|Home | About | Journals | Submit | Contact Us | Français|
DNA damage checkpoint and recombinational repair are both important for cell survival of replication stress. Because these two processes influence each other, isolation of their respective contributions is challenging. Research in budding yeast shows that removal of the DNA helicase Mph1 improves survival of cells with defective Smc5/6 complex under replication stress. mph1 is known to reduce the levels of recombination intermediates in smc6 mutants. Here, we show that mph1 also hyperactivates the Mec1 checkpoint. We dissect the effects of recombination regulation and checkpoint hyperactivation by altering the checkpoint circuitry to enhance checkpoint signaling without reducing recombination intermediate levels. We show that these approaches, similar to mph1, lead to better survival of smc6 cells upon transient replication stress, likely by ameliorating replication and chromosomal segregation defects. Unlike mph1, however, they do not suppress smc6 sensitivity to chronic stress. Conversely, reducing the checkpoint response does not impair survival of smc6 mph1 mutants under chronic stress. These results suggest a two-phase model in which smc6 mutant survival upon transient replication stress can be improved by enhancing Mec1 checkpoint signaling, whereas smc6 sensitivity to chronic stress can be overcome by reducing recombination intermediates.
Homologous recombination (HR) facilitates genome duplication under replication stress by repairing DNA strand breaks or single-strand DNA (ssDNA) gaps and restarting stalled replication forks (Aguilera and Gómez-González, 2008 ; Li and Heyer, 2008 ; Branzei and Foiani, 2010 ). During these processes, the strand exchange protein Rad51 coats ssDNA and enables ssDNA pairing with a homologous sequence to template new DNA synthesis. This leads to the formation of HR intermediates, such as D-loop and Holliday junction structures. A number of other proteins also play important roles in HR intermediate metabolism under these situations. In Saccharomyces cerevisiae, these include factors that promote HR intermediate formation or maintenance, such as the DNA helicase Mph1 and the Rad51 paralogue Shu complex (Mankouri et al., 2007 , 2009 ; Chen et al., 2009 ; Choi et al., 2010 ), and those that promote intermediate dissolution and resolution. The main dissolution factor is the STR complex, composed of the Sgs1 helicase, the topoisomerase Top3, and its partner, Rmi1 (Liberi et al., 2005 ; Cejka and Kowalczykowski, 2010 ; Cejka et al., 2010 ; Hickson and Mankouri, 2011 ). In addition, the octameric Smc5/6 complex, composed of Smc5, Smc6, and six other subunits (Nse1, Mms21, and Nse3–6; Zhao and Blobel, 2005 ; Kegel and Sjogren, 2010 ), also contributes to HR intermediate processing. The Smc5/6 complex is essential in budding yeast; like STR deletion mutants, its hypomorphic alleles show increased levels of HR intermediates that can be visualized as X-shaped structures (X-mols) on two-dimensional gel electrophoresis (2D gel; Zhao and Blobel, 2005 ; Branzei et al., 2006 ; Chen et al., 2009 ; Sollier et al., 2009 ; Bermudez-Lopez et al., 2010 ; Chavez et al., 2010 ).
Although the aforementioned proteins have been primarily studied in the recombination context, they also affect other aspects of the replication stress response, particularly the DNA damage checkpoint. DNA structures generated during perturbed replication can be bound by checkpoint sensor proteins, such as the Rad17-Mec3-Ddc1 complex, referred to as 9-1-1 based on its homologues (RAD9-HUS1-RAD1). The 9-1-1 complex and other sensor proteins recruit and activate the apical checkpoint kinase Mec1 in budding yeast (Putnam et al., 2009 ; Branzei and Foiani, 2010 ). Activated Mec1 in turn phosphorylates and activates the main effector kinase, Rad53. Further phosphorylation of a large number of substrates by Mec1 and Rad53 leads to changes promoting replication stress tolerance, such as replication fork stabilization, activation of DNA repair processes, and delayed cell cycle progression (Putnam et al., 2009 ; Branzei and Foiani, 2010 ). Links between the DNA damage checkpoint and HR have been documented. Of most relevance is that proteins involved both in recombination intermediate formation and dissolution (or resolution) influence the DNA damage checkpoint but in an opposite manner. In budding yeast, sgs1 cells are defective in Rad53 activation (Frei and Gasser, 2000 ; Liberi et al., 2005 ; Mankouri et al., 2009 ), and in fission yeast, an smc6 mutant fails to maintain the DNA damage checkpoint (Harvey et al., 2004 ). In contrast, the lack of upstream HR factors, such as Rad51 and Shu, results in increased Rad53 activation, presumably due to increased ssDNA levels (Lee et al., 2003 ; Mankouri et al., 2007 , 2009 ).
Because the foregoing mutants simultaneously affect HR and checkpoint, deconvoluting the mechanism underlying their genetic interactions is difficult. For example, removing Rad51 and the Shu complex improves the tolerance of smc6 and sgs1 cells to replication stress (Shor et al., 2005 ; Mankouri et al., 2007 ; Ball et al., 2009 ; Chen et al., 2009 ; Choi et al., 2010 ). This suppression could be interpreted as rad51 or shu reducing levels of recombination intermediates or X-mols (Mankouri et al., 2007 ; Chen et al., 2009 ; Choi et al., 2010 ). This interpretation would imply that X-mol accumulation is more toxic than the failure to initiate HR. However, because rad51 or shu also increases the checkpoint response, the observed suppression could also be attributed to enhanced DNA damage checkpoint signaling. Thus far, it has been difficult to elucidate how recombination and DNA damage checkpoint separately affect the replication stress tolerance of smc6 and sgs1 mutants. Lack of this information prevents clear interpretation of the genetic observations and impedes our understanding of the physiological consequences of X-mol accumulation.
To address these issues, we examined a mutant allele of budding yeast Smc6, smc6-P4, which contains the K239R mutation. We previously showed that smc6-P4 cells are extremely sensitive to replication stress and display an elevated level of X-mols when replicating in the presence of methyl methanesulfonate (MMS; Chen et al., 2009 ). Both defects are suppressed by the removal of Mph1, Shu, or the proliferating cell nuclear antigen–polyubiquitinating enzyme Mms2, with mph1 having the strongest effect (Chen et al., 2009 ; Choi et al., 2010 ). Here we show that smc6-P4 and mph1 exert opposite effects on the DNA damage checkpoint: mph1 increases it, whereas smc6-P4 decreases it, and the smc6-P4 mph1 double mutant behaves like mph1. To assess the contribution of increased checkpoint response to the replication stress tolerance of smc6-P4, we used two strategies that alter the checkpoint circuitry to enhance the DNA damage checkpoint. Both corrected Rad53 phosphorylation defects in smc6-P4 cells without reducing X-mol levels. They also increased smc6-P4 tolerance to transient, but not chronic, replication stress, whereas mph1 conferred tolerance to both. Furthermore, we reduced the checkpoint response in smc6-P4 mph1 double mutants by removing the checkpoint sensor protein Mec3 and found that mph1 can still suppress the sensitivity of smc6-P4 cells to chronic replication stress. These results suggest that, whereas enhanced DNA damage checkpoint promotes tolerance to transient replication stress, X-mol removal is required for the survival of smc6 mutants under persistent exposure to such stress.
The mph1 mutation strongly suppresses a number of smc6-mutant defects, notably conferring three orders of magnitude more resistance to the replication-blocking agent MMS (Chen et al., 2009 ). Although our previously reported decrease in X-mol levels might be one cause (Chen et al., 2009 ) of the strong suppression, additional mechanisms might also contribute. Because HR mutants influence the DNA damage checkpoint response, we examined whether mph1 and smc6-P4 also alter this important replication stress tolerance mechanism and, if so, how this is related to the observed suppression.
We first examined how mph1 and smc6-P4 affect Rad53 phosphorylation, a standard readout of the activation of Rad53 and DNA damage checkpoint. Rad53 phosphorylation is indicated by the appearance of a higher–molecular weight band on immunoblots and can be seen in wild-type cells after 0.03% MMS treatment (Figure 1A). After the same treatment, mph1 resulted in a complete upward shift of Rad53, a characteristic feature of Rad53 hyperphosphorylation (Figure 1A). In contrast, smc6-P4 cells exhibited less Rad53 phosphorylation, as the phosphorylated Rad53 band (Rad53-P) is weaker in intensity than that of wild-type cells (Figure 1A). smc6-P4 mph1 double mutants behaved similarly to mph1, indicating that mph1 results in Rad53 hyperphosphorylation in both wild-type and smc6-P4 cells.
To determine whether the altered Rad53 phosphorylation levels in smc6 and mph1 mutants reflect a change in the initial activation or maintenance of Rad53 modification, we performed time course experiments in which G1-synchronized cells were released into MMS-containing media (Figure 1B). In wild-type cells, the Rad53-P band appeared at 20 min postrelease, peaked at 40 min, and diminished at 180 min, when most cells had finished replication, as judged by flow cytometry (fluorescence-activated cell sorting [FACS]; Figure 1, C and andD).D). In smc6-P4 cells, Rad53-P band was also visible 20 min postrelease, but the magnitude of phosphorylation did not reach the maximum level seen in wild-type (WT) cells (Figure 1C). This difference between the two strains could not be caused by cell-cycle progression changes, since their FACS profiles were similar (Figure 1D). We conclude that smc6-P4 cells are defective in maximal Rad53 phosphorylation and, by extension, its activation.
In mph1 cells, Rad53 phosphorylation appeared to be stronger than wild type at 20 min and reached the maximum level at 40 min postrelease. Of importance, mph1 cells failed to attenuate Rad53 phosphorylation even at 180 min (Figure 1C). Consistent with persistent Rad53 phosphorylation, mph1 cells also exhibited a delay in S-phase progression compared with wild-type cells (Figure 1, C and D). mph1 smc6-P4 cells behaved similarly to mph1 cells in terms of Rad53 phosphorylation level and S-phase progression (Figure 1, C and D). Therefore data from both asynchronous and time course experiments show that smc6-P4 and mph1 have opposite effects on Rad53 phosphorylation and that mph1 is epistatic to smc6-P4 for this phenotype.
Because mutations of key residues in the Mph1 helicase domain (mph1-hd) that abolish its helicase activity suppress smc6-P4’s MMS sensitivity and X-mol accumulation similarly to mph1 (Chen et al., 2009 ), we asked whether mph1-hd also affects the DNA damage checkpoint. We found that mph1-hd behavior resembled that of mph1 in both asynchronous and synchronized experiments. mph1-hd cells showed Rad53 hyperphosphorylation and slower S-phase progression, regardless of Smc6 status (Figure 1A and Supplemental Figure S1). Thus the lack of Mph1 helicase activity accounts for the observed effects on the DNA damage checkpoint. These results raise the possibility that the mounting of a more robust DNA damage checkpoint response is partly responsible for mph1 suppression of smc6-P4 MMS sensitivity. This effect could serve to stabilize stalled replication forks and provide more time for repairing DNA lesions.
In budding yeast, Mec1 is the main checkpoint kinase that controls Rad53 activation and S-phase progression, and its homologue, Tel1, makes minor contributions (Putnam et al., 2009 ; Branzei and Foiani, 2010 ). To assess whether the observed mph1 effect on the DNA damage checkpoint is due to a change in the Mec1-dependent pathway, we examined mec1 cells containing sml1, a suppressor of mec1 lethality that does not affect checkpoint function (Zhao et al., 1998 ). As shown in Figure 2A and consistent with the literature, mec1 cells contain unphosphorylated Rad53 and progress through S phase more rapidly than wild-type cells after MMS treatment. The removal of Mec1 in mph1 or mph1 smc6-P4 cells largely abolished Rad53 phosphorylation and the observed S-phase delay (Figures 2, A and B). We conclude that the increased Rad53 phosphorylation and delayed replication seen in both mph1 and mph1 smc6-P4 cells are dependent on Mec1-mediated checkpoint activities. Delayed replication in wild-type cells under genotoxic stress is due to Mec1-mediated inhibition of late replication origin firing (Santocanale and Diffley, 1998 ; Shirahige et al., 1998 ). Thus this is more likely accountable for the S-phase delay in mph1 mutants than an inability to repair damaged DNA.
Because smc6-P4 mph1 cells exhibit higher Rad53 phosphorylation levels than smc6-P4 cells, we asked whether enhancing the DNA damage checkpoint alone could improve the replication stress tolerance of smc6-P4 cells. To this end, we used two different approaches that directly alter checkpoint circuitry. The first approach used the TEL1-hy909 gain-of-function allele, which results in elevated Tel1 kinase activity and Rad53 hyperphosphorylation (Baldo et al., 2008 ). We confirmed Rad53 hyperphosphorylation in TEL1-hy909 cells after MMS treatment in a time course experiment (Supplemental Figure S2A). The degree of Rad53 hyperphosphorylation caused by TEL1-hy909 is similar to that seen with mph1, although only the latter slows S-phase progression (Supplemental Figure S2A). As reported previously, TEL1-hy909 greatly improved the survival of mec1 cells during chronic exposure to MMS (Supplemental Figure S2A; Baldo et al., 2008 ). These results together indicate that TEL1-hy909 augments a critical aspect of the Mec1-mediated checkpoint response to increase viability in MMS-containing media. That TEL1-hy909 did not significantly affect late replication origin firing as reflected by FACS analysis is consistent with the notion that this aspect of checkpoint control is not essential for cell survival upon replication stress (Tercero et al., 2003 ).
After confirming that TEL1-hy909 can hyperactivate checkpoint under our experimental conditions, we examined its effect on the checkpoint response, recombination intermediate levels, and MMS sensitivity of smc6-P4 cells in a time course experiment. First, we found that TEL1-hy909 increased the level of Rad53 phosphorylation in smc6-P4 cells, albeit less strongly than mph1 (Figure 3A). TEL1-hy909 did not affect S-phase progression as seen in wild-type cells (Figure 3A). Consistent with its observed Rad53 hyperphosphorylation and like mph1, TEL1-hy909 resulted in a greater degree of degradation of the ribonucleotide reductase inhibitor Sml1, another frequently used readout of DNA damage checkpoint function (Figure 3A and Supplemental Figure S3A; Zhao et al., 2001 ). Second, TEL1-hy909 did not reduce X-mol levels in smc6-P4 cells throughout the time course, suggesting that increased checkpoint response does not grossly decrease HR intermediate levels (Figure 3B). Third, TEL1-hy909 improved the viability of smc6-P4 cells to a similar degree as mph1 when cells were withdrawn at each time point to assess survival (Figure 3C). Thus hyperactivation of the DNA damage checkpoint alone without reducing X-mol levels is sufficient for improving the tolerance of smc6-P4 cells to transient replication stress.
Next we examined how TEL1-hy909 affects smc6-P4 cell survival during chronic MMS exposure. We found that, unlike mph1, TEL1-hy909 did not improve the viability of smc6-P4 cells during chronic exposure to MMS, even at a concentration lower than the one at which it suppresses the lethality of mec1 (Figure 3D). Taken together, the results indicate that increasing Rad53 phosphorylation levels by TEL1-hy909 promotes the survival of smc6-P4 cells after transient but not chronic exposure to MMS.
To understand how DNA damage checkpoint hyperactivation improves smc6-P4 tolerance to transient replication stress, we examined both chromosomal replication and segregation. Because smc6-P4 cells began to lose viability in S phase when treated with MMS, we first examined chromosomal replication using pulsed-field gel electrophoresis (PFGE). We treated G1 cells with a pulse of MMS and then released them into the cell cycle in normal media (Figure 4A). Based on the criterion that only fully replicated chromosomes can enter the gel, wild-type cells appeared to complete replication at around 60 min (Figure 4, A–C). In contrast, smc6-P4 cells failed to finish chromosomal replication even at 240 min postrelease. Introduction of TEL1-hy909 increased extent of replication completion in smc6-P4 cells (Figure 4, A–C). These results suggest that enhanced checkpoint response can improve replication capacity in smc6-P4 cells after transient MMS treatment.
We also assessed chromosome segregation at 240 min postrelease in the foregoing experiment. Compared with wild type, smc6-P4 strains exhibited fewer normal anaphase and telophase cells but more large-budded cells with missegregated or mispositioned nuclei (p < 0.05; Figure 4, D and E). TEL1-hy909 increased the former populations and decreased the latter (p < 0.05; Figure 4, D and E). One interpretation is that correction of checkpoint defect in smc6-P4 cells by TEL1-hy909 is sufficient to improve chromosome segregation, leading to better survival.
We also used another strategy to increase checkpoint response, on the basis of the observation that induced proximity of the DNA damage checkpoint sensor protein Ddc1 and the Mec1 binding partner Ddc2 is sufficient to activate checkpoint (Bonilla et al., 2008 ). In this system, Ddc1 and Ddc2 are fused to LacI–green fluorescent protein (GFP) modules, and their targeting to chromosomal LacO arrays results in Rad53 phosphorylation and checkpoint activation even without DNA damage in S and G2/M phases. This likely occurs via Mec1-Ddc2 recruitment to chromatin by the 9-1-1 complex (Bonilla et al., 2008 ). We tested how this system affects DNA damage checkpoint responses, X-mol levels, and replication stress tolerance in smc6-P4 cells.
The expression of Ddc1 and Ddc2 fusion constructs was induced by a pulse of galactose in G1-arrested cells before the cells were released into MMS-containing glucose media. Because this system activates Rad53 even without DNA-damaging agents, we used a lower concentration of MMS (0.005%). Time course experiments show that this induction led to Rad53 hyperphosphorylation in both WT and smc6-P4 backgrounds (Figure 5A; compare the lanes with and without expression of the constructs). This effect on Rad53 phosphorylation is similar to that induced by mph1 (Figure 5A). Consistent with this, lower levels of the Sml1 protein were detected when Ddc1 and Ddc2 fusions were induced in both WT and smc6-P4 cells (Figure 5A and Supplemental Figure S3B). Examination of replication intermediates by 2D gel found that the engineered Ddc1-Ddc2 juxtaposition did not alter X-mol levels in smc6-P4 cells (Figure 5B). Finally, this system was as effective as mph1 in improving smc6-P4 survival after transient exposure to MMS (Figure 5B). These results are consistent with those obtained for TEL1-hy909; taken together, they strongly suggest that increasing Rad53 activation in smc6-P4 cells is sufficient for increasing their resistance to transient replication stress. Because prolonged expression of the Ddc1 and Ddc2 fusions can impair replication (Bonilla et al., 2008 ), this system cannot be used to evaluate responses to chronic MMS exposure.
The results obtained so far suggest that correcting Rad53 phosphorylation in smc6-P4 cells by TEL1-hy909 or the juxtaposition of Ddc1 and Ddc2 can increase cellular tolerance to transient but not chronic MMS exposure. To determine directly whether mph1-mediated checkpoint hyperactivation contributes to the viability of smc6-P4 mph1 cells upon chronic MMS treatment, we reduced the checkpoint response in this double mutant by removing the checkpoint sensor protein Mec3.
Unlike mec1, which exhibited strong synthetic sickness with mph1 in the presence of MMS (Supplemental Figure S2C), mec3 did not appear to affect mph1 survival on MMS-containing media (Supplemental Figure S4). Thus it is possible to determine whether reducing checkpoint response via mec3 affects the suppression of smc6-P4 cells by mph1. As shown in Figure 6A, mec3 diminished Rad53 phosphorylation in mph1, smc6-P4, and smc6-P4 mph1 cells. Quantification showed that mec3 effectively reduced Rad53 phosphorylation levels in smc6-P4 mph1 cells such that the triple mutant displays a similar degree of Rad53 phosphorylation as smc6-P4 cells (Figure 6B). Of importance, mec3 did not alter X-mol levels significantly in either smc6-P4 or smc6-P4 mph1 cells in time course experiments when G1 cells were released into MMS-containing media (Figure 6C). Finally, we found that mph1 still conferred robust suppression of smc6-P4 MMS sensitivity in the absence of Mec3 (Figure 6D). A similar effect was also seen with removal of the Mec3 loader, Rad24 (Figure 6D and Supplemental Figure S4). Taken together, these results suggest that improvement of smc6-P4 survival of chronic MMS exposure by mph1 is more a consequence of reduced X-mol levels than of enhanced checkpoint response.
Unraveling the genetic relationships between some recombinational repair proteins is complicated by their additional effects on the DNA damage checkpoint. We previously showed that removing the DNA helicase Mph1 or mutating its key enzymatic residues reduces X-mol levels and strongly suppresses the MMS sensitivity of smc6-P4 cells (Chen et al., 2009 ). Our new data show that smc6-P4 and mph1 (or mph1-hd) have opposing effects on the Mec1 checkpoint. smc6-P4 cells failed to phosphorylate Rad53 to wild-type levels upon MMS treatment, whereas mph1 (or mph1-hd) exhibited persistent Rad53 phosphorylation (Figure 1 and Supplemental Figure S1). Although the underlying reason for these effects is not entirely clear, it likely pertains to alteration of ssDNA levels or replisome stability as suggested for similar situations (Cobb et al., 2005, Harvey et al., 2004 ; Alabert et al., 2009 ; Yeung et al., 2011). Besides Rad53 hyperphosphorylation, mph1 (or mph1-hd) also resulted in a strong Mec1-dependent S-phase delay, and smc6-P4 mph1 behaved like mph1 (Figure 2). These novel observations of DNA damage checkpoint effects raised the question of whether an enhanced checkpoint response is sufficient to improve smc6-P4 survival upon replication stress.
To address this question, we used two different approaches to augment the DNA damage checkpoint response without affecting X-mol levels. Both the hyperactive TEL1-hy909 allele and the induced proximity of Ddc1 and Ddc2 increased Rad53 phosphorylation levels in cells with normal and defective Smc6 upon MMS treatment, with a stronger effect seen in the latter (Figures 3A and and5A5A and Supplemental Figure S2A). Both resulted in a greater degree of Sml1 degradation, consistent with an enhanced checkpoint response (Figures 3A and and5A5A and Supplemental Figure S3). Neither TEL1-hy909 nor the Ddc1-Ddc2 system lowered the level of X-mols in smc6-P4 cells, suggesting that checkpoint hyperactivation does not affect HR intermediate levels and that these alleles can be used to isolate checkpoint- from HR-dependent effects (Figures 3B and and5B).5B). We found that both strategies improved the replication stress tolerance of smc6-P4 cells during a time course of 2-h exposure to MMS (Figures 3C and and5C).5C). We note that TEL1-hy909 also improved the checkpoint response and survival of another smc6 mutant, smc6-56, after exposure to transient replication stress (Supplemental Figure S5, A and B). Thus improved smc6 mutant resistance to acute replication stress can be achieved solely by DNA damage checkpoint hyperactivation. Our results further show that an enhanced checkpoint response can improve replication capacity and chromosomal segregation in smc6-P4 cells (Figure 4, A–D). This is presumably achieved by promoting replication fork stability and allowing more time for nuclear segregation. However, other effects, such as those involving previously reported effects on kinetochore and spindle functions (Yong-Gonzales et al., 2012 ), may also contribute.
Although TEL1-hy909 suppressed smc6-P4 sensitivity during transient MMS exposure to a similar extent as mph1, only mph1 promoted the survival of smc6-P4 cells during chronic MMS treatment (Figure 3D). This argues that the observed suppression by mph1 involves more than just checkpoint hyperactivation. This idea is further supported by the observation that mph1 still confers suppression to smc6-P4 cells under chronic and transient MMS treatment upon MEC3 deletion, which reduced the checkpoint response without affecting X-mol level (Figure 6, A–D, and data not shown). Thus the observed mph1 suppression is at least partly due to a reduction in recombination intermediate levels. We note that a recent study reports the ability of mec3 to partially reduce X-mol levels in sgs1 cells (Karras et al., 2013 ), whereas we detected no such effect in smc6-P4 cells (Figure 6C). This difference is consistent with previously noted differences in the effect of mph1 on X-mol levels in smc6-P4 versus sgs1 cells (Chen et al., 2009 ; Mankouri et al., 2009 ), thus arguing for differential pathway involvement of the Smc5/6 complex versus Sgs1, in addition to their common functions.
On the basis of our results, we propose a two-phase model to explain the severe sensitivity of smc6 mutant cells to replication stress (Figure 6E). When replication forks are stalled due to transient stress, a strong DNA damage checkpoint response is advantageous to smc6 mutants. On chronic exposure to replication stress, however, preventing HR intermediate accumulation becomes the dominant factor for mutant cell survival. In this model, the checkpoint-related and X-mol regulation functions of the Smc5/6 complex are separable. This model may also be applicable to other mutants, such as sgs1 and esc2, as they also exhibit X-mol accumulation and checkpoint defects and their MMS sensitivity is suppressed by the removal of recombination factors such as Rad51 and Shu (Liberi et al., 2005 ; Mankouri et al., 2009 ; Sollier et al., 2009 ; Choi et al., 2010 ). The tools used here to dissect the contributions of checkpoint hyperactivation and recombination may be useful for evaluating these cases as well. Our observation that neither hyperactivation nor reduction of checkpoint in smc6-P4 cells affected HR intermediate levels suggests that checkpoint does not affect at least one branch of recombination-mediated damage bypass. This extends previous observations that checkpoint does not inhibit all modes of recombinational repair under replication stress, although it hinders those at chromosomal breaks, as measured by Rad52 foci levels (Lisby et al., 2004 ; Alabert et al., 2009 ; Barlow and Rothstein, 2009 ). In addition, our findings may be related to those in higher eukaryotes, in which the regulation of HR products is important for prolonged but not transient exposure to replication stress (Petermann et al., 2010 ). Compounded, these studies begin to unravel the complex interplay between checkpoint and recombinational repair. Further investigation into the underlying mechanisms of this interplay will provide insight into how these two important genotoxic tolerance mechanisms are coordinated at the molecular level.
The yeast strains used in this study are listed in Table 1 and Supplemental Table 1. They are derivatives of W1588-4C, a RAD5 derivative of W303 (MATa ade2-1 can1-100 ura3-1 his3-11,15 leu2-3112 trp1-1 rad5-535; Thomas and Rothstein, 1989 ). Only one strain for each genotype is listed, but at least two independent spore clones of each genotype were used in each of the experiments. Standard yeast protocols were used for strain construction, growth, and medium preparation. The construction of smc6-P4 and smc6-56 strains was described previously (Chen et al., 2009 ).
Cell synchronization was performed by adding α factor (Memorial Sloan-Kettering Cancer Center Proteomics Core) to cells growing in log phase to a final concentration of 5 μg/ml for ~2 h and evaluating the percentage of unbudded cells in the culture. Galactose induction was carried out as previously described (Bonilla et al., 2008 ). In brief, cells were first arrested with α factor for 2 h, and then galactose was added for 2 h in the presence of α factor. The release from α factor was performed in the presence of MMS at a final concentration of 0.03 or 0.005% as indicated. Cell cycle progression was analyzed by FACS as performed previously (Zhao and Rothstein, 2002 ). One representative result is presented for each genotype.
Spot assays for detecting DNA damage sensitivity were carried out as described previously (Chen et al., 2009 ). Briefly, log-phase cultures were serially diluted 10-fold and spotted onto agar plates containing yeast extract/peptone/dextrose media with the addition of the indicated doses of MMS. Plates were incubated at 30°C and photographed after at least 48 h. For killing curves, cultures challenged by MMS were taken at intervals, sonicated, and serially diluted before plating. An equal volume of 10% sodium thiosulfate was used to quench the effect of MMS in the sample before serial dilutions. The percentage of viable colonies was calculated by dividing the number of colonies by the number of cells plated based on the optical density readings of a spectrophotometer (Biomate 3; Thermo Scientific, Waltham, MA). Because the plating efficiency of each genotype varies in normal growth conditions, the viability of each genotype in MMS at each time point was determined by normalizing the percentage of the viable colonies to its plating efficiency. Unpaired Student's t test was used for statistical analysis.
The 2D gel electrophoresis was carried out and X-mols were quantified as described (Chen et al., 2009 ). To assay Rad53 phosphorylation and Sml1 protein levels, the trichloroacetic acid protein extraction method was used as originally described (Foiani et al., 1994 ). The extracts were separated on standard SDS–PAGE gels and Western blotted, followed by probing with anti-Flag (Sigma-Aldrich, St. Louis, MO) or anti-hemagglutinin (Memorial Sloan-Kettering Cancer Center Monoclonal Antibody Core Facility) antibodies to detect Rad53, anti-Sml1 antibody to detect Sml1, and YL1/2 antibody (AbD Serotec, Raleigh, NC) to detect tubulin. The abundance of a protein was quantified by measuring the intensity of its band using Image Gauge (Fujifilm, Tokyo, Japan). The percentage of Rad53 phosphorylation was calculated using the signal of Rad53-P divided by total Rad53 signal. At least two independent spore clones per genotype were examined for each genotype, and the representative results are shown.
Chromosome plugs were prepared and PFGE was performed as previously described (Cremona et al., 2012 ). For microscopy, cells were fixed by addition of formaldehyde to a final concentration of 3.7% in the culture for 10 min, followed by washing with 0.1 M potassium phosphate, pH 8.1. Cells were then resuspended in a buffer of 1.2 M sorbitol and 0.1 M potassium phosphate, pH 8.1, and aliquots were stained with 4 μg/ml of Hoechst 33258 dye and processed for microscopy as previously described (Yong-Gonzales et al., 2012 ). The exposure times used for Tub1-GFP and Hoechst were 2 and 0.2 s respectively. All imaging was captured on an Axio Imager microscope (Carl Zeiss, Jena, Germany) with a 100× objective lens (numerical aperture 1.4). From 8 to 10 Z-sections with a 0.5-μm step size were taken to cover the whole yeast cell. Unpaired Student's t test was used for statistical analysis.
We thank Maria Longhese (Università degli Studi di Milano-Bicocca, Milan, Italy) for the TEL1-hy909 allele and David Toczyski (University of California at San Francisco, San Francisco, CA) for the strain containing the LacI-GFP fusions of Ddc1 and Ddc2; Koyi Choi for her help in generating some of the strains; and members of the Zhao lab for comments on the manuscript. This work was supported by National Institute of General Medical Science Grant R01 GM080670, American Cancer Society Research Scholar Grant RSG-12-013-01-CCG, and a Lymphoma Society Scholar Award to X.Z. and by Associazione Italiana per la Ricerca sul Cancro Grant IG 10637 and European Research Council Grant 242928 to D.B.
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E12-11-0836) on June 19, 2013.
*Present address: Department of Biochemistry and Molecular Pharmacology, New York University School of Medicine, New York, NY 10016.