Base Excision Repair Intermediates Induce p53-independent Cytotoxic and Genotoxic Responses*

DNA alkylation damage is primarily repaired by the base excision repair (BER) machinery in mammalian cells. In repair of the N-alkylated purine base lesion, for example, alkyl adenine DNA glycosylase (Aag) recognizes and removes the base, and DNA polymerase β (β-pol) contributes the gap tailoring and DNA synthesis steps. It is the loss of β-pol-mediated 5′-deoxyribose phosphate removal that renders mouse fibroblasts alkylation-hypersensitive. Here we report that the hypersensitivity of β-pol-deficient cells after methyl methanesulfonate-induced alkylation damage is wholly dependent upon glycosylase-mediated initiation of repair, indicating that alkylated base lesions themselves are tolerated in these cells and demonstrate that β-pol protects against accumulation of toxic BER intermediates. Further, we find that these intermediates are initially tolerated in vivo by a second repair pathway, homologous recombination, inducing an increase in sister chromatid exchange events. If left unresolved, these BER intermediates trigger a rapid block in DNA synthesis and cytotoxicity. Surprisingly, both the cytotoxic and genotoxic signals are independent of both the p53 response and mismatch DNA repair pathways, demonstrating that p53 is not required for a functional BER pathway, that the observed damage response is not part of the p53 response network, and that the BER intermediate-induced cytotoxic and genotoxic effects are distinct from the mechanism engaged in response to mismatch repair signaling. These studies demonstrate that, although base damage is repaired by the BER pathway, incomplete BER intermediates are shuttled into the homologous recombination pathway, suggesting possible coordination between BER and the recombination machinery.

environmental exposure (1,2). In many cases, these genetic changes induce a "mutator" phenotype due to sequence changes in DNA repair or DNA damage checkpoint genes, among others (3). The importance of maintaining the structural and informational integrity of the genome is further highlighted by the plethora of DNA repair and DNA damage checkpoint pathways (2). Consequently, it is important to understand the diverse cellular systems for DNA damage repair available for protection from an enormous array of DNA lesions (4).
The base excision repair (BER) 1 pathway is considered the predominant DNA repair system in mammalian cells for eliminating small DNA lesions generated either exogenously or endogenously at DNA bases (4 -6). Such DNA damage can be caused by exposure to environmental agents or by normal cellular metabolic processes that produce alkylating molecules, reactive oxygen species, and other reactive metabolites capable of modifying DNA. In the mammalian BER pathway, the damaged base residue is removed by a lesion-specific DNA glycosylase. Subsequently, the resulting abasic site is recognized by apurinic/apyrimidinic endonuclease (APendo), which incises the damaged strand, leaving a single-nucleotide gap with 3Ј-OH and 5Ј-deoxyribose phosphate (5Ј-dRP) groups at the margins. A DNA polymerase ␤ (␤-pol)-mediated DNA synthesis step extends from the 3Ј-OH (7)(8)(9), and the 5Ј-dRP group is removed by the 5Ј-dRP lyase activity of ␤-pol (10 -14). DNA ligase I or a complex of DNA ligase III and x-ray cross-complementing factor 1 conducts the final, nick sealing, step in the pathway. In addition, several proteins have been observed to form functional partnerships with these BER proteins, including p53, poly(ADP-ribose) polymerase, p300, and proliferating cell nuclear antigen (15)(16)(17)(18).
In vitro studies suggested a role for ␤-pol in varying types of DNA repair (19). Transgenic mice with a homozygous null mutation in the ␤-pol gene are nonviable after birth, thereby preventing studies on the in vivo function of ␤-pol (20). However, we utilized ␤-pol (ϩ/Ϫ) mice to establish embryonic fibroblast (MEF) cell lines homozygous for a null mutation in the ␤-pol gene (21). ␤-pol (Ϫ/Ϫ) cells are normal in viability and growth characteristics but are more sensitive to monofunctional alkylating agents such as methyl methanesulfonate (MMS) than wild type cells (21).
␤-pol possesses both polymerase and 5Ј-dRP lyase activity, but we found previously that only the 5Ј-dRP lyase activity of ␤-pol is essential for resistance to MMS and that ␤-pol appears to be the major, if not only, protein capable of efficiently removing 5Ј-dRP groups formed as BER intermediates (22). However, these earlier studies did not determine the source of these repair intermediates, and it was not yet determined if the 5Ј-dRP groups were formed as a result of BER initiation by a lesion-specific glycosylase. Further, no study to date has begun to unravel the mechanism of cytoxicity induced by the BER intermediate 5Ј-dRP.
In this report, we describe the development and characterization of a series of single and double knockout cell lines and mouse models to detail the cytotoxic and genotoxic effects of both alkylated base lesions and the resultant BER intermediate 5Ј-dRP. We show that 5Ј-dRP formation (the lethal lesion causing the MMS-sensitive phenotype of ␤-pol (Ϫ/Ϫ) cells) is critically dependent on the alkyladenine DNA glycosylase (Aag), known to remove 3-methyladenine (3-MeA), 7-methylguanine, and other alkylated bases from DNA. Surprisingly, the unrepaired N-alkylated purine base lesions are well tolerated in these cells. Interestingly, our results suggest that BER intermediates are associated with an increase in homologous recombination and eventually trigger a block in DNA synthesis. Both the cytotoxic and genotoxic effects in these cells are independent of p53 and the mismatch DNA repair pathway. These studies suggest a possible coordination between BER and the recombination machinery.
Cytotoxicity Assays-Cytotoxicity was determined by growth inhibition assays as described previously (22). Briefly, cells were exposed to the DNA-damaging agent, MMS or N-methyl-NЈ-nitro-N-nitrosoguanidine (MNNG), 24 h after seeding (40,000 cells/well) in 6-well dishes. Cells (triplicate wells) were then treated for 1 h at 37°C in a 10% CO 2 incubator with serial dilutions of mutagen in growth medium. For each experiment, cells were counted after 3 days (untreated cells are found to be Ͻ80% confluent). Cell numbers will be determined by a cell lysis protocol. Results are the mean of at least two separate experiments.
Percentage of control growth is as follows: (number of treated cells)/ (number of control cells)⅐100.
Flow Cytometric Cell Cycle Analysis-Cell cycle and DNA synthesis was analyzed simultaneously by staining with propidium iodide and incorporation of bromodeoxyuridine (BrdUrd) with some modifications (30). Briefly, cells were seeded in 100-mm dishes at a density of 327,000 cells/dish. (equivalent to the cell density used in the cytotoxicity experiments). The following day, cells were treated for 1 h with MMS. At 2 or 8 h after MMS treatment (as indicated), 10 M BrdUrd (Sigma) was added to the dishes for 30 min to pulse-label the cells. Cells were then washed with PBS, harvested by trypsinization, and washed a second time with PBS. The cell pellet obtained after centrifugation was resuspended in 100 l of cold PBS, and the cells were dropped slowly into 70% ethanol and allowed to fix at 4°C overnight. The samples were washed, suspended in 2 N HCl containing 0.5% Triton X-100, and incubated for 30 min at room temperature to denature the DNA. The cell samples were pelleted, resuspended in 0.1 M sodium borate (pH 8.5) to neutralize the acid and then washed with PBS. Cells were then incubated at 4°C overnight with 20 l of anti-BrdUrd-fluorescein isothiocyanate-conjugated antibody (BD Biosciences) in PBS containing 0.5% Tween 20 and 1% bovine serum albumin and 5 l of 10 mg/ml RNase (Sigma) stock solution. The following day, the cells were pelleted, washed with PBS, and resuspended in 1 ml of PBS containing 5 g/ml propidium iodide (Sigma). The samples were analyzed by flow cytometry using Cell Quest software (BD Biosciences). Cell cycle populations are designated G 0 /G 1 (2 N DNA content with no BrdUrd incorporation), S (variable DNA content with BrdUrd incorporation), and G 2 /M (4 N DNA content without BrdUrd incorporation).
Sister Chromatid Exchange Assay-For sister chomatid exchange (SCE) measurements, 1 ϫ 10 6 cells were seeded onto 75-cm 2 tissue culture dishes 8 h before drug treatment. The cells were treated with MMS in complete McCoy's 5A medium (Invitrogen) supplemented with 10 M BrdUrd for 1 h at 37°C in 5% CO 2 . Following drug treatment, cells were incubated in McCoy's 5A medium supplemented with 10 M BrdUrd for 20 h. Colcemid (0.1 g/ml; Invitrogen) was included for the last 2 h of incubation, and the cells were subsequently harvested by mitotic shakeoff, resuspended, and incubated for 15 min at 37°C in hypotonic solution (0.2% potassium chloride, 0.2% sodium citrate, and 10% fetal bovine serum) and then fixed in Carnoy's solution. To produce "harlequin" chro-FIG. 1. a, genotype analysis. Genomic DNA from each cell line was isolated and analyzed by allele-specific PCR (specific to either the wild type (wt) or null allele) to confirm the ␤-pol and Aag genotype. For ␤-pol, the appearance of the ethidium bromide-stained band indicates the presence of either the wt or null allele, as indicated. For Aag, the upper band corresponds to the presence of the null allele, whereas the lower band indicates the presence of the wt allele, as indicated by the arrows. b, mRNA expression. Total RNA was isolated from each cell line and analyzed by reverse transcriptase-PCR for expression of mouse ␤-pol, mouse Aag, and control mouse actin mRNA. mosomes, a modified fluorescence plus Giemsa technique was used (31). Slides were stained in Hoechst 33258 (5 g/ml) for 20 min, mounted in 0.067 M Sorensen's buffer with a coverslip, and exposed to a General Electric 15-watt black light bulb at 65°C for 20 min. Slides were then heated at 65°C in 20ϫ SSC for 20 min, rinsed, and stained in a 5% Giesma solution in 0.067 M Sorensen's buffer. Twenty second-division metaphase spreads were counted per data point.

Aag Is Required to Initiate BER of MMS-and MNNG-induced
Lesions to Generate the Toxic 5Ј-dRP Repair Intermediate-To study the cytotoxicity and genotoxicity of both alkylated DNA bases and BER repair intermediates, we developed a set of isogenic MEF cell lines with homozygous null mutations in the Aag gene, in the ␤-pol gene or in both the Aag and ␤-pol genes (Fig. 1). To ensure that no significant backup or compensatory enzymatic activity was present, extracts from each were analyzed for Aag-specific activity (Fig. 2a) and ␤-pol-specific BER gap-filling activity (Fig. 2b). Wild type and ␤-pol (Ϫ/Ϫ) cells showed the expected levels of Aag activity, with little or no Aag-specific activity in the Aag (Ϫ/Ϫ) and ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells (Fig. 2a), whereas each had similar levels of uracil-DNA glycosylase activity (data not shown). Conversely, the wild type and Aag (Ϫ/Ϫ) cells reported robust ␤-pol-specific BER activity with little or no activity seen in extracts from ␤-pol (Ϫ/Ϫ) or ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells (Fig. 2b).
The cytotoxicity of methylating agent-induced DNA base damage is increased in both Aag (Ϫ/Ϫ) embryonic stem (ES) cells (32) and ␤-pol (Ϫ/Ϫ) MEFs (21), implying that both methylated bases and BER intermediates are cytotoxic. We therefore compared the cytotoxicity of MMS in wild type and isogenic ␤-pol (Ϫ/Ϫ) (Fig. 2c) and Aag (Ϫ/Ϫ) MEFs (Fig. 2d). Consistent with the known phenotype of these cells and the hypothesis that the 5Ј-dRP lyase activity of ␤-pol is essential for reversing MMS-induced cytotoxicity (22), ␤-pol (Ϫ/Ϫ) MEFs were more sensitive to MMS than wild type MEFs, as expected (Fig. 2c). We next tested Aag (Ϫ/Ϫ) MEFs to directly evaluate the cytotoxicity of Aag substrates, since MMS causes a significant buildup of the principal Aag substrate 3-MeA in Aag (Ϫ/Ϫ) ES cells (33,34). Surprisingly, Aag substrates generated by MMS exposure conferred no increase in cytotoxicity (Fig. 2d). This finding suggests that lesion avoidance mechanisms are more robust in MEFs than ES cells (see below). The lack of MMS-induced cytotoxicity allowed us to characterize the cytotoxic and genotoxic effects of BER intermediates in these cells.
Next, we compared ␤-pol (Ϫ/Ϫ) cells with ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells for MMS-induced cytotoxicity (Fig. 2e). We found that the MMS-sensitive phenotype of ␤-pol (Ϫ/Ϫ) cells is critically dependent on Aag activity, because ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells were much more resistant to MMS than ␤-pol (Ϫ/Ϫ) cells (Fig. 2e). ␤-pol (Ϫ/Ϫ) MEFs have also been reported to be hypersensitive to alkylation damage caused by MNNG (21). Using the same set of four isogenic cell lines, we compared the cytotoxicity profile induced by MNNG to determine whether this induced cytotoxic response is also mediated by repair intermediates initiated by Aag. Identical to the results for MMS treatment, ␤-pol (Ϫ/Ϫ) MEFs but not Aag (Ϫ/Ϫ) MEFs were hypersensitive to MNNG, and the genetic loss of Aag prevented the formation of repair intermediates that cause the cytotoxicity in the absence of ␤-pol (Fig. 3). It should be noted that all four cell lines exhibited similar cytotoxic profiles following UV exposure (data not shown). Therefore, both MMS and MNNG generate Aag substrates in DNA; if left unrepaired, these Aag substrates are not cytotoxic in these cells, yet accumulation of the 5Ј-dRP repair intermediate (formed following Aag-mediated lesion removal and APendo hydrolysis of the abasic site) induces a dramatic cytotoxic response.
To further demonstrate that Aag-mediated lesion removal contributes to the generation of cytotoxic ␤-pol substrates, we transduced wild type, ␤-pol (Ϫ/Ϫ) , Aag (Ϫ/Ϫ) , and ␤-pol (Ϫ/Ϫ) / Aag (Ϫ/Ϫ) cells with the MMP retrovirus expressing either GFP or human alkyladenine DNA glycosylase (hAAG). To ensure uniform hAAG expression levels, individual clones were isolated and selected for similar AAG activity. As shown in Fig.  2a, all hAAG overexpressing lines had activity levels approaching 5 times that seen in wild type cells. Consistent with the model that Aag-mediated lesion removal is required for the formation of cytotoxic ␤-pol substrates, hAAG overexpression increased the MMS-induced cytotoxicity in a ␤-pol (Ϫ/Ϫ) background. As shown in Fig. 2f, ␤-pol (Ϫ/Ϫ) [ϩhAAG] cells were significantly more sensitive to MMS treatment than the control ␤-pol (Ϫ/Ϫ) [ϩGFP] cells. Further, hAAG overexpression in ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells restored the MMS-induced hypersensitivity that is the hallmark of the ␤-pol (Ϫ/Ϫ) phenotype (21, 22) (Fig. 2g). mediated by a similar phenomenon. Clearly, spontaneous generation of Aag substrates does not lead to lethality in mice, since Aag (Ϫ/Ϫ) mice are viable and are without an obvious phenotype (23,37). We therefore bred a colony of ␤-pol (ϩ/Ϫ) / Aag (Ϫ/Ϫ) mice to determine if genetic loss of Aag would rescue the ␤-pol (Ϫ/Ϫ) -associated lethality. As shown in Table I, the Aag (Ϫ/Ϫ) mutation did not rescue the ␤-pol (Ϫ/Ϫ) lethality. At a minimum, this reveals that the ␤-pol (Ϫ/Ϫ) lethal phenotype is not due to Aag-initiated repair intermediates. Similar studies with APendo (ϩ/Ϫ) /Aag (Ϫ/Ϫ) mice suggest that the APendo (Ϫ/Ϫ) lethal phenotype is also not due to Aag-initiated repair intermediates. It remains to be determined whether genetic loss of any or all of the damage-specific DNA glycosylases (38) would result in genetic rescue of the ␤-pol (Ϫ/Ϫ) or the APendo (Ϫ/Ϫ) lethal phenotypes in mice (20,24,35,36). The BER Intermediate 5Ј-dRP Induces a Rapid and Severe Block to Replication-MMS has long been used to study DNA damage-induced cell cycle checkpoints and inhibition of DNA replication (39,40). For example, in recent studies with wild type strains of budding yeast, MMS-induced DNA damage was shown to reduce the rate of replication fork progression (39). By examining the effect of MMS on S-phase-associated DNA synthesis in the isogenic cell lines described above, we could dissect the impact of MMS-induced base damage versus 5Ј-dRP BER intermediates on DNA replication. Using a combination of BrdUrd incorporation, propidium iodide staining, and flow cytometry, we discovered that both MMS-induced base damage (accumulated in Aag (Ϫ/Ϫ) and ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells) and MMS-induced BER intermediates (accumulated in ␤-pol (Ϫ/Ϫ) cells) blocked DNA synthesis (Fig. 4). However, the inhibition was both modest and transient in the Aag (Ϫ/Ϫ) and ␤-pol (Ϫ/Ϫ) / Aag (Ϫ/Ϫ) cells compared with the ␤-pol (Ϫ/Ϫ) cells, where the block was both severe and persistent. Fig. 4 shows that 3-MeA (plus other Aag substrates) presents a mild replication block visible at 2 h after MMS exposure but that synthesis fully recovers by 8 h. In contrast, 5Ј-dRP BER intermediates severely block replication by 2 h, and the block is almost complete by 8 h. Thus, it would seem that the alkylated bases acted upon by Aag can ultimately be tolerated or avoided in these cells, presumably by the action of mechanisms such as DNA lesion bypass or recombination (39,40); such avoidance would account for the fact that 3-MeA lesions are not toxic to these cells (Figs. 2-4). Apparently, 5Ј-dRP lesions are not tolerated in these cells and thus block replication and lead to cell death (Figs. 2-4).
Cytotoxicity and Genotoxicity of 5Ј-dRP Is Independent of p53 and the Mismatch DNA Repair Pathway-A hallmark response to DNA damage is the activation of the tumor suppressor gene p53 and the subsequent induction of the p53 network (30,(41)(42)(43). Stabilized p53 participates in either halting cell cycle progression to allow completion of DNA repair (30,(41)(42)(43) or activating apoptotic programmed cell death (30,(41)(42)(43). However, the cells described here are SV40 T-antigen-transformed MEFs and hence are p53-defective (44), suggesting that the biological effects of 5Ј-dRP DNA lesions are p53-independent. To test this further, we prepared p53 (Ϫ/Ϫ) and ␤-pol (Ϫ/Ϫ) / p53 (Ϫ/Ϫ) primary MEFs and analyzed each for MMS-induced cytotoxicity. As shown in Fig. 5a, cell death induced by 5Ј-dRP lesions is clearly not dependent on p53 function, since the loss of ␤-pol in a p53 (Ϫ/Ϫ) background still renders the cells hypersensitive to MMS. Likewise, the 5Ј-dRP-induced block to DNA replication also appears to be a p53-independent phenomenon (Fig. 5b). Interestingly, p53 has been implicated directly in alkylation damage repair by stimulating BER through its interaction with ␤-pol and APendo (18). However, the results shown here indicate that whereas p53 may stimulate BER, it is not required for BER function. In addition, we have previously demonstrated that the mismatch DNA repair pathway recognizes alkylated bases (O 6 -MeG) and that such recognition initiates an apoptotic response (45). We developed mismatch DNA repair-deficient MEFs (PMS-2 (Ϫ/Ϫ) and ␤-pol (Ϫ/Ϫ) /PMS-2 (Ϫ/Ϫ) cells) to determine whether the 5Ј-dRP-induced cytotoxicity is related to this mismatch repair-dependent apoptotic repair mechanism. As shown in Fig. 6, the MMS-mediated hypersensitivity of ␤-pol (Ϫ/Ϫ) cells is not altered by a deficiency in the mismatch repair gene PMS-2 (Fig. 6).
5Ј-dRP Lesions Are Initially Tolerated by Homologous Recombination-DNA damage-induced arrest of the replication fork may be overcome by homologous recombination mechanisms of lesion avoidance (46,47). Since ␤-pol substrates or BER intermediates (5Ј-dRP) block replication (Fig. 4), we determined whether these BER intermediates can also induce recombination. At minimally toxic doses of MMS (up to 0.2 mM), we compared the genotoxic effect of both methylated base damage and BER intermediates by measuring the induction of SCEs. As shown in Figs. 7 and 8, MMS treatment of wild type, Aag (Ϫ/Ϫ) , and ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) MEFs generated a similar dosedependent increase in SCE events. However, MMS-induced BER intermediates (accumulating in ␤-pol (Ϫ/Ϫ) cells) generated relatively more SCE events, indicating that 5Ј-dRP lesions can indeed stimulate homologous recombination. Similar to the pattern seen in the cytotoxicity assays, ␤-pol (Ϫ/Ϫ) cells require Aag activity to give rise to the higher than wild type levels of MMS-induced SCE events (Fig. 7). Further, retroviral expres-sion of hAAG in ␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) cells restored the ␤-pol-dependent increase in SCEs (Figs. 7 and 8). DISCUSSION In this paper, we further pursued the hypothesis that BER intermediates can be cytotoxic, in and of themselves, especially in the absence of efficient BER. We designed experiments to ask whether the MMS hypersensitivity observed for ␤-pol (Ϫ/Ϫ) / BER-deficient MEF cells is dependent upon initiation of this repair pathway by the Aag DNA glycosylase. The results clearly show that the alkylating agent (MMS) hypersensitivity of the ␤-pol (Ϫ/Ϫ) cell was not observed in the absence of Aag expression. Thus, the observation of hypersensitivity in this cell system requires the combination of Aag expression and the absence of ␤-pol. Overall, the results are consistent with the model that the 5Ј-dRP-containing BER intermediate is far more toxic to these cells than the alkylated DNA resulting from the MMS exposure. Thus, these cells are able to tolerate MMSinduced DNA lesions without triggering a cell death response. As shown in Fig. 4, the Aag null cell detects MMS-induced damage; the cell temporarily interrupts DNA synthesis but then returns to DNA synthesis again after a short period. The mechanism of this lesion detection and bypass process is unknown at the moment. However, our data support a model whereby the replication fork is blocked or broken down by a DNA strand break containing the 5Ј-dRP lesion because ␤-pol (Ϫ/Ϫ) cells experience prolonged inhibition of DNA synthesis and an induction of recombination-mediated DNA repair. However, failure to resolve the 5Ј-dRP-mediated block in replication fork progression is ultimately toxic.
Previous results from several laboratories have revealed a range of MMS sensitivity responses upon Aag deletion. For example, Roth and Samson found that myeloid progenitor bone marrow cells from Aag (Ϫ/Ϫ) mice were less MMS-sensitive than corresponding cells from wild type animals, whereas other groups found that Aag (Ϫ/Ϫ) ES cells and MEFs were more MMS-sensitive or Me-lex-sensitive than Aag (ϩ/ϩ) counterparts (23,37,48,49). The results described here reveal that the ␤-pol expression status of cells can be a variable in the MMS-sensitivity phenotype upon Aag gene deletion. Thus, in our study with MEFs carrying the ␤-pol (Ϫ/Ϫ) background, MMS sensitivity was observed in Aag (ϩ/ϩ) cells (␤-pol (Ϫ/Ϫ) /Aag (ϩ/ϩ) ) but not for Aag (Ϫ/Ϫ) cells (␤-pol (Ϫ/Ϫ) /Aag (Ϫ/Ϫ) ). In the ␤-pol (ϩ/ϩ) background, the effect of Aag gene deletion was negligible. Another important variable in the MMS sensitivity phenotype of Aag (Ϫ/Ϫ) cells may be the DNA lesion bypass and recognition system in the cells. As discussed above, our MEF lines clearly have a means of bypassing the replication-blocking 3-MeA lesion. Various Aag (Ϫ/Ϫ) cells deficient in such mechanisms may well be more MMS-sensitive than their Aag (ϩ/ϩ) counterpart, assuming that ␤-pol expression is strong. Although these additional factors of ␤-pol expression and lesion bypass were not evaluated in the previous studies noted above, the fundamental role of Aag gene expression in determining the MMS sensitivity phenotype was nonetheless apparent.
In the absence of Aag, compensatory pathways (alternative DNA glycosylases or the nucleotide excision repair pathway) (27,34) did not effectively remove MMS-induced or MNNGinduced alkylation damage. Accumulated alkylation damage then slows progression of the replication fork in a checkpointindependent fashion (40). The studies described here, together with our earlier report on MMS-induced mutagenesis (28), support a model whereby a mutagenic lesion bypass mechanism allows for survival after alkylation damage, and this lesion tolerance mechanism is independent of p53. Conversely, the cytotoxicity of 3-MeA is probably a p53-dependent effect (33). However, once BER is initiated, damage-specific glycosy- lases give rise to accumulation of ␤-pol substrates (5Ј-dRP). 5Ј-dRP causes a complete block to replication fork progression and possibly induces fork regression and resolution by BLM and WRN (40,46). ␤-pol deficiency therefore results in an increase in damage-induced recombination and toxicity in a p53-independent and mismatch DNA repair-independent mechanism. The 5Ј-dRP-mediated cytotoxicity described herein may well represent a novel mechanism of checkpoint activation and replication arrest. Finally, it remains to be determined whether the presence of p53 will alter this DNA damage response paradigm and, further, whether BER and homologous recombination repair complexes, in a coordinated effort, facilitate survival following base damage.