double mutants.
We previously demonstrated that deletion of the high-affinity zinc transporter gene ZRT1 results in a dramatic increase in the amount of blebbing in an ade2-min3 strain (6, 7). To determine if MRC1, RAD17, MEC1, and RAD53 function within a ZRT1-dependent pathway, we constructed mrc1Δ zrt1Δ, rad17Δ zrt1Δ, mec1-21 zrt1Δ, and rad53-1 zrt1Δ double mutants. The mrc1Δ zrt1Δ mutant had an average of 31.1 blebs/colony, higher than the average number of blebs/colony for each single mutant (for mrc1Δ, 23.9, and for zrt1Δ, 24.5 blebs/colony) (Fig. 5a). However, blebs were difficult to quantify after 6 days at room temperature with this particular strain due to an enhanced blebbing phenotype of the double mutant. We therefore quantified the average number of blebs for each of these strains at an earlier time point. After 3 days at room temperature, the mrc1Δ zrt1Δ mutant produced an average of 28.5 blebs/colony, a value significantly greater than that for either of the single mutants (for mrc1Δ, 8.0, for zrt1Δ, 15.9 blebs/colony). This additive effect and the enhanced blebbing phenotype of the double mutant indicate that Mrc1p regulates minisatellite stability in a pathway that is at least partially independent of Zrt1p.
To determine if the enhanced blebbing phenotype of the mrc1Δ zrt1Δ double mutant was due to an increase in the number of minisatellite alteration events or an increase in the growth rate of the white cells comprising the blebs, we performed a time course analysis of white cells isolated from blebs of the wild-type parent or the mrc1Δ, zrt1Δ, or mrc1Δ zrt1Δ mutant (data not shown). We found that blebs from the mrc1Δ zrt1Δ double mutant did not grow at a significantly higher rate than those from the wild-type strain or either single mutant. This result, as well as the results of the blebbing quantification experiment, suggests that the enhanced blebbing phenotype found in the double mutant is due to an increase in the number of minisatellite alteration events at an earlier time rather than an increase in the growth rate of the cells forming the blebs.
The rad17Δ zrt1Δ double mutant produced an average of 16.0 blebs/colony, a value not significantly greater than that for the rad17Δ (12.5 blebs/colony) or zrt1Δ (24.5 blebs/colony) single mutant, placing RAD17 in a ZRT1-dependent pathway. The mec1-21 zrt1Δ strain had an average of 24.3 blebs/colony, while the rad53-1 zrt1Δ mutant averaged 23.5 blebs/colony. Both values were not significantly greater than those for the corresponding single mutants and, unlike the MRC1 result, place MEC1 and RAD53 in the same pathway as ZRT1 (Fig. 5a). Our results suggest that Mrc1p functions in a pathway separate from Rad53p and Mec1p during stationary phase. This is in contrast to their S-phase checkpoint roles, in which each gene product functions within the same checkpoint signaling pathway (Fig. 1) (14, 15).
To examine this relationship further, we constructed mrc1Δ mec1-21, mrc1Δ rad53-1, and mec1-21 rad53-1 double mutants (Fig. 5b and andc).c). Analysis of the mrc1Δ mec1-21 mutant revealed that while the level of blebbing of the double mutant (29.1 blebs/colony) was not significantly additive compared to that for the mrc1Δ (23.9 blebs/colony) or mec1-21 (14.2 blebs/colony) strain, the blebbing phenotype of the mrc1Δ mec1-21 mutant itself was drastically different from that of the parental strains. Double mutant colonies had a range of bleb sizes, unlike all other strains in which the bleb size on the colony is homogeneous. This difference suggests that MRC1 and MEC1 do indeed comprise different pathways and that the minisatellite alterations in the double mutant occur at a variety of time points during stationary phase.
In keeping with this trend, the mrc1Δ rad53-1 mutant (38.3 blebs/colony) produced a significantly additive level of blebbing compared to that for either single mutant, demonstrating that Mrc1p and Rad53p are functioning in separate pathways during stationary phase (Fig. 5b and andc).c). Interestingly, while the mec1-21 rad53-1 mutant (21.9 blebs/colony) did not have a level of blebbing that was significantly additive compared to that of either single mutant, at 14.2 (mec1-21) or 17.8 (rad53-1) blebs per colony, the resulting blebbing phenotype of the double mutant was drastically different. Like the mrc1Δ zrt1Δ and mrc1Δ mec1-21 strains, the double mutant produced blebs that were significantly larger than those produced by the parental strains, suggesting that the minisatellite alterations in a mec1-21 rad53-1 strain occur at an earlier time point than those in the mec1-21 or rad53-1 strain. Together, our data suggest that MEC1 and RAD53 comprise different pathways in stationary-phase cells. This is in sharp contrast to their roles in actively dividing cells, in which Mec1p and Rad53p function together in well-characterized checkpoint signaling pathways (46).
The division of MEC1 and RAD53 into different pathways is further supported by the analysis of RAD17 double mutants (Fig. 5b). The mec1-21 rad17Δ mutant had 34.1 blebs/colony, a value significantly greater than that of either single mutant, at 14.2 (mec1-21) or 12.5 (rad17Δ) blebs/colony. In contrast, the rad53-1 rad17Δ strain (7.0 blebs/colony) produced a level of blebbing significantly lower than that of either single mutant, at 17.8 (rad53-1) and 12.5 (rad17Δ) blebs/colony, which supports our previous data that MEC1 and RAD53 fall into separate pathways during stationary phase. The combination of the mrc1Δ and rad17Δ mutations was synthetically lethal in our strain background; therefore, we are unable to assess the pathway relationship of MRC1 and RAD17.
Our lab found that RAD27, PKC1, and END3 represent several other distinct pathways that mediate minisatellite instability in stationary-phase cells (8). RAD27 encodes a nuclease involved in processing Okazaki fragments (51). PKC1 codes for an essential serene/threonine protein kinase C homolog, while END3 encodes a protein involved in endocytosis and actin organization (52, 53). The mrc1Δ rad27Δ strain and the mec1-21 rad27Δ strain produced averages of 59.5 and 46.9 blebs/colony, respectively, values significantly higher than those of the mrc1Δ (23.9 blebs/colony), rad27Δ (38.8 blebs/colony), and mec1-21 (14.2 blebs/colony) single mutants (Fig. 5d). The rad53-1 rad27Δ double mutant averaged 27.3 blebs/colony, which was not significantly additive compared to findings for the parental single mutants. In contrast, the rad17Δ rad27Δ strain had an average of 30.9 blebs/colony, a value that falls between those of the single mutants, at 12.5 (rad17Δ) and 38.8 (rad27Δ) blebs/colony. From these data, we conclude that Rad17p and Rad53p but not Mrc1p or Mec1p function in a Rad27p-dependent pathway.
Double mutants with PKC1 (mrc1Δ pkc1-4, 25.7 blebs/colony; rad17Δ pkc1-4, 15.1 blebs/colony) produced levels of blebbing that were not significantly elevated over those of the mrc1Δ (23.9 blebs/colony), rad17Δ (12.5 blebs/colony), and pkc1-4 (14.8 blebs/colony) single mutants (Fig. 5e). Similarly, the mec1-21 pkc1-4 (13.1 blebs/colony) and rad53-1 pkc1-4 (14.6 blebs/colony) strains had levels of blebbing that did not significantly differ from those of the parental strains (mec1-21, 14.2 blebs/colony; rad53-1, 17.8 blebs/colony). Mrc1p, Rad17p, Mec1p, and Rad53p all appear to function within a Pkc1p-dependent pathway. Also, the blebbing frequencies in the end3Δ (8.0 blebs/colony) double mutants with mrc1Δ end3Δ (15.9 blebs/colony), rad17Δ end3Δ (7.7 blebs/colony), mec1-21 end3Δ (12.0 blebs/colony), and rad53-1 end3Δ (9.8 blebs/colony) genotypes were not enhanced in comparison to those of the corresponding single mutants (Fig. 5f).
Together, our data suggest that checkpoint proteins participate in at least three independent pathways to maintain the stability of minisatellites during stationary phase. One pathway involves Rad53p, Zrt1p, Rad17p, and Rad27p, while a second has Mrc1p, Csm3p, and Tof1p. Both pathways contain Pkc1p and End3p. The final pathway utilizes Mec1p, Zrt1p, Pkc1p, and End3p but does not contain Rad17p or Rad27p.
In actively dividing cells, Mrc1p has been shown to promote Mec1p- and Rad53p-dependent checkpoint signaling and maintain proper replicative function (11, 13–15). The Elledge group has developed MRC1 mutants that allow one to study the effects of the checkpoint or replicative functions of Mrc1p separately. These include the mrc1AQ mutant, a strain bearing mutations at the Mec1p phosphorylation sites, and the mrc1-C14 strain, a C-terminal truncation mutant that displays a prolonged S phase (13, 15).
To determine if the minisatellite alterations in the mrc1Δ strain are due to loss of the checkpoint function of Mrc1p or the replicative function of Mrc1p, we quantified the average number of blebs/colony in strains bearing the ade2-min3 allele and the mrc1AQ or mrc1-C14 separation-of-function alleles (Fig. 6a). The mrc1AQ phosphorylation site mutant produced 7.1 blebs/colony, a value that, while still significantly greater than that of the wild-type strain (4.7 blebs/colony), is not a drastic increase such as that seen in the mrc1Δ deletion mutant (23.9 blebs/colony). The mrc1-C14 terminal truncation mutant averaged 18.5 blebs/colony, a value significantly greater than that of the wild-type strain. From these data we conclude that the minisatellite stability in stationary-phase cells is mediated predominately by the replicative function of Mrc1p, rather than the checkpoint signaling function of Mrc1p that utilizes Mec1p and Rad53p. This finding complements our pathway analysis data that showed that MRC1, MEC1 and RAD53 lie in distinct genetic pathways in stationary-phase cells (Fig. 5b and andcc).
Surprisingly, a strain bearing both the AQ and C14 mutations did not exhibit a level of blebbing equivalent to that of the mrc1Δ strain (17.7 versus 23.9 blebs/colony, respectively) (Fig. 6a). This result could be due to the continued presence of the bulk of the Mrc1p protein or to a possible backup checkpoint pathway, either or both of which could contribute to continued minisatellite stability in the absence of the well-characterized functions of Mrc1p.
Mrc1p and Rad9p were previously shown to represent functionally redundant checkpoint pathways (11). Further analysis from the Elledge lab suggests that Rad9p may act as a backup pathway in actively dividing cells when Mrc1p function is compromised (11, 15). To address the question of whether or not Rad9p acts as a backup checkpoint pathway in stationary phase, we compared the level of blebbing of an mrc1AQ rad9Δ double mutant to those of the parental single mutants (Fig. 6a). The combined mutations produced a synergistic blebbing effect (13.9 blebs/colony) compared to results for the mrc1AQ (7.1 blebs/colony) and rad9Δ (3.2 blebs/colony) single mutants. Thus, when the checkpoint signaling function of Mrc1p is absent, a Rad9p-dependent checkpoint signaling pathway prevents further minisatellite instability. The mrc1Δ rad9Δ and mrc1-C14 rad9Δ mutation combinations exhibit synthetic lethality in our background, in agreement with prior observations that Rad9p is required for cell viability in mrc1Δ mutants that accrue DNA damage due to replication stress (11, 15).
Our quiescent and nonquiescent population analysis of the mrc1Δ strain revealed that the majority of stationary-phase minisatellite alterations were specific to quiescent cells. However, we also found that ~25% of alterations appeared to occur in nonquiescent cells (Fig. 4). To determine if the different functions of Mrc1p accounted for these results, we deleted POR1 and ETR1 in the mrc1AQ, mrc1-C14, and mrc1AQ-C14 mutant backgrounds (Fig. 6b). Deletion of POR1 in both the mrc1AQ and mrc1-C14 backgrounds did not result in a decrease in blebbing (8.2 and 17.1 blebs/colony, respectively). However, a significant decrease in blebbing was observed in the mrc1AQ-C14 por1Δ strain (14.0 blebs/colony). In comparison, deletion of ETR1 in each of the mutant backgrounds resulted in a dramatic decrease in the level of blebbing compared to that for each of the single mutants (for mrc1AQ etr1Δ, 0.4 bleb/colony; for mrc1-C14 etr1Δ, 0.4 bleb/colony; for mrc1AQ-C14 etr1Δ, 0.6 bleb/colony). We conclude that minisatellite alterations associated with the loss of the well-characterized checkpoint signaling and replication functions of Mrc1p occur primarily within the quiescent population of stationary-phase cells.
To investigate the mechanism giving rise to minisatellite alterations and the blebbing phenotype in our yeast strains, we quantified the average number of blebs per colony in strains bearing the ade2-min3 minisatellite tract and rfa1-t11, a mutant allele of the RFA1 subunit of replication protein A (RPA) (39). RPA is an ssDNA binding protein that is highly conserved in eukaryotes, where it functions to promote effective DNA replication, recombination, and repair (reviewed in reference 54). Yeast strains carrying the rfa1-t11 mutation are defective in recombination during DNA repair and during meiosis but have no measurable defects in DNA replication (39, 55, 56). Thus, examination of the rfa1-t11 allele allows us to separate the effects of repair-associated recombination and DNA replication on minisatellite stability in stationary-phase cells.
A strain bearing the rfa1-t11 and ade2-min3 alleles produced a level of blebbing (5.2 blebs/colony) that was not significantly different from that of the wild-type strain (4.7 blebs/colony). Thus, the rfa1-t11 mutation by itself does not appear to affect minisatellite stability. We next investigated the level of blebbing in strains bearing rfa1-t11 and deletion of either MRC1 or ZRT1. The mrc1Δ rfa1-t11 strain produced a level of blebbing of 10.2 blebs/colony, a value approximately 50% of that produced by the mrc1Δ mutant alone (23.9 blebs/colony). A similar result occurred in the zrt1Δ rfa1-t11 mutant, which had an average of 11.3 blebs/colony, compared to 24.5 blebs/colony produced by the zrt1Δ mutant. This reduction is equivalent to that seen previously with various recombination gene deletion combinations, such as rad50Δ zrt1Δ (6, 7). We conclude that while rfa1-t11 does not have an effect on the low level of spontaneous minisatellite instability displayed in the wild-type strain, minisatellite alterations occurring within mrc1Δ or zrt1Δ mutant strains are at least partially dependent upon the recombination function associated with RFA1.
In this study, we used a modified version of the SGA analysis (31, 32) to screen for checkpoint components that play an active role in stabilizing minisatellites in stationary-phase yeast cells. We find that a specific subset of checkpoint genes, including MRC1, CSM3, TOF1, DDC1, RAD17, MEC3, TEL1, MEC1, and RAD53, act to prevent minisatellite alterations in stationary-phase cells. We show that minisatellite alterations within these checkpoint mutant strains occur predominantly in the quiescent cell population of stationary-phase colonies. Finally, we demonstrate that MRC1, MEC1, and RAD53 function in separate pathways in stationary-phase cells, a significant difference from the checkpoint pathways characterized in actively dividing cells. Our data suggest that novel checkpoint pathways are in place in stationary-phase cells to prevent genomic instability and that components of these pathways are interacting differently in stationary-phase cells than in actively dividing cells.
The role of checkpoint proteins in actively growing cells is to sense a cellular state and transmit a signal under the appropriate conditions. While quiescent G0 cells are not actively going through a cell cycle, the stationary-phase checkpoint proteins identified in our work likely are still sensing a cellular state and transmitting a signal when problematic conditions are present (possibly ssDNA, as discussed below). This signal may activate DNA repair activities, for example. In addition, it is possible that the checkpoint proteins in stationary-phase cells are establishing a block to progression, as they would in actively growing cells, to prevent the cell from exiting G0 before the problem has been resolved.
Our SGA screen of nonessential checkpoint genes for genes whose deletion destabilized a minisatellite tract inserted into ADE2 revealed that specific components of checkpoint signaling pathways are active in stationary-phase cells. We identified several that are known to function together in well-characterized signaling pathways in actively growing cells. These include MRC1, CSM3, and TOF1, whose gene products form a complex that associates with the replication fork and mediates replication checkpoint signaling (11, 15, 57, 58). Also included are the kinase encoded by TEL1 (18, 19) and PCNA-like clamp complex proteins encoded by DDC1, RAD17, and MEC3, which mediate DNA damage checkpoint signaling (21, 23) and are considered to be upstream of Rad9p and Chk1p (Fig. 1). However, CHK1 and RAD9, which are central components of the intra-S and DNA damage checkpoint pathways (Fig. 1), were not identified as hits in our screen. This surprising result was confirmed by analysis in a separate strain background. DDC1, RAD17, MEC3, and TEL1, while not strong hits from our SGA screen, were verified to have a significantly elevated level of blebbing when also deleted in this background. Thus, while we have demonstrated that some checkpoint proteins mediate DNA stability in stationary-phase cells, our data argue that specific checkpoint pathways, or the interactions between those checkpoint components, are significantly different in stationary-phase cells from those actively growing cells.
A subsequent analysis of essential checkpoint gene mutants reinforced this finding. In actively growing cells, the essential Mec1 protein is upstream of the essential Rad53 protein in the replication checkpoint pathway and in the intra-S and DNA damage pathways (Fig. 1) (9, 46). However, our results suggest that MEC1 and RAD53 do not fall within the same genetic pathway (or at least provide unique contributions to minisatellite stability), implying that Mec1p is not actively signaling a Rad53p checkpoint pathway in stationary-phase cells with respect to minisatellite stability (Fig. 5b and andc).c). This observation is unlike the previously characterized roles of Mec1p in dividing cells and even in oxidative repair-deficient stationary-phase cells, in which Mec1p is required to activate Rad53p by phosphorylation (Fig. 1) (30, 46). Some studies have shown that Rad53p does have a Mec1p-independent role in maintaining genomic stability, in which Rad53p monitors histone levels in order to prevent delays in DNA repair synthesis (59, 60). Our work suggests that Rad53p maintains minisatellite stability in stationary-phase cells independently of Mec1p. Therefore, Rad53p could be part of an otherwise unidentified checkpoint signaling pathway in stationary-phase cells or it could be monitoring histone levels, either of which then influences minisatellite instability.
Pathway analysis of strains bearing mutations in MRC1, CSM3, or TOF1 provides evidence that all three genes act within the same genetic pathway. This result was not unexpected, since the protein products of each gene form a heterotrimeric complex that travels with replication forks (57, 58). Based upon the results from our screen and pathway analysis data, we predict that Mrc1p, Csm3p, and Tof1p form a complex that associates with replication sites in stationary-phase cells. While bulk DNA synthesis does not occur in stationary-phase cells, recent work has demonstrated that discrete areas of DNA replication do occur within the genome of stationary-phase yeast cells (61), presumably at regions of localized DNA repair. Therefore, the Mrc1p/Csm3p/Tof1p complex may associate with replication sites at areas of DNA repair and prevent genomic instability through mediating checkpoint signaling in stationary-phase cells.
To address whether or not minisatellites are stabilized by the Mrc1p checkpoint or replication function during stationary phase, we assessed the level of blebbing in strains bearing either the mrc1AQ allele or the mrc1-C14 allele (Fig. 6a) (13, 15). Our data show that the level of blebbing in a strain bearing the Mrc1p checkpoint loss-of-function allele mrc1AQ is not drastically greater than that of the WT strain. This indicates that even in the absence of proper checkpoint signaling, the minisatellite is still predominantly stable. In contrast, the level of blebbing in a strain bearing the Mrc1p loss-of-replicative-function allele, mrc1-C14, is significantly greater than that in the WT strain. Together, we find that Mrc1p stabilizes minisatellites in stationary-phase cells largely through its replicative function rather than through mediating checkpoint signaling.
In support of this hypothesis, our pathway analysis of MRC1, MEC1, and RAD53 shows that MRC1 does not fall within the same genetic pathway as MEC1 or RAD53 in stationary-phase cells (Fig. 5b and andc).c). Similarly, our data demonstrating that MEC1 and RAD53 do not lie in the same pathway support the possibility that the characterized replication checkpoint, in which Mec1p is recruited to the replication fork and activates Rad53p via Mrc1p (Fig. 1) (11, 13–15), is not participating in stabilizing minisatellites in stationary-phase cells. Therefore, we suggest Mrc1p prevents minisatellite rearrangements in stationary-phase cells by maintaining correct replicative function rather than by promoting checkpoint signaling via a Rad53p signaling pathway (Fig. 6a); we cannot rule out the possibility that Mrc1p is mediating checkpoint signaling through an as yet unidentified pathway, however.
Interestingly, a strain bearing both the mrc1AQ and mrc1-C14 deletions did not result in a level of blebbing equivalent to that of the mrc1Δ deletion strain (Fig. 6a). This could be due to one of the following possibilities: (i) the bulk of the Mrc1 protein is still intact and stabilizes the minisatellite through being present physically; (ii) a secondary backup pathway is activated when Mrc1p function is aberrant, thereby preventing the minisatellite from undergoing further alterations; or (iii) residual checkpoint or replication function remains in either the mrc1AQ or mrc1-C14 mutant. The second possibility could account for the suppression of minisatellite instability displayed by our strain bearing the mrc1AQ allele.
It was previously demonstrated that Mrc1p and the DNA damage checkpoint protein Rad9p are functionally redundant, suggesting that the Rad9p pathway is activated in strains bearing mutant Mrc1p (11, 15). To determine if a Rad9p pathway is initiated in stationary-phase cells bearing checkpoint-defective Mrc1p, we assessed the level of blebbing produced by the mrc1AQ rad9Δ mutant strain. Despite the finding that RAD9 deletion by itself does not appear to affect minisatellite stability, deletion of RAD9 in an mrc1AQ background results in a synergistic increase in the level of blebbing, demonstrating that a secondary Rad9p-dependent signaling pathway is activated in stationary-phase cells to prevent minisatellite instability when Mrc1p checkpoint signaling is compromised.
To determine how many independent pathways regulate minisatellite instability in stationary-phase cells, we constructed strains bearing mutations in MRC1, MEC1, RAD53, or RAD17 and several genes we previously identified as stabilizing the ade2-min3 construct (6–8). Our results reveal that there are three distinct pathways that stabilize minisatellites in yeast cells in stationary phase (Fig. 1b). The first two of these pathways are the MRC1/CSM3/TOF1 pathway and the RAD53/ZRT1/RAD17/RAD27 pathway. Interestingly, PKC1 and END3 appear to act in both pathways. The final pathway consists of MEC1, ZRT1, PKC1, and END3 but does not contain RAD17 or RAD27. These data provide further evidence that MRC1, MEC1, and RAD53 act in separate pathways in stationary-phase cells and suggest that Mec1p is not activated through the presence of an intact clamp as in the established checkpoint pathway. Thus, like Rad53p, Mec1p may not be functioning in a checkpoint-dependent signaling pathway or may be participating in an as yet unidentified pathway in stationary-phase yeast cells.
A common element of all of the mutants with a strong blebbing phenotype is an involvement in ssDNA metabolism; accumulation of ssDNA could act as a template for minisatellite instability in stationary-phase cells. MRC1, CSM3, and TOF1 promote genomic stability by preventing aberrant replication and mediating checkpoint signaling in the presence of ssDNA (13–15, 57). RAD27, ZRT1, END3, RAD53, and PKC1 are also important in preventing the accumulation of ssDNA. Deletion of RAD27 results in unprocessed Okazaki fragments and an increase in ssDNA (62, 63). Strains bearing a deletion of END3 have been shown to produce an increase in reactive oxygen species (ROS) (64). Similarly, mutations in RAD53 or PKC1 result in cells that are susceptible to oxidative stress that can then accumulate ROS (65–67). Previous studies have shown that an increase in ROS in yeast cells results in the accumulation of ssDNA and repetitive DNA instability (64, 68). Our lab recently demonstrated that increased ade2-min3 blebbing in an END3 mutant is correlated with an increase in ROS, and a suppressor mutation that reduces the level of blebbing also reduces the level of ROS in the cells (8). Two models are possible from our results—either the MRC1/CSM3/TOF1, MEC1/ZRT1, and the RAD53/ZRT1/RAD27/RAD17 pathways are independent but acted upon by PKC1 and END3 (as shown in Fig. 1b) or they are branches of a uniform pathway containing PKC1 and END3. In either case, our data are consistent with the idea that the primary lesion leading to minisatellite instability in stationary-phase cells is ssDNA formation and that G0 cells possess multiple methods for dealing with ssDNA.
To address whether or not an accumulation of ssDNA could be an intermediate of minisatellite instability in stationary-phase cells, we quantified the level of blebbing in strains bearing a recombination-defective allele of the ssDNA-binding RPA complex subunit RFA1 (rfa-t11) (39). We found that the rfa1-t11 mutation by itself does not appear to affect minisatellite instability. However, strains bearing the rfa1-t11 allele along with deletions in either MRC1 or ZRT1 displayed a 50% reduction in the level of minisatellite alterations compared to findings for either single mutant. Thus, loss of the recombination-associated function of RFA1 results in a partial suppression of the minisatellite alterations in mrc1Δ or zrt1Δ mutant strains. Based upon our results, we suggest that Rfa1p-initiated recombination mediates a portion of the minisatellite alterations exhibited in strains bearing a deletion of MRC1 or ZRT1. The rfa1-t11 results support our proposal that stationary-phase minisatellite alteration events could result from a failure to repair accumulated ssDNA nicks or gaps.
A few other checkpoint genes exhibited a moderate phenotype in our screens but have not been analyzed further yet, including IPL1, CEP3, and RFC4. IPL1 encodes the central component of the aurora kinase complex that is involved in the spindle checkpoint, along with SLI15, BIR1, and NBL1 (reviewed in reference 69). While the essential genes BIR1 and NBL1 were not present in our set, the SLI15 deletion did not exhibit a blebbing phenotype (Table 3). The role that IPL1 is playing in stationary-phase cells is not apparent, although CEP3 (an essential kinetochore component [43, 44]) may be functioning in a manner similar to that of IPL1. Intriguingly, MEC1 has been implicated in activation of the spindle checkpoint, and this may occur in a RAD53-independent manner (e.g., see references 70 and 71); double mutant analysis with MEC1 and the spindle checkpoint components we have identified may provide insight into this aspect of stationary-phase minisatellite stability. Since RFC4 encodes a component of replication factor C (45), it may be involved in a replication step of DNA repair events or it may interact with the MRC1/CSM3/TOF1 pathway. Given the outcome of our analysis of BUB3, it is possible that any or all of these are false positives.
Finally, we were interested in the state of the stationary-phase cells that were undergoing minisatellite alterations. Stationary-phase cultures of yeast cells consist of a nondividing quiescent cell population and a slow-dividing nonquiescent population of cells (47, 48). To determine if minisatellite alterations occurred within the true quiescent cell population of stationary-phase cells, we constructed double mutants bearing a deletion of ETR1 or POR1 in a checkpoint mutant background, since deletion of ETR1 reduces the reproductive capacity of quiescent cells while deletion of POR1 has the same effect in nonquiescent cells (48). We found that loss of ETR1 in all checkpoint strain backgrounds eliminated blebbing in each of these strains, indicating that the majority of the stationary-phase alterations are occurring in the quiescent cell population. In comparison, deletion of POR1 in the majority of strain backgrounds resulted in only a partial decrease in blebbing. The small effect of POR1 deletion on blebbing in the checkpoint mutants may be explained by the role that checkpoint proteins play in dividing cells, even slowly dividing nonquiescent cells that are ostensibly in G0. Intriguingly, the effect of POR1 deletion was stronger in the mec1-21 and rad53-1 checkpoint mutants, in which the level of blebbing was similar to that of the wild-type strain. This result suggests that minisatellite alterations occurring within a mec1-21 or rad53-1 strain arise in both nondividing quiescent cells and dividing cells, consistent with the color segregation phenotype of each strain in which blebbing and sectoring are present together (Fig. 2b).
In conclusion, our work demonstrates that several checkpoint signaling components may be active during stationary phase and function to prevent the instability of minisatellites. However, the pathways and interactions found in stationary-phase cells differ significantly from those found in actively dividing cells. We suggest that the stationary-phase checkpoint pathways act to prevent the accumulation of ssDNA that leads to genomic instability in stationary-phase cells.
We thank P. Jauert for technical assistance. We also thank D. Clarke, S. Elledge, D. Koepp, R. Kolodner, and L. Symington for yeast strains or plasmids. We are grateful to C. Boone and R. Wright for yeast haploid sets.
This work was funded by a grant from the National Institutes of Health (5RO1-GM072598) to David T. Kirkpatrick.
Published ahead of print 12 November 2012