Skip to main content
NCBI home page
Genetics logoLink to Genetics
. 2008 Aug;179(4):1835–1844. doi: 10.1534/genetics.108.089979

Interplay of DNA Repair Pathways Controls Methylation Damage Toxicity in Saccharomyces cerevisiae

Petr Cejka *, Josef Jiricny *,†,1
PMCID: PMC2516062  PMID: 18579505

Abstract

Methylating agents of SN1 type are widely used in cancer chemotherapy, but their mode of action is poorly understood. In particular, it is unclear how the primary cytotoxic lesion, O6-methylguanine (MeG), causes cell death. One hypothesis stipulates that binding of mismatch repair (MMR) proteins to MeG/T mispairs arising during DNA replication triggers cell-cycle arrest and cell death. An alternative hypothesis posits that MeG cytotoxicity is linked to futile processing of MeG-containing base pairs by the MMR system. In this study, we provide compelling genetic evidence in support of the latter hypothesis. Treatment of 4644 deletion mutants of Saccharomyces cerevisiae with the prototypic SN1-type methylating agent N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) identified MMR as the only pathway that sensitizes cells to MNNG. In contrast, homologous recombination (HR), postreplicative repair, DNA helicases, and chromatin maintenance factors protect yeast cells against the cytotoxicity of this chemical. Notably, DNA damage signaling proteins played a protective rather than sensitizing role in the MNNG response. Taken together, this evidence demonstrates that MeG-containing lesions in yeast must be processed to be cytotoxic.

SINCE their discovery at the end of World War II, SN1-type alkylating agents have been in constant use in cancer therapy. Unfortunately, the response to these chemotherapeutics is variable, primarily due to resistance, both endogenous and acquired. Because these chemicals are also mutagens, their clinical use should be restricted to patients who are likely to respond. However, to predict their efficacy, it is imperative that the mode of action of these drugs be fully understood. During the past 2 decades, two important pieces of experimental evidence came to light. First, it was shown that the primary cytotoxic lesion is O6-methylguanine (MeG). This evidence came initially from studies with Escherichia coli, which showed that the product of the ada gene encodes a methyltransferase that detoxifies MeG by removing the methyl group to regenerate unmodified guanine. Orthologs of this protein, termed methylguanine methyltransferase (MGMT) or alkylguanine methyltransferase (AGT), are found in all living organisms and it can be generally stated that the cytotoxicity of methylating agents of the SN1 type, such as N-methyl-N′-nitro-N-nitrosoguanidine (MNNG), which generate ∼5–10% of MeG residues (Beranek 1990), is inversely proportional to the levels of MGMT in the treated cells. In contrast, the toxicity of SN2-type agents such as methyl methanesulphonate (MMS), which give rise almost exclusively to bases modified on the ring nitrogens, is largely independent of MGMT.

The second piece of evidence showed that the sensitivity of ada− dam− E. coli to N-methyl-N-nitrosourea (MNU), another SN1-type alkylating agent, could be rescued by inactivation of mismatch repair (Karran and Marinus 1982). This phenomenon implicated the mismatch repair (MMR) system in the processing of MeG. Unfortunately, this genetic evidence yielded no mechanistic insight as to the type of damage that was addressed by MMR or how it led to cell death. We recently showed that in MGMT-deficient, MMR-proficient mammalian cells, the cell-cycle arrest induced by clinically relevant MNNG concentrations is activated only after the second S phase (Stojic et al. 2004a), which suggested that MMR-dependent MeG processing resulted in the generation of secondary, cytotoxic lesions. We postulated that these lesions could be intermediates of DNA recombination, given that the presence of MeG in DNA of different organisms was shown to induce recombination (Ryttman and Zetterberg 1976; Hastings 1984; Zhang et al. 1996; Nowosielska et al. 2006) and that the toxicity of SN1-type methylating agents is controlled not only by the MGMT levels and MMR status of the cells, but also by the efficiency of homologous recombination (HR) (Nowosielska et al. 2006; Tsaryk et al. 2006). Indeed, we could show that the high efficiency of recombination in yeast cells apparently masks their MMR-dependent response to methylating agents: a recombination-deficient rad52 strain of Saccharomyces cerevisiae could be shown to be hypersensitive to killing by MNNG, but its sensitivity could be rescued by mutations in the MMR genes MSH2 or MLH1 (Cejka et al. 2005).

The above experimental evidence strongly argues that activation of the cell-cycle arrest induced by methylating agents requires processing of MeG-containing DNA. However, it was recently suggested that solely the recognition of MeG-containing base pairs by MMR proteins is sufficient to induce cell death by activating DNA signaling pathways and apoptosis (Lin et al. 2004; Yang et al. 2004; Yoshioka et al. 2006). In an attempt to resolve this dilemma and to throw some light on this clinically important problem, we carried out high-throughput genetic screens in S. cerevisiae. In the first series of experiments, designed to test whether factors other than Msh2 and Mlh1 mediate the toxicity of methylating agents, we crossed 4644 deletion mutants with an mgt1 rad52 strain and tested for resistance to MNNG. In the second screen, we searched for proteins involved in the rescue of intermediates generated by MMR during the processing of MeG-containing DNA. To this end, we crossed the MGMT-deficient mgt1 strain of S. cerevisiae with the deletion library and screened for sensitivity to MNNG. While we failed to find experimental support for the direct signaling hypothesis, we provide compelling genetic evidence that MeG residues in yeast DNA must be processed to be cytotoxic.

MATERIALS AND METHODS

Yeast media and construction of strains:

Yeast media were prepared as described previously (Cejka et al. 2005). Synthetic defined (SD) medium lacking arginine supplemented with canavanine (50 μg/ml, Sigma) was used for the selection of can1 cells. Disruptions of the genes of interest were performed with KANMX, NAT1, and URA3 replacement cassettes, with specifically designed primers (sequences available on request) and pUG6, pAG35, and pUG72 plasmids used as templates for PCR, as described (Guldener et al. 1996). The transformations were performed by the lithium acetate method. The genotypes of all strains were verified by PCR and/or Southern blotting. The ordered arrays of gene deletion mutants were constructed as described (Tong et al. 2001). We succeeded in obtaining 90–95% of all double/triple mutants; the remaining 5–10% failed to produce multiple mutants, principally due to synthetic lethality, extremely slow growth, and/or problems with mating and sporulation. Close linkage did not represent a significant problem, since we failed to obtain only ∼0.15% of strains due to the short distance between the bait and the respective mutation.

Spot tests:

The screens for MNNG sensitivity or resistance were performed manually in 96-well or 96-colonies-per-plate format. Because of the short half-life of MNNG in aqueous solution, all treatments were done in liquid cultures rather than on plates. Briefly, the cells were inoculated into YPD medium and cultivated overnight to early stationary phase. They were then diluted 40-fold and cultivated for 90 min in YPD medium, and subsequently treated or mock-treated two times for 1 hr with 1.5 μm MNNG (Sigma) in case of the SLP1 array and 0.5 μm MNNG in case of the SLP19 array. After 2 hr in the presence of MNNG, the cells were spotted on YPD medium. One drop of 3 μl corresponded to about 105 cells. The plates were then photographed and evaluated qualitatively by comparing treated and mock-treated cultures after 1 day of cultivation at 30° (supplemental Figure 1B). The assay selected for mutants either that were killed or whose growth was strongly inhibited by MNNG treatment. Prospective resistant or sensitive mutants were collected and reordered into 96-well format. More detailed quantitative analysis was carried out with these mutants. The spot tests were done similarly, except that the cultures were serially diluted prior to plating and the results were assayed after 2 days of cultivation to score for the number of colonies. The data sets presented in Table 2 were obtained in the following manner. The number of colonies from mock-treated culture (NM) was divided by the number of colonies from treated culture (NT). The resulting value (NM/NT) was in responding strains >1 and represented the killing induced by MNNG. Finally, to compare the effect of the tested mutations in various genetic backgrounds, these relative values were divided by those obtained for the reference strain, indicated at the top of each column in Table 2. For example, the data for the rad52 mutation were calculated as (NM/NT)MGT1rad52/(NM/NT)MGT1 (first column); (NM/NT)mgt1rad52/(NM/NT)mgt1 (second column), and (NM/NT)rad52msh2mgt1/(NM/NT)msh2mgt1 (third column). The results indicate the average of at least three independent experiments. Mutants were included in Table 2 when their MNNG sensitivity was at least twofold higher than that of the relevant reference strains. The sole exceptions were the cac1, cac2, and cac3 mutants, which showed a dramatic growth arrest upon MNNG treatment and were thus picked up in the original screen, but which showed no difference in overall survival when scored quantitatively (see text for details).

TABLE 2.

Mutations identified in the screen for MNNG sensitivity in SLP1 (mgt1) background

3 μm MNNG
10 μm MNNG
MGT1 mgt1 msh2 mgt1 MGT1 mgt1 msh2 mgt1 Notes
Homologous recombination
rad52 1.3 55.8 2.3 4.1 >20 10.0
rad51 1.2 10.7 1.8 1.7 14.3 11.5
rad57 1.5 4.8 1.6 1.8 13.5 4.2
rad55 1.4 5.1 2.0 1.7 29.8 17.4
rad59 1.1 2.2 1.2 1.5 6.6 2.2 a
rad54 1.6 6.5 1.5 1.5 24.1 10.0
mms4 0.8 3.1 2.1 2.0 11.9 8.2
mus81 1.0 2.4 2.3 2.1 8.1 12.7
Mms22
mms22 1.0 21.8 1.0 2.7 >20 >20
rtt107 1.3 1.4 0.7 1.1 3.6 0.7
rtt101 0.7 3.2 1.6 1.1 4.9 4.7
Postreplicative repair
pol32 0.7 3.5 ND 0.9 >20 ND a
srs2 1.2 3.7 1.8 1.2 5.4 2.9
rad5 5.5 21.2 17.4 >20 >20 42.5
ubc13 1.0 1.5 1.3 1.4 2.0 2.3
mms2 1.2 2.0 0.9 1.0 2.0 0.8
DNA damage signaling, checkpoint
rad50 1.8 56.6 9.4 13.6 >20 >20
xrs2 2.0 107.3 8.7 13.8 >20 >20
rad9 0.9 2.7 1.3 2.1 6.8 5.0 a
dun1 1.3 1.0 1.3 1.1 3.5 1.2 a
elg1 1.4 8.9 1.4 1.2 >20 6.3
ctf18 1.3 2.2 2.0 0.9 8.6 6.1
rad24 1.0 1.7 1.6 1.3 3.0 7.2
slx4 1.1 1.2 1.0 0.9 2.8 1.0 a
ctf8 1.4 2.2 2.4 1.4 6.1 10.6
dcc1 1.1 1.9 0.9 1.4 1.6 4.9
ddc1 0.9 2.8 1.4 1.8 3.4 2.7
Sgs1 and associated proteins
sgs1 2.1 3.5 12.0 2.4 4.6 >20
rmi1 1.7 5.5 3.8 2.1 >20 7.1
top3 1.2 10.2 7.0 2.4 >20 24.2
Other repair factors
doa1 0.8 1.5 0.9 0.9 3.7 1.5 a
met18 0.7 2.7 2.1 0.7 22.1 11.1 a
rad27 1.1 2.9 1.3 4.0 18.5 >20
mms1 1.3 3.8 2.2 1.9 >20 14.3 a
Chromatin maintenance
rtt109 1.0 7.4 1.0 1.2 >20 0.3
ctf4 0.9 6.5 2.8 1.6 >20 7.8 a
asf1 1.5 13.4 2.0 3.5 6.9 0.7
cac2 1.0 1.1 0.9 0.9 1.4 1.4 b
cac1 0.9 1.5 1.1 0.7 0.7 1.0 b
cac3 1.4 1.4 1.0 1.2 1.7 0.8 b
RNA and protein metabolism, mitosis, meiosis, other
cse2 0.6 3.1 1.4 1.0 7.6 0.4 a
cik1 0.9 2.5 3.5 1.4 2.5 8.8 a
spt21 1.2 2.1 1.7 1.1 8.5 2.1
lsm1 1.0 2.1 ND 0.8 5.2 ND a
lsm6 1.0 1.5 1.4 0.8 2.0 3.1 a
med1 0.9 1.4 1.1 0.9 2.7 2.4 a
ubp6 1.0 1.3 1.0 1.0 2.8 2.1 a
cdc50 0.8 3.8 1.1 0.7 >20 2.9
erg3 0.7 1.7 3.2 0.6 >20 >20 a
ybr094w 1.0 2.5 1.3 3.3 3.9 9.7
sro9 1.0 1.3 ND 0.8 4.6 ND c
rpl2b 0.5 1.7 1.4 1.0 2.1 1.6 c
ume1 1.0 1.4 1.2 0.9 2.3 0.9 c
rtg1 0.5 1.6 1.9 0.6 19.8 5.9 c
uba4 1.0 1.3 1.4 1.2 5.7 1.3 c
rhb1 1.0 1.3 1.0 0.8 2.9 1.6 c
rtg2 0.7 1.4 2.2 0.7 2.0 5.2 c
ncs6 0.7 1.4 0.9 0.9 5.1 4.7 c
ybr099c 1.5 2.3 1.5 3.5 6.9 4.6 c
ylr376c 2.0 1.0 2.0 1.1 1.9 3.4 ac
ynl171c 1.0 1.8 ND 1.6 6.3 ND c
yor275c 2.2 0.9 4.7 1.1 1.0 4.0 ac
Open in a new tab

The strains were treated two times for 1 hr with the indicated MNNG concentrations. The values indicate fold increase in MNNG sensitivity conferred by the respective mutation (column one) in the indicated genetic backgrounds (above) and represent a mean of at least three independent experiments. See materials and methods for details. ND, not determined.

a

Mutants not originally identified in our screen, but ascertained as MNNG-sensitive after comparison with (Chang et al. 2002).

b

Mutants that transiently arrest upon MNNG treatment, but do not show difference in survival.

c

Factor showing no protein–protein or synthetic lethal interaction within the group identified in our screen as determined by the GRID database or other literature data.

Sensitivity to MMS was assayed by plating untreated cultures on regular YPD medium supplemented or not with 0.01% MMS (Sigma). All other spot tests were performed as indicated and/or described previously (Cejka et al. 2005).

Gel-shift experiments:

The oligonucletide substrates (Cejka et al. 2005) and the preparation of nuclear extracts (Iaccarino et al. 1996) were described previously. Six micrograms of the indicated nuclear extract, 300 ng of poly-dIdC competitor, and the labeled substrate (6.6 nm) in binding buffer [10% glycerol, 25 mm Hepes (pH 8.0), 0.5 mm EDTA, 0.5 mm DTT, 1× complete protease inhibitory cocktail without EDTA (Roche)] were incubated at room temperature in a total volume of 30 μl. The reaction products were separated on native 6% polyacrylamide gels at 150 V for 85 min. The gels were dried and scanned with Typhoon 9400 PhosphoImager and quantified with ImageQuant software.

RESULTS

Experimental setup of the genomewide screen:

Given the amenability of S. cerevisiae to genetic manipulation and high throughput screening, we deployed this organism in our search for metabolic pathways involved in the processing of MNNG-induced DNA damage. In our earlier experiments, we showed that inactivation of homologous recombination by deleting the RAD52 gene brought about a significant increase in MNNG sensitivity in cells lacking Mgt1. Because this increase was fully dependent on the MMR genes MLH1 or MSH2, we proposed that the hypersensitivity of the mgt1 rad52 mutant to MNNG is linked to MMR-mediated processing of MeG-containing DNA, which gives rise to cytotoxic intermediates that cannot be resolved in this strain (Cejka et al. 2005). As the latter study was carried out in the FF18733/4 strain, we first had to confirm that a mgt1 rad52 mutant in the BY5563 background, on which the array of 4644 haploid yeast gene deletion mutants used in this study was based, displayed a similar phenotype. This was indeed the case (Figure 1A).

Figure 1.—

Figure 1.—

Open in a new tab

Experimental set up of the genomewide screen designed to identify factors involved in MNNG-induced MeG-DNA damage processing. (A) MNNG-sensitivity of SLP1 (mgt1), SLP15 (mgt1 msh2), SLP19 (mgt1 rad52), and SLP17 (mgt1 msh2 rad52) mutant strains. Graphic representation of the sensitivity at the indicated MNNG concentrations as determined by counting the surviving colonies 48 hr after treatment (left). A representative experiment showing MNNG-treated cells plated at indicated dilutions 24 hr after treatment is shown (right). (B and C) Scheme of genomewide screens for factors contributing to MNNG-sensitivity in the mgt1 rad52 background (B) or required for MNNG-resistance in the mgt1 background (C). Active hypothetical pathways in the bait strain are depicted with bold arrows. See text for details.

To find out whether factors other than MMR proteins are involved in the generation of MNNG-induced cytotoxic intermediates, we carried out the first screen, in which the mgt1 rad52 bait strain was crossed with the ordered array of 4644 viable haploid yeast strains and where we screened for triple mutants resistant to MNNG (Figure 1B).

Conversely, factors that help resolve the cytotoxic intermediates generated by MMR-dependent processing of MeG-containing DNA should contribute to MNNG resistance. Thus, inactivation of genes encoding these factors was expected to lead to sensitization to the drug. We set out to search for these factors in the second screen, in which the MMR-proficient mgt1 bait strain was crossed with the above library and where we scored for MNNG sensitivity (Figure 1C).

Both of the above libraries were generated using an automated system (Tong et al. 2001; supplemental Figure 1A and supplemental Methods). MNNG resistance in the mgt1 rad52-derived array and MNNG sensitivity in the mgt1-derived array were initially assayed qualitatively and later quantitatively, as described in materials and methods. Our approach was validated by the identification of the mgt1 rad52 msh2 and mgt1 rad52 mlh1 triple mutants in the former and of the mgt1 rad52 mutant in the latter screen (supplemental Figure 1B and data not shown).

MMR is the only dominant pathway contributing to MNNG toxicity:

Our screen for factors mediating the toxicity of MNNG identified 11 mutants (supplemental Table 1), which included strains lacking the known MMR genes MSH2, MSH6, MLH1, and PMS1. The deletion of the MSH3 gene, in agreement with our previous study (Cejka et al. 2005), did not bring about MNNG resistance. In six of the seven remaining strains, we performed genetic analyses to confirm that the MNNG-resistant phenotype was indeed linked to the annotated deletion (supplemental Figure 2A). In five cases, we could show that the observed alleviation of MNNG-induced toxicity was not caused by deletion of the identified genes and thus that the resistance to MNNG must have been linked to other mutations in these strains, most likely in the MMR genes. The only exception was the ymr166c deletion, which clearly segregated with the MNNG-resistant phenotype. YMR166C encodes a predicted inner mitochondrial membrane transporter, and its role in MNNG processing and possibly MMR seemed unlikely. Importantly, this gene is directly adjacent to MLH1 and its deletion was previously shown to affect the expression of Mlh1 (Huang et al. 2003). The seventh mutant, bio2, exhibited a severe growth defect that did not allow us to perform the backcrossing control experiments; however, it seems highly unlikely that the product of this gene, biotin synthase, is involved in MeG processing. Our results thus imply that MMR is the only nonessential and nonredundant pathway required for sensitization of S. cerevisiae cells to MNNG in the absence of Mgt1 and HR (Figure 2A). Significantly, this process does not appear to require checkpoint proteins.

Figure 2.—

Figure 2.—

Open in a new tab

Mutations influencing cellular response to MNNG and cross-sensitivity to MMS. (A) Factors mediating MNNG sensitivity. SLP19 (mgt1 rad52), sensitive strain; SLP17 (mgt1 rad52 msh2), resistant strain. The genomewide screen and subsequent verification experiments identified exclusively the MMR factors Mlh1, Pms1, Msh2, and Msh6. Approximately 105 cells were treated two times for 1 hr with 0.5 μm MNNG, plated, and evaluated after 24 hr. (B) Cross-sensitivity of MNNG-sensitive deletion mutants to MMS. While the majority of the mutants identified were sensitive to both MNNG and MMS (e.g., rtt101 and cdc50), a few exhibited a significantly differential response. Approximately 102 (left and middle) or 103 cells (right) were mock treated, treated two times for 1 hr with 3 μm MNNG, or plated on MMS-containing plates and evaluated after 48 hr.

We failed to identify the deoxycytidine deaminase dcd1 mutant, recently reported to confer MNNG resistance to a rad52 mgt1 strain, which was of similar magnitude to that seen with the rad52 mgt1 msh2 or rad52 mgt1 mlh1 strains (Liskay et al. 2007). Examination of the library revealed that this deletion strain was missing from our collection. However, this does not alter our conclusion that the key pathway conferring resistance to MNNG in the rad52 mgt1 genetic background is MMR; the dcd1 mutation alters the dCTP/dTTP ratio in mutant cells in favor of dCTP and thus reduces the likelihood that MeG residues will mispair with T during replication. This will in turn reduce the involvement of the MMR system in the cytotoxic processing of MeG/T mispairs.

Protection against MMR-dependent MNNG toxicity is mediated by homologous recombination and associated pathways:

In the mgt1 background, we initially identified 44 genes, the deletion of which sensitized cells to MNNG. In a subsequent screen, in which we compared the sensitivities of the mutants to MNNG and MMS (see below), we identified 18 additional mutants with a weak MNNG-sensitive phenotype (Table 2). The majority of the 62 identified gene products are known players in the maintenance of genomic stability.

We next compared the sensitivities of the identified mutants in the mgt1 and MGT1 backgrounds. As shown in Table 2 and in supplemental Figure 2B, the presence of Mgt1 dramatically alleviated MNNG toxicity in most tested mutants. Since MeG is the preferred substrate of Mgt1 (Xiao et al. 1991), we concluded that the factors we identified in the above screen are required to resolve cytotoxic structures arising in DNA containing explicitly this modified base. However, because our primary interest was to identify factors involved in the repair of lesions arising through erroneous processing of MeG by the MMR system, we carried out an additional set of experiments. According to our hypothesis, inactivation of MMR should rescue the MNNG-sensitive phenotype of the mutant strains. We therefore compared the sensitivities of the mutants in mgt1 and mgt1 msh2 genetic backgrounds. As shown in Table 2 and in supplemental Figure 2B, a large number of the above mutants were less sensitive to MNNG in the latter background, which pays witness to the fact that the MMR system generates cytotoxic lesions during MeG processing, which have to be repaired by the factors identified in the above screen. Similar to mammalian cells (Stojic et al. 2005), the rescue was incomplete, which suggests that MeG is slightly toxic even in the absence of processing by MMR, especially at higher MNNG concentrations (10 μm, Table 2).

The above genetic analysis placed a large subset of the genes identified in our screen downstream of MMR-dependent metabolism of MeG-containing DNA. In an attempt to identify the pathways involved in this process, we divided the above-identified genes into functional groups according to the biological roles played by their respective protein products (Saccharomyces Genome Database and Krawitz et al. 2002; Mullen et al. 2005). As shown in Table 2, the strongest sensitization to MNNG was seen upon inactivation of the Mre11-Rad50-Xrs2 (MRX) complex. (Note that the rad52 mre11 mutant was not present in our library.) This complex is involved in the initial steps of recombination, both HR and nonhomologous end joining (NHEJ) (see Krogh and Symington 2004 for review). In addition, the MRX complex is required for checkpoint activation after the formation of double-strand breaks (DSBs) (D'amours and Jackson 2002). These multiple functions of the heterotrimeric MRX complex might thus explain the extreme MNNG sensitivity of the mrx mutants.

Pairing of resected ends of DSBs requires the HR machinery. That our screen identified most HR genes further supported our hypothesis that the methylation damage was generating DSBs. Rad51-mediated homologous pairing and strand exchange appears to be the preferred mechanism of resolving structures downstream of MMR, as the loss of this pathway substantially sensitizes cells to MNNG and as this effect is most effectively rescued by inactivation of MMR. The Rad51-independent, Rad59-mediated pathway also appears to be involved, albeit to a lesser extent. However, both HR subpathways clearly participate in the repair of the MMR-dependent MeG processing, as loss of Rad52, which is involved in both Rad51- and Rad59-dependent mechanisms (reviewed in Krogh and Symington 2004) gives rise to MNNG hypersensitivity.

The role of Mms22 is not known at present, but this polypeptide is believed to act in a DSB repair pathway distinct from HR (Baldwin et al. 2005). Rtt101, its interacting partner, has E3-ubiquitin-ligase activity that was proposed to modify components of the replication machinery in response to DNA damage (Luke et al. 2006).

The RAD6-dependent postreplicative repair pathway also seems to be involved in the processing of MeG lesions, but deletions of genes in this pathway gave rise to only weak phenotypes. (Note that the rad6 and rad18 mutations were missing in our screen.) The notable exception was RAD5, the loss of which increased the sensitivity to MNNG quite dramatically. This might be explained by the involvement of Rad5 in DSB repair in addition to its function in the error-free MMS2/UBC13-dependent pathway activated by polyubiquitylation of PCNA (Chen et al. 2005). Mutations in genes encoding the Sgs1 helicase and its associated factors Top3 and Rmi1 also increased MNNG cytotoxicity quite substantially. However, in both these groups, the lesions addressed by these pathways were apparently not MMR induced, as loss of Msh2 failed to rescue the sensitivities to a substantial extent (Table 2).

We also identified all three subunits of the chromatin assembly factor 1 (CAF-1), Cac1, Cac2, and Cac3, as well as Rtt109, Asf1, and several other mutants showing synthetic lethal interactions with mutations in these chromatin maintenance factors (Table 2). The mutants lacking CAF-1 subunits were not significantly sensitized to MNNG; rather, they displayed a prolonged MMR-dependent cell-cycle arrest (compare Table 2 and supplemental Figure 2B). Loss of CAF-1 or Asf1 was shown previously to impair HR- and NHEJ-dependent repair of DSBs (Lewis et al. 2005); these factors were therefore proposed to play key roles in chromatin assembly during DSB repair via HR (Linger and Tyler 2005). More recently, RTT109 was shown to encode a histone acetyltranferase that interacts with Asf1 and acetylates lysine 56 of histone H3. Given that this modification was shown to be critical for cell survival when DNA damage arises during S phase (Driscoll et al. 2007), the present evidence further strengthens our hypothesis that the cytotoxic lesions arising through the MMR-dependent processing of MeG arise during DNA replication. Moreover, the identification of chromatin assembly factors in our screen further substantiated the involvement of HR in the processing of this type of methylation damage.

The finding that the greatest sensitization resulted from the loss of factors involved in the processing of damaged DNA and in the signaling of these events provided further experimental evidence in support of the hypothesis that MeG-containing DNA has to be actively processed by MMR before the cell-cycle arrest is activated (Stojic et al. 2004a; Mojas et al. 2007) and showed that DNA damage-induced signaling protects yeast cells from the cytotoxic effect of MNNG. Attempts to characterize the molecular mechanisms involving the products of the above-identified genes in the processing of methylation damage are currently in progress in our laboratory.

Genes required for survival upon MNNG and MMS treatment are in part different:

In addition to generating high amounts of MeG, SN1-type methylating agents also produce O4-methythymine (∼0.5%) and substantial amounts (∼10%) of phosphotriesters. However, the predominant products of methylation with both SN1- and SN2-type agents are N3-methyladenine and N7-methylguanine, which account for 70–90% of all modifications (Beranek 1990). In an attempt to distinguish between factors on our list that are involved in the processing of methylation damage common to both types of agents and those required specifically for the processing of MeG-dependent intermediates, we compared the relative sensitivities of our mutants to MNNG and MMS, which produce only very small amounts of O-methylated bases. Surprisingly, and as in the case of MNNG, most mutants were more sensitive to MMS in the absence of Mgt1, even though the observed difference was only small (data not shown). It therefore seems that even the low amounts of MeG generated by MMS contribute to the toxicity of this chemical. We then compared the results of our experiment with those of Chang et al. (2002), who carried out a genomewide search for factors causing MMS sensitivity (Chang et al. 2002). As anticipated, most of the mutants identified in the two screens were similarly sensitive to MMS and MNNG. However, there were several notable exceptions (Figure 2B). The loss of Mag1 sensitized cells preferentially to MMS. Mag1 is a DNA glycosylase responsible primarily for the removal from DNA of N3-methyladenine, which is believed to block DNA replication. The mms2/ubc13 mutants were also more sensitive to MMS. The heterodimer encoded by these genes is believed to act primarily on collapsed replication forks, which might arise frequently upon MMS treatment when the Mag1-dependent base excision repair of N3-methyladenine is saturated (Boiteux et al. 1984).

We also identified mutants that were hypersensitive selectively to MNNG. The alternative RFC-1 subunit Elg1 may be of special interest in this respect. Mutations in ELG1 were shown to cause only a very low MMS-sensitivity, less than that caused by mutations in the other alternative RFC-1 subunits Ctf18 and Rad24 (Kanellis et al. 2003). Upon MNNG treatment, the requirement for these alternative RFC-1 complexes was reversed: deletion of ELG1 dramatically sensitized cells to MNNG, while absence of CTF18 and RAD24 gave rise to much weaker phenotypes (Table 2). This is in line with recent findings that Elg1 may play a role in HR-mediated repair of double-strand breaks (Ogiwara et al. 2007) implicated in methylation damage processing.

Processing rather than mere binding of MeG lesions by MMR proteins mediates toxicity of MNNG in S. cerevisiae:

In agreement with results obtained in E. coli (Rasmussen and Samson 1996; Calmann and Marinus 2005), our data suggest that toxicity of MNNG in yeast is linked with the processing of MeG-containing mispairs by MMR and HR. Although the situation in higher eukaryotes appears to be somewhat more complex, a large body of experimental evidence accumulated over the years (for review see Kaina 2004; Stojic et al. 2004a) supports the processing hypothesis. However, recent work by Yoshioka et al. (2006) suggested that the binding of MeG/T by the human MSH2/MSH6 mismatch recognition factor is sufficient to activate the Chk1-dependent cell-cycle arrest and apoptosis in a MLH1/PMS2-dependent manner (Yoshioka et al. 2006). This result was in agreement with earlier work (Lin et al. 2004; Yang et al. 2004), where missense mutations in murine Msh2 or Msh6 were reported to cause loss of MMR function, but did not affect the apoptotic response of the cells to several different alkylating agents. Moreover, the mutant proteins were suggested to act in a dominant negative manner, through inhibiting the function of the wild-type factor in heterozygous cells by forming longer-lived, but unproductive complexes with mismatch-containing substrates (Hess et al. 2002, 2006).

Although the mutator phenotypes of the corresponding missense mutations in MSH6 were first characterized in yeast (Yang et al. 2004), their response to MNNG was not tested, since the involvement of yeast MMR in MeG-dependent killing was unknown at that time. We therefore replaced the wild-type MSH6 gene with the msh6G1067I mutation (Table 1) and sequenced the entire mutant gene and its flanking regions, to exclude the presence of additional mutations within the msh6 locus (data not shown). The msh6G1067I mutant displayed an increased spontaneous forward mutation rate in the CAN1 assay, which was similar to that reported previously (Yang et al. 2004). This increase was also similar to that seen with the msh6 null mutant (Figure 3A). Nuclear extracts of these cells possessed mismatch binding activity that was less sensitive to ATP than that present in extracts of wild-type cells (Figure 3B and supplemental Figure 3A), in agreement with previous studies (Hess et al. 2002, 2006). However, when this mutation was crossed into the mgt1 rad52 background, it rescued the sensitivity of the mgt1 rad52 mutant to the same extent as the msh6 null mutation (Figure 3C and supplemental Figure 3B). This result demonstrates that the toxicity of MNNG in S. cerevisiae requires processing of MeG-containing base pairs by the MMR machinery and that mismatch recognition alone is insufficient to induce cell killing.

TABLE 1.

Saccharomyces cerevisiae strains used in this study

Strain Relevant genotype Source
PY5563 MATα can1∷Mfa1pr-HIS3 lyp1 his3 leu2 ura3 met15 M. Peter
SLP1 PY5563 with mgt1∷NAT This study
SLP15 PY5563 with mgt1∷NAT msh2∷KANMX This study
SLP17 PY5563 with rad52∷(T)9URA3 mgt1∷NAT msh2∷KANMX This study
SLP19 PY5563 with rad52∷(T)9URA3 mgt1∷NAT This study
FF18733 MATa; leu2-3, 112; ura3-52; his7-2; lys1-1; trp1-289 F. Fabre
EP54 FF18733 with msh6∷KANMX E. Papouli
FPC96 FF18733 with msh6G1067I This study
FPC92 FF18733 with mgt1∷KANMX; msh6∷KANMX This study
FPC97 FF18733 with mgt1∷KANMX; msh6G1067I This study
FPC99 FF18733 with mgt1∷KANMX; rad52∷URA3; msh6∷KANMX This study
FPC100 FF18733 with mgt1∷KANMX; rad52∷URA3; msh6G1067I This study
Open in a new tab

Figure 3.—

Figure 3.—

Open in a new tab

Analysis of the msh6 G1067I mutation. (A) CAN1 forward mutation rates of indicated strains. The numbers given are the mutation rates followed by the fold increase relative to wild type in parentheses. (B) Mismatch binding activity in nuclear extracts prepared from wild-type (MSH6), Msh6G1067I (msh6GI), or msh6 null (msh6Δ) cells. The protein/DNA complexes were analyzed by a gel-shift assay and visualized by autoradiography. The specific complex is indicated by an arrow. (C) MNNG-induced killing of strains with indicated genetic backgrounds. Mid-log phase cells were treated with the indicated concentrations of MNNG, spotted at proper serial dilutions on YPD plates, and evaluated after 48 hr as described in materials and methods. The results show a representative experiment.

DISCUSSION

Sensitization of mammalian cells to SN1-type methylating agents has been suggested to involve futile processing of MeG-containing base pairs by the MMR system (see Jiricny 2006; Stojic et al. 2004a for reviews). An alternative, direct signaling hypothesis suggested that the recognition of MeG-containing mismatches by MMR proteins may be sufficient to activate the cell-cycle checkpoint and apoptosis (Yang et al. 2004; Yoshioka et al. 2006). This latter model presupposes that MMR proteins communicate directly with the checkpoint machinery and that they distinguish between normal mismatches and those containing MeG. Recent crystallographic evidence argues against the latter point (Warren et al. 2007), and our electron microscopic data, which directly demonstrated the presence of single-stranded gaps in newly replicated DNA of both yeast and human cells after MNNG treatment (Mojas et al. 2007), further supports the futile processing rather than the direct signaling hypothesis. Thus, in eukaryotic cells, the inability of MMR to remove MeG residues from the template strand results in replication restarting downstream from the modified base, leaving short single-stranded gaps opposite these modifications in the newly replicated DNA. Such gaps would be expected to be highly cytotoxic, unless resolved by HR prior to the following round of DNA replication. A similar mechanism appears to be in place also in dam mutants of E. coli, as shown in the Marinus laboratory (Calmann et al. 2005).

The mechanism of action of methylating agents in yeast appeared to be different at first, because the MMR status of the S. cerevisiae did not affect its response to MNNG (Xiao et al. 1995). We showed, however, that the MMR-mediated cytotoxicity was masked by the highly efficient HR in this organism, supporting a model in which HR is required to resolve MeG-containing structures downstream of MMR (Calmann et al. 2005; Cejka et al. 2005). Moreover, as mentioned above, we were also able to directly demonstrate the presence of single-stranded gaps in the DNA of MNNG-treated yeast cells (Mojas et al. 2007). Despite these similarities to the bacterial and mammalian systems, we noted that yeast cells readily recover from the MNNG-induced cell-cycle arrest (Mojas et al. 2007), in contrast to human cells, which mostly die (Stojic et al. 2004b). To gain more insight into the processing of methylation damage in S. cerevisiae, we turned to high-throughput assays. In our search for factors mediating MNNG sensitivity, we identified solely the known MMR factors Msh2, Msh6, Mlh1, and Pms1. The four MMR proteins, together with Dcd1 identified in the Liskay laboratory (Liskay et al. 2007), thus appear to be the only nonredundant and nonessential factors required for MNNG-induced killing in S. cerevisiae. Incidentally, the fact that we did not find any DNA damage signaling mutants showed that checkpoint pathways do not mediate the toxicity of methylating agents in yeast. In contrast, several signaling molecules were identified in the search for factors, the lack of which confers MNNG sensitivity. This suggested that DNA damage signaling protects S. cerevisiae from MNNG-induced cytotoxicity.

What is the nature of the MeG- and MMR-dependent intermediates, and which mechanisms resolve them? Our data provide several important insights. The pathways identified in our screen for MNNG resistance that act downstream of MMR, namely HR, the MMS22 pathway, the CAF-1 and MRX complexes, and others are believed to act specifically in various steps of double-strand break repair (Baldwin et al. 2005; Lewis et al. 2005; Linger and Tyler 2005). In addition, our results often overlapped with those performed with other DSB-inducing agents such as ionizing radiation (Game et al. 2003) or topoisomerase poisons (Baldwin et al. 2005; R. Rothstein, personal communication). In contrast, proteins believed to rescue stalled replication polymerases, such as the Mms2/Ubc13 complex or translesion polymerases, which were shown to be required for survival upon MMS or UV treatments (Birrell et al. 2001; Chang et al. 2002), were either not identified in our screen or found to resolve lesions that were not dependent on MMR (Table 2). This suggests that structures ultimately formed upon MMR-dependent processing of MeG-containing mispairs are (or at least closely resemble) double-strand breaks.

Moreover, the analysis of the msh6G1076I mutation provided independent and additional evidence that processing of MeG adducts by MMR is required for MNNG-induced killing in yeast. This mutant, although able to recognize and bind mismatches, was shown to be defective in MMR due to the slow turnover of the yeast MutSα on its mispair-containing DNA substrate (Hess et al. 2002, 2006). In murine cells, binding to damaged DNA was reported to be sufficient to activate the cell-cycle arrest and apoptosis, separating thus the repair function from DNA damage-induced killing (Lin et al. 2004; Yang et al. 2004). We show that this is not the case in yeast; the yeast mutant responded to MNNG treatment identically to the msh6 deletion strain, showing that damage recognition in this organism is not sufficient to activate the cell killing process.

The high evolutionary conservation of methylguanine methyltransferases suggests that MeG represents a considerable threat to genomic integrity of all living organisms. Because these modifications become cytotoxic in the presence of another highly conserved pathway, MMR, it is of interest to review how different organisms cope with this unique type of DNA damage under circumstances, where endogenous levels of MGMT are insufficient to remove all MeG residues from their genomes. In E. coli, the ada gene is inducible and MGMT is thus produced in larger amounts when the cells are exposed to methylating agents. In eukaryotes, the MGMT-encoding genes are not damage inducible, so the cells developed alternative strategies to deal with MeG. In higher eukaryotes, where it is preferable that a cell die rather than mutate, MMR-dependent processing of MeG-containing DNA leads to activation of cell-cycle arrest and apoptosis. In monocellular eukaryotes such as S. cerevisiae, survival at the cost of mutagenesis is preferable to cell death. It relies on its highly efficient recombination pathways to deal with the DSBs arising at replication forks that have collapsed after having reached a single-stranded DNA gap left opposite a MeG. This rescue would be expected to be highly mutagenic and preliminary data from our laboratory (P. Cejka and J. Jiricny, unpublished results) indicate that this is indeed the case. What our results fail to provide is evidence implicating DNA damage signaling proteins in MMR-dependent MNNG toxicity. The results of this extensive genetic screen thus provide further support for the hypothesis that the cytotoxocity of SN1-type methylating agents is dependent on the futile processing of MeG-containing mispairs by the MMR system.

Acknowledgments

We express our gratitude to Margaret Fäsi, Julia Mawer, and Marc Sohrmann for excellent technical assistance, to Matthias Peter for providing us with access to the yeast deletion library, and to Peter Ahnesorg, Primo Schär, Massimo Lopes, and Pavel Janscak for suggestions and critical reading of the manuscript. This work was funded by grants from the Bonizzi-Theler Stiftung, the Sixth Framework Program of the European Community, the Swiss National Science Foundation, and UBS AG.

References

  1. Baldwin, E. L., A. C. Berger, A. H. Corbett and N. Osheroff, 2005. Mms22p protects Saccharomyces cerevisiae from DNA damage induced by topoisomerase II. Nucleic Acids Res. 33 1021–1030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beranek, D. T., 1990. Distribution of methyl and ethyl adducts following alkylation with monofunctional alkylating agents. Mutat. Res. 231 11–30. [DOI] [PubMed] [Google Scholar]
  3. Birrell, G. W., G. Giaever, A. M. Chu, R. W. Davis and J. M. Brown, 2001. A genome-wide screen in Saccharomyces cerevisiae for genes affecting UV radiation sensitivity. Proc. Natl. Acad. Sci. USA 98 12608–12613. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Boiteux, S., O. Huisman and J. Laval, 1984. 3-Methyladenine residues in DNA induce the SOS function sfiA in Escherichia coli. EMBO J. 3 2569–2573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Calmann, M. A., and M. G. Marinus, 2005. Differential effects of cisplatin and MNNG on dna mutants of Escherichia coli. Mutat. Res. 578 406–416. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Calmann, M. A., J. E. Evans and M. G. Marinus, 2005. MutS inhibits RecA-mediated strand transfer with methylated DNA substrates. Nucleic Acids Res. 33 3591–3597. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cejka, P., N. Mojas, L. Gillet, P. Schar and J. Jiricny, 2005. Homologous recombination rescues mismatch-repair-dependent cytotoxicity of S(N)1-type methylating agents in S. cerevisiae. Curr. Biol. 15 1395–1400. [DOI] [PubMed] [Google Scholar]
  8. Chang, M., M. Bellaoui, C. Boone and G. W. Brown, 2002. A genome-wide screen for methyl methanesulfonate-sensitive mutants reveals genes required for S phase progression in the presence of DNA damage. Proc. Natl. Acad. Sci. USA 99 16934–16939. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen, S., A. A. Davies, D. Sagan and H. D. Ulrich, 2005. The RING finger ATPase Rad5p of Saccharomyces cerevisiae contributes to DNA double-strand break repair in a ubiquitin-independent manner. Nucleic Acids Res. 33 5878–5886. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. D'Amours, D., and S. P. Jackson, 2002. The Mre11 complex: at the crossroads of dna repair and checkpoint signalling. Nat. Rev. Mol. Cell Biol. 3 317–327. [DOI] [PubMed] [Google Scholar]
  11. Driscoll, R., A. Hudson and S. P. Jackson, 2007. Yeast Rtt109 promotes genome stability by acetylating histone H3 on lysine 56. Science 315 649–652. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Game, J. C., G. W. Birrell, J. A. Brown, T. Shibata, C. Baccari et al., 2003. Use of a genome-wide approach to identify new genes that control resistance of Saccharomyces cerevisiae to ionizing radiation. Radiat. Res. 160 14–24. [DOI] [PubMed] [Google Scholar]
  13. Guldener, U., S. Heck, T. Fielder, J. Beinhauer and J. H. Hegemann, 1996. A new efficient gene disruption cassette for repeated use in budding yeast. Nucleic Acids Res. 24 2519–2524. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hastings, P. J., 1984. Measurement of restoration and conversion: its meaning for the mismatch repair hypothesis of conversion. Cold Spring Harbor Symp. Quant. Biol. 49 49–53. [DOI] [PubMed] [Google Scholar]
  15. Hess, M. T., R. D. Gupta and R. D. Kolodner, 2002. Dominant Saccharomyces cerevisiae msh6 mutations cause increased mispair binding and decreased dissociation from mispairs by Msh2-Msh6 in the presence of ATP. J. Biol. Chem. 277 25545–25553. [DOI] [PubMed] [Google Scholar]
  16. Hess, M. T., M. L. Mendillo, D. J. Mazur and R. D. Kolodner, 2006. Biochemical basis for dominant mutations in the Saccharomyces cerevisiae MSH6 gene. Proc. Natl. Acad. Sci. USA. 103 558–563. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Huang, M. E., A. G. Rio, A. Nicolas and R. D. Kolodner, 2003. A genomewide screen in Saccharomyces cerevisiae for genes that suppress the accumulation of mutations. Proc. Natl. Acad. Sci. USA 100 11529–11534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Iaccarino, I., F. Palombo, J. Drummond, N. F. Totty, J. J. Hsuan et al., 1996. MSH6, a Saccharomyces cerevisiae protein that binds to mismatches as a heterodimer with MSH2. Curr. Biol. 6 484–486. [DOI] [PubMed] [Google Scholar]
  19. Jiricny, J., 2006. The multifaceted mismatch-repair system. Nat. Rev. Mol. Cell Biol. 7 335–346. [DOI] [PubMed] [Google Scholar]
  20. Kaina, B., 2004. Mechanisms and consequences of methylating agent-induced SCEs and chromosomal aberrations: a long road traveled and still a far way to go. Cytogenet. Genome Res. 104 77–86. [DOI] [PubMed] [Google Scholar]
  21. Kanellis, P., R. Agyei and D. Durocher, 2003. Elg1 forms an alternative PCNA-interacting RFC complex required to maintain genome stability. Curr. Biol. 13 1583–1595. [DOI] [PubMed] [Google Scholar]
  22. Karran, P., and M. G. Marinus, 1982. Mismatch correction at O6-methylguanine residues in E. coli DNA. Nature 296 868–869. [DOI] [PubMed] [Google Scholar]
  23. Krawitz, D. C., T. Kama and P. D. Kaufman, 2002. Chromatin assembly factor I mutants defective for PCNA binding require Asf1/Hir proteins for silencing. Mol. Cell. Biol. 22 614–625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Krogh, B. O., and L. S. Symington, 2004. Recombination proteins in yeast. Annu. Rev. Genet. 38 233–271. [DOI] [PubMed] [Google Scholar]
  25. Lewis, L. K., G. Karthikeyan, J. Cassiano and M. A. Resnick, 2005. Reduction of nucleosome assembly during new DNA synthesis impairs both major pathways of double-strand break repair. Nucleic Acids Res. 33 4928–4939. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lin, D. P., Y. Wang, S. J. Scherer, A. B. Clark, K. Yang et al., 2004. An Msh2 point mutation uncouples DNA mismatch repair and apoptosis. Cancer Res. 64 517–522. [DOI] [PubMed] [Google Scholar]
  27. Linger, J., and J. K. Tyler, 2005. The yeast histone chaperone chromatin assembly factor 1 protects against double-strand DNA-damaging agents. Genetics 171 1513–1522. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Liskay, R. M., L. J. Wheeler, C. K. Mathews and N. Erdeniz, 2007. Involvement of deoxycytidylate deaminase in the response to S(n)1-type methylation DNA damage in budding yeast. Curr. Biol. 17 R755–R757. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Luke, B., G. Versini, M. Jaquenoud, I. W. Zaidi, T. Kurz et al., 2006. The cullin Rtt101p promotes replication fork progression through damaged DNA and natural pause sites. Curr. Biol. 16 786–792. [DOI] [PubMed] [Google Scholar]
  30. Mojas, N., M. Lopes and J. Jiricny, 2007. Mismatch repair-dependent processing of methylation damage gives rise to persistent single-stranded gaps in newly replicated DNA. Genes Dev. 21 3342–3355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mullen, J. R., F. S. Nallaseth, Y. Q. Lan, C. E. Slagle and S. J. Brill, 2005. Yeast Rmi1/Nce4 controls genome stability as a subunit of the Sgs1-Top3 complex. Mol. Cell. Biol. 25 4476–4487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Nowosielska, A., S. A. Smith, B. P. Engelward and M. G. Marinus, 2006. Homologous recombination prevents methylation-induced toxicity in Escherichia coli. Nucleic Acids Res. 34 2258–2268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ogiwara, H., A. Ui, T. Enomoto and M. Seki, 2007. Role of Elg1 protein in double strand break repair. Nucleic Acids Res. 35 353–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rasmussen, L. J., and L. Samson, 1996. The Escherichia coli MutS DNA mismatch binding protein specifically binds O(6)-methylguanine DNA lesions. Carcinogenesis 17 2085–2088. [DOI] [PubMed] [Google Scholar]
  35. Ryttman, H., and G. Zetterberg, 1976. Induction of mitotic recombination with N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) in Saccharomyces cerevisiae. A comparison between treatment in vitro and in the host-mediated assay. Mutat. Res. 34 201–216. [DOI] [PubMed] [Google Scholar]
  36. Stojic, L., R. Brun and J. Jiricny, 2004. a Mismatch repair and DNA damage signalling. DNA Repair (Amst) 3 1091–1101. [DOI] [PubMed] [Google Scholar]
  37. Stojic, L., N. Mojas, P. Cejka, M. Di Pietro, S. Ferrari et al., 2004. b Mismatch repair-dependent G2 checkpoint induced by low doses of SN1 type methylating agents requires the ATR kinase. Genes Dev. 18 1331–1344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stojic, L., P. Cejka and J. Jiricny, 2005. High doses of SN1 type methylating agents activate DNA damage signaling cascades that are largely independent of mismatch repair. Cell Cycle 4 473–477. [DOI] [PubMed] [Google Scholar]
  39. Tong, A. H., M. Evangelista, A. B. Parsons, H. Xu, G. D. Bader et al., 2001. Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294 2364–2368. [DOI] [PubMed] [Google Scholar]
  40. Tsaryk, R., K. Fabian, J. Thacker and B. Kaina, 2006. Xrcc2 deficiency sensitizes cells to apoptosis by MNNG and the alkylating anticancer drugs temozolomide, fotemustine and mafosfamide. Cancer Lett. 239 305–313. [DOI] [PubMed] [Google Scholar]
  41. Warren, J. J., T. J. Pohlhaus, A. Changela, R. R. Iyer, P. L. Modrich et al., 2007. Structure of the human MutSalpha DNA lesion recognition complex. Mol. Cell 26 579–592. [DOI] [PubMed] [Google Scholar]
  42. Xiao, W., B. Derfler, J. Chen and L. Samson, 1991. Primary sequence and biological functions of a Saccharomyces cerevisiae O6-methylguanine/O4-methylthymine DNA repair methyltransferase gene. EMBO J. 10 2179–2186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Xiao, W., L. Rathgeber, T. Fontanie and S. Bawa, 1995. DNA mismatch repair mutants do not increase N-methyl-N′-nitro-N-nitrosoguanidine tolerance in O6-methylguanine DNA methyltransferase-deficient yeast cells. Carcinogenesis 16 1933–1939. [DOI] [PubMed] [Google Scholar]
  44. Yang, G., S. J. Scherer, S. S. Shell, K. Yang, M. Kim et al., 2004. Dominant effects of an Msh6 missense mutation on DNA repair and cancer susceptibility. Cancer Cell 6 139–150. [DOI] [PubMed] [Google Scholar]
  45. Yoshioka, K., Y. Yoshioka and P. Hsieh, 2006. ATR kinase activation mediated by MutSalpha and MutLalpha in response to cytotoxic O6-methylguanine adducts. Mol. Cell 22 501–510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Zhang, H., T. Tsujimura, N. P. Bhattacharyya, V. M. Maher and J. J. McCormick, 1996. O6-methylguanine induces intrachromosomal homologous recombination in human cells. Carcinogenesis 17 2229–2235. [DOI] [PubMed] [Google Scholar]

Articles from Genetics are provided here courtesy of Oxford University Press

RESOURCES