DNA Damage and Repair Methods.

AP sites were measured by the aldehyde reactive probe-slot-blot method as described (21, 22). Heat labile sites were determined as follows: 10 μg DNA extracted from control cells or cells exposed to MMS was incubated at 100°C for 20 min in PBS. The AP sites induced by this heat incubation were measured by the aldehyde reactive probe-slot-blot method. For QPCR, high molecular weight DNA was isolated, and quantitative amplification of the β-globin target sequence was performed as described (23). Single-cell gel analysis (comet assay) was conducted as described (24).

Determination of Mutant Frequency at λ cII.

Cells (40,000) were added to each of six wells in a 6-well dish. After 24 h, the media were replaced with growth media alone or media supplemented with MMS or MNNG, and the cells were incubated at 34°C for 1 h. Alternatively, the media were removed and the cells were washed twice with PBS. After removal of the PBS, the cells were exposed to UV (15 J/m²) essentially as described (15). After genotoxicant exposure (MMS, MNNG, or UV), cells were washed once with growth media, refed with growth media, and incubated until the cells reached confluence. The cells were then seeded into 150-mm dishes and grown to confluence again. Normal culture time after exposure was 5–8 days. Genomic DNA was extracted from the cells (5–10 × 10⁶ cells per sample) with Recoverease DNA Isolation Kits (Stratagene) and stored at 4°C. λ-LIZ shuttle vectors were rescued from genomic DNA with Transpack packaging extract according to the manufacturer's protocol (Stratagene). Plaques with mutations in the cII gene were determined essentially as described (19). Detailed procedures are described in additional Materials and Methods, which are published as supporting information on the PNAS web site, www.pnas.org. Briefly, mutant frequencies were determined by using genomic DNA isolated from at least three separate cell samples. For each DNA sample, a minimum of 300,000 plaques were analyzed from at least three packaging reactions to ensure an accurate mutant frequency determination.

Spontaneous and MMS-Induced λ cII Mutation Spectra.

The spontaneous and MMS-induced mutation spectra were determined at the cII locus in wild-type (MB36.3) or β-pol null (MB38Δ4) embryonic fibroblasts with a modified fluctuation test. Forty-eight individual cell populations (24 from each cell line) were established by plating 1,000 cells into each well of two 24-well dishes. Cell cultures were then expanded and DNA was isolated as described above. If the same mutation was observed more than once in each of the cell populations, then that mutation was counted only once in the spectra.

Status of AP Sites in Genomic DNA After Methylation.

MEFs deleted in the β-pol gene are known to be hypersensitive to the cytotoxic effects of MMS (5, 6, 15, 18, 20, 37), and this hypersensitivity is linked to a failure to remove the 5′ dRP group in the BER intermediate (20). However, it is not known whether a β-pol deficiency can impact levels of earlier intermediates in the repair mechanism, such as AP sites. We found that AP sites did not increase in DNA from either wild-type or β-pol null cells immediately after exposure to MMS (Fig. ​(Fig.11c). In addition, when assayed after a repair incubation, i.e., 4 h after exposure, there was still no difference in AP sites in genomic DNA from MMS-treated or untreated wild-type and β-pol null cells (not shown). As expected, MMS exposure in both cell lines gave rise to methylated bases, as revealed by a large increase in measured AP sites after heat treatment of the purified DNA (Fig. ​(Fig.11c). These results indicate that for both cell lines most of the MMS-induced methylated purine bases remained in genomic DNA (38) rather than being repaired or accumulated as AP sites.

Spontaneous Mutations at the λ cII Gene.

DNA lesions and the absence of DNA repair functions can result in mutations and/or unique mutation spectra. To evaluate these possibilities in wild-type and β-pol null cells, mutations in the chromosome-borne λ cII transgene were quantified, and independent mutants were selected and characterized by sequencing. Wild-type and β-pol null cells had spontaneous mutant frequencies of 1.4 ± 0.2 × 10⁻⁴ and 2.3 ± 0.6 × 10⁻⁴, respectively (Fig. ​(Fig.22a). A total of 130 and 102 independent spontaneous mutations from wild-type and β-pol null cells, respectively, were sequenced. The frequencies of mutations observed in these spontaneous mutant spectra are summarized in Table ​Table1.1. The relative spontaneous mutation frequencies of G:C → A:T and A:T → G:C transitions were comparable in both cell lines. The frequencies of all other mutations were elevated in β-pol null cells by between 1.8- and 3-fold.

Open in a separate window

Figure 2

MMS-induced, UV-induced, and MNNG-induced mutations at the λ cII gene in β-pol null cells. (a) Spontaneous (filled bar) and MMS-induced (open bar) mutant frequencies at λ cII in wild-type or β-pol null MEFs. Bars represent the SD from three experiments. Mutant frequencies were calculated after analysis of 2.2 × 10⁶ plaque-forming units (pfu) (wild type; spontaneous), 1.6 × 10⁶ pfu (β-pol null; spontaneous), 1.1 × 10⁶ pfu (wild type; +1 mM MMS), and 1.0 × 10⁶ pfu (β-pol null; +1 mM MMS), respectively. (b) Spontaneous (filled bar) and UV-induced (open bar) mutant frequencies at λ cII in wild-type or β-pol null MEFs. Mutant frequencies were calculated after analysis of 4.7 × 10⁵ pfu (wild type; spontaneous), 3.9 × 10⁵ pfu (β-pol null; spontaneous), 4.5 × 10⁵ pfu (wild type; +UV), and 4.2 × 10⁵ pfu (β-pol null; +UV), respectively. (c) Spontaneous (filled bar) and MNNG-induced (open bar) mutant frequencies at λ cII in wild-type or β-pol null MEFs. Mutant frequencies were calculated after analysis of 4.9 × 10⁵ pfu (wild type; spontaneous), 6.1 × 10⁵ pfu (β-pol null; spontaneous), 5.3 × 10⁵ pfu (wild type; +6 μM MNNG), and 6.7 × 10⁵ pfu (β-pol null; +6 μM MNNG), respectively. (d) Induced mutant frequencies at λ cII in wild-type (filled bar) or β-pol null (open bar) MEFs. The induced mutant frequency value was determined by subtracting the value for the spontaneous mutant frequency from the value for the indicated genotoxicant-induced mutant frequency as shown in a–c.

Mutation Spectral Analysis for the λ cII Gene After Methylation.

MMS was mutagenic in β-pol null cells, inducing an increase in mutant frequency when comparing MMS-exposed wild-type and β-pol null cells (Fig. ​(Fig.22a). This increase was observed for MMS treatment at 1 mM, as well as for a MMS dose of equivalent (75%) cytotoxicity (not shown). A β-pol-dependent increase in mutant frequency was not observed when cells were exposed to either UV or MNNG; wild-type and β-pol null cells demonstrated identical UV-induced λ cII mutant frequency (Fig. ​(Fig.22b) and similar results were obtained for MNNG-induced mutant frequency (Fig. ​(Fig.22c).

The spectra for the MMS-induced mutations in wild-type and β-pol null cells are summarized in Table ​Table2.2. Wild-type cells showed little or no increase in cII mutation frequency as a function of MMS treatment (see Table 3, which is published as supporting information on the PNAS web site). In contrast, MMS-treated β-pol null cells had higher mutation frequencies for all base substitution mutations than MMS-treated wild-type cells (Table ​(Table2).2). There are several options for analysis of these mutational data, e.g., plus versus minus MMS, plus versus minus β-pol, changes within class of mutation, and changes as a function of DNA sequence in the cII gene (see Tables 3 and 4, which are published as supporting information on the PNAS web site). As seen in Table ​Table2,2, β-pol null cells showed greater MMS-induced mutation frequency than wild-type cells for both transitions and transversions. Yet, some of the increase in transversions in MMS-treated β-pol null cells (Table ​(Table2)2) could have been caused by the modest increase (≈2-fold) in spontaneous transversions seen in these cells. In contrast, none of the increase in A:T to G:C transitions seen in MMS-treated β-pol null cells (Table ​(Table2)2) appears to be explained by any increase in spontaneous mutations of this type (see Table 3). Finally, the mutation spectra contained six mutational hotspots, defined as sites where five or more mutations were detected in each cell type, but these mutations did not appear to depend on β-pol expression or MMS treatment (data not shown; see Fig. 5, which is published as supporting information on the PNAS web site).

Mutagenesis After MMS Exposure.

Our comparison of spontaneous mutant frequency in wild-type and β-pol null cells demonstrated a weak spontaneous mutator phenotype for the β-pol null cells (Table ​(Table11 and Fig. ​Fig.2).2). Mutant frequency in the MMS-treated β-pol null cells was higher than that in untreated β-pol null cells and much higher than in MMS-treated wild-type cells (Table ​(Table2).2). These results indicate that these BER-deficient MEF cells are hypermutable after MMS exposure. A similar observation had been made for the XRCC1-defective cell line EM-C11, in that mutations at the hprt locus increased severalfold after MMS treatment, as compared with a wild-type Chinese hamster ovary cell line (40). These XRCC1-defective cells, like the β-pol null cells, are hypersensitive to MMS and appear to be deficient in BER.

Sequencing of individual mutants revealed that the frequency of A:T → G:C mutations increased 9-fold and all other base substitution mutations were elevated from 3.2- to 7.1-fold in β-pol null cells after MMS treatment, as compared with wild-type cells (Table ​(Table2).2). The frequency of G:C → T:A and A:T → T:A mutations were 4.5- and 5-fold higher, respectively, in MMS-treated β-pol null cells (Table ​(Table2)2) than in wild-type cells. We note that these two base substitution mutations were also found to predominate in AAG null mice after exposure to MMS (41).

Methylation-Induced Mutagenesis as a Function of β-Pol Status.

Studies of several transgenic and endogenous genes in wild-type mice or cell lines have shown that MMS does not increase mutant frequency (40, 42–45). However, in repair-deficient rodent cells, MMS can be mutagenic (40, 42, 46, 47). This finding, along with the results reported here, suggested the use of the S_N2 methylating agent MMS as the DNA damaging agent of choice for the mutagenesis studies described. Except as a reference, S_N1 methylating agents such as MNNG, producing relatively more O⁶-methyl guanine (O⁶meG), were avoided in this study because of their strong tendency to produce G → A mutations and also because O⁶-meG is repaired by the methyl transferase mechanism, rather than by β-pol-dependent BER.

MMS exposure of cells carrying a genetic deficiency in AAG would be expected to give rise to methylation-induced DNA mutations resulting from replicative miscoding by MMS-induced lesions (41). We presented evidence here that a deficiency in β-pol, a downstream enzyme in the BER pathway, also imparts a methylation-induced increase in DNA mutations. Further, our evidence obtained with cells expressing different levels of AAG suggests that these mutations were caused by mutagenic AAG substrates. These results suggest a defect in processing methylated purine bases caused by the deficiency in the downstream BER enzyme β-pol.

Mechanism of Increase in MMS-Induced Mutants.

MMS is known to produce a number of methylated DNA lesions in addition to the predominant lesion N7-methylguanine, which is not miscoding during DNA replication. Several of these minor lesions are known to be capable of miscoding during DNA synthesis and could be repaired slower in β-pol null cells than in the wild-type cells, leading to replication-based mutations in the null cells, as noted above. To evaluate this idea, we examined the effect of overexpression of mammalian AAG. Such overexpression of AAG was found to suppress the increase in MMS-induced mutant frequency in the β-pol null cells (Fig. ​(Fig.33c), suggesting that substrates for this enzyme are responsible for the increased mutagenesis. We also found that MMS was mutagenic in β-pol null/AAG null fibroblasts (Fig. ​(Fig.4),4), again suggesting that AAG substrates left unrepaired are mutagenic in this system. Thus, it is likely that quantitatively minor methylated bases miscode with incoming dNTPs during DNA replication and that the β-pol deficiency somehow influences the processing of these mutagenic AAG substrates. In addition to miscoding, unrepaired methylpurines may undergo depurination, leading to AP sites that are mutagenic in the β-pol null cell. These mechanisms could explain an increase in mutant frequency, such as that observed in MMS-treated β-pol null cells.

We thank Dr. Roel Schaaper (Laboratory of Molecular Genetics, National Institute of Environmental Health Sciences) and Dr. Barry Gold (University of Nebraska Medical Center, Omaha) for valuable discussions and critical evaluation of the manuscript. We also thank Drs. Raymond Tice and Barbara Shane (Integrated Laboratory Systems, Durham, NC) for assistance with the Comet and cII mutation assays, respectively. L.S. is the Ellison American Cancer Society Research Professor, and this research was supported by RO1s CA75576 and CA55042.