|Home | About | Journals | Submit | Contact Us | Français|
The oral pathogen, Streptococcus mutans, possesses inducible DNA repair defenses for protection against pH fluctuations and production of reactive oxygen metabolites such as hydrogen peroxide (H2O2), which are present in the oral cavity. DNA base excision repair (BER) has a critical role in genome maintenance by preventing the accumulation of mutations associated with environmental factors and normal products of cellular metabolism. In this study, we examined the consequences of compromising the DNA glycosylases (Fpg and MutY) and endonucleases (Smx and Smn) of the BER pathway and their relative role in adaptation and virulence. Enzymatic characterization of the BER system showed that it protects the organism against the effects of the highly mutagenic lesion, 7,8-dihydro-8-oxo-2’-deoxyguanine (8-oxo-dG). S. mutans strains lacking a functional Fpg, MutY, or Smn showed elevated spontaneous mutation frequencies; and, these mutator phenotypes correlated with the ability of the strains to survive killing by acid and oxidative agents. In addition, in the G. mellonella virulence model, strains of S. mutans deficient in Fpg, MutY and Smn showed increased virulence as compared to the parent strain. Our results suggest that, for S. mutans, mutator phenotypes, due to loss of BER enzymes, may confer an advantage to virulence of the organism.
Streptococcus mutans is a major oral pathogen that infects more than half of the world’s human population (Becker et al., 2002, Mitchell, 2003). The bacterium inhabits the oral biofilm referred to as dental plaque, found on tooth surfaces, and contributes to the initiation and progression of biofilm-related diseases, varying in severity from tooth decay to sub-acute bacterial endocarditis (Lemos & Burne, 2008). The ability of S.mutans to form biofilms, generate organic acids from dietary carbohydrate sources (acidogenicity), and tolerate acidic environments (acidurance) contributes to the cariogenic potential of the organism and confers an ecological advantage in the competition with other oral bacterial species (Banas, 2004, Kuramitsu, 1993, Napimoga et al., 2005).
A variety of microenvironments exist in oral biofilms, including those containing oxidative agents such as hydrogen peroxide, nutrient gradients, and organic acids produced and secreted as a result of microbial metabolism, all of which can contribute to the disruption of microbial homeostasis. Oral commensal bacteria such as S. gordonii and S. sanguinis are able to compete with S. mutans, a catalase-negative organism, by producing hydrogen peroxide (H2O2), which is known to be toxic to S. mutans (Kreth et al., 2005). Reactive oxygen species (ROS), such as hydrogen peroxide (H2O2), hydroxyl radicals (OH•), and superoxide (O2•), are by-products of metabolism that can damage most cellular components, including DNA (Gros et al., 2002, Miller & Britigan, 1997, Cooke et al., 2003). Although S. mutans has a variety of ROS detoxification enzymes such as superoxide dismutase, alkyl hydroperoxide reductase, and glutathione oxidoreductase, the organism is still susceptible to continuous oxidative DNA damage caused by the constant exposure of exogenous agents and production of endogenous oxygen-derived species. As part of its ability to adapt to environmental stress, including oxidative attack on the DNA backbone, S. mutans is equipped with DNA repair mechanisms as part of a second line of defenses that enable it to survive and thrive as a pathogenic organism (Hahn et al., 1999, Quivey et al., 2000).
Among the established DNA repair pathways, base excision repair (BER) is the most versatile and frequently found pathway for coping with DNA damage in nature (David et al., 2007, Nilsen & Krokan, 2001, Cadet et al., 2000). The BER pathway is responsible for removing one of the most common forms of oxidative DNA damage formed in cells, 7,8-dihydro-8-oxo-2’-deoxyguanine (8-oxo-dG) (Boiteux & Radicella, 1999, Michaels & Miller, 1992, David et al., 2007). This oxidized form of guanine is a potent mutagenic lesion, as it can lead to mismatch mutations by pairing with adenine instead of cytosine, resulting in G to T and C to A substitutions in the genome.
In E. coli, both Fpg (MutM) and MutY are known to be DNA glycosylases/lyases that excise oxidized bases from the DNA backbone; whereas, Exo III (Xth) and Nth are apurinic/apyrimidinic endonucleases (APE), catalyzing the removal of blocked sites left by the action of DNA glycosylases (Dizdaroglu, 2003, Krokan et al., 1997, Pope et al., 2002, Warner et al., 1980). In S. mutans, formamidopyramidine DNA-glycosylase (Fpg; SMU.1614), A/G-specific DNA glycosylase (MutY; SMU.1865), exonuclease (Smx; SMU.1649) (Faustoferri et al., 2005, Hahn et al., 1999), and endonuclease (Smn; SMU.1650) are the orthologous members of the BER pathway (Ajdic et al., 2002).
Inactivation of BER-encoding genes has been linked to the generation of mutator phenotypes in several prokaryotes and eukaryotes (Glassner et al., 1998, Huang et al., 2006, Jain et al., 2007). Deletions in any of the BER-encoding genes have been shown to elevate spontaneous mutation frequencies, mainly by the accumulation of 8-oxo-dG levels within the cell. Changes in mutation frequencies can generate different effects on the fitness of an organism, ranging from deleterious through neutral to beneficial (Denamur & Matic, 2006, Gibson et al., 1970). The beneficial results of mutations include the ability to more quickly adapt to changing and stressful environments by increasing pathogenicity and antibiotic resistance (Aertsen & Michiels, 2005, Jayaraman, 2009).
While BER, and other forms of DNA repair, have been studied extensively in bacteria (Nilsen & Krokan, 2001, Bjelland & Seeberg, 2003, Richardson et al., 2009, Jain et al., 2007), the role of BER proteins in pathogenesis and mutability in S. mutans has not been reported, though it is clear that the organism can mount a vigorous response to oxidative stress (Derr et al., 2012, Lemos & Burne, 2008). Here, we examined the role of the BER system, in stress response and survival of S. mutans, following exposure to acidic and oxidative stress conditions. Further, we have determined the biochemical function of the BER enzymes, their role in oxidative DNA damage repair, and their involvement in spontaneous mutagenesis. The results of this study suggest that deficiencies in BER enzymes contribute to an elevated responsiveness to acidic and oxidative conditions, which are important aspects of dental plaque ecology and virulence of S. mutans.
Study of the physiological role of BER-encoding genes in S. mutans, began with an evaluation of whether mutation of the fpg, mutY, smx, and smn genes resulted in growth defects under stress-inducing conditions. Insertional mutation of fpg did not result in phenotypic differences from the parent strain during growth in BHI medium at 37°C (Fig. 1A). However, fpg exhibited a faster growth rate and final growth yield when compared to the parent strain in media titrated to pH 5 (Fig. 1B). In contrast, the mutY and smn strains, grown in BHI medium and in medium titrated to pH 5, exhibited reduced doubling times, indicating faster growth, when compared to the parent strain (Fig. 1A, B). Results from growth experiments with the smx strain, in BHI medium and in media titrated to pH 5, showed a statistically significant decrease in growth kinetics when compared to the parent strain in BHI medium at 37°C (Fig. 1A, B). The data indicates that loss of fpg, mutY, and smn resulted in enhanced resistance to acid stress.
Sensitivity of the BER mutant strains grown under oxidative stress conditions was also examined. When compared to the parent strain, only the smx-deficient strain displayed impaired growth in the presence of oxidative stress (Fig. 1C). This was consistent with previous literature, in which xthA endonuclease mutants were extremely sensitive to H2O2 treatment (Demple et al., 1983, Suvarnapunya et al., 2003). The growth data suggested that Smx plays a significant role in the repair of lethal DNA lesions generated by H2O2.
Here, we determined promoter activities for the BER genes, in strains grown in defined environmental conditions, using chloramphenicol acetyl-transferase (CAT) genetic reporter fusions. CAT activity, in all fusion-bearing strains, was elevated in cells grown in atmospheric levels of oxygen, compared to the activity in cells grown in enriched-carbon dioxide atmospheres (5% CO2, considered to be microaerobic) or under anaerobic conditions (Table 1). However, the presence of exogenous hydrogen peroxide had the greatest influence on activity for all BER gene promoters, when compared to the activity observed in cultures grown in shaking conditions. Expression levels for all promoter fusions showed up-regulation in the presence of exogenous H2O2, suggesting a robust cellular response to the compound that included expression of the BER genes (Table 1). Activity for the BER gene promoters also showed a strong pH-dependence, as determined by CAT activity measurements in cells grown at fixed pH values of 5 versus 7, suggesting that the acid stress-response included DNA repair. Promoter activity of smx showed significant up-regulation in cells grown at pH 7, compared to the activity observed in cultures grown in an enriched-carbon dioxide atmosphere (5% CO2), suggesting that the smx promoter may be regulated somewhat differently than the other promoters, though the mechanism is unclear at this time. Nevertheless, taken together, the reporter fusion data indicates that acidic and oxidative stresses are sufficiently strong environmental cues to induce transcription of the BER genes.
Hydrogen peroxide (H2O2) and its derivative, hydroxyl radical (OH•), are known to be among the most powerful oxidizers that can damage macromolecules, including DNA (Gros et al., 2002, Miller & Britigan, 1997). To establish the contribution of BER in the protection of S. mutans from the damage inflicted by H2O2, we estimated the relative levels of damage in BER strains using a genomic DNA fragmentation assay. Cultures of the BER-deficient strains, grown to exponential phase (OD600 ~0.4), were harvested and treated with 5mM H2O2 for specified periods of time. Samples of DNA from each group were fractionated by gel electrophoresis to visualize fragmented DNA (Fig. 2A). The pre-treatment control cells (0 hr) showed little, if any, DNA fragmentation, indicating undetectable levels of single-stranded breakage. Fragmentation became more apparent in DNA isolated from cultures of the parent strain exposed to H2O2 for 0.5 to 3 hrs (Fig. 2A, B). Significant amounts of genomic DNA were fragmented after 6 hrs of exposure and continued up to 12 hrs for all strains (Fig. 2A). The smx strain showed a significant increase in the amount of fragmented DNA in all time points compared to cell cultures from the parent strain. Major differences were observed at 24 hrs post-exposure, at which point the parent strain had, apparently, begun to repair fragmented DNA. All BER mutant strains displayed a significant percentage of DNA fragmentation compared to the parent strain, suggesting a reduced rate of DNA repair capability on the BER-deficient strains after 24 hrs of H2O2-exposure (Fig. 2A,B). We also determined cell survival after exposure to 5mM H2O2 by estimating colony-forming units (CFUs) at each of the time points (Fig. 2C). Overall, the BER mutant strains exhibited reduced survival following H2O2-treatment, compared to the parent strain, UA159. Of the mutant strains, mutY was the least sensitive, followed by smn and fpg. No smx colonies were recovered after 12 hours (Fig. 2C). Data from the DNA fragmentation assay provided strong evidence that BER enzymes play a role in the protection and repair of H2O2-mediated DNA damage.
A vast majority of promutagenic lesions in cells arise by the action of ROS, generated either by cellular metabolism or by exposure to exogenous agents. One of the most common and deleterious oxidative DNA lesions, is an oxidized form of guanine, 7,8-dihydro-8-oxo-2’-deoxyguanine (8-oxo-dG) (Grollman & Moriya, 1993). To determine whether BER enzymes contribute to protection from the mutagenic effect of 8-oxo-dG, a competitive ELISA was employed to measure the accumulation of 8-oxo-dG levels in S. mutans chromosomal DNA. Elevated accumulation of 8-oxo-dG was observed in the DNA isolated from Fpg, MutY, and Smn-defective strains, when compared with the parent strain, UA159 (Fig. 3). No significant differences were observed in samples from the smx mutant strain. In support of the fragmentation data, we conclude from the 8-oxo-dG measurements that there is significant involvement of BER enzymes in minimizing the amount of oxidized bases in the genome of S. mutans.
To further our understanding of the recognition and excision of nucleotides mismatched with oxidized bases, we tested the ability of the BER enzymes to cleave 8-oxo-dG paired with A, C, G or T and/or mismatched DNA bases (A/C, A/G). In addition, we searched for the activity responsible for the removal of the AP site, the intermediate enzymatic step in BER that allows the initiation of DNA repair synthesis. Results from DNA-nicking assays showed that 8-oxo-dG was cleaved when base-paired with C, G or T, using extracts prepared from UA159, the parent strain (Fig. 4A, lanes 4, 6, and 8). The activity in the parent strain was absent in extracts from the Fpg-deficient strain (Fig. 4A, lanes 5, 7, and 9), indicating that Fpg is necessary for processing of this specifically oxidized base. Identical results were obtained using extracts of the genetically complemented fpgc strain, suggesting that excision of 8-oxo-dG base-paired with C, G, or T is intrinsic to Fpg (Fig. 4D, lane 1). Further, Fpg in S. mutans exhibited the same property reported in E. coli (Leipold et al., 2000, Fromme & Verdine, 2002): the enzyme does not show detectable activity toward 8-oxo-dG when paired with A residues (Fig. 4A, lane 3). We also observed that Fpg, from S. mutans, does not have the ability to remove adenine from A/C and A/G mismatches (data not shown). Base-excision activity of nucleotides (A, C, or T) mispaired with 8-oxo-dG, was absent in the extracts from the MutY-deficient strain, when compared to extracts from the parent strain (Fig. 4B, lanes 2–5 and 8–9). We could not identify MutY excision of DNA substrates containing G base-paired to 8-oxo-dG residues, suggesting this particular lesion is processed exclusively by the Fpg enzyme (Fig. 4A,B, lanes 6 and 7). However, a functional MutY was necessary for the recognition and processing of adenine bases mismatched with cytosine or guanine (Fig. 4B, lanes 11 and 13). Control reactions, with cell extracts of the genetically complemented mutYc strain, showed that the product bands in Fig. 4B correspond to processing by MutY of A, C or T base-paired with 8-oxo-dG (Fig. 4D, lane 2).
With the aim of identifying the AP endonuclease activity, we analyzed substrate specificities of the parent strain; using whole-cell extracts to treat a substrate carrying the AP site analog, tetrahydrofuran (THF) (Fig. 4C, lanes 1 and 5). Using extracts of the Smx-deficient strain, we observed low levels of AP endonuclease activity with the THF-containing substrate (Fig. 4C, lane 2), suggesting partial overlap with the second endonuclease in the organism (Smn). In contrast, in experiments in which the THF-containing oligomer was annealed to a radiolabeled, complementary oligomer, we saw that a functional Smn, and not Smx, was necessary for the recognition and processing of abasic sites (Fig. 4C, lanes 6 and 7). Use of genetically complemented smxc and smnc whole-cell extracts substantiated the THF-excision activity seen in Figure 4C; suggesting that the excision activities on the THF-containing oligomer are intrinsic to Smx and Smn (Fig. 4D, lane 3,4). Collectively, these results confirm that Fpg, MutY, Smx, and Smn repair proteins are part of BER in S. mutans, a multistep pathway that participates in the removal of oxidized bases.
It is well-established that accumulation of the oxidative stress product, 8-oxo-dG, leads to the generation of mutator phenotypes in several organisms (Glassner et al., 1998, Huang et al., 2006, Jain et al., 2007, Davidsen et al., 2005). Hence, to determine the biological effects of the loss of BER-encoding genes, and their contribution to mutation avoidance in the S. mutans chromosome, spontaneous mutation frequencies of the BER mutant strains and the parent strain were evaluated using rifampicin-resistance as the reporter. The loss of fpg, mutY, or smn in S. mutans resulted in elevated spontaneous mutagenesis (Fig. 5A). The loss of fpg resulted in an approx. 5-fold increase in mutation frequency over the parent strain, and the smn mutant strain was approx. 15-fold more mutable than the parent strain. Disruption of mutY resulted in a large increase in spontaneous mutation frequency (over 24-fold). The mutability of the smx mutant strain was similar to the parent strain. Finally, genetically complemented strains were created carrying functional copies of BER genes used in this study. The increase in mutation frequency was largely reversed in the genetically complemented BER strains (fpgc, mutYc, smxc, and smnc), as we observed no major differences compared to the parent strain (Fig. 5B). These data demonstrate that, with the apparent exception of smx, the BER genes of S. mutans are involved in mutation avoidance.
Rifampicin-resistance arises because mutations occur in the rifampicin-binding site of the DNA-dependent RNA polymerase, thus rifampicin cannot bind, allowing transcription of RNA and subsequent translation to proteins. Resistance to rifampicin has been linked to amino acid alterations found in three regions of rpoB, termed cluster I through III in E. coli (Aubry-Damon et al., 2002, Jin & Gross, 1988). Using mutations in the rpoB gene of S. mutans as our assay, we determined the contribution of the BER-encoding genes to mutations in the genome. Loss of the DNA glycosylases (Fpg, MutY) led to an increase in transversion mutations, nearly 20% more than observed in the parent strain (Fig. 5C). The result was consistent with the observation that Fpg and MutY were responsible for removing 8-oxo-dG (Fig. 4A, B). Loss of smx led to an approx. 30% increase in transition mutations when compared to the parent strain (Fig. 5C). On the other hand, deficiency in smn resulted in mutations similar to those caused by disruption of the DNA glycosylases (Fig. 5C). The loss of Smn activity resulted in an increase in transversion mutations, approx. 15% more than observed in the parent strain (Fig. 5C). The results correlated with the elevated 8-oxo-dG levels shown above (Fig. 3). The results indicate that, in S. mutans, BER enzymes play a major role in modulating the accumulation of mutations.
Generation of a mutator phenotype, from the loss of DNA repair function, has been shown to lead to more rapid adaptation to stressful environments (Jayaraman, 2009). To evaluate the link between a mutator phenotype and stress-induced adaptation in S. mutans, we determined the ability of the BER-deficient strains to survive long-term acid challenge. Cultures of the parent strain and BER strains were incubated over prolonged periods, in the presence of excess glucose, to determine the ability of the strains to survive their own sugar metabolism and resulting acidification. The results showed that strains with a mutator phenotype, fpg, mutY, and smn, survived longer than the parent strain, UA159 (Fig. 6A). For smx, a non-mutator strain, relative fitness was reduced significantly, with viable cells undetectable after 2 days (Fig. 6A). The genetically complemented fpgc, mutYc, and smnc strains reverted the increased long-term viability to that of the parent strain, suggesting that deficiencies in these genes were directly involved with this phenotype (Fig. 6B). Furthermore, the genetically complemented smx strain (smxc) showed a relatively modest improvement in the ability to survive long-term acid challenge. However, we found significant differences among the stationary-phase survival profiles of the smxc strain and the parent strain (Fig. 6B).
We further assessed the mutator phenotypes of the BER-deficient strains with regard to survival after exposure to H2O2. Disc inhibition assays showed that the smx strain was sensitive to H2O2 (Fig. 6C, D). Disruption of fpg showed no differences in sensitivity when compared to the parent strain (Fig. 6C, D). On the other hand, loss of smn and mutY resulted in 2 to 3-fold greater resistance to oxidative damage than the parent strain, UA159 (Fig. 6C, D). Interestingly, we noticed the presence of resistant colonies inside the growth inhibition zone (indicated by black arrows in Fig. 6C) for fpg, mutY, and smn strains, strongly correlating with changes in mutation frequencies and the adaptation of the organism to the stress presented by H2O2. Overall, the results indicated that deficiencies in specific BER enzymes affected survival in extreme acid and oxidative conditions, an important aspect of the ecology in dental plaque and in the virulence of S. mutans.
To further establish the link between BER mutator phenotypes and stress adaptation in S. mutans, a competition assay was performed to determine the ability of the BER mutant strains to resist the growth-inhibiting effects of H2O2 produced by the commensal oral streptococci S. gordonii and S. sanguinis. Strains disrupted in fpg and mutY exhibited an elevation in resistance to the peroxidogenesis of S. gordonii and S. sanguinis (Fig. 7A). Loss of smn did not result in significant differences in growth inhibition when compared to the parent strain (Fig. 7B). These observations correlated with the results from the mutability assays that showed greater adaptation to stressful conditions (Fig. 5). In contrast, loss of smx resulted in growth inhibition in the presence of the H2O2 produced by both commensals (Fig. 7B). To confirm that the inhibition of BER mutant strains of S. mutans was caused by H2O2, we added catalase to cultures of S. mutans and repeated the assay with both S. gordonii and S. sanguinis. As predicted, all strains were able to grow, and coexist, with both commensals, indicating that H2O2, secreted from the commensal species, was the growth-inhibiting compound (Fig. 7C). Overall, these results demonstrate that the peroxide producers, S. gordonii and S. sanguinis, were unable to antagonize growth of S. mutans when the organism was carrying deletions in BER genes. The data suggest that BER-derived mutations allowed the organism to outcompete major oral commensals in this in vitro assay.
The Galleria mellonella larval model of infection has been very useful in evaluating the ability of bacteria, and derived mutant strains, to resist oxidative bursts, in vivo, (Bergin et al., 2005). Studies with G. mellonella larvae have included Bacillus cereus, Enterococcus faecalis, Staphylococcus aureus, and S. mutans (Kajfasz et al., 2010). Recent studies with this model have provided insights into different aspects of microbial infection, colonization, and pathogenicity, including S. mutans (Kajfasz et al., 2010, Mukherjee et al., 2010, Aperis et al., 2007). The host response of G. mellonella larvae to pathogens includes mechanisms found in mammalian systems (e.g. lysozymes), as well as reactive oxygen species, and antimicrobial peptides (Seitz et al., 2003). Based on our preceding observations regarding the effects of the BER mutations on cell viability in oxidative, and acidic, conditions, we tested the ability of the S. mutans parent strain, UA159, and the BER mutant strains to colonize and kill larvae of G. mellonella. We found that rates of killing were significantly higher in larvae infected with strains expressing a mutable phenotype (fpg, mutY, and smn) than in those infected with the parent strain (Fig. 8A). Disruption of smx failed to show a significant difference in virulence, as compared to the parent strain. The increased rate of mortality in the caterpillar model seen with the BER mutant strains was reversed in the genetically complemented strains (fpgc, mutYc, smxc, and smnc), as no significant differences were observed in virulence as compared to the parent strain (Fig. 8B). These findings suggested that the mutator strains were more resistant, than the parent strain, at overcoming the innate immune defenses of G. mellonella, leading to a faster larval death following infection.
Streptococcus mutans, a nearly universal infectious agent of the human oral cavity, is found on tooth surfaces in an environmental niche known as dental plaque, in which organisms are present that are capable of producing mM levels of hydrogen peroxide (Ryan & Kleinberg, 1995) and reducing extracellular pH to values below which tooth enamel dissolves (Jensen & Schachtele, 1983). Thus, the conditions on tooth surfaces are such that damage to DNA is a constant challenge to the organisms present in the milieu. The present study describes the characterization of the BER system in a major cariogenic pathogen, Streptococcus mutans, and its role in the stress-response system of the organism.
Initial experiments with BER mutant strains, characterizing growth and resistance to hydrogen peroxide, showed that loss of smx resulted in hypersensitivity to oxidative stress. The observation is consistent with earlier reports showing that xthA (Exo III) mutants are usually sensitive to oxidizing agents, including H2O2 (Demple et al., 1983, Galhardo et al., 2000, Cunningham et al., 1986). It has been suggested that ROS, generated from H2O2, can cause double-stranded breaks and chromosomal fragmentation, generating 3’ phosphate termini from the deoxyribose ring (Demple et al., 1986, Milcarek & Weiss, 1972). In E.coli, XthA accounts for 99% of the 3’ phosphatase activity that allows the removal of the phosphate termini (Barzilay & Hickson, 1995, Ljungquist et al., 1976, Milcarek & Weiss, 1972, Samuni et al., 1991). The data in this study suggests the strong possibility that the S. mutans Smx enzyme possesses a similar phosphatase activity. Consequently, loss of Smx would lead to accumulation of 3- termini, which would block the initiation of DNA repair synthesis, leading to genetic instability and/or cell death.
Results from substrate-specificity experiments confirmed that the Fpg, MutY, Smx, and Smn enzymes are responsible for removal of damaged bases, and further processing of AP sites, with significant differences and functional redundancy of enzymes. Fpg specifically removes 8-oxo-dG, base-paired with C, G, or T, from the deoxyribose sugar in DNA; whereas, MutY excised A,C, and T, when base-paired with 8-oxo-dG, similar to previous reports (Michaels & Miller, 1992, Slupska et al., 1996, Fowler et al., 2003).
Evidence obtained with fpg and mutY mutant strains was consistent with their involvement in the repair of oxidized bases in S. mutans. Interestingly, measurements of DNA oxidation in both DNA glycosylase strains (fpg and mutY) showed a linear correlation between the levels of 8-oxo-dG and mutation frequencies. The relatively higher mutation frequencies observed with the fpg and mutY mutant strains demonstrated that levels of internal oxidation of DNA, in particular formation of 8-oxo-dG, could be an important source of spontaneous mutagenesis. We also noticed that cells defective in mutY led to a much more pronounced mutator phenotype than loss of Fpg. This could be explained by a model in which MutY is responsible not only for removal of bases incorporated opposite an 8-oxo-dG (A-, C-, or T-8-oxo-dG), but also for mismatch repair as well (A/C, A/G); compared to Fpg, which repairs only oxidatively-damaged bases. Alternatively, it remains possible that the loss of MutY impairs the efficacy of the BER pathway, increasing the chance that errors will be retained throughout the genome, and to a greater extent, spontaneous mutations.
Previous data had shown that loss of Smx partially eliminates the ability to excise AP substrates, suggesting the presence of other endonucleases in the organism (Faustoferri et al., 2005). Here, we confirmed that Smn plays the role of the secondary AP endonuclease. The substrate-specificity data showed that both Smx and Smn have AP endonuclease activity, nicking the DNA strand on the 5’ side of the AP site. Overall, we have shown that both enzymes confer partially overlapping activities, acting essentially as potential back-ups.
In addition to H2O2 sensitivity, disruption of Smx markedly diminished survival following long-term acid exposure, when compared with smn and the parent strain. Previous reports have indicated that depurination and depyrimidination of DNA can occur at high rates at acidic and neutral pH (Lindahl, 1993). In the present study, elevated levels of transcription for smx were observed under low as well as neutral pH, suggesting detection, by the organism, of damaged DNA during growth in general. The data suggests that the enhanced susceptibility of the smx strain to low pH, and oxidative stress, may be a consequence of accumulating AP sites (either spontaneously or via DNA glycosylases) during the repair of acid-mediated damage. This also suggests that Smx plays a major role in counteracting lesions induced by acid, while Smn has a subordinate role. The concept is supported, in part, by previous reports showing that E. coli XthA (Exo III) constitutes approx. 90% of the AP endonuclease activity in crude extracts; whereas endonuclease (Nth) accounts for most of the residual activity (Ljungquist et al., 1976, Cunningham et al., 1986). In S. mutans, Smn appears to account for a similar proportion of the total AP endonuclease activity, and was unable to compensate for loss of Smx.
Analysis of the AP endonuclease-deficient strains revealed that the smn strain experienced greater spontaneous mutagenesis than the smx and parental strain, indicating that Smn has the dominant role in mutagenic avoidance. The mutation spectrum of the smn and smx strains revealed a specific increase in transversion mutations in the smn strain. The mutagenic profile in the smn mutant can be ascribed to another Smn activity, other than the AP endonuclease. In fact, we showed that Smn is capable of recognition and cleavage of DNA sites containing mismatched bases paired with THF residues. We interpret the findings to indicate that Smn has multifunctional capability. Studies are presently under way to extend the understanding of substrate recognition and catalysis in both endonucleases.
In addition to the clear role of BER in managing genomic damage caused by environmental factors, it is also possible that BER-deficient strains may have enhanced tolerance to stress by fluctuation of mutation frequencies. This possibility is based on the findings that the rise in mutation frequency in the Fpg, MutY, and Smn-deficient strains coincided with the development of resistant subpopulations to both acid and oxidative stressors. Recent studies, in other bacterial systems, have correlated increments in both mutation rates and virulence of certain pathogens (Taddei et al., 1997, Wright, 2004), including studies with mutator strains of uropathogenic E. coli (Labat et al., 2005); epidemic serogroup A Neisseria meningitis (Richardson et al., 2002, Davidsen et al., 2007), Salmonella sp. (LeClerc et al., 1996); Pseudomonas aeruginosa and Helicobacter pylori (Oliver et al., 2000, Bjorkholm et al., 2001, respectively). This may also be the case in S. mutans, where mutator phenotypes could enhance bacterial fitness and virulence. Data supporting this perspective were observed when the fpg, mutY, and smn strains were exposed to long-term acid and H2O2. The resulting cultures contained elevated populations of stress-resistant cells, compared to the parent strain, suggesting an advantage in survival in an environment with selective pressure. Differences between mutator and non-mutator strains were observed in competition experiments between S. mutans BER strains and the oral commensal, peroxide-producing strains, S. gordonii and S. sanguinis. Survival of mutator strains (fpg, mutY, and smn) was substantially favored over the non-mutator strain (smx). Moreover, we found that the strains exhibiting elevated mutation frequencies also exhibited greater virulence in the G. mellonella larvae model of infection, strongly suggesting that induction of the global oxidative stress response in the BER strains benefitted the organism in an infection model. The interpretation of these data is supported by a previous report concerning the global oxidative stress response in S. mutans and the larval model (Kajfasz et al., 2010).
Although mutations are generally deleterious, spontaneous lesions can create subpopulations of mutators that can generate beneficial mutations under appropriate selective conditions. Therefore, persistent BER strains can be viewed as comprising two populations, mutator and non-mutator, living in an equilibrium determined by the fitness cost and benefit of particular mutations. Nevertheless, we can stipulate that this would only be a short-term solution, since accumulation of lesions may result in the progressive loss of functional proteins required for the survival of the organism.
In summary, the present work demonstrates that oxidative DNA damage occurs in S. mutans, despite a variety of tolerance mechanisms, and that the BER system plays a role in repair of DNA damage. We demonstrated an association between oxidative DNA modifications in S. mutans and the development of mutator phenotypes, which included stress-resistance. These observations suggest that the major role of BER in S. mutans is mutagenic avoidance and repair of oxidatively damaged DNA. Furthermore, it appears that functional BER-encoded genes are not required for virulence in S. mutans. In fact, deficiency in specific BER-encoded genes leads to a short-termed beneficial effect on cell viability. Taken together, the physiological and genetic evidence shows that BER genes participate not only in maintaining the genetic stability of S. mutans, but also, play a major role in the stress-responsiveness of the organism to acidic and oxidative environments.
S. mutans strains used in this study are derivatives of S. mutans UA159 and are listed in Table S1. For routine applications, cultures in brain heart infusion medium (BHI; BD/Difco, Franklin Lakes, NJ) were grown overnight at 37°C in a 5% (vol/vol) CO2/ 95% air. Where appropriate, erythromycin (5 µg ml−1), kanamycin (1 mg ml−1), or spectinomycin (1 mg ml−1) was added to the medium. For growth curves, a Bioscreen C (Growth Curves USA, Haverhill, MA) system was employed to measure cell growth for 24 h at 37°C. OD600 was measured every hour with a 10-second shaking period before each reading, to evenly suspend cultures. Overnight cultures were diluted 1:20 in fresh BHI medium and incubated at 37°C in an atmosphere of 5% (vol/vol) CO2/ 95% air to an optical density of 600 nm (OD600) of 0.3. An aliquot (10 µl) of cell suspension was inoculated into each well containing 300 µl of the appropriate medium.
Escherichia coli was grown on LB medium shaking at 37°C. The following selective antibiotics were added, where needed: ampicillin (100 µg ml−1), erythromycin (500 µg ml−1), chloramphenicol (20 µg ml−1), and kanamycin (50 µg ml−1).
Strains utilized or created in this study are described below and listed in Table S1. In this study, we refer to the NCBI SMU gene designation (http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genome&Cmd=ShowDetailView&TermToSearch) (Ajdic et al., 2002). Cloning and subsequent mutation of the smx locus (S. mutans exonuclease) was essentially as described previously (Faustoferri et al., 2005). The UR101 strain from that study was re-engineered by transformation of S. mutans UA159 with pSMexoΔSpec17 (Table S2) and the new strain was named S. mutans UR111.
The endonuclease III gene, denoted as SMU.1650, was amplified from the S. mutans genome using the primer pair EndoIIIFwd and EndoIIIRev. The amplicon, containing 326bp of the 5' flanking region and 201bp of the 3' flanking region, was cloned into pGEM-T (Promega, Madison, WI) to create pGEMsmn and its identity was confirmed by sequencing. We designated the gene smn for S. mutans endonuclease. Insertional mutagenesis of the smn gene was performed using the EZ::TN transposase system from Epicentre (Madison, WI). First, we created a transposon containing an antibiotic resistance marker selectable in S. mutans. An erythromycin (ErmR) resistance cassette was excised from pTS19E (Aoki et al., 1986) by BamHI digestion and inserted into the BamHI site in pMOD-2<MCS> (Epicentre, Madison, WI), creating pMODErm. The resulting construct contained an ErmR cassette flanked by 19bp mosaic ends for transposition into the gene of choice. Plasmid DNAs, pGEMsmn plus pMODErm, were allowed to react with EZ::TN Transposase according to the manufacturer's directions. A portion of the resultant reaction mixture was used to transform E. coli DH10B. Transformants were initially plated on erythromycin-containing solid LB medium to select for transposition events into pGEMsmn. Isolates were plated on ampicillin-containing solid medium to eliminate those transposition events that would have occurred in the AmpR gene of the pGEMsmn plasmid, thereby enriching the population for transposon insertion into the smn gene fragment. Colonies that were ErmR and AmpR were sequenced to determine the precise location of the ErmR cassette. One isolate, named pGEMsmnErm, contained sufficient DNA on either side of the ErmR cassette to allow homologous recombination into the S. mutans genome.
S. mutans UA159 was transformed to ErmR by electroporation (Loimaranta et al., 1998) with plasmid pGEMsmnErm. Erythromycin-resistant transformants were screened by Southern hybridization for a double crossover event (data not shown). An isolate was identified that displayed the appropriate increase in size from the native smn gene and the resultant strain was named S. mutans UR112.
The formamidopyrimidine glycosylase (fpg) gene, denoted as SMU.1614, was PCR amplified from S. mutans genomic DNA using the primer pair SFapyFwd and SFapyRev (Table S3). The fpg gene and flanking region (200bp on the 5' end and 198bp on the 3' end) was cloned into pGEM-T and named pFapy. Next, the ErmR cassette from pTS19E (Aoki et al., 1986) was inserted into a HindIII site located at position +570 bp of the cloned fragment contained in pFapy. The resulting plasmid, pFapyErm, was used to transform S. mutans UA159 to ErmR using published procedures (Perry & Kuramitsu, 1981, Quivey & Faustoferri, 1992). Strains arising from the transformations were screened by Southern hybridization to verify the appropriate construction (data not shown) and one such strain was designated S. mutans UR118 and used in this study.
Construction of the insertional mutation in the A/G-specific adenine glycosylase (mutY) gene, denoted as SMU.1865, was obtained following a similar procedure as above. Essentially, the mutY gene and flanking region (282bp on the 5' end and 118bp on the 3' end) was PCR amplified from S. mutans genomic DNA using the primer pair MutYFwd and MutYRev and cloned into pCR-Blunt using a Zero Blunt® PCR Cloning Kit (Invitrogen, Carlsbad, CA), resulting in plasmid pCRmutY (Table S3). A BglII restriction site was engineered in the cloned mutY, using SOE PCR (Faustoferri et al., 2005, Horton et al., 1989, Hu, 1993) with primers MutYSOEFwd and MutYSOERev and named pCRmutYBgl. A gel-isolated DNA fragment containing an ErmR cassette (Aoki et al., 1986) was ligated into the engineered BglII site in the mutY gene. The resultant plasmid was named pCRmutYErm (Table S2), which was then used to transform S. mutans UA159. Transformants were screened via PCR to verify appropriate construction and one strain was designated S. mutans UR195.
Complementation cloning techniques were performed as previously described (Derr et al., 2012). Briefly, primers FpgPFwd and FpgGRev (Table S3) were used to amplify the fpg coding region and cognate promoter fragment for subsequent cloning into pSUGK-Bgl (Table S2). The correct integration of the fpg promoter, plus coding region, into the BglII site of pSUGKBgl, in the opposite orientation to gtfA, was determined by colony PCR. The correct construct, pSUGKfpg, was transformed into S. mutans UR118 and selected on BHI agar medium containing kanamycin. The complement strain was named S. mutans UR291 (fpgc). The mutY, smx, and smn complemented strains were created using similar methods. Primers used for amplification of the coding regions and cognate promoters are listed in Table S3. All other steps were completed as described above. The complement derivatives were designated as S. mutans UR303 (mutYc), S. mutans UR289 (smxc), and S. mutans UR290 (smnc).
Constructions of the CAT-reporter fusion strain were performed essentially as previously described (Derr et al., 2012). The smx promoter region was PCR amplified using a primer pair containing unique restriction sites: SmxUP17Fwd and SmxUPNotIRev (Table S3). The smx promoter amplicon was ligated into pGEM-T and named pGEMsmxp. The pGEMsmxp plasmid was digested with BamHI and NotI and ligated to pENTRCAT to create the plasmid pEsmxCAT. The resulting plasmid carrying an smx promoter-CAT fusion was cloned into the integration vector pBGKGW (-) using LR clonase (Invitrogen, Carlsbad, CA) as previously described (Derr et al., 2012) resulting in the plasmid pBsmxCAT. Correct construction of the smx promoter-CAT fusion, in the opposite orientation to the sucrose phosphorylase gtfA gene, a non-essential gene for cell viability, was screened by colony PCR and was verified by sequencing with the gtfseqKan primer. Plasmid pBsmxCAT was transformed into S. mutans UA159 to create strain UR114.
Cloning and subsequent creation of the smn-promoter CAT fusion was obtained following a similar procedure to the one detailed above. The smn promoter was PCR amplified using a primer pair containing unique restriction sites: SmnUPBamHIFwd and SmnUPNotIRev (Table S3). The smn promoter amplicon was ligated into pGEM-T and named pGEMsmnp. The pGEMsmnp plasmid was digested with BamHI and NotI and ligated to similarly digested pENTRCAT to create the plasmid pEsmnCAT. The resulting plasmid carrying a smn promoter-CAT fusion was combined with pBGKGW(-), in the presence of LR clonase, resulting in the plasmid pBsmnCAT. Correct construction of the smn promoter-CAT fusion in the opposite orientation to the gtfA locus was screened by colony PCR and verified by sequencing with the gtfseqKan primer. Plasmid pBsmnCAT was transformed into S. mutans UA159 to create strain UR115.
Background activity was determined using the strain S. mutans UR113 (Derr et al., 2012) containing a copy of the promoterless CAT integrated into the chromosome in the gtfA locus. CAT assays were performed essentially as previously described (Kuhnert et al., 2004). Results are represented as nmols chloramphenicol acetylated min−1 mg total protein−1, as detailed (Shaw, 1975).
The assay was performed as previously described, with some modifications (Kreth et al., 2008, Zirkle & Krieg, 1996). Single colonies of S. mutans BER and parent strains were used to inoculate 5 ml of BHI medium and incubated at 37°C in a 5% (vol/vol) CO2/ 95% air atmosphere. Overnight cultures were diluted 1:20 in fresh BHI medium and incubated at 37°C to an optical density of 600 nm (OD600) of 0.4. Two 10 ml samples from each flask were harvested by centrifugation at 3452 × g for 15 minutes at 4°C, washed, and pellets were placed at −80°C until processed. These samples served as the controls (pre-H2O2 exposure). Hydrogen peroxide was added to one of the flasks to a final concentration of 5mM. After 30 minutes, two 10 ml samples from both flasks were collected and treated as described above. Additional samples were removed at 3, 6, 12, 24 hours after H2O2 exposure. Final pH value of all cultures was similar, ranging from values of 5.2 to 5.5. Genomic DNA was isolated using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI). DNA samples of 2.5 µg in a volume of 20 µL were analyzed by gel electrophoresis on a 1.8% agarose gel. The gels were quantified with the available software NIH Image J (http://rsb.info.nih.gov/ij/). The intensity of the bands representing the fragmented DNA is expressed as a percentage of the total intensities relative to the 0 hr time point on the same gel.
Mutant and parent strains of S. mutans were grown overnight in BHI at 37°C in an atmosphere of 5% (vol/vol) CO2/ 95% air. Genomic DNA was isolated from parent (UA159) or BER-deficient strains (fpg, mutY, smx, and smn) as described above. A competitive ELISA assay for 8-oxo-dG was performed according to the manufacturer’s protocol (Cayman Chemical, Ann Arbor, MI). Reactions were conducted by adding the 8-oxo-dG antibody, tracer and either standards or DNA samples to an ELISA plate, which had been pre-coated with 8-oxo-dG antibody. After washing the plate to remove all unbound reagents, Ellman’s Reagent was added. Absorbance was measured at 412 nm. Controls without added DNA and appropriate blanks were also incorporated into the experiments.
The assay was performed as previously described (Faustoferri et al., 2005). The oligonucleotides used in this study are shown in Table S3. DNA substrates for the Fpg activity assay containing 8-oxo-dG were created by end-labeling oligonucleotides with [γ32P] ATP (Perkin Elmer, Waltham, MA) at the 5’ end using T4 polynucleotide kinase (Invitrogen, Carlsbad, CA) as per manufacturer’s directions. Unincorporated nucleotides were removed from labeled DNA using a QIAquick Nucleotide Removal kit (Qiagen, Chatsworth, CA) and eluted in TE buffer. Labeled strands of 8-oxo-dG were annealed to either Comp-A, -C, -G, or -T oligomers in a 1:2 molar ratio by heating to 90°C for 2 minutes and then cooling to room temperature over 3 hours. DNA substrates for the MutY activity assay containing Base A and/or Comp-A, -C, -G, or -T oligomers were created by end-labeling oligonucleotides with [γ32P] ATP (Perkin Elmer, Waltham, MA) at the 5’ end using T4 polynucleotide kinase (Invitrogen, Carlsbad, CA) as per manufacturer’s directions and annealed to either 8-oxo-dG or Comp-C, or -G, oligomers in a 1:2 molar ratio by heating to 90°C for 2 minutes and then cooling to room temperature over 3 hours.
DNA substrates for the Smx activity assay containing the 17mer THF oligomer were created by end-labeling oligonucleotides with [γ32P] ATP (Perkin Elmer, Waltham, MA) at the 5’ end using T4 polynucleotide kinase (Invitrogen, Carlsbad, CA) as per manufacturer’s directions and annealed to Comp-17mer oligomer in a 1:2 molar ratio by heating to 90°C for 2 minutes and then cooling to room temperature over 3 hours. DNA substrates for the Smn activity assay containing Comp-17mer oligomer were created by end-labeling oligonucleotides with [γ32P] ATP (Perkin Elmer, Waltham, MA) at the 5’ end using T4 polynucleotide kinase (Invitrogen, Carlsbad, CA) as per manufacturer’s directions and annealed to 17mer THF oligomer in a 1:2 molar ratio by heating to 90°C for 2 minutes and then cooling to room temperature over 3 hours.
Crude protein extracts were prepared, as previously described, from S.mutans cells grown overnight in BHI broth (Faustoferri et al., 2005). Reactions were conducted by mixing 25 µg crude extracts from parent (UA159), BER-deficient strains (fpg, mutY, smx, and smn) or genetically complemented BER strains (fpgc, mutYc, smxc, and smnc), DNA substrates (10 pmol), and assay buffer (200mM Tris-HCl, pH 7.8, 100 mM EDTA, 2.5 M KCl) in a volume of 50 µl. Reactions were incubated at 37°C for 30 minutes before the addition of 10 µl of dye (98% (v/v) formamide, 10 mM EDTA, 0.025% (w/v) bromophenol blue, and 0.025% (w/v) xylene cyanol). Samples were loaded onto a 20% denaturing polyacrylamide gel and electrophoresed for 2 hours at 60 W. Images were analyzed by phosphorimaging in a BioRad FX phosphorimager (BioRad, Hercules, CA).
Single colonies of S. mutans BER mutant strains were used to inoculate 5 ml BHI medium. Overnight cultures were diluted 1:20 in fresh BHI medium and incubated at 37°C in a 5% (vol/vol) CO2/ 95% air atmosphere to an optical density of 600 nm (OD600) ~1.0. An aliquot (100 µl) of the cultures was plated onto BHI agar medium containing 10 µg ml−1 rifampicin, and serial dilutions were plated onto BHI agar plates to determine viable cells. Frequency of spontaneous mutation to RifR was calculated by dividing the cfu ml−1 on Rif-containing plates by the total cfu ml−1. Number of replicates was as indicated in the legend.
Rifampicin-resistant mutations were determined by mapping and sequencing the regions of the rpoB gene to identify mutation specificity between smx, smn, fpg and mutY mutant strains (data not shown). The isolated colonies were picked from rifampicin-containing plates as described above, resuspended in 20 µl water, incubated at 90°C for 10 min, and spun at 17,900 × g for 1 min in a table-top centrifuge. The supernatant was used as a template to amplify the rpoB region using RpoB1 and RpoB12 primers (Table S3). The sequences obtained from each of the BER-gene-encoding mutant strains were aligned with sequences obtained from the parent strain, UA159, to identify mutations that had occurred. A total of 45 samples per strain were sequenced.
Survival after a long-term acid challenge was assayed as previously described (Kajfasz et al., 2010). Briefly, an overnight culture was diluted 1:20 in tryptone-yeast extract (TY) medium containing excess glucose (50 mM). The growth of the cultures was monitored until stationary phase was reached, at which point an aliquot was removed for serial dilution and plating on BHI agar medium. The cultures were incubated in TY medium containing 50 mM glucose at 37°C in a 5% (vol/vol) CO2/ 95% air atmosphere for several days, with serial dilutions of the cultures plated daily until growth was no longer detected. The plates were incubated for 48 h before colonies were counted.
Mutant and parent strains of S. mutans were grown overnight in BHI at 37°C in 5% (vol/vol) CO2/ 95% air atmosphere. Overnight cultures were diluted 1:20 in fresh BHI medium and incubated at 37°C in a 5% (vol/vol) CO2/ 95% air atmosphere to an optical density of 600 nm (OD600) of 0.3. Aliquots (100 µl) of cultures were spread on 30 ml BHI agar medium plates. Sterile Whatman paper discs (Whatman, Piscataway, NJ) were wetted with a 1 M H2O2 solution before being placed on a lawn of cells. Plates were incubated at 37°C in 5% (vol/vol) CO2/ 95% air for 24 hours before diameters of zones of inhibition were measured. Experiments were repeated four independent times with similar results.
The interspecies competition assays between S. mutans, S. gordonii and S. sanguinis were performed as previously described, with some modifications (Kreth et al., 2008). Overnight cultures of S. gordonii and S. sanguinis were diluted 1:20 in fresh BHI medium and incubated at 37°C in 5% (vol/vol) CO2/ 95% air atmosphere to an optical density of 600 nm (OD600) of ~0.6. Aliquots (8 µl) of cultures were inoculated on “nutrient-rich” agar media (BHI plus 1% sucrose), which favors H2O2 production. After a 24 hour incubation period, BER mutant strains and the parent strain, grown to similar OD600 values, were inoculated next to S. gordonii and S. sanguinis, and plates were incubated for an additional 24 hr and photographed. For the catalase control, 8 µg ml−1 of catalase (bovine liver, Sigma, St. Louis, MO) was added to BER mutant strains and parent strain prior to plating.
G. mellonella killing assays were performed as previously described (Kajfasz et al., 2010). Groups of 20 larvae, ranging from 0.2 to 0.3 g in weight, with no signs of melanization, were randomly chosen and used for subsequent infection. BER mutant strains and the parent strain, S. mutans UA159, were grown overnight in Brain Heart Infusion broth with 5% (v/v) horse serum at 37°C in 5% (vol/vol) CO2/95% air atmosphere. Cells were diluted in the morning, washed and resuspended to 107 cells ml−1 in 0.9% NaCl. For the heat-killed control, cells were exposed to 75°C for 50 min and treated as cells above. A Hamilton syringe was used to inject 5 µl aliquots of the inoculum into the haemocoel of each caterpillar via the last proleg. After injection, larvae were incubated at 37°C. Caterpillars were monitored at various time intervals and considered dead when they showed no movement in response to touch. Kaplan-Meier killing curves were plotted and estimation of differences in survival were compared using the log-rank test. Experiments were repeated three independent times with similar results.
This study was supported by the Training Program in Oral Sciences T32 DE-07165 (K.G.) and NHI/NIDCR DE-13683 (R.G.Q.). We thank Andy Cardillo, Tracey Householder, Benjamin Metcalf, and Kelly Monahan for technical assistance.