Effects of Substrate Specificity on Initiating the Base Excision Repair of

Dec 22, 2004 - ... Journal of Chemical Information and Computer Sciences .... Effects of Substrate Specificity on Initiating the Base Excision Repair ...
0 downloads 0 Views 240KB Size
Download PDF
Chem. Res. Toxicol. 2005, 18, 87-94

87

Effects of Substrate Specificity on Initiating the Base Excision Repair of N-Methylpurines by Variant Human 3-Methyladenine DNA Glycosylases Ellen E. Connor, Jacqueline J. Wilson, and Michael D. Wyatt* Department of Basic Pharmaceutical Sciences, College of Pharmacy, University of South Carolina, Columbia, South Carolina 29208 Received July 1, 2004

The human 3-methyladenine (AAG, ANPG, MPG) DNA glycosylase excises alkylated purines from DNA. In previous studies, we determined the importance of an active site amino acid (asparagine 169) in the recognition of substrates by AAG. In this study, we characterize the consequences of expressing the AAG variants bearing amino acid substitutions at position 169 in Saccharomyces cerevisiae that lack endogenous 3-methyladenine DNA glycosylase. Survival, mutation induction, and DNA double strand break formation were determined in response to methyl methanesulfonate. The ability of purified wild-type and AAG variants to remove 3-methyladenine and 7-methylguanine, the two most abundant adducts produced by methyl methanesulfonate, was also determined. The N169D AAG variant displayed a ∼100fold lower activity for 3-methyladenine as compared to wild-type and did not detectably remove 7-methylguanine. When expressed in S. cerevisiae, the N169D variant provided better protection against methyl methanesulfonate toxicity than wild-type. Fewer strand breaks in vivo were also seen in the presence of the N169D variant following MMS exposure. In contrast, the N169A and N169S AAG variants displayed ∼30-fold lower activity for 3-methyladenine and 7-methylguanine. Expression of the N169A and N169S AAG variants in S. cerevisiae during methyl methanesulfonate exposure resulted in greater sensitivity, greater mutation induction following MMS exposure, and more strand breaks in vivo. Strand breaks seen in S. cerevisiae that express wild-type AAG or the N169 variants were resolved to varying extents during recovery. In contrast, strand breaks formed in S. cerevisiae that expressed a catalytically inactive AAG variant were not resolved during the recovery times examined. Taken together, the results provide evidence that 3-methyladenine adducts not repaired by base excision repair cause double strand breaks that are not rapidly resolved. Evidence is also provided that the BER intermediates resulting from excision of 7-methylguanine by wild-type AAG contributes to the mutagenicity and cytotoxicity of alkylating agents.

Introduction Cellular DNA repair pathways target inappropriate DNA structures and thus play a vital role in maintaining genomic integrity (1). Base excision repair (BER)1 is the major repair pathway that targets DNA base damage resulting from alkylation, deamination, oxidation, or base loss. DNA glycosylases initiate the BER pathway by excising damaged or inappropriate bases (2). DNA glycosylases that remove the alkylated base 3-methyladenine (3-MeA) have been identified in all species examined for such an activity. Studies in Escherichia coli and Saccharomyces cerevisiae demonstrated the importance of alkylpurine DNA glycosylases in protecting against the mutagenic and toxic effects of alkylating agents. However, loss of the Schizosaccharomyces pombe Mag1 homologue does not confer sensitivity to alkylating agents (3), indicating that different species utilize 3-MeA DNA glycosylase-initiated BER to different extents. * To whom correspondence should be addressed. Tel: 803-777-0856. Fax: 803-777-8356. E-mail: [email protected]. 1 Abbreviations: BER, base excision repair; AAG, alkyladenine DNA glycosylase; 3-MeA, 3-methyladenine; 7-MeG, 7-methylguanine; MMS, methyl methanesulfonate; N169, asparagine at amino acid position 169 in wild-type AAG; DSB, DNA double strand break.

Studies of 3-MeA DNA glycosylases in mammalian systems also provide a complex picture of the importance of alkylpurine DNA glycosylase activity. Humans possess a 3-MeA DNA glycosylase that is similar to mice, rats, and arabidopsis (4). This 3-MeA DNA glycosylase family excises a broad range of substrates that include 7methylguanine (7-MeG), hypoxanthine (Hx), 1,N6-ethenoadenine (A), and 1,N2-ethenoguanine (G) (5). Mice homozygous null for murine Aag are viable and do not suffer from accelerated tumorigenesis (6, 7). Interestingly, the Aag null mice and primary embryonic fibroblasts do not display sensitivity to alkylating agents (7). Whereas Aag -/- embryonic stem (ES) cells are more sensitive to MMS than Aag +/+ ES cells (8, 9), Aag -/bone marrow (BM) cells of myeloid lineage display resistance to methylating agents as compared to Aag +/+ myeloid BM cells (10). The studies highlight the complexity in the cellular response to alkylation damage. Previous studies in which 3-MeA DNA glycosylases were overexpressed in mammalian cells provided evidence that 3-MeA DNA glycosylase overexpression in response to alkylating agents can be detrimental (1113). An appreciation of the balance of the BER pathway has grown in recent years with the realization that

10.1021/tx049822q CCC: $30.25 © 2005 American Chemical Society Published on Web 12/22/2004

88

Chem. Res. Toxicol., Vol. 18, No. 1, 2005

excessive DNA glycosylase activity can lead to deleterious biological consequences (14, 15). Two complexities in interpreting the relative importance of mammalian alkyladenine DNA glycosylase (AAG)-initiated BER in response to DNA methylating agents are the relative formation of 7-MeG and 3-MeA by methylating agents and their relative removal by alkylpurine DNA glycosylases or other repair pathways. Specifically, 3-MeA constitutes ∼10% of the total adduct burden produced by methyl methanesulfonate (MMS) while 7-MeG is the predominant adduct (>70%) (16). In the absence of AAGinitiated BER, 3-MeA and 7-MeG appear to be substrates for global genomic nucleotide excision repair (17-20). Another complexity has been the substrate range of the alkylpurine DNA glycosylases, which is broad. Substrate recognition is particularly intriguing for 3-MeA DNA glycosylases, as the diverse range of adducts does not share common features that could be exploited. Interestingly, 3-MeA DNA glycosylases can remove normal bases at low but detectable levels (21-24). We engineered a series of human AAG variants that were designed to test steric and electrostatic considerations within the active site of AAG that might facilitate substrate recognition and exclude nonsubstrates from the active site (23). In the crystal structure of AAG bound to A containing DNA, asparagine 169 is located in the floor of the active site pocket in close proximity to the C-2 position of the A (25). Variants of AAG bearing different amino acid substitutions at N169 each displayed some degree of deviation in activity as compared to wild-type. Altering the Asn to serine or alanine increases glycosylase activity for Hx and A in vitro but at a cost of acquiring an ability to mistakenly excise guanine from mispairs (23). When the wild-type Asn was replaced with an aspartate, the N169D variant lost the ability to excise Hx and A. In this study, we have investigated the consequences of expressing wild-type AAG, a catalytically inactive AAG variant, and the N169 variants of AAG that have different substrate specificities in S. cerevisiae that lack endogenous 3-MeA DNA glycosylase activity.

Experimental Procedures Methylated Base Release in Vitro. Salmon testes DNA was treated with N-[3H]Methyl-N-nitrosourea (1 mCi, Amersham Pharmacia) to produce 3H 7-MeG and 3-MeA adducted DNA as previously described (26). The purification of wild-type AAG and AAG variants bearing substitutions at N169 was previously described (23). The release of 7-MeG and 3-MeA (over time) was determined by incubating the tritiated DNA (0.25 µg/ µL) with glycosylase at 37 °C in 100 µg/mL BSA, 50 mM Tris (pH 8.3), and 5 mM EDTA. Glycosylase concentrations were as follows: 30 and 300 nM for wild-type AAG and 3 µM for the AAG variants (N169A, N169D, and N169S). DNA was incubated under the same conditions without glycosylase to determine the background levels of spontaneous release, while DNA was heated to 90 °C in the presence of 0.1 M HCl for 1 h to determine the total amount of tritiated N-methylpurines in the DNA. Following time points of 1, 2, 5, 10, 30, or 60 min, reactions were stopped with 0.1 vol of sodium acetate and 3 vol of ethanol. Following chilling at -70 °C for 15 min, the samples were spun at 13 000 rpm for 10 min to precipitate the DNA. The supernatant containing released bases was transferred to fresh tubes and dried down in a CentriVap Concentrator (LabConco). The samples were dissolved in 5 µL of 0.1 M HCl and spotted on Whatman paper in order to separate the tritiated 7-MeG and 3-MeA by paper chromatography. Spiking the spots with 7-MeG and 3-MeA allows for visualization under UV light. The chro-

Connor et al. matography was performed for 10 h in running buffer containing 70% propanol and 10% ammonium hydroxide in water. After the Whatman paper dried, the regions containing 7-MeG (rd ) 0.15-0.35) and 3-MeA (rd ) 0.40-0.60) were excised and scintillation counted. The background in the absence of glycosylase was ∼3 fmol per reaction. The maximum amount of 7-MeG that the wild-type protein removed was 120 fmol. The N169D protein did not produce 7-MeG excision above background at protein concentrations up to 10 times that was used for the wild-type protein. S. cerevisiae Survival following Acute Treatment with Alkylating Agent. The RS1 MAG1∆ S. cerevisiae strain was a gift from Drs. Leona Samson and Brian Glassner (14). Fulllength DNA glycosylases were expressed from the pYES2 vector (23). S. cerevisiae expressing wild-type AAG and N169 AAG variants were treated with the methylating agent MMS in order to determine the ability of the glycosylases to protect cells from the acute effects of MMS. The generation of AAG cDNA bearing the amino acid substitutions N169A, N169D, N169S, and E125Q has been previously reported (23, 25). Switching Glu 125 to Gln produced a catalytically inactive AAG that retained the ability to bind substrate containing DNA (25). S. cerevisiae were grown to confluence at 30 °C in selective media under noninducing conditions (glucose). The cells were resuspended in selective media at a cell concentration of 3 × 107 cells per mL and grown under inducing conditions (galactose) for 3 h. Following induction, the cultures were split and incubated in the presence of 0, 0.1, 0.3, and 0.5% MMS (0.4% MMS for E125Q) for 30 min at 30 °C. The cells were then pelleted and placed in fresh galactose media to remove the MMS. Dilutions of the untreated and MMS-treated cultures were plated on selective media agar containing galactose and incubated at 30 °C for 4 days. Colonies were counted, and colony forming units/mL were determined using an aCOLylte colony counter (Synbiosis, United Kingdom). S. cerevisiae survival following glycosylase expression in the absence of alkylating agent was determined using the control untreated samples. The statistical significance of the results was determined by subjecting at least four replicates of each strain to a two-tailed t-test assuming unequal variance. S. cerevisiae Alkylation Treatment for Analysis of DNA Damage in Vivo. S. cerevisiae in log phase growth were challenged with MMS to determine DNA double strand break (DSB) formation following expression of wild-type AAG, the N169 AAG variants, and E125Q AAG. S. cerevisiae were grown to confluence in selective media under noninducing conditions (YNB glucose). Following removal of the media, the cells were placed in fresh media under inducing conditions (YNB galactose) at a 1:6 dilution. After cell growth reached the log phase, the cultures were treated with 0, 0.06, 0.08, or 0.1% MMS for 30 min. Following MMS treatment, cells were collected, washed with 50 mM EDTA, and counted. To determine the repair of the DNA DSBs, the cells treated with 0.1% MMS were then placed in fresh inducing media to remove the MMS. Following repair periods of 0.5, 1, 3, and 6 h, cells were collected and washed with 50 mM EDTA. DSB Analysis. The prepared S. cerevisiae were embedded in 1% agarose according to the CHEF Genomic DNA Plug Kit protocol (Bio-Rad, Hercules, CA). Subsequent enzymatic treatments resulted in cellular lysis and purification of chromosomesized DNA in agarose plugs. One-third of a plug was placed on each tooth of the comb. Pulsed Field Certified Agarose (BioRad) at a 1% concentration in 0.5× TBE was prepared and brought to 50 °C. The gel was then cast around the plugs and allowed to cool for 30 min. Gel electrophoresis was carried out in chilled 0.5× TBE for 24 h at 6 V/cm and a switch time of 60-120 s using a CHEF-DR II Pulsed Field Electrophoresis System (Bio-Rad). To visualize the DNA, the gel was stained in 0.5 µg/mL ethidium bromide for 30 min and destained in distilled water for 1 h. The gel was analyzed using a Molecular FX imager and Quantity One software (Bio-Rad). The total DNA in each lane was calculated and normalized to account for lane-

AAG Excision of Methyl Purines

Chem. Res. Toxicol., Vol. 18, No. 1, 2005 89

Table 1. Excision Rates (min-1) for 3-MeA and 7-MeG by Wild-Type AAG and the N169 Mutants wild-type N169A N169D N169S

3-MeA

7-MeG

20.458 ( 9.407a 0.567 ( 0.093 0.220 ( 0.018 0.607 ( 0.088

0.706 ( 0.182 0.022 ( 0.008 NDb 0.025 ( 0.005

a fmol base released/pmol protein/min. b ND, not detectable under the conditions used.

to-lane variation in the loading procedure. Following normalization, DSB formation was calculated as a loss of high molecular weight (HMW) DNA (above 750 kb). Data are presented as percentage DNA lost following MMS treatment as compared to untreated samples. S. cerevisiae Mutation Induction following Acute Treatment with Alkylating Agent. Mutation induction was determined by following S. cerevisiae resistance to the lethal arginine analogue, canavanine, as previously described (27). S. cerevisiae expressing wild-type AAG, N169A, N169D, N169S, and E125Q were grown under the conditions used for the MMS survival assay. Following glycosylase induction, the S. cerevisiae cultures were split and treated with the following concentrations of MMS for 30 min: wild-type and N169D (0, 0.1, 0.2, and 0.3%), N169A and N169S (0, 0.06, 0.1, and 0.2%), and E125Q (0, 0.01, 0.03, and 0.06%). The cells were then pelleted and placed in fresh galactose media to remove the MMS. Dilutions of the cultures were plated on selective galactose media agar and on selective galactose media agar containing 60 µg/mL canavanine. Following 4 days of incubation at 30 °C, colonies were counted using an aCOLylte colony counter (Synbiosis). Mutation induction was calculated by comparing the number of cells at each MMS dose in the presence of canavanine to the number of cells surviving the MMS treatment in the absence of MMS. Analysis of the Mutational Spectrum. To determine the type of mutations that are being introduced, the CAN1 gene from canavanine resistant colonies was sequenced. Genomic DNA was isolated from S. cerevisiae that survived the MMS treatment and grew in the presence of canavanine using the YeasStar genomic DNA kit (Zymo Research, Orange, CA). The CAN1 gene was amplified by PCR using the following primers: 5′-CTTAACTCCTGTAAAAAC and 5′-GAAATGGCGTGGGAATGTG. The PCR products were isolated with a MinElute PCR purification kit (Qiagen, Valencia, CA), and sequence analysis was performed by SeqWright (Houston, TX). A χ-square test and Fisher exact test analyses were performed to determine statistical significance. Because of the small frequencies of occurrence, the N169 variants were combined, as were the A:T targeted substitutions and deletion events. Once combined, A χ-square test and Fisher exact test revealed p values of 0.007 and 0.0137, respectively, indicating that the N169 variants show a significantly different proportion of mutational events as compared to wild-type.

Results Methylated Purine Release in Vitro. Our previous work demonstrated that the biochemical and biological phenotypes of each of the N169 variants deviated to some degree from wild-type. Specifically, the in vitro activity for Hx and A did not associate with the relative survival in a MMS gradient plate assay of S. cerevisiae expressing the wild-type and N169 variants (23). We therefore determined the ability of purified wild-type and variant AAG proteins to release 3-MeA and 7-MeG from methylated DNA over time. The purified proteins were wildtype AAG, N169A, N169D, and N169S (23). The wildtype glycosylase efficiently removed 3MeA (20.5 ( 9.41 fmol base release/pmol protein/min) (Table 1). Altering the amino acid at the N169 position significantly reduced

Figure 1. Effect of DNA glycosylase expression on S. cerevisiae colony formation. To induce glycosylase expression, confluent S. cerevisiae were placed in inducing media (YNB gal) at a concentration of 3 × 107 cells/mL and incubated at 30 °C for 3 h. Following incubation, the cultures were diluted and spread on inducing plates and incubated at 30 °C. The controls were transferred to fresh noninducing media for 3 h and then diluted and plated. After 4 days of growth, colonies were counted and resulting colony forming units/mL were calculated. Data are an average of three independent experiments and are presented as percent decrease as compared to colony forming units/mL for the cells under noninducing conditions.

the ability of AAG to release 3-MeA. Specifically, substituting the Asn with either Ala or Ser led to a substantial reduction in 3-MeA excision (0.567 ( 0.094 and 0.607 ( 0.088, respectively). The switch to Asp reduced 3-MeA excision slightly further (0.220 ( 0.018). Consistent with previous reports, wild-type AAG removed 7-MeG less efficiently (0.706 ( 0.182) as compared to 3-MeA (Table 1). As with 3-MeA excision, Ala and Ser substitutions at the N169 position led to similar reductions in 7-MeG excision (0.022 ( 0.008 and 0.025 ( 0.005, respectively). Interestingly, the Asp variant did not display any detectable activity for 7-MeG under the conditions used, which is estimated to be, at minimum, 400-fold lower than wildtype (Experimental Procedures). Effect of DNA Glycosylase Expression in S. cerevisiae in the Absence of MMS. During the course of our initial investigation, we noted that S. cerevisiae expressing the N169S variant appeared to grow poorly. We therefore determined the plating efficiency of S. cerevisiae following expression of wild-type AAG and the N169 variants. Western analysis demonstrated that induction of protein expression produced very similar levels for AAG and the AAG variants in S. cerevisiae (less than 2-fold difference) (23). S. cerevisiae were grown to confluence in noninducing selective media, transferred to inducing selective media for 3 h, then diluted and plated under inducing conditions. The controls were transferred to fresh noninducing media for 3 h, then diluted, and plated. Interestingly, S. cerevisiae expressing the N169S variant show a statistically significant decrease in relative plating efficiency as compared to S. cerevisiae expressing wild-type or the other variants, with p values of 0.0046 (wild-type), 0.049 (N169A), and 0.047 (N169D) (Figure 1). S. cerevisiae expressing the N169A variant show a statistically significant decrease in relative plating efficiency as compared to S. cerevisiae expressing wild-type, with a p value of 0.033 (Figure 1). Effect of Glycosylase Activity on S. cerevisiae Survival following MMS Exposure. The ability of wild-type AAG, the N169 variants, and E125Q AAG to rescue S. cerevisiae lacking the endogenous MAG1 DNA glycosylase from acute exposure to MMS was determined

90

Chem. Res. Toxicol., Vol. 18, No. 1, 2005

Figure 2. S. cerevisiae survival following acute treatment with MMS. Wild-type (closed triangle), N169A (closed circle), N169D (closed square), N169S (open triangle), and E125Q (open circle) AAG glycosylases were examined for the ability to protect S. cerevisiae from the killing effects of MMS. Confluent S. cerevisiae were grown under inducing conditions (YNB gal) at a starting concentration of 3 × 107 cells/mL for 3 h. Following the addition of 0, 0.1, 0.3, or 0.5% MMS and incubation for 30 min, the cultures were diluted and spread on inducing plates and incubated at 30 °C. After 4 days of growth, colonies were counted and resulting colony forming units/mL were calculated. Data are presented as percent survival as compared to an untreated control and are an average of at least three independent experiments.

by colony formation. S. cerevisiae expressing the catalytically inactive E125Q variant were dramatically sensitive at a dose of 0.3% MMS and failed to produce any colonies at 0.5% MMS (Figure 2). The estimated dose that induced 90% death was 0.083% MMS. However, S. cerevisiae expressing wild-type AAG were significantly more resistant to MMS (90% death at 0.24% MMS). Interestingly, S. cerevisiae expressing N169D AAG were more resistant than wild-type, with 90% death at 0.35% MMS (Figure 2). We previously observed a modest increase in the resistance of S. cerevisiae expressing N169D as compared to wild-type when chronically exposed to MMS in a gradient plate assay (23). In contrast, S. cerevisiae expressing either N169A or N169S displayed an intermediate sensitivity between wild-type and E125Q expressing S. cerevisiae (Figure 2). The results demonstrate that the N169D variant, which does not detectably remove 7-MeG, provides substantially better protection against the toxic effects of MMS than wild-type AAG, despite an approximately 100-fold reduced ability to remove 3-MeA as compared to wild-type AAG. Effect of Glycosylase Activity on Strand Break Formation in S. cerevisiae following MMS Exposure. We determined the extent of DSB formation in vivo resulting from MMS treatment and subsequent glycosylase activity in S. cerevisiae lacking MAG1 and expressing wild-type AAG, the N169 variants, and E125Q AAG. Following an acute 30 min exposure of S. cerevisiae to MMS, genomic DNA was subjected to pulsed field gel electophoresis (Figure 3). The formation of DSBs was measured as a loss of HMW genomic DNA. There was an MMS dose-dependent increase in the formation of DSBs, as indicated by a loss of HMW DNA, regardless of which glycosylase was expressed (Figure 4A). In S. cerevisiae expressing E125Q AAG, the highest dose of MMS (0.1%) led to a loss of ∼10% of HMW DNA, indicating that MMS can induce DSBs that are not initiated by an alkylpurine DNA glycosylase. A similar pattern of HMW DNA loss was noted in S. cerevisiae expressing wild-type AAG (Figure 4A). The similarity in

Connor et al.

Figure 3. Representative pulsed field electrophoresis gel showing DNA double strand formation and repair. DSB formation was determined as a loss of the HMW chromosomal DNA. In each gel, the first lane is an untreated control sample. Lanes 2-4 are DNA samples isolated following a 30 min exposure to MMS concentrations of 0.06, 0.08, and 0.10%, respectively. Lanes 5-7 are DNA samples that were isolated from yeast treated with 0.1% MMS for 30 min and followed by repair times of 30 min, 1 h, and 3 h, respectively. (A) A representative gel of DNA isolated from S. cerevisiae expressing N169D. The cells expressing N169D are comparatively resistant to the formation of DSBs following MMS exposure, and the repair of the incurred DSBs approaches completion by 6 h. (B) A representative gel of DNA isolated from S. cerevisiae expressing N169S. The cells expressing N169S display a more dramatic DSB formation with increasing MMS concentration. In addition, repair of these DSBs does not reach completion.

DSB formation between wild-type and E125Q indicates that initiating the BER pathway does not lead to a substantial increase in the accumulation of DNA DSBs. In support of this, fewer DSBs occurred in S. cerevisiae expressing N169D (∼6% loss of HMW DNA) as compared to E125Q or wild-type (Figure 4A). Interestingly, the loss of HMW DNA following MMS exposure in S. cerevisiae expressing N169A and N169S was greater (∼14 and ∼18%, respectively) as compared to S. cerevisiae expressing wild-type, N169D, or E125Q AAG (Figure 4A). Note that the E125Q expressing S. cerevisiae were significantly more sensitive to MMS than S. cerevisiae expressing N169S or N169A AAG. The results show that following MMS treatment DSBs are formed independently of DNA glycosylase activity, in addition to the DSBs that directly result from DNA glycosylase activity. The resolution of DNA DSBs was determined at various time points following removal of 0.1% MMS. Repair of the DSBs seen in wild-type expressing S. cerevisiae is near completion at a 6 h time point, as compared to the untreated sample (Figure 4B). Similar to wild-type, the DSBs seen in N169D expressing S. cerevisiae are resolved at the 6 h time point. In contrast, the DSBs formed in E125Q expressing S. cerevisiae are not resolved at 6 h following MMS treatment. In S. cerevisiae expressing the N169A and N169S variants, repair of DSBs was seen at early time points, but further repair was not seen after the 1 h time point and DSB repair failed to reach completion by 6 h. Taken together, the results suggest that DSBs resulting from DNA glycosylase activity can be repaired more rapidly, whereas DSBs directly induced by MMS require longer to repair.

AAG Excision of Methyl Purines

Chem. Res. Toxicol., Vol. 18, No. 1, 2005 91

Figure 5. S. cerevisiae mutation induction following acute treatment with MMS. S. cerevisiae expressing wild-type (closed triangle), N169A (closed circle), N169D (closed square), and N169S (open triangle) AAG glycosylases were examined for mutation induction, as determined by resistance to the lethal arginine analogue canavanine, following MMS treatment. Confluent S. cerevisiae were grown under inducing conditions (YNB gal) at a starting concentration of 3 × 107 cells/mL for 3 h. Following the addition of MMS at the indicated concentrations and incubation for 30 min, the cultures were diluted and spread on inducing plates containing and lacking canavanine and incubated at 30 °C. The number of canavanine resistant colonies per 108 cells was calculated as a percentage as compared to the number of surviving cells in the absence of canavanine for each MMS dose. Data are an average of at least four independent experiments.

Figure 4. In vivo DSB formation and repair in S. cerevisiae following MMS treatment. (A) S. cerevisiae grown to log phase in inducing media (YNB gal) were treated with 0, 0.06, 0.08, or 0.1% MMS for 30 min. Following treatment, the formation of DSBs was analyzed by pulsed field gel electrophoresis. The data are an average of at least two independent experiments and are presented as a percentage of HMW DNA as compared to the untreated sample. (B) S. cerevisiae treated with 0.1% MMS were placed in fresh media and incubated at 30 °C for the indicated time points to follow the repair of DSBs. Data are an average of at least two independent experiments and are presented as a percentage of HMW DNA as compared to the untreated sample.

Effect of Glycosylase Activity on Mutation Induction following MMS Exposure. The induction of mutations following MMS treatment was determined by the induction of resistance to the lethal arginine analogue canavanine. The AAG variants were expressed in S. cerevisiae lacking MAG1 and exposed to various concentrations of MMS for 30 min as above. In the absence of MMS, cells expressing wild-type AAG showed 14.1 canr colonies per 108 cells, while cells expressing the N169 variants all showed higher mutation inductions of 49.7, 78.6, and 107.7 per 108 cells for N169A, N169D, and N169S, respectively. In S. cerevisiae that express wildtype AAG, MMS exposure led to a substantial increase in DNA mutations (Figure 5). S. cerevisiae expressing N169D AAG survived better in the presence of MMS as compared to those expressing wild-type AAG (Figure 2). However, N169D expression in the presence of MMS doses of 0.2 and 0.3% led to greater mutation induction as compared to wild-type (Figure 5). The N169A and N169S AAG glycosylases protected S. cerevisiae from MMS killing to a similar extent, although not as well as wild-type AAG (Figure 2). At a MMS dose of 0.1%, N169A and N169S AAG expression in S. cerevisiae led to a

similar mutation induction, which was greater than the mutation induction in S. cerevisiae expressing wild-type AAG. However, the mutation induction in S. cerevisiae expressing N169S AAG was elevated to approximately twice the mutation induction in S. cerevisiae expressing N169A at a dose of 0.2% MMS (Figure 5). S. cerevisiae expressing E125Q AAG were highly susceptible to the killing effects of MMS; therefore, mutation analysis was carried out at lower MMS doses. Under these conditions, only a modest increase in DNA mutations (∼5-fold) is detected in S. cerevisiae expressing E125Q AAG (data not shown). To determine the type of mutation induced, individual colonies that survived on the canavanine plates were grown and genomic DNA was isolated. The CAN1 gene was PCR amplified and sequenced for 10 variants for each S. cerevisiae strain expressing the different AAG glycosylases. We set out to compare the induction of mutations at A:T base pairs (presumably caused by 3-MeA or its excision) vs mutations at G:C base pairs (presumably caused by 7-MeG or its excision). It was presumed that the source of a base pair substitution mutation originated from the alkylpurine or subsequent BER intermediate resulting from excision of the alkylpurine for the following reasons. Greater than 98% of the adducts that MMS forms occur on purines (16), and human AAG (and the N169 variants) exclusively excises purines. For S. cerevisiae expressing wild-type AAG, base substitution mutations occurred at three G:C base pairs and three A:T base pairs, while two A:T deletions and one large internal deletion were detected (Table 2). Interestingly, for S. cerevisiae expressing the N169S glycosylase, all of the base substitution mutations (nine total) occurred at G:C base pairs with one deletion mutant at an A:T base pair. Note that we previously reported that this variant has the ability to excise G opposite T or U. For S. cerevisiae expressing the N169A

92

Chem. Res. Toxicol., Vol. 18, No. 1, 2005

Connor et al.

Table 2. Spectrum of Mutations Following MMS Treatment in the CAN1 Gene in S. cerevisiae Expressing Wild-Type AAG and the N169 Variants wild-type N169A A:T targeted substitutions 3 (33.3)a G:C targeted substitutions 3 (33.3) deletion events 3 (33.3) a

1 (10) 8 (80) 1 (10)

N169D

N169S

3 (21.4) 0 10 (71.4) 9 (90) 1 (7.1) 1 (10)

Number of mutants (% of the total).

glycosylase, which can also excise G opposite T or U, the mutations were almost all base substitutions at G:C base pairs (eight of ten) instead of A:T base pairs (one), with one C-terminal deletion mutant. For S. cerevisiae expressing the N169D glycosylase, mutations occurred at three A:T base pairs, ten G:C base pairs, and a deletion of an (A:T)2 dinucleotide. The mutation spectra for the N169 variants were compared to the wild-type for statistical significance (Experimental Procedures). A χ-square test and Fisher exact test revealed p values of 0.007 and 0.0137, respectively, indicating that the N169 variants show a significantly different proportion of mutational events as compared to wild-type.

Discussion The N169 variant AAG enzymes utilized in this study were designed to test the role of N169 in substrate recognition by AAG. Our previous studies demonstrated that the N169A and N169S variants excised Hx and A at a faster rate than wild-type (23). Furthermore, the N169A and N169S variants were able to excise guanine in a mispair, leading to the conclusion that N169 acts to prevent the aberrant excision of guanine. We had therefore suspected that 7-MeG might also be excised at a faster rate by the N169A and N169S variants as compared to wild-type. Interestingly, the purified variants excised 7-MeG poorly in vitro as compared to wild-type (Table 1). Although the serine and alanine substitutions decrease the potential steric clash with the N-2 position of guanine, the substitutions have disrupted the recognition of methylated purines. Excision of the positively charged substrates has been compromised by the serine/ alanine substitutions, whereas excision of the uncharged Hx and A by N169S and N169A surpassed wild-type activity. The N169D variant was designed to test the effect of introducing a negative charge into the active site of AAG. The aspartate substitution appeared to greatly compromise the ability of AAG to excise all of its substrates; excision of 7-MeG, Hx, and A was not detectable (this work and ref 23). The different abilities of the N169 variants to excise 3-MeA and 7-MeG were exploited to determine the in vivo consequences of altering the substrate specificity of AAG, which initiates BER for alkylated purines. Early studies of 3-MeA DNA glycosylases in E. coli and S. cerevisiae suggested the importance of these enzymes in protecting cells from the mutagenic and toxic effects of alkylating agents (28, 29). However, subsequent studies led to the understanding that the imbalanced expression of 3-MeA DNA glycosylases can be deleterious in the presence or absence of alkylating agents (11-15, 30). We have measured survival, mutation induction, and DSB formation in S. cerevisiae that express either a wild-type, a catalytically inactive variant, or the N169 variants of human AAG. S. cerevisiae expressing the catalytically inactive E125Q AAG were extremely sensitive to MMS.

In the absence of catalytically active AAG, MMS induced DSBs that were independent of BER initiated by alkylpurine DNA glycosylases. The BER-independent DSBs were not rapidly resolved. Using an agent that produces 3-MeA almost exclusively, it has been shown that 3-MeA can induce mutations in the absence of 3-MeA DNA glycosylase (31). It has also been shown that 3-MeA can induce sister chromatid exchanges, chromosome gaps, and breaks (8). Expression of the N169D variant of AAG, which appears to only excise 3-MeA, provided a much greater degree of protection against MMS toxicity in S. cerevisiae (lacking endogenous alkylpurine DNA glycosylase activity) as compared to expressing the inactive E125Q variant. Expression of N169D AAG also provided greater protection against MMS toxicity than expressing wild-type AAG, despite a greatly compromised ability of N169D AAG to remove 3-MeA. Furthermore, there were fewer DSBs in the N169D expressing cells. However, there was an induction of mutations in N169D expressing cells, indicating that the mutagenic damage occurring was not likely resulting from the resolution of DSBs. The differences observed in survival, DSB formation, and mutation are not thought to be caused by differences in glycosylase expression levels, as we have shown that the levels of AAG and AAG variants differ by less that 2-fold following induction (23). Others have shown that expression of the E. coli tag glycosylase, which is specific for 3-MeA, can suppress recombination induced by MMS (32). Taken together, the results suggest that unrepaired 3-MeA induces DSBs in the absence of BERs that are not rapidly resolved. We demonstrate that expressing N169D AAG reduced DSBs, although base substitutions were induced. It is possible that the reduced capacity of N169D AAG to remove 3-MeA and 7-MeG was insufficient to prevent mutagenic consequences of methylpurines but represented a better balance (for survival) in the relative level of BER initiated in this model. The expression in S. cerevisiae of wild-type AAG resulted in substantial resistance to MMS as compared to the catalytically inactive E125Q variant. DSBs were formed, but these were rapidly resolved as compared to DSBs in cells incapable of initiating BER at methylated purines. However, cells expressing wild-type AAG were more sensitive to MMS as compared to cells expressing N169D. These results suggest that the ability of wildtype AAG to remove 7-MeG contributes to toxicity in this model. It is possible that the compromised alkylpurine excision activity of the N169D variant represents a better balance for survival than wild-type by simply initiating BER at fewer alkylpurines. However, the poorer survival of yeast expressing the N169A and N169S variants as compared to wild-type (and N169D) suggests against simply the overall level of BER initiation, because these variants excise both alkylpurines worse than wild-type but significantly better than the N169D variant. Recent work by others has also demonstrated that AAG-initiated removal of 7-MeG specifically can be detrimental. The MMS sensitivity of DNA polymerase β null mouse embryonic fibroblasts is hypothesized to derive from AAG initiating the removal of 7-MeG, leading to an increased accumulation of toxic dRp repair intermediates (33). The activity of AAG for 7-MeG may simply be a byproduct of the evolution of the mammalian AAG enzyme active site for more biologically deleterious adducts. Interestingly, the Bacillus subtilis homologue of AAG appears to excise Hx and A preferentially, while not excising 7-MeG (34).

AAG Excision of Methyl Purines

Alternatively, there might be a biological benefit to enzymatically removing 7-MeG if spontaneous depurination results in an unprotected AP site, a benefit that is not detected with toxic doses of MMS. In this regard, the N169S variant proved interesting. S. cerevisiae expressing the N169S variant suffered a decrease in plating efficiency relative to S. cerevisiae expressing the other AAG glycosylases. Expression of the N169S variant in the presence of MMS did not protect S. cerevisiae to the same extent as wild-type or the N169D variant. Expression of the N169S variant resulted in more DSBs immediately following MMS exposure. We surmise that the increased DSBs seen in S. cerevisiae expressing N169S AAG result from a combination of an increased ability of N169S to remove 7-MeG (as compared to E125Q and N169D) and a compromised ability to remove 3-MeA (as compared to wild-type). Alternatively, the increased DSBs resulting from N169S activity might represent an aberrant form of BER. Interestingly, these DSBs appeared to be resolved more rapidly, as compared to the strand breaks formed in the absence of active glycosylase. The mutation induction was greatest in cells expressing N169S. The aberrant excision of guanine opposite uracil (deaminated cytosine, ref 23) by the N169S variant might further contribute to the mutation burden. The endogenous hydrolytic deamination of cytosine to uracil in double-stranded DNA causes G:U mispairs (35). S. cerevisiae strains deficient in uracil DNA glycosylase activity suffer from ∼20-fold elevated spontaneous mutation rates at G:C base pairs (36). The elevated mutation induction and decrease in survival of yeast expressing the N169S might in part stem from its aberrant excision activity for guanine opposite deaminated cytosines, an activity that would be promutagenic. It is noteworthy that in the absence of MMS, cells expressing N169S show a 7.5-fold increase in mutations as compared to cells expressing WT. The above studies highlight the importance of the balance of BER, as unrepaired BER intermediates can signal apoptosis (37, 38). There is a balance between initiating adduct removal (and the ensuing BER intermediates) and allowing the adduct to remain unrepaired. The N169D variant may prove particularly interesting in this regard as a human DNA glycosylase capable of removing the toxic 3-MeA adduct yet is compromised for 7-MeG removal so that its activity does not overproduce BER intermediates. One might also consider that excessive stimulation of BER could constitute an approach to sensitizing tumor cells to DNA damaging agents. Indeed, it was very recently shown that transiently overexpressing AAG sensitizes human breast cancer cells to alkylating agents (39). Defining the biological consequences of initiating and completing BER in vivo in different mammalian cell types will better define the importance of balance in the BER pathway.

Acknowledgment. Drs. Leona Samson and Brian Glassner are gratefully acknowledged for the S. cerevisiae RS1 mag∆ strain and pYES2 vector containing the fulllength wild-type AAG cDNA. Dr. Edsel Pena (Department of Statistics, University of South Carolina) is gratefully acknowledged for statistical help. This work was made possible by Grant K22 ES00333 from the National Institute for Environmental Health Sciences, NIH (to M.D.W.).

Chem. Res. Toxicol., Vol. 18, No. 1, 2005 93

References (1) Hoeijmakers, J. H. (2001) Genome maintenance mechanisms for preventing cancer. Nature 411, 366-374. (2) Lindahl, T., and Wood, R. D. (1999) Quality control by DNA repair. Science 286, 1897-1905. (3) Memisoglu, A., and Samson, L. (2000) Contribution of base excision repair, nucleotide excision repair, and DNA recombination to alkylation resistance of the fission yeast Schizosaccharomyces pombe. J. Bacteriol. 182, 2104-2112. (4) Wood, R. D., Mitchell, M., Sgouros, J., and Lindahl, T. (2001) Human DNA repair genes. Science 291, 1284-1289. (5) Wyatt, M. D., Allan, J. M., Lau, A. Y., Ellenberger, T. E., and Samson, L. D. (1999) 3-Methyladenine DNA glycosylases: Structure, function, and biological importance. BioEssays 21, 668-676. (6) Engelward, B. P., Weeda, G., Wyatt, M. D., Broekhof, J. L. M., De Wit, J., Donker, I., Allan, J. M., Gold, B., Hoeijmakers, J. H. J., and Samson, L. D. (1997) Base excision repair deficient mice lacking the Aag alkyladenine DNA glycosylase. Proc. Natl. Acad. Sci. U.S.A. 94, 13087-13092. (7) Elder, R. H., Jansen, J. G., Weeks, R. J., Willington, M. A., Deans, B., Watson, A. J., Mynett, K. J., Bailey, J. A., Cooper, D. P., Rafferty, J. A., Heeran, M. C., Wijnhoven, S. W., van Zeeland, A. A., and Margison, G. P. (1998) Alkylpurine-DNA-N-glycosylase knockout mice show increased susceptibility to induction of mutations by methyl methanesulfonate. Mol. Cell. Biol. 18, 58285837. (8) Engelward, B. P., Allan, J. M., Dreslin, J. A., Kelly, J. D., Gold, B., and Samson, L. D. (1998) A chemical and genetic approach together define the biological consequences of 3-methyladenine lesions in the mammalian genome. J. Biol. Chem. 273, 54125418. (9) Engelward, B. P., Dreslin, A., Christensen, J., Huszar, D., Kurahara, C., and Samson, L. (1996) Repair-deficient 3-methyladenine DNA glycosylase homozygous mutant mouse cells have increased sensitivity to alkylation-induced chromosome damage and cell killing. EMBO J. 15, 945-952. (10) Roth, R. B., and Samson, L. D. (2002) 3-Methyladenine DNA glycosylase-deficient Aag null mice display unexpected bone marrow alkylation resistance. Cancer Res. 62, 656-660. (11) Calleja, F., Jansen, J. G., Vrieling, H., Laval, F., and van Zeeland, A. A. (1999) Modulation of the toxic and mutagenic effects induced by methyl methanesulfonate in Chinese hamster ovary cells by overexpression of the rat N-alkylpurine-DNA glycosylase. Mutat. Res. 425, 185-194. (12) Coquerelle, T., Dosch, J., and Kaina, B. (1995) Overexpression of N-methylpurine-DNA glycosylase in Chinese hamster ovary cells renders them more sensitive to the production of chromosomal aberrations by methylating agentssA case of imbalanced DNA repair. Mutat. Res. 336, 9-17. (13) Ibeanu, G., Hartenstein, B., Dunn, W. C., Chang, L. Y., Hofmann, E., Coquerelle, T., Mitra, S., and Kaina, B. (1992) Overexpression of human DNA repair protein N-methylpurine-DNA glycosylase results in the increased removal of N-methylpurines in DNA without a concomitant increase in resistance to alkylating agents in Chinese hamster ovary cells. Carcinogenesis 13, 1989-1995. (14) Glassner, B. J., Rasmussen, L. J., Najarian, M. T., Posnick, L. M., and Samson, L. D. (1998) Generation of a strong mutator phenotype in yeast by imbalanced base excision repair. Proc. Natl. Acad. Sci. U.S.A. 95, 9997-10002. (15) Posnick, L. M., and Samson, L. D. (1999) Imbalanced base excision repair increases spontaneous mutation and alkylation sensitivity in Escherichia coli. J. Bacteriol. 181, 6763-6771. (16) Beranek, D. T. (1990) Distribution of methyl and ethyl adducts following alkylation with monofunctional alkylating agents. Mutat. Res. 231, 11-30. (17) Wang, W., Sitaram, A., and Scicchitano, D. A. (1995) 3-Methyladenine and 7-methylguanine exhibit no preferential removal from the transcribed strand of the dihydrofolate reductase gene in Chinese hamster ovary B11 cells. Biochemistry 34, 1798-1804. (18) Scicchitano, D. A., and Hanawalt, P. C. (1989) Repair of Nmethylpurines in specific DNA sequences in Chinese hamster ovary cells: Absence of strand specificity in the dihydrofolate reductase gene. Proc. Natl. Acad. Sci. U.S.A. 86, 3050-3054. (19) Plosky, B., Samson, L. D., Engelward, B. P., Gold, B., Schlaen, B., Millas, T., Magnotti, M., Schor, J., and Scicchitano, D. A. (2002) Base excision repair and nucleotide excision repair contribute to the removal of N-methylpurines from active genes. DNA Repair 1, 683-696. (20) Xiao, W., and Chow, B. L. (1998) Synergism between yeast nucleotide and base excision repair pathways in the protection against DNA methylation damage. Curr. Genet. 33, 92-99.

94

Chem. Res. Toxicol., Vol. 18, No. 1, 2005

(21) O’Brien, P. J., and Ellenberger, T. (2004) The Escherichia coli 3-methyladenine-DNA glycosylase AlkA has a remarkably versatile active site. J. Biol. Chem. 4, 26876-26884. (22) O’Brien, P. J., and Ellenberger, T. (2004) Dissecting the broad substrate specificity of human 3-methyladenine-DNA glycosylase. J. Biol. Chem. 279, 9750-9757. (23) Connor, E. E., and Wyatt, M. D. (2002) Active-site clashes prevent the human 3-methyladenine DNA glycosylase from improperly removing bases. Chem. Biol. 9, 1033-1041. (24) Berdal, K. G., Johansen, R. F., and Seeberg, E. (1998) Release of normal bases from intact DNA by a native DNA repair enzyme. EMBO J. 17, 363-367. (25) Lau, A. Y., Wyatt, M. D., Glassner, B. J., Samson, L. D., and Ellenberger, T. (2000) Molecular basis for discriminating between normal and damaged bases by the human alkyladenine glycosylase, AAG. Proc. Natl. Acad. Sci. U.S.A. 97, 13573-13578. (26) Samson, L., Derfler, B., Boosalis, M., and Call, K. (1991) Cloning and characterization of a 3-methyladenine DNA glycosylase cDNA from human cells whose gene maps to chromosome 16. Proc. Natl. Acad. Sci. U.S.A. 88, 9127-9131. (27) Tran, P. T., and Liskay, R. M. (2000) Functional studies on the candidate ATPase domains of Saccharomyces cerevisiae MutLalpha. Mol. Cell. Biol. 20, 6390-6398. (28) Karran, P., Lindahl, T., Ofsteng, I., Evensen, G. B., and Seeberg, E. (1980) Escherichia coli mutants deficient in 3-methyladenineDNA glycosylase. J. Mol. Biol. 140, 101-127. (29) Chen, J., Derfler, B., Maskati, A., and Samson, L. (1989) Cloning a eukaryotic DNA glycosylase repair gene by the suppression of a DNA repair defect in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 86, 7961-7965. (30) Kaina, B., Fritz, G., Ochs, K., Haas, S., Grombacher, T., Dosch, J., Christmann, M., Lund, P., Gregel, C. M., and Becker, K. (1998) Transgenic systems in studies on genotoxicity of alkylating agents: Critical lesions, thresholds and defense mechanisms. Mutat. Res. 405, 179-191. (31) Monti, P., Campomenosi, P., Ciribilli, Y., Iannone, R., Inga, A., Shah, D., Scott, G., Burns, P. A., Menichini, P., Abbondandolo, A., Gold, B., and Fronza, G. (2002) Influences of base excision

Connor et al.

(32)

(33)

(34)

(35) (36)

(37)

(38)

(39)

repair defects on the lethality and mutagenicity induced by Melex, a sequence-selective N3-adenine methylating agent. J. Biol. Chem. 277, 28663-28668. Hendricks, C. A., Razlog, M., Matsuguchi, T., Goyal, A., Brock, A. L., and Engelward, B. P. (2002) The S. cerevisiae Mag1 3-methyladenine DNA glycosylase modulates susceptibility to homologous recombination. DNA Repair 1, 645-659. Horton, J. K., Joyce-Gray, D. F., Pachkowski, B. F., Swenberg, J. A., and Wilson, S. H. (2003) Hypersensitivity of DNA polymerase beta null mouse fibroblasts reflects accumulation of cytotoxic repair intermediates from site-specific alkyl DNA lesions. DNA Repair 2, 27-48. Aamodt, R. M., Falnes, P. O., Johansen, R. F., Seeberg, E., and Bjoras, M. (2004) The Bacillus subtilis counterpart of the mammalian 3-methyladenine DNA glycosylase has hypoxanthine and 1,N6-ethenoadenine as preferred substrates. J. Biol. Chem. 279, 13601-13606. Lindahl, T. (1993) Instability and decay of the primary structure of DNA. Nature 362, 709-715. Impellizzeri, K. J., Anderson, B., and Burgers, P. M. J. (1991) The spectrum of spontaneous mutations in a Saccharomycescerevisiae uracil-DNA-glycosylase mutant limits the function of this enzyme to cytosine deamination repair. J. Bacteriol. 173, 6807-6810. Ochs, K., Sobol, R. W., Wilson, S. H., and Kaina, B. (1999) Cells deficient in DNA polymerase beta are hypersensitive to alkylating agent-induced apoptosis and chromosomal breakage. Cancer Res. 59, 1544-1551. Ochs, K., Lips, J., Profittlich, S., and Kaina, B. (2002) Deficiency in DNA polymerase beta provokes replication-dependent apoptosis via DNA breakage, Bcl-2 decline and caspase-3/9 activation. Cancer Res. 62, 1524-1530. Rime, M., Caldwell, D., and Kelley, M. R. (2004) Transient adenoviral N-methylpurine DNA glycosylase overexpression imparts chemotherapeutic sensitivity to human breast cancer cells. Mol. Cancer Ther. 3, 955-967.

TX049822Q