8-Oxo-7,8-Dihydroguanine Is Removed by a Nucleotide Excision Repair-Like Mechanism in Porphyromonas gingivalis W83

Abstract

A consequence of oxidative stress is DNA damage. The survival of Porphyromonas gingivalis in the inflammatory microenvironment of the periodontal pocket requires an ability to overcome oxidative stress caused by reactive oxygen species (ROS). 8-Oxo-7,8-dihydroguanine (8-oxoG) is typical of oxidative damage induced by ROS. There is no information on the presence of 8-oxoG in P. gingivalis under oxidative stress conditions or on a putative mechanism for its repair. High-pressure liquid chromatography with electrochemical detection analysis of chromosomal DNA revealed higher levels of 8-oxoG in P. gingivalis FLL92, a nonpigmented isogenic mutant, than in the wild-type strain. 8-OxoG repair activity was also increased in cell extracts from P. gingivalis FLL92 compared to those from the parent strain. Enzymatic removal of 8-oxoG was catalyzed by a nucleotide excision repair (NER)-like mechanism rather than the base excision repair (BER) observed in Escherichia coli. In addition, in comparison with other anaerobic periodontal pathogens, the removal of 8-oxoG was unique to P. gingivalis. Taken together, the increased 8-oxoG levels in P. gingivalis FLL92 could further support a role for the hemin layer as a unique mechanism in oxidative stress resistance in this organism. In addition, this is the first observation of an NER-like mechanism as the major mechanism for removal of 8-oxoG in P. gingivalis.

Porphyromonas gingivalis, a black-pigmented gram-negative anaerobic organism, is a major etiological agent implicated in adult periodontitis. The survival of this organism in the periodontal pocket would require an ability to overcome oxidative stress induced by reactive oxygen species generated by neutrophils and macrophages or exposure to air (8, 34). In general, oxidative stress resistance in bacteria involves antioxidant enzymes and/or endonucleases (3, 6, 17, 33, 40). Recent reports on P. gingivalis have documented the use of antioxidant enzymes such as superoxide dismutase or rubrerythrin as a mechanism to overcome oxidative stress (2, 3, 27, 33, 39). Additionally, P. gingivalis possesses a heme layer that has been reported to play a role in oxidative stress resistance (37, 38). This heme layer acts as an oxidative sink to eliminate the effects of hydrogen peroxide and other reactive oxygen species, thus protecting the cell membrane and other cellular components of the organism from oxidative damage (37).

Oxidative DNA damage is a major consequence of oxidative stress. This damage is induced by the modification of nucleotide bases by the reactive oxygen species. 8-Oxo-7,8-dihydroguanine (8-oxoG), one such example, is produced abundantly in DNA exposed to free radical and reactive oxygen species (40). This modification, if not repaired, can result in mutagenesis and lethality to the organism (14, 15, 22). Base excision repair (BER) and nucleotide excision repair (NER) are two known mechanisms for the repair of oxidative DNA damage that are conserved in many organisms including eukaryotes. Removal of 8-oxoG appears to occur mostly by BER, which in Escherichia coli involves the formamidopyrimidine glycosylase (Fpg) enzyme encoded by the mutM gene (5, 6, 11, 24). NER is unique due to its ability to repair a wide spectrum of DNA lesions. Proteins including UvrA, -B, -C, and -D are involved in the recognition of the lesion, cleavage on the 3′ and 5′ ends of that lesion, and the release of the patch of DNA including the damaged base (4). While these DNA repair mechanisms have been described for many organisms, there is a gap in our knowledge on similar mechanisms in P. gingivalis.

In the vimA (virulence modulating)-defective mutant P. gingivalis FLL92, there was alteration in gingipain maturation-activation and distribution (30). Thus, this isogenic mutant was non-black pigmented and missing cell-associated gingipain activity. Because the gingipains RgpA and Kgp are involved in the acquisition and binding of heme on the bacterial cell surface (19, 36), any defect in these proteases would affect heme binding and accumulation on the surface of P. gingivalis. This could raise important questions concerning the ability of P. gingivalis to survive in the periodontal pocket, in the presence of reactive oxygen species, where increased temperature, due to inflammation, can downregulate gingipain activity (31). Taken together, this would suggest that an absent and/or decreased accumulated heme layer could significantly affect the ability of P. gingivalis to withstand oxidative stress.

Under conditions of oxidative stress, we have examined DNA damage in P. gingivalis mutant FLL92 compared with the wild-type strain. In the presence of hydrogen peroxide, there was an increase in 8-oxoG in P. gingivalis FLL92 over that in the wild-type strain W83. Furthermore, P. gingivalis FLL92 was more resistant to oxidative stress than the wild type was, which appears to correlate with an increase in the repair activity of 8-oxoG. Analysis of the complete genomic sequence of P. gingivalis did not reveal an Fpg homologue for the removal of 8-oxoG (28) (http://www.oralgen.lanl.gov/). Here we report that the enzymatic removal of 8-oxoG is catalyzed by an NER-like mechanism. To our knowledge, these data represent the first report of a repair mechanism for 8-oxoG in P. gingivalis.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

All anaerobes (P. gingivalis W83 and FLL92, Fusobacterium nucleatum ATCC 25866, Prevotella intermedia ATCC 25611, and Tannerella forsythensis) were grown in brain heart infusion (BHI) broth (Difco Laboratories, Detroit, Mich.) supplemented with hemin (5 μg/ml), vitamin K (0.5 μg/ml), and cysteine (1%). T. forsythensis was supplemented with N-acetylmuramic acid (0.002 mg/ml). All anaerobes were maintained in anaerobic conditions (10% H₂, 10% CO₂, and 80% N₂) in an anaerobic chamber (Coy Manufacturing, Ann Arbor, Mich.) at 37°C. E. coli (TOP10) was grown in Luria-Bertani broth in aerobic conditions at 37°C.

Sensitivity testing.

P. gingivalis W83 and P. gingivalis FLL92, the vimA-defective isogenic mutant strain, were tested for sensitivity to hydrogen peroxide. P. gingivalis W83 and FLL92 were grown to early log phase (optical density at 600 nm [OD₆₀₀] of 0.2) in BHI broth supplemented with hemin and vitamin K. H₂O₂, at concentrations of 0.1, 0.25, 0.5, or 1 mM, was then added to the cell cultures, which were further incubated for 16 h. At 4-h intervals, the OD₆₀₀ of the cells was determined. Cell cultures without peroxide were used as controls.

Synthesis of 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG).

All reactions followed the method of Kasai and Nishimura (16). The concentration of the solution was calculated using the extinction coefficient for 8-oxodG (ɛ = 10,300) by determining the UV absorbance at 293 nm on a UV-visible light lambda 3B spectrophotometer (Perkin-Elmer).

DNA isolation and evaluation of 8-oxoG.

P. gingivalis strains W83 and FLL92 were grown to mid-log phase (OD₆₀₀, 0.6 to 0.7) and treated with 0.25 mM hydrogen peroxide or left untreated for 20 min while being incubated anaerobically at 37°C. The bacteria were then removed and centrifuged at 9,000 × g for 15 min at 4°C. The supernatant was removed immediately, and the chromosomal DNA was extracted as described elsewhere (23). DNA (100 μg) of each of the treated and untreated P. gingivalis strains was digested with 10 μg of nuclease P1 (Roche Biochemicals, Indianapolis, Ind.)/μl for 15 min at 65°C. The samples were cooled on ice, and then 6 U of shrimp alkaline phosphatase (Roche Biochemicals) was added and the mixture was further incubated at 37°C overnight. 8-OxodG and 2′-deoxyguanosine (2dG) were resolved by high-pressure liquid chromatography (HPLC) and quantified by electrochemical detection (ECD) by using a CoulArray ECD system (ESA Model 5600A) 3-μm, 150- by 4.6-mm C₁₈ column. The nucleosides were eluted from the column with an isocratic mobile phase consisting of 100 mM sodium acetate (pH 5.2)-5.0% methanol. The mobile phase was filtered and degassed prior to application followed by filtration with an 0.45-μm-pore-size, HV Durapore filter. The HPLC-ECD system was calibrated with 6.25 fmol to 1.25 pmol of 8-oxodG (synthesized in our laboratory) and 5 pmol to 5 nmol of 2dG standard (Sigma Chemical Company, St. Louis, Mo.). The retention times for 8-oxodG and 2dG standards were 6.0 and 8.0 min, respectively. Subsequent identification and quantitation of 8-oxodG were performed by comparison with retention time and by the method of peak area measurement with a linear regression curve for standard solutions. All samples were assayed in triplicate in two independent experiments.

Oligonucleotide labeling and annealing procedures.

Oligonucleotide fragments used in this study (see Fig. 2) were made by Synthegen, Houston, Tex. All oligonucleotides were 5′ end labeled with [γ-³²P]ATP (ICN) and T4 polynucleotide kinase (Promega Corp.) under conditions recommended by the enzyme supplier. The labeled oligonucleotides were purified to remove excess unincorporated γ-³²P by centrifugation through Micro Bio Spin 6 Sephadex columns (Bio-Rad Laboratories, Hercules, Calif.). The purified labeled single strand was annealed to a twofold molar excess of an unlabeled complementary strand for 5 min at 95°C and cooled slowly to room temperature.

FIG. 2.

Oligonucleotide sequences used in this study. X, 8-oxoG (O1 to O5), C (N1), or U (U1); P, C (O1 to O5) or G (U1 and N1).

Preparation of crude bacterial extracts.

P. gingivalis, F. nucleatum, P. intermedia, and T. forsythensis were grown overnight in BHI. A 1/10 dilution of each bacterial strain was made in fresh BHI medium and grown to an OD₆₀₀ of 0.6. E. coli was grown in a similar manner under aerobic conditions. The cell pellets were collected by centrifugation at 9,000 × g for 10 min at 4°C. The cell pellet was treated with protease inhibitors, resuspended in 5 ml of 50 mM Tris-HCl (pH 8.0)-1 mM EDTA lysis buffer, and subjected to eight freeze-thaw cycles. Cell debris was removed by centrifugation at 12,000 × g for 20 min at 4°C. The protein concentration of the supernatant was determined using the DC protein assay kit (Bio-Rad).

Glycosylase assay.

Labeled and annealed oligonucleotides (2 pmol) were incubated with the anaerobe bacterial extract (20 μg), E. coli extract (1 μg), or P. gingivalis extract (2 μg) in 1× enzyme buffer supplied with the uracil DNA glycosylase (Ung) or Fpg enzyme at 37°C for 1 h. An equal volume of loading buffer (98% formamide, 0.01 M EDTA, 1 mg of xylene cyanol/ml, and 1 mg of bromophenol blue/ml) was added to stop the reaction. Fifty picomoles of competitor oligonucleotide was added to each reaction mixture, which was heated to 95°C for 5 min to denature the duplex, which was afterward resolved by gel electrophoresis. Fpg and Ung (Trevigen Inc., Gaithersburg, Md.) control reactions were performed according to the method of Liu et al. (20). Briefly, 2 pmol of the specific oligonucleotide was incubated with 1 U of the enzyme at 37°C for 1 h in reaction buffers provided by the manufacturers. Cleavage of abasic sites after glycosylase treatment with Ung was performed by adding 5 μl of 0.1 M NaOH for 30 min at 37°C.

Gel electrophoresis and analysis of cleavage.

Reaction samples were loaded onto a 20% denaturing polyacrylamide gel (7 M urea) and run for 80 min at 550 V. The resulting bands corresponding to the cleavage products and uncleaved substrate were visualized using a Molecular Dynamics PhosphorImager (Amersham) and ImageQuant 5.0 software.

Statistical analysis.

Statistical analysis was performed using the Student two-tailed t test.

RESULTS

HPLC-ECD analysis of DNA extracted from P. gingivalis.

To detect the presence of 8-oxodG in P. gingivalis, chromosomal DNA from P. gingivalis W83 and the vimA-defective nonpigmented isogenic mutant FLL92, grown in the presence or absence of hydrogen peroxide, was extracted and analyzed by HPLC-ECD. As shown in Table 1, there was a statistically significant (P = 0.02) increase in 8-oxodG in P. gingivalis FLL92 compared to the wild type in the presence of hydrogen peroxide. There was no detectable 8-oxodG in W83 DNA in the absence of hydrogen peroxide. Taken together, these data suggest that there is more oxidative damage in cellular DNA of both P. gingivalis strains under conditions of oxidative stress. However, there was no statistically significant difference in oxidative DNA damage between the nonpigmented P. gingivalis mutant FLL92 and the wild type under normal physiological conditions.

TABLE 1.

8-OxodG detection by HPLC-ECD

P. gingivalis strain and treatment	8-oxodG concn (mol)
W83 without H2O2	None
W83 with H2O2	7.74 × 10−16
FLL92 without H2O2	9.52 × 10−16
FLL92 with H2O2	2.0 × 10−14a

^aP < 0.05.

Sensitivity of P. gingivalis FLL92 to hydrogen peroxide.

P. gingivalis W83 and the isogenic mutant P. gingivalis FLL92 were assessed for sensitivity to hydrogen peroxide. In contrast to the parent strain, P. gingivalis FLL92 demonstrated a greater resistance to hydrogen peroxide at a concentration of 0.1 mM (Fig. 1). Statistical analysis revealed that the increased resistance was significant. At higher concentrations (0.25 mM) or concentrations lower than 0.1 mM, the two strains showed similar profiles of sensitivity to those peroxides (data not shown). Taken together, these data suggest that P. gingivalis FLL92 has an increased resistance to hydrogen peroxide compared to the wild-type W83.

FIG. 1.

Sensitivity of P. gingivalis FLL92 to hydrogen peroxide. P. gingivalis was grown to early log phase (OD₆₀₀ of 0.2) in BHI broth. A 0.1 mM concentration of H₂O₂ (•, W83; ▪, FLL92) was then added to the cell cultures, which were further incubated for 16 h. Cell cultures without H₂O₂ (▴, W83; ⧫, FLL92) were used as controls. The results shown are representative of three independent experiments.

8-OxoG repair activity in P. gingivalis.

Because P. gingivalis FLL92 showed increased resistance to oxidative stress and had more detectable 8-oxodG than the wild-type W83 did, we investigated the repair activity of the mutagenic DNA lesion. Bacterial extracts from the P. gingivalis isogenic strains grown in the presence or absence of hydrogen peroxide were used in glycosylase assays with a [γ-³²P]ATP-5′-end-labeled 8-oxodG:C-containing oligonucleotide (24-mer) (Fig. 2, O1). If 8-oxoG is removed by a BER mechanism, a cleavage product corresponding to a 12-mer would be observed. As shown in Fig. 3A, the Fpg enzyme generated the expected 12-mer fragment; however, a fragment of a similar size was missing in P. gingivalis strains W83 and FLL92. Instead, a cleavage product of approximately 17 bases was observed (Fig. 3A, lanes 1 to 4). In addition, there was a greater intensity of the cleavage product generated by the P. gingivalis FLL92 (lanes 3 and 4) extract compared to the wild type. As a control experiment, the removal of uracil was also examined by using the same extracts. As shown in Fig. 3B, the level of activity for Ung was similar for the two P. gingivalis strains. Also, the cleavage products generated by the extracts from these strains were similar in size to that from the E. coli control. To rule out the possibility that a restriction enzyme in the crude extract was responsible for the 17-base cleavage product, glycosylase assays were performed with an oligomer where the 8-oxoG was replaced by guanine. There was no cleavage product similar in size to the 17-mer with extracts from P. gingivalis FLL92 (Fig. 3C, lane 3). As expected, there was no cleavage of this oligomer by the Fpg enzyme (Fig. 3C, N1 positive control). Collectively, these data indicate that the repair mechanism of 8-oxoG is different in P. gingivalis than in E. coli and suggest a mechanism of NER. These data also suggest greater 8-oxoG repair activity in the nonpigmented mutant P. gingivalis FLL92 than in the wild type.

FIG. 3.

8-OxoG and uracil removal activities of P. gingivalis strain W83 and FLL92 cell extracts. [γ-³²P]ATP-5′-end-labeled oligonucleotide O1 (A), U1 (B), or N1 (C) was incubated with P. gingivalis and E. coli cell extracts for 1 h (Fpg) or 1 min (Ung), electrophoresed, and visualized. (A) Lanes with controls (negative and positive) contain O1 oligonucleotides and the E. coli cloned Fpg enzyme (indicated above respective lanes); lane 1, O1 and P. gingivalis W83 extract (without H₂O₂); lane 2, P. gingivalis W83 extract (with H₂O₂); lane 3, O1 and P. gingivalis FLL92 extract (without H₂O₂); lane 4, O1 and P. gingivalis FLL92 extract (with H₂O₂). (B) Lanes with controls (negative and positive) contain U1 oligonucleotides and the E. coli cloned Ung enzyme (indicated above respective lanes); lane 1, U1 and P. gingivalis W83 extract (without H₂O₂); lane 2, U1 and P. gingivalis W83 extract (with H₂O₂); lane 3, U1 and P. gingivalis FLL92 extract (without H₂O₂); lane 4, U1 and P. gingivalis FLL92 extract (with H₂O₂); lane 5, U1 and E. coli cell extract. (C) Lanes with controls (negative and positive) contain O1 or N1 oligonucleotides and the E. coli cloned Fpg enzyme; lane 1, O1 and P. gingivalis FLL92 extract (1-h incubation); lane 2, O1 and P. gingivalis FLL92 extract (2-h incubation); lane 3, N1 and P. gingivalis FLL92 extract. The cleavage product sizes are clearly indicated on the right with arrows. All lanes (except enzyme controls) contain 2 μg of bacterial protein cell extract.

8-OxoG repair activity of P. gingivalis in different oligonucleotides.

In order to further characterize the repair believed to be occurring by an NER-like mechanism, we investigated the effect that different locations of 8-oxoG in the same oligonucleotide would have on the NER activity. Oligonucleotides were synthesized with 8-oxoG in four different positions, 5′ end labeled with [γ-³²P]ATP, and annealed with a complementary strand with cytosine (C) paired to 8-oxoG. As shown in Fig. 4, Fpg generated the expected cleavage fragments; however, there was no change in the cleavage fragment observed in P. gingivalis when 8-oxoG was at a different position in the oligomer. Taken together, these data suggest that the cleavage pattern for the 24-mer oligonucleotide containing 8-oxoG is the same regardless of the position of 8-oxoG within the fragment.

FIG. 4.

8-OxoG removal activity of P. gingivalis cell extract from different oligonucleotides. [γ-³²P]ATP-5′-end-labeled oligonucleotides O1, O2, and O3 (A) and O1, O4, and O5 (B) were incubated with P. gingivalis cell extracts for 1 or 2 h, electrophoresed, and visualized. (A) Lanes with controls (negative and positive) contain O1, O2, and O3 oligonucleotides and the E. coli cloned Fpg enzyme (indicated above respective lanes); lanes 1 to 3, O1, O2, and O3 oligonucleotides, respectively, incubated with P. gingivalis FLL92 cell extracts. (B) Lanes with controls (negative and positive) contain O1, O4, and O5 oligonucleotides and the E. coli cloned Fpg enzyme (indicated above respective lanes); lanes 1 to 3, O1, O4, and O5 oligonucleotides, respectively, incubated with P. gingivalis FLL92 cell extracts. All lanes (except enzyme controls) contain 2 μg of bacterial protein cell extract.

8-OxoG removal activity in other anaerobes.

To determine if other anaerobic periodontal pathogens have a similar mechanism for removal of 8-oxoG, crude extracts from P. intermedia, F. nucleatum, and T. forsythensis were used in glycosylase assays with oligonucleotides O1 and U1 (Fig. 2). There was no similar band indicative of NER observed in the three anaerobes (Fig. 5A, lanes 2 to 4) compared to controls or P. gingivalis FLL92 (lane 1). In addition, all three anaerobes possessed similar Ung activities (Fig. 5B, lanes 2 to 7). These data suggests that the NER-like mechanism for the removal of 8-oxoG might be unique to P. gingivalis.

FIG. 5.

8-OxoG and uracil removal activities of P. gingivalis, P. intermedia, F. nucleatum, and T. forsythensis cell extracts. [γ-³²P]ATP-5′-end-labeled oligonucleotides O1 (A) and U1 (B) were incubated with P. gingivalis cell extracts for 1 h (Fpg) or 1 or 5 min (Ung), electrophoresed, and visualized. (A) Lanes with controls (negative and positive) contain O1 oligonucleotides and the E. coli cloned Fpg enzyme (indicated above respective lanes); lane 1, O1 and P. gingivalis W83 extract (without H₂O₂), lane 2, O1 and P. intermedia extract; lane 3, O1 and F. nucleatum extract; lane 4, O1 and T. forsythensis extract. (B) Lanes with controls (negative and positive) contain U1 oligonucleotides and the E. coli cloned Ung enzyme (indicated above respective lanes); lane 1, U1 and P. gingivalis W83 extract (without H₂O₂); lanes 2 and 3, U1 and P. intermedia extract incubated for 1 and 5 min, respectively; lanes 4 and 5, U1 and F. nucleatum cell extract incubated for 1 and 5 min, respectively; lanes 6 and 7, U1 and T. forsythensis extract incubated for 1 and 5 min, respectively. Sizes of cleavage products are clearly indicated on the right. All lanes (except enzyme controls) contain 2 μg of bacterial protein cell extract.

DISCUSSION

Damage to bacterial membranes and DNA is a major consequence of oxidative stress (25). DNA base derivatives are the most abundant kind of lesions formed from oxidative stress (7). Guanines are especially susceptible to modification by the reactive oxygen species, yielding mainly 8-oxoguanine, which results in GC-to-TA transversion mutagenesis (9). Several reports have indicated that the primary repair of 8-oxoG is performed by BER (6, 10, 35, 40). Because oxidative stress represents a significant challenge to anaerobes such as P. gingivalis, it is imperative that this organism develop mechanisms of oxidative stress resistance that will allow for persistence in the inflammatory condition of periodontitis. Currently there is little information available on anaerobes addressing the role of DNA repair as a result of oxidative stress.

In P. gingivalis there is an established role for antioxidant enzymes and the hemin layer in oxidative stress resistance (27, 37-39). Because P. gingivalis is constantly exposed to oxidative stress, we evaluated the hypothesis that the nonpigmented mutant P. gingivalis FLL92 would be more sensitive to oxidative stress than the wild type. A comparison of the sensitivities of P. gingivalis FLL92 and the wild-type strain W83 to hydrogen peroxide suggests that the nonpigmented P. gingivalis mutant FLL92 is more resistant to oxidative stress than the wild type is. This unexpected finding raised the possibility that there might be other redundant mechanisms to prevent the entry of reactive oxygen species into the cells and/or to induce increased repair activity. HPLC-ECD analysis of chromosomal DNA extracted from P. gingivalis strains grown under oxidative stress indicated an increased accumulation of 8-oxoG in P. gingivalis FLL92 compared to the wild type. It is unclear if expression of the antioxidant enzymes is altered in P. gingivalis FLL92; however, these data could be consistent with the function of the heme layer (37, 38), which appears to be missing in P. gingivalis FLL92. Thus, this further suggests that the increased resistance of P. gingivalis FLL92 to oxidative stress may require increased repair activity of the oxidative damage base.

There was no detectable evidence of the E. coli mutM homologue (22, 26) in the P. gingivalis genome (28) (http://www.oralgen.lanl.gov/). Instead, a sequence homologue of the E. coli mutY gene, known to function in the removal of adenine paired with 8-oxoG (29), was identified. While we cannot rule out the possibility of an unidentified functional homologue of Fpg in P. gingivalis as observed in other organisms (32, 35), and the possibility of this function being masked, these observations may indicate that this organism may use a mechanism other than BER to remove 8-oxoG lesions. In this study the removal of the modified uracil base in P. gingivalis occurred by BER. This further confirms that Ung activity is highly conserved throughout bacteria and eukaryotes. However, removal of 8-oxoG in P. gingivalis occurred by an NER-like mechanism, in contrast to BER as observed in E. coli and other species (6, 10, 12, 21). Further, there was an increase in 8-oxoG removal-repair activity in P. gingivalis FLL92 compared to that in the wild type. This is consistent with the increased resistance of P. gingivalis to oxidative stress and may suggest the upregulation of an 8-oxoG removal-repair enzyme(s). It is unclear what gene(s) may play a role in this activity. A preliminary survey of the P. gingivalis genome has revealed genes encoding the UvrA, -B, and -C proteins, which in E. coli have been demonstrated to repair DNA lesions by NER (1, 10, 13). In addition, methyl-directed mismatch repair has also been implicated for its role in removing 8-oxoG in E. coli (41). Homologues of mutS and mutL have been identified in the P. gingivalis genome, and a role for these genes in the removal-repair of 8-oxoG lesions in P. gingivalis is under further investigation in the laboratory.

In this study we were unable to define a precise mechanism for the recognition and removal of 8-oxoG from duplex DNA. Altering the location of the 8-oxoG in different oligonucleotide fragments generated a similar cleavage pattern. The cleavage, however, of the oligonucleotide fragment to generate a 17-mer band, by using the P. gingivalis extracts, was specific for 8-oxoG-containing oligomers. This made it unlikely that a specific endonuclease restriction enzyme was responsible for the cleavage pattern. The presence of 8-oxoG in the oligonucleotide fragment reduced its degradation under our assay conditions. While these data support the presence of exonucleases in the bacterial extracts, they also suggest that multiple proteins may be involved in the 8-oxoG removal-repair activity. Thus, it is possible that DNA binding proteins in the 8-oxoG-containing oligomers prevented its degradation by the bacterial extracts, in contrast to oligomers without the modified base that were quickly degraded. These observations would be consistent with NER in E. coli, where the nucleotides around the lesion are protected by proteins participating in the removal-repair cascade (4). The mechanism for removal of the 8-oxoG appears to be unique to P. gingivalis compared to other periodontal pathogens including P. intermedia, F. nucleatum, and T. forsythensis. A functional homologue of the mutM gene has been identified in the F. nucleatum genome (http://www.oralgen.lanl.gov/).

This report, to our knowledge, has provided the first evidence for the presence of 8-oxoG in P. gingivalis exposed to oxidative stress. We have also shown that the mechanism for removal in this organism may involve an NER-like mechanism. The increased 8-oxodG and its corresponding higher repair activity in the nonpigmented isogenic mutant of P. gingivalis have raised interesting questions concerning mechanisms of oxidative stress resistance in P. gingivalis that facilitate its survival in the periodontal pocket. A possible scenario could occur in which the absence or reduction of the hemin layer, due to a decrease in protease expression (31) or increased hydrogen peroxide (18), may make P. gingivalis more vulnerable to oxidative damage. As a result P. gingivalis may upregulate DNA repair enzymes to compensate for the reduced or absent hemin layer. Further studies are needed to delineate this unique mechanism in P. gingivalis and determine its implications for the pathogenicity of this organism. Moreover, further investigation into the mechanisms involved in DNA repair in anaerobes will aid in our understanding of oxidative stress resistance, and these mechanisms may serve as potential targets for novel therapeutic strategies.