E. coli mismatch repair enhances AT-to-GC mutagenesis caused by alkylating agents

Mutation Research 815 (2017) 22–27

Contents lists available at ScienceDirect

Mutation Research/Genetic Toxicology and Environmental Mutagenesis journal homepage: www.elsevier.com/locate/gentox Community address: www.elsevier.com/locate/mutres

E. coli mismatch repair enhances AT-to-GC mutagenesis caused by alkylating agents Kota Nakano a , Yoko Yamada a , Eizo Takahashi a,b , Sakae Arimoto a , Keinosuke Okamoto a , Kazuo Negishi b , Tomoe Negishi a,b,∗ a b

Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Tsushima-naka, Kita-ku, Okayama, 700-8530, Japan Nihon Pharmaceutical University, Ina, Kita-Adachi-Gun, Saitama, 362-0806, Japan

a r t i c l e

i n f o

Article history: Received 26 October 2016 Received in revised form 16 January 2017 Accepted 13 February 2017 Available online 15 February 2017 Keywords: O4 -alkylthymine Mismatch repair (MMR) E. coli O6 -alkylguanine-DNA alkyltransferase (AGT) Nucleotide excision repair (NER) Mutagenesis Mutagenic repair

a b s t r a c t Alkylating agents are known to induce the formation of O6 -alkylguanine (O6 -alkG) and O4 -alkylthymine (O4 -alkT) in DNA. These lesions have been widely investigated as major sources of mutations. We previously showed that mismatch repair (MMR) facilitates the suppression of GC-to-AT mutations caused by O6 -methylguanine more efficiently than the suppression of GC-to-AT mutations caused by O6 -ethylguanine. However, the manner by which O4 -alkyT lesions are repaired remains unclear. In the present study, we investigated the repair pathway involved in the repair of O4 -alkT. The E. coli CC106 strain, which harbors prolac in its genomic DNA and carries the F’CC106 episome, can be used to detect AT-to-GC reverse-mutation of the gene encoding ␤-galactosidase. Such AT-to-GC mutations should be induced through the formation of O4 -alkT at AT base pairs. As expected, an O6 -alkylguanine-DNA alkyltransferase (AGT) -deficient CC106 strain, which is defective in both ada and agt genes, exhibited elevated mutant frequencies in the presence of methylating agents and ethylating agents. However, in the UvrAdeficient strain, the methylating agents were less mutagenic than in wild-type, while ethylating agents were more mutagenic than in wild-type, as observed with agents that induce O6 -alkylguanine modifications. Unexpectedly, the mutant frequencies decreased in a MutS-deficient strain, and a similar tendency was observed in MutL- or MutH-deficient strains. Thus, MMR appears to promote mutation at AT base pairs. Similar results were obtained in experiments employing double-mutant strains harboring defects in both MMR and AGT, or MMR and NER. E. coli MMR enhances AT-to-GC mutagenesis, such as that caused by O4 -alkylthymine. We hypothesize that the MutS protein recognizes the O4 -alkT:A base pair more efficiently than O4 -alkT:G. Such a distinction would result in misincorporation of G at the O4 -alkT site, followed by higher mutation frequencies in wild-type cells, which have MutS protein, compared to MMR-deficient strains. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Alkylation occurs at various sites on DNA bases following treatment with alkylating agents. It is well known that O6 -alkylguanine lesions induce GC-to-AT transitions, while O4 -alkylthymine lesions induce AT-to-GC transitions. From bacteria to humans, such damage is corrected by a wide range of DNA repair systems, including nucleotide excision repair (NER), base excision repair (BER), and systems that remove alkyl groups directly from modified bases such

∗ Corresponding author at: Nihon Pharmaceutical University, Ina, Kita-AdachiGun, Saitama, 362-0806, Japan. E-mail address: [email protected] (T. Negishi). http://dx.doi.org/10.1016/j.mrgentox.2017.02.001 1383-5718/© 2017 Elsevier B.V. All rights reserved.

as that involving O6 -alkylguanine-DNA alkyltransferase (AGT; E.C. 2.1.1.63) [1,2]. AGT is thought to play a major role in protecting organisms from alkylation damage [3,4]. The E. coli AGT system consists of two enzymes, where the activity of one is inducible (Ada) while the other is constitutive (Ogt). Both enzymes efficiently remove alkylated bases by transferring the alkyl group from DNA to themselves in a suicidal process that results in enzyme inactivation. NER also is known to repair a broad range of DNA lesions, including alkylated bases [5]. The mismatch repair system (MMR) is also involved in the repair of DNA damage, including alkylated base pairs, thereby facilitating the high fidelity of genomic DNA [6,7]. Base pairs containing O6 -alkylguanine are recognized by MutS, a protein that detects mismatch lesions, and thereby facilitates routing to the MMR pathway [7]. We previously demon-

K. Nakano et al. / Mutation Research 815 (2017) 22–27

strated that deficiency of MutS increases the mutations induced by methylating agents more efficiently than that induced by ethylating agents, a finding that correlated well with higher MutS binding to methylated guanine-containing base pairs than to ethylated guanine-containing base pairs [8]. This result is also consistent with the observation that the rates of involvement of the repair pathways differ between O6 -methylguanine and O6 -ethylguanine [9]. However, it remains unclear which repair system is primarily involved in the repair of alkylation damage at the O4 -site of thymine created by various methylating and ethylating agents. Fang et al. reported that O4 -alkylthymine is repaired by AGT, although the efficacy is still unclear [10]. In the present study, we measured the mutant frequencies induced by alkylating agents in various repair backgrounds to estimate the relevant contributions of these pathways to AT-to-GC mutation. To estimate the relationship between each repair system, we constructed MMR-deficient strains that were additionally deficient in AGT or NER activity, and measured the mutant frequencies of alkylating agents. In all cases, the frequency of mutations induced by alkylating agents was unexpectedly lower in MutS-deficient E. coli than in wild-type. These results suggested that the MutS protein may recognize the base pair O4 -alkT:A more efficiently than O4 -alkT:G as a repairable mismatch lesion, thereby resulting in higher mutation frequencies in wild-type strains compared to MMR-deficient strains. 2. Materials and methods 2.1. Chemicals N-Methyl-N-nitrosourea (MNU) (CAS 684-93-5) and N-ethyl(ENNG) (CAS 63885-23-4) were purchased from Nacalai Tesque (Kyoto, Japan). N-Ethyl-N-nitrosourea (ENU) (CAS 759-73-9) and N-methyl-N -nitro-N-nitrosoguanidine (MNNG) (CAS 70-25-7) were purchased from Sigma–Aldrich Chemicals (St. Louis, MO).

N -nitro-N-nitroso-guanidine

2.2. Bacterial strains The bacterial strains used in this study are shown in Table 1. All strains used in the mutagenesis experiments were derived from strain KA796 (ara, thi, pro-lac) [11] and also containing the F’prolac episomes from strains NR10831 (CC101) to NR10836 (CC106) that permit scoring of the respective six base substitutions in the lac gene, as previously described [12]. The wild-type strain, NR10836; the MMR-deficient derivatives, NR12898 (mutS201:Tn 5), NR11106 (mutL211:Tn 5), and NR12899 (mutH471:Tn 5); and the NER-deficient derivative NR13000 (uvrA277:Tn 10), have been previously described by Negishi et al. [13]. AGT-deficient strains GW5352 (ada-10:Tn 10) [14] and KT233 (ada:kan, ogt:cat) [15] were gifts from Drs. S. Cohen (Massachusetts Institute of Technology) and M. Sekiguchi (Kyushu University), respectively. Mutants doubly-deficient in MMR and either AGT (KO11141, KO21181 and KO31131) or NER (KO10021U, KO20021U and KO30013U) were constructed in the present study according to the methods previously described, where AGT (ada and ogt) or NER (uvrA) genes were disrupted in MMR-deficient strains by P1 transduction [9,16]. Disruption of each gene (ada, ogt or uvrA) in the final strains was confirmed by observed loss of PCR product amplification, as described by Taira et al. [9] (data not shown). The spontaneous mutant frequencies of the E. coli strains used in the present study are shown in Table 1. These mutant frequencies represent an average value calculated from the data obtained in each experiment (Tables 2 and 3, Figs. 1 and 2). The lack of repair function in these strains could be confirmed by the observed elevated spontaneous

23

mutant frequencies. MMR-deficient stains show higher spontaneous mutant frequencies than the MMR-proficient strains. 2.3. Mutation assay The lac allele of F’CC106 reverts to lac+ exclusively by AT-toGC transition [12]. We performed the lac reversion assay according to methods previously described [8]. To estimate mutant frequencies, 0.1 ml overnight cultures of bacteria were incubated for 1 h at 37 ◦ C with 0.1 ml of test compound solution containing 0.5 ml of 0.1 M sodium phosphate buffer (pH 7.4). Then 0.1 ml of the treated cultures were spread onto three minimal lactose plates to determine the number of revertants (lac+), and adequately diluted cultures were also spread onto three minimal glucose plates to determine the total viable cell numbers. Mutant frequencies were calculated by dividing the number of lac+ revertants by the number of total viable cells screened. The mutagens were used at the doses where the survivals are over 20 percent. As AGT-deficient strains are very sensitive to methylating agents, the doses of MNU and MNNG employed were lower in assays utilizing AGT-deficient strains than in other experiments. Experiments were performed in triplicate and independently with each mutagen, and typical results are shown in Figs. 1 and 2. Statistical analysis was performed using the Student’s t test. Significant P values are P > 0.01 and P > 0.05 compared with the mutant frequency for wild-type. 3. Results 3.1. Mutation spectra induced by MNU and ENU We investigated the base substitutions induced by methylation or ethylation using the CC101 to CC106 series of strains. As shown in Table 2, the mutagenicity of MNU and ENU was significantly elevated in strains CC102 and CC106 compared to the other strains tested. The frequency of MNU-induced mutations was about 10fold higher than that induced by ENU in CC102, in which GC-to-AT transitions where detected, although the highest mutant frequency of ENU was also detected in CC102. In contrast, ENU was more mutagenic than MNU in CC106, in which AT-to-GC transitions were detected. These results suggest that O6 -alkylated guanine is the predominant cause of mutations in E. coli, followed by O4 -alkylated thymine. 3.2. Mutagenic activity of alkylating agents in repair-deficient strains AT-to-GC transitions are expected to result from modification of thymine. In an effort to investigate which repair systems contribute to the repair of alkylated thymine, we measured mutant frequencies induced by methylating agents (MNU and MNNG) and ethylating agents (ENU and ENNG) in E. coli CC106 derivatives deficient in AGT, NER, or MMR. As shown in Fig. 1(A) and (D), mutations were clearly induced by each methylating agent in the AGT-deficient strain while negligible mutations were shown for wild-type. However, the mutant frequencies unexpectedly decreased in MMR-deficient strains (mutS, mutL and mutH) (Fig. 1(C) and (F)). Notably, the mutant frequency in the mutS strain was reduced to a level similar to that observed in the absence of treatment. The NER deficiency appeared not to be responsible for the mutagenesis induced by MNU or MNNG (Fig. 1(B) and (E)). AGT played a smaller role in the repair of damage caused by ethylating agents compared to that caused by methylating agents, and in contrast NER worked preferentially on ethylation-induced damage (Fig. 2). ENNG-induced mutant frequencies were significantly elevated in the uvrA (NER-deficient) strain at every dose tested (Fig. 2(B)). The frequency of ENU-induced mutagenesis also

24

K. Nakano et al. / Mutation Research 815 (2017) 22–27

Table 1 E. coli strains used in this study and their spontaneous mutant frequencies. Straina

Relevant genotype

Episome

Spontaneous mutant frequencyb (x10−7 )

NR10831 NR10832 NR10833 NR10834 NR10835 NR10836 NR12898 NR11106 NR12899 NR13000 GW5352 KT233 KO01126 KO01121 KO11141 KO21181 KO31131 KO10021U KO20021U KO30013U

KA796, wild type KA796, wild type KA796, wild type KA796, wild type KA796, wild type KA796, wild type KA796, mutS201:Tn 5 KA796, mutL211:Tn 5 KA796, mutH471:Tn 5 KA796, uvrA277:Tn 10 AB1157, ada-10:Tn 10 AB1157, ada:kan, ogt:cat NR10836, ada-10:Tn 10 NR10836, ada-10:Tn 10, ogt:cat NR12898, mutS201:Tn 5, ada-10:Tn 10, ogt:cat NR11106, mutL211:Tn5, ada-10:Tn10, ogt:cat NR12899, mutH471:Tn5, ada-10:Tn10, ogt:cat NR12898, mutS201:Tn 5, uvrA277:Tn 10 NR11106, mutL211:Tn 5, uvrA277:Tn 10 NR12899, mutH471:Tn 5, uvrA277:Tn 10

CC101 CC102 CC103 CC104 CC105 CC106 CC106 CC106 CC106 CC106 – – CC106 CC106 CC106 CC106 CC106 CC106 CC106 CC106

0.16 ± 0.25 0.29 ± 0.14 <0.01 0.78 ± 0.45 0.29 ± 0.33 <0.01 0.63 ± 0.65 2.79 ± 1.29 2.79 ± 1.04 <0.01 – – <0.01 <0.01 0.98 ± 0.05 3.57 ± 1.90 2.67 ± 1.21 0.21 ± 0.23 3.68 ± 3.47 1.50 ± 0.68

a b

All strains, except GW5352 and KT233, contain the F’ episome [12,13]. Spontaneous mutant frequencies as measured in the present study.

Table 2 Nitrosourea-induced mutability of different types of E. coli strains. Mutant frequency (x10−7 )

Mutagen Conc.(mM)

Strain Episome Base change

NR10831 CC101 AT → CG

NR10832 CC102 GC → AT

NR10833 CC103 GC → CG

NR10834 CC104 GC → TA

NR10835 CC105 AT → TA

NR10836 CC106 AT → GC

MNU

0 1 2.5 4 5

0 0.5 1.6 ± 1.6 0.8 1.3 ± 0.8

0 47.7 ± 16 169 ± 66 272 ± 44 441 ± 108

0 0 0 0 0

0 0 0 ± 0.3 0.8 1.5 ± 1.2

0 0 0 ± 2.2 0 0 ± 0.9

0 0 2.4 ± 1.5 1.3 3.6 ± 2.7

ENU

0 1 2.5 4 5

0 0.4 0.5 ± 0.4 0.6 0 ± 0.6

0 2.5 ± 1.8 11.5 ± 7.4 13.5 ± 4.7 42.6 ± 20

0 0 0 0 0

0 0 0.2 ± 1.3 0 1.1 ± 1.8

0 0 0 ± 1.1 0 0.3 ± 1.3

0 0 1.3 ± 1 3.2 15.2 ± 12.8

Table 3 Nitrosourea-induced mutability of different types of E. coli strains. Mutagen

Strain Conc.(mM)

(+)

mutS a

−7

MF (x 10 uvrA MNU MNNG ENU ENNG ada, ogt MNU MNNG ENU ENNG

)

mutL −7

Survival (%)

MF (x 10

)

mutH −7

Survival (%)

MF (x 10

)

Survival (%)

MF (x 10−7 )

Survival (%)

0 25 0 0.2 0 25 0 1

0 10.4 ± 5.61 0 7.65 ± 2.39 0 123.0 ± 9.22 0 103.9 ± 27.2

100 12.7 ± 3.5 100 63.4 ± 1.2 100 23.7 ± 1.9 100 50.9 ± 10.5

0.17 ± 0.30 0 0.27 ± 0.24 2.60 ± 0.10 0.23 ± 0.20 12.0 ± 1.22 0.18 ± 0.30 8.51 ± 7.7

100 65.7 ± 30.2 100 49.0 ± 7.9 100 31.8 ± 3.8 100 36.0 ± 9.1

1.20 ± 1.0 3.17 ± 0.17 2.16 ± 0.95 5.16 ± 1.28 2.45 ± 0.38 65.8 ± 14.5 8.90 ± 2.92 88.9 ± 11.3

100 51.4 ± 3.5 100 54.9 ± 1.5 100 52.2 ± 5.8 100 48.5 ± 9.2

0.97 ± 0.39 7.76 ± 2.17 1.60 ± 0.46 13.7 ± 3.16 1.87 ± 0.99 71.0 ± 15.3 1.59 ± 0.72 135.9 ± 10.8

100 33.0 ± 9.2 100 64.9 ± 13 100 54.5 ± 6.4 100 39.0 ± 2.0

0 5 0 0.01 0 25 0 1

0 7.08 ± 5.98 0.08 ± 0.14 5.60 ± 1.66 0 77.7 ± 15.7 0 119.5 ± 35.4

100 12.9 ± 0.9 100 50.7 ± 1.3 100 41 ± 12.9 100 43.9 ± 8.5

1.16 ± 0.75 1.84 ± 1.77 0.71 ± 0.07 1.60 ± 1.17 1.02 ± 0.01 48.6 ± 19.5 1.01 ± 0.46 25.5 ± 6.38

100 44.5 ± 13.8 100 108 ± 18.3 100 36.1 ± 11.3 100 41.2 ± 4.7

3.26 ± 0.94 5.05 ± 0.22 3.98 ± 1.37 8.34 ± 0.91 1.23 ± 0.65 97.8 ± 6.29 5.83 ± 0.63 110.7 ± 3.43

100 37.5 ± 5.3 100 69.8 ± 13.6 100 74.7 ± 3.8 100 51.9 ± 2.7

2.68 ± 0.59 5.61 ± 1.03 2.59 ± 1.21 5.52 ± 2.93 1.99 ± 0.37 63.92 ±2.58 3.42 ± 2.13 112.4 ± 19.8

100 68.8 ± 6.4 100 73.7 ± 17.4 100 72.9 ± 5.3 100 67.8 ± 1.1

Strains are NR13000 (uvrA), KO01121 (ada ogt), KO10021U (uvrA mutS), KO20021U (uvrA mutL), KO30013U (uvrA mutH), KO11141 (ada ogt mutS), KO21181 (ada ogt mutL), and KO31131 (ada ogt mutH). a MF is mutant frequency calculated as “Materials and Methods”. Survival is calculated from the colony number of glucose plate.

K. Nakano et al. / Mutation Research 815 (2017) 22–27

25

Fig. 1. Mutant frequencies induced by methylating agents: (A)–(C) MNNG, and (D)–(F) MNU. Strains are NR10836 (wild-type), closed circle; KO01121 (ada, ogt), dark square; NR13000 (uvrA), dark triangle; NR12898 (mutS), open circle; NR11106 (mutL), open triangle; NR12899 (mutH), open square. Mutant frequencies are corrected by subtracting spontaneous mutant frequencies. Statistical analysis was performed using the Student’s t-test. **P > 0.01 and *P > 0.05 compared with the mutant frequency for wild-type.

Fig. 2. Mutant frequencies induced by ethylating agents: (A)–(C) ENNG, and (D)–(F) ENU. Strains are NR10836 (wild-type), closed circle; KO01121 (ada, ogt), dark square; NR13000 (uvrA), dark triangle; NR12898 (mutS), open circle; NR11106 (mutL), open triangle; NR12899 (mutH), open square. Mutant frequencies are corrected by subtracting spontaneous mutant frequencies. Statistical analysis was performed using the Student’s t test. ***P > 0.001, **P > 0.01 and *P > 0.05 compared with the mutant frequency for wild-type.

tended to increase (Fig. 2(E)). MMR deficiency provided similar suppression of mutations induced by ethylating or methylating agents (Fig. 2(C) and (F)). Taken together, these results indicated that alkylated thymine residues are repaired primarily via the AGT system, and that alkylated thymine residues are more mutagenic in the presence of a functional MMR system.

3.3. Mutagenic activity of alkylating agents in mutants doubly-deficient in MMR and AGT/NER The involvement of MMR in response AT-to-GC mutations induced by alkylating agents was unexpected. In an effort to examine the potential involvement of MMR on AT-to-GC mutations even in the absence of AGT or NER, we constructed E. coli strains that

26

K. Nakano et al. / Mutation Research 815 (2017) 22–27

were doubly-deficient in the MMR and AGT or MMR and NER systems, and then repeated the mutation assays using these strains (Table 3). The mutant frequencies induced by all mutagens were lower in the AGT/NER-deficient mutS strain compared to the AGTor NER-deficient strains. Similar decreases in mutant frequencies were not observed in the corresponding mutL or mutH strains. Thus, in lieu of preventing mutation, the presence of a functional mutS allele appears to promote the frequency of mutation even in cells deficient in other repair pathways.

4. Discussion Various sites within DNA are substrates for modification by alkylating agents. Among the resulting alkylated products, O6alkylguanine, O4-alkylthymine and O2-alkylthymine residues are considered to be responsible for the observed mutation effects [17]. The repair pathway of O6-alkylguanine is well documented and includes contributions from our research [8,9]. However, the pathways employed by prokaryotes in repairing O4-alkylthymine remain unknown. In the present study, we assessed the repair of O4 -alkylthymine in E. coli by examining the rate of AT-to-GC transitions in the E. coli CC106 strain (and repair-deficient derivatives thereof) following treatment with alkylating agents. We presumed that the transitions should be induced by the formation of O4 -alkylthymine. Our results showed that AGT-deficiency elevated the AT-to-GC mutation rate in a manner similar to that observed with O6 -alkylguanine, in that AGT removed the methyl residue more effectively than the ethyl residue [9]. We found that NER was involved in the repair of ethylation but displayed little or no effect in the repair of methylation, in agreement with previous reports. In mammalian cells, it has been reported that O4 -alkylthymine is preferentially repaired by excision systems, and that MMR is also involved in a similar manner with respect to O6 -alkylguanine [18–21]. However, in bacteria, the involvement of MMR in the repair of O4-alkylthymine seemed to differ from cases involving O6 -alkylguanine. From our results, MMR-deficiency appeared to decrease the mutation rate induced by O4 -alkylthymine, as demonstrated in strains deficient for MMR alone or for MMR in combination with either AGT or NER. Previously, we showed (as expected) that the lack of MMR elevated the mutation rate driven by O6 -alkylguanine, as demonstrated in CC102-derived strains deficient for MMR alone or for MMR in combination with either AGT or NER [9]. Mutation rates were lower in mutS strains compared to wild-type, mutL or mutH strains in all experiments employing CC106-derivative single- or double-mutant strains. The MutS protein functions as an apparatus for recognizing DNA lesions [6]. Based on the results obtained in the present study, we hypothesize that MutS protein preferentially recognizes the O4 -alkylT:A base pair as a mismatch unlike the case with O4 -alkylT:G. When the O4 -alkylT:A base pair is formed in DNA, this pair would be recognized by MutS without replication, as shown in Fig. 3 (upper pathway, bold lines), and in MMR pathway the adenine residue is removed and guanine subsequently incorporated opposite to O4 -alkylT during repair. Therefore, the resulting post-replication mutation that occurs with O4 -alkylT:G might recognized by MutS at a lower frequency than that involving the O4 -alkylT:A base pair. If O4 -alkylT:A is replicated, O4-alkylT:G will be formed, as shown in Fig. 3 (lower pathway, thin lines). In this case, as MMR functions to a lesser degree than in the case with O4 -alkylT:A, mutation is induced. Pauly et al. showed that in wild-type E. coli, O4 -methylT is more mutagenic than O6 -methylG and O6 -ethylG, although in cells defective for both AGT and MMR the mutagenicity of O4 methylT was similar to that of O6 -methylG [22]. Feitsma et al. reported that MMR deficiency did not enhance the rate of muta-

Fig. 3. Proposed scheme for the repair pathway of O4 -alkylthymine and process leading to mutation.

tion by ENU [23]. Their results might support our data, where in wild-type O6 -alkylG is repaired efficiently with AGT and MMR and/or NER, whereas O4 -alkylT is repaired at a lower frequency due to the more frequent recognition of the O4 -alkylT:A pair by MutS than the O4 -alkylT:G pair. As a result, O6 -methylG is more mutagenic in an MMR-deficient strain, while O4 -alkylT is more mutagenic in a wild-type (MMR+ ) background. Georgiadis et al. suggested that O4 -alkylT:A has greater effects on DNA conformation than O4 -alkylT:G [24]. Swann and collaborators showed that the O4 -alkylT:G pair could retain the Watson-Crick alignment, while O4 -alkylT:A induced a “wobble” alignment [25]. As the structural change in DNA containing the O4 -alkylT:A pair appears to be large and flexible, MutS protein could conceivably recognize the O4 alkylT:A pair better than the O4 -alkylT:G pair. In a primer extension experiment, O4 -alkylT was shown to preferentially incorporate G but not A at the opposite site [26]. These reports are consistent with our finding that the O4 -alkylT:A pair might be better recognized as a defect in the DNA structure and hence become a target of MMR repair. Taken together, the presence of MutS seems to be mutagenic with the alkylation of thymine residues. It is difficult to account for the observation that mutant frequencies in mutL and mutH strains were higher than that in the mutS strain and similar to that in wild-type, except for ENU. We propose that the binding of MutS might recruit other systems where MMR-like functions operate even when MutL or MutH is lacking. Further investigations will be necessary, including efforts to determine which base pairs are preferentially bound by MutS. In order to confirm the involvement of MutS, a complementation test in which MutS is expressed in the mutS strain seems to be necessary. Nonetheless, we propose the possibility that the presence of a MMR system or MutS is mutagenic to some DNA lesions. Assuming that similar events occur in human cells, the results obtained in the present study provide important information for the design of anti-cancer drugs, given that various cancer cells are known to harbor defects in MMR. Acknowledgements This work was supported by a Grant-in-Aid for Challenging Exploratory Research (25670064) from the Japan Society for the Promotion of Science. References [1] N. Kondo, A. Takahashi, K. Ono, T. Ohnishi, DNA dam age induced by alkylating agents and repair pathway, J. Nucleic Acids 2010 (2010), http://dx. doi.org/10.4061/2010/543531. [2] T.S. Dexheimer, DNA repair pathway and mechanism, in: L.A. Mathews, S.M. Cabarcas, E. Hurt (Eds.), DNA Repair of Cancer Stem Cells, 2013, pp. 19–32. [3] S. Mitra, MGMT: a personal perspective, DNA Repair (Amst.) 6 (2007) 1064–1070.

K. Nakano et al. / Mutation Research 815 (2017) 22–27 [4] B. Kaina, M. Christmann, S. Naumann, W.P. Roos, MGMT: Key node in the battle against genotoxicity carcinogenicity and apoptosis induced by alkylating agents, DNA Repair (Amst.) 6 (2007) 1079–1099. [5] T. Nouspikel, Nucleotide excision repair: variations on versatility, Cell. Mol. Live Sci. 66 (2009) 994–1009. [6] T.A. Kunkel, D.A. Erie, DNA mismatch repair, Annu. Rev. Biochem. 74 (2005) 681–710. [7] L.J. Rasmussen, L. Samson, The Escherichia coli MutS DNA mismatch binding protein specifically binds O6-methylguanine DNA lesions, Carcinogenesis 17 (1996) 2085–2088. [8] K. Taira, S. Nakamura, K. Nakano, D. Maehara, K. Okamoto, S. Arimoto, D. Loakes, L. Worth, R.M. Schaaper, K. Seio, M. Sekine, K. Negishi, T. Negishi, Binding of MutS protein to oligonucleotides containing a methylated- or an ethylated guanine residue, and correlation with mutation frequency, Mutat. Res. 640 (2008) 107–112. [9] K. Taira, S. Kaneto, K. Nakano, S. Watanabe, E. Takahashi, S. Arimoto, K. Okamoto, R.M. Schaaper, K. Negishi, T. Negishi, Distinct pathways for repairing mutagenic lesions induced by methylating and ethylating agents, Mutagenesis 28 (2013) 341–350. [10] Q. Fang, S. Kanugula, J.L. Tubbs, J.A. Tainer, A.E. Pegg, Repair of O4 -alkylthymine by O6 -alkylguanine-DNA- alkyltransferases, J. Biol. Chem. 285 (2010) 8185–8195. [11] R.M. Schaaper, B.N. Danforth, B.W. Glickman, Rapid repeated cloning of mutant lac repressor genes, Gene 39 (1985) 181–189. [12] C.G. Cupples, H. Miller, A set of lacZ mutations in Escherichia coli that allow rapid detection of each of the six base substitutions, Proc. Natl. Acad. Sci. U. S. A. 86 (1989) 5345–5349. [13] K. Negishi, D. Loakes, R.M. Schaaper, Saturation of DNA mismatch repair and error catastrophe by a base analogue in Escherichia coli, Genetics 161 (2002) 1363–1371. [14] P.K. Lemotte, G.C. Walker, Induction and autoregulation of ada: a positively acting element regulating the response of Escherichia coli K-12 to methylating agents, J. Bacteriol. 161 (1985) 888–895. [15] K. Takano, T. Nakamura, M. Sekiguchi, Roles of two types of O6-methylguanine-DNA methyltransferases in DNA repair, Mutat. Res. 254 (1991) 37–44.

27

[16] E. Takahashi, K. Okamoto, S. Arimoto, H. Yamanaka, T. Negishi, Involvement of the drug efflux protein TolC in mutagenicity induced by MNNG or Trp-P-2, Mutat. Res. 605 (2006) 42–50. [17] D.K. La, J.A. Swenberg, DNA adducts: biological markers of exposure and potential applications to risk assessment, Mutat. Res. 365 (1996) 129–146. [18] A.A. Wani, G. Wani, S.M. D’Amborosio, Repair of O4 -alkylthymine damage in human cells, Basic Life Sci. 53 (1990) 417–435. [19] C.W. Op het Veld, S. van Hees-Stuivenberg, A.A. van Zeeland, J.G. Jansen, Effect of nucleotide excision repair on hprt gene mutations in rodent cells exposed to DNA ethylating agents, Mutagenesis 12 (1997) 417–424. [20] L. Samson, S. Han, J.C. Marquis, L.J. Rasmussen, Mammalian DNA repair methyltransferases shield O4 MeT from nucleotide excision repair, Carcinogenesis 18 (1997) 919–924. [21] N. Claij, A. van der Wal, M. Dekker, L. Jansen, H. te Riele, DNA mismatch repair deficiency stimulates N-ethyl-N-nitrosourea-induced mutagenesis and lymphomagenesis, Cancer Res. 63 (2003) 2061–2066. [22] G.T. Pauly, S.H. Hughes, R.C. Moschel, Comparison of mutagenesis by O6 -methyl- and O6 -ethylguanine and O4 -methylthymine in Escherichia coli using double-stranded and gapped plasmids, Carcinogenesis 19 (1998) 457–461. [23] H. Feitsma, E. de Bruijn, J. van de Belt, I.J. Nijman, E. Cuppen, Mismatch repair deficiency does not enhance ENU mutagenesis in the zebrafish germ line, Mutagenesis 23 (2008) 325–329. [24] P. Georgiadis, Y.-Z. Xu, P.F. Swann, Nitrosamine-induced cancer: O4 -alkylthymine produces sites of DNA hyperflexibility, Biochemistry 30 (1991) 11725–11732. [25] P.F. Swann, Why do O6 -alkylguanine and O4 -alkylthymine miscode? The relationship between the structure of DNA containing O6 -alkylguanine and O4 -alkylthymine and the mutagenic properties of these bases, Mutat. Res. 233 (1990) 81–94. [26] M.K. Dosanjh, P. Menichini, R. Eritja, B. Singer, Both O4 -methylthymine and O4 -ethylthymine preferentially form alkyl T-G pairs that do not block in vitro replication in a defined sequence, Carcinogenesis 14 (1993) 1915–1919.