INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Microorganisms have developed highly efficient mechanisms that allow them to adapt rapidly and survive a variety of physical and chemical stress conditions such as oxygen availability, redox potential, temperature, pH, and osmolarity (33). One of these adaptations is the utilization of molecular oxygen as a final electron acceptor when facultative bacteria are shifted from anaerobic to aerobic conditions (45). Consequently, during aerobic growth, generation of the reactive oxygen species (ROS) superoxide anion (O₂) and of hydrogen peroxide (H₂O₂) is unavoidable, and the highly reactive oxidant hydroxyl radical (OH^·) also may be formed through the Fenton reaction of H₂O₂ with free transition metals such as ferrous iron (14). ROS are potent cellular oxidizing agents that damage proteins, membrane lipids, and DNA (14, 19, 40). To minimize this damage, microorganisms eliminate the harmful effect of oxygen by-products by the synthesis of ROS-scavenging enzymes such as superoxide dismutase, catalases, and peroxidases, oxidative damage-repairing enzymes, and other proteins with unknown functions (14, 40).

One important aspect of this ROS response is that treatment of facultative and aerobic bacteria with sublethal concentrations of H₂O₂ induces a protection of the cells against levels of H₂O₂ that would be otherwise lethal (4, 9, 10, 25). This peroxide response in Salmonella typhimurium and Escherichia coli results in the synthesis of at least 30 proteins, of which 9 are under OxyR regulatory control. These include catalase (KatG), alkyl hydroperoxide reductase (AhpCF), glutathione reductase (GorA), and a nonspecific DNA-binding protein (Dps/PexB) involved in protection of DNA against oxidative damage (1, 14, 20). Similarly, in response to both oxidative stress and stationary phase, Bacillus subtilis induces the synthesis of KatA, AhpCF, Dps and MrgA, an oxidative stress and metalloregulated Dps homologue (3, 8, 17).

Generally there is a paucity of information about the adaptive mechanisms that confer aerotolerance and survival of anaerobic bacteria in the presence of either oxygen or ROS. However, induction of an oxidative stress response has been shown to occur in the aerotolerant opportunistic human pathogen Bacteroides fragilis (29, 35, 37). B. fragilis is among the most aerotolerant of anaerobic bacteria and is able to resist the presence of molecular oxygen for up to 2 to 3 days without a significant loss of viability (32, 42). This aerotolerance is dependent on the ability to synthesize new proteins immediately following a shift to aerobic conditions or treatment with sublethal concentrations of H₂O₂ (29, 35, 37). At least 28 newly synthesized proteins are induced upon oxygen exposure or addition of exogenous H₂O₂ to mid-log-phase anaerobic cultures. Among these oxidative stress induced proteins are the ROS-scavenging enzymes catalase KatB (29) and superoxide dismutase (16), but the mechanism(s) regulating the synthesis of these proteins is not understood. Recently, we have shown that expression of the B. fragilis katB gene is transcriptionally regulated by oxidative stress and by carbon and energy limitation in the absence of oxygen (30). Investigations of the regulatory mechanisms as well as the roles that these proteins play in the aerotolerance of B. fragilis have led us to the isolation of a constitutive H₂O₂-resistant mutant. In this study, we present physiological and genetics analysis of this mutant which contribute to our understanding of the gene(s) involved and the regulation of inducible protection against peroxides in aerotolerant anaerobic bacteria.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and growth conditions. B. fragilis 638R (27) was grown anaerobically in brain heart infusion broth supplemented with hemin, cysteine, and NaHCO₃ (BHIS) for routine cultures (38). Cultures were also grown in chemically defined medium (44) for some enzyme analysis in crude extracts.

Killing assays. Induction of the peroxide stress response with a sublethal concentration of H₂O₂ was carried out as follows. Mid-log-phase cells grown in BHIS to an A₅₅₀ of 0.3 (approximately 2 × 10⁸ to 4 × 10⁸ cells/ml) were pretreated with 150 µM 2,2'-bipyridyl and 50 µM H₂O₂ for 15 min followed by a second addition of 50 µM H₂O₂ for 15 min. Treatment with bipyridyl was shown to be necessary for accurate viable counts (29). Then the cultures were split in 10-ml aliquots and challenged with H₂O₂ ranging from 0 to 100 mM for 15 min. For studies with cumene hydroperoxide (CHP), the cultures were pretreated as described above, and then CHP (in dimethyl sulfoxide [DMSO]) was added to final concentration of 0 to 5 mM. To determine viable counts, treated cultures were diluted in brain heart infusion broth containing 150 µM 2,2'-bipyridyl and 10 µg of bovine liver catalase per ml, plated on BHIS agar, and incubated for 3 to 5 days at 37°C. All procedures were performed within an anaerobic chamber. Exposure of anaerobic cultures to aerobic conditions was carried out as previously described (29).

Mutant selection. The procedure for isolation of the H₂O₂-resistant mutant described in this study was a modification of the continuous enrichment/selection method described by Hartford and Dowds (17). Briefly, B. fragilis 638R was grown in 10 ml of BHIS to approximately 2 × 10⁸ to 4 × 10⁸ cells/ml, treated with 10 mM H₂O₂ for 15 min, and then plated on BHIS agar for 24 to 48 h. One of the surviving clones was grown overnight in BHIS containing 10 mM H₂O₂. The resulting culture then was successively passaged overnight in BHIS containing 20, 30, 40, and then 50 mM H₂O₂. The B. fragilis culture resistant to 50 mM H₂O₂ was isolated on BHIS agar, and a single colony (designated IB263) was selected for further experiments.

Construction of a katB::cat transcriptional fusion. A katB::cat transcriptional fusion was constructed and integrated into the B. fragilis chromosome in order to study trans-acting regulation of the katB promoter. Briefly, a 0.8-kb blunt-ended TaqI DNA fragment containing the ribosome-binding site and coding region of the Tn9 chloramphenicol acetyltransferase (cat) gene was ligated into the internal EcoRV and MscI sites of katB in pFD567 (28). Restriction digestion of the new construct pFD605(katB'::cat) showed that the cat gene was in the same orientation as katB. Subsequently, a 2.5-kb SmaI/NarI fragment from pFD605(katB'::cat) was cloned into the unique SmaI/NarI sites of the suicide vector pFD516 (39). Then a 3.8-kb NarI fragment from pFD620 carrying B. fragilis -glucosidase bglA gene (Genbank accession no. AF006658) was cloned in the unique NarI site of the construct to provide a target for single-crossover-targeted insertion in the B. fragilis chromosome. The bglA gene is not essential for growth in BHIS, and there is at least one additional -glucosidase gene in B. fragilis (31). The new construct, pFD688(katB'::cat) (Fig. 1), was mobilized from Escherichia coli DH5 by triparental mating into B. fragilis 638R and IB263, using standard filter mating protocols (36). Southern blotting hybridization analysis revealed that pFD688 was integrated into the B. fragilis bglA locus, creating strains 638R (katB⁺ katB'::cat) and IB263 (katB⁺ katB'::cat). Total RNA extraction and Northern blot analysis of katB mRNA were carried out exactly as previously described (30). The densitometry analysis of the autoradiograph was normalized to the relative intensity of total 23S and 16S rRNAs detected on the ethidium bromide-stained agarose gel.

View larger version (10K): [in this window] [in a new window]   FIG. 1.   Schematic diagram of the suicide shuttle vector pFD688 containing the transcriptional fusion katB'::cat and the bglA locus for targeted insertion into the chromosome of B. fragilis. Construction of the strains 638R (katB+ katB'::cat) and IB263 (katB+ katB'::cat) by using pFD688 is described in Materials and Methods.

Catalase and CAT enzyme assays. Catalase activity was measured spectrophotometrically as previously described (28). One unit of catalase is the amount of enzyme which decomposes 1 µmol of H₂O₂ per min at 25°C. The spectrophotometric assay for CAT was performed in crude extracts essentially as described by Brosius and Lupski (7).

Partial protein purification and PAGE. Partial purification of oxidative stress proteins was obtained by electroelution of the bacterial crude extract on a polyacrylamide gel at 40-mA constant current, using a Prep-Cell model 491 (Bio-Rad Laboratories, Inc., Melville, N.Y.). Fractions containing the proteins of interest were pooled and concentrated. Preparative and analytical nondenaturing polyacrylamide gel electrophoresis (PAGE) and denaturing sodium dodecyl sulfate (SDS)-PAGE were performed as described by Laemmli (21). Following electrophoresis, proteins were detected by staining the gel with either Coomassie blue R250 or Ponceau S. Protein concentration was determined by the Bradford method (6), using lysozyme as a standard.

N-terminal amino acid sequence and database comparison. The proteins resolved by SDS-PAGE were blotted to a polyvinylidene difluoride membrane in 10 mM CAPS (3-[cyclohexylamino-1-propanesulfonic acid])-10% methanol, pH 11.0 (23). The blotted proteins were subjected to fully automated solid-phase Edman degradation to determine the N-terminal amino acid sequence. The N-terminal sequencing was performed by D. Klapper, University North Carolina, Chapel Hill. Computer analyses of N-terminal amino acid sequences were performed with the University of Wisconsin Genetics Computer Group DNA sequence analysis software (11).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation and characterization of a hydrogen peroxide-resistant mutant. A B. fragilis hydrogen peroxide-resistant strain, IB263, was isolated following enrichment of the 638R cultures in increasing concentrations of H₂O₂. After several passages in BHIS in the absence of H₂O₂, cell viable counts were performed to determine whether resistance to H₂O₂ was constitutively expressed in IB263. The mutant strain IB263 was no longer killed by passage in media containing as much as 50 mM H₂O₂, suggesting that there was constitutive resistance to H₂O₂ due to a stable mutation(s) rather than a reversible and temporary physiological adaptation. This possibility was supported by the results in Fig. 2A, where it is clearly shown that IB263 was highly resistant up to 100 mM H₂O₂ with or without H₂O₂ induction. In contrast, the parent strain lost about 2 orders of magnitude in viability in the presence of 5 mM H₂O₂ when the peroxide response was induced, and there was essentially no protection in untreated cells.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Survival of mid-log-phase cells of B. fragilis 638R and IB263 following addition of hydrogen peroxide (A) and CHP (B) for 15 min. Data points represent B. fragilis 638R cultures pretreated () and not pretreated () with H2O2 and B. fragilis IB263 cultures pretreated () and not pretreated () with H2O2.

View larger version (13K): [in this window] [in a new window]   FIG. 3.   Survival of anaerobic B. fragilis 638R () and IB263 () mid-log-phase (A) and stationary-phase (B) cells shifted to aerobic conditions ( and , respectively). Cultures of mid-log-phase cells (A550 = 0.3) or early-stationary-phase cells (A550 = 1.1) were divided at time zero; one half was shaken at 250 rpm in air, and the other half was maintained anaerobically. Viable cell counts were determined at the times shown.

Deregulation of catalase activity and evidence for a trans-acting regulatory mechanism. Previous work had shown that induction of catalase was essential for full protection against exogenous H₂O₂ (29). Thus, to investigate whether constitutive resistance to H₂O₂ was correlated to an increase in catalase activity, both parent and mutant strains were tested by catalase assays (Fig. 4A). Catalase activity of approximately 5,100 U/mg of protein was found in crude extracts of the mutant strain grown in chemically defined medium, indicating that it was no longer repressed under anaerobic conditions. There was a >180-fold increase over the much lower activity (27 U/mg protein) found in the anaerobic cultures of the parent strain, suggesting that the mechanism(s) that controls catalase expression in the parent strain is no longer functional in the mutant. In addition, there was no further significant induction of catalase activity when anaerobic cultures of IB263 were exposed to oxygen or treated with hydrogen peroxide (6,700 and 5,600 U/mg of protein, respectively). In contrast there was approximately a 15-fold-higher activity induced in the parent strain following exposure to the same oxidative stress conditions (170 and 360 U/mg of protein, respectively). As the catalase regulation in the parent strain occurs at the transcriptional level (30), Northern blot hybridization analysis of the mutant was carried out. The results revealed that the level of katB mRNA present in strain IB263 under anaerobic conditions was approximately 20-fold higher than in the anaerobic culture of the parent strain (Fig. 5). katB mRNA obtained from the mutant exposed to oxygen or treated with hydrogen peroxide showed a slight increase in expression which was similar to the induction seen for the parent strain (between 30- and 35-fold increase). These results strongly suggest that the high levels of catalase in the mutant strain were due to the transcriptional deregulation of katB mRNA expression.

View larger version (27K): [in this window] [in a new window]   FIG. 4.   Catalase (A) and CAT (B) activities in crude extracts of mid-log-phase cells of B. fragilis strains grown in chemically defined medium under different oxidative stress conditions. The insets show amplified scales of the B. fragilis 638R catalase and CAT activities.

View larger version (80K): [in this window] [in a new window]   FIG. 5.   Northern hybridization analysis of mid-log-phase B. fragilis 638R and IB263 cells under different oxidative stress conditions. (A) Autoradiograph of a Northern blot probed with the katB internal fragment. (B) Ethidium bromide-stained agarose gel loaded with 30 µg of total RNA in each lane. The 23S and 16S rRNAs are also indicated. Lanes: 1, anaerobic cultures; 2, cultures treated with 500 µM H2O2; 3, culture treated with 1 mM potassium ferricyanide; 4, cultures exposed to aeration for 1 h.

Identification of overexpressed proteins in IB263. The data from the H₂O₂ and CHP survival studies suggested that the regulatory mutation in IB263 may affect the global oxidative stress response. Therefore, additional studies were performed to see if other oxidative stress proteins were constitutively expressed in the mutant strain. Nondenaturing PAGE of crude extracts from B. fragilis 638R and IB263 exposed to oxygen and hydrogen peroxide revealed several candidate oxidative stress proteins (Fig. 6). One overexpressed protein had an electrophoretic mobility identical to that of B. fragilis catalase KatB previously purified (28), which together with the results shown above confirms our findings showing that catalase is overexpressed in the mutant strain.

View larger version (106K): [in this window] [in a new window]   FIG. 6.   Nondenaturing PAGE (7.5 to 20% gradient polyacrylamide gel) of mid-log-phase crude extracts of B. fragilis 638R and IB263. Lanes: 1, anaerobic cultures; 2, anaerobic cultures pretreated with H2O2; 3, anaerobic cultures exposed to oxygen for 1 h; 4, approximately 2.5 µg of purified B. fragilis catalase KatB (28). Lanes 1 to 3 were loaded with approximately 50 µg of total protein; the gel was stained with Coomassie blue. Arrows show the major protein bands that are overexpressed in the mutant strain.

View larger version (35K): [in this window] [in a new window]   FIG. 7.   (A) Alignment of the N-terminal amino acid sequence of a B. fragilis Dps homologue (Bf-Dps) with those of Synechococcus strain PCC7942 (Synec-DpsA) (26), nonheme iron-binding ferritin of L. innocua (Lisin-Fer) (5), B. subtilis MrgA (Bacsu-MrgA) (8), H. pylori NapA (Helpy-NapA) (13), H. influenzae (Haein-YD49; GenBank accession no. P45173), A. variabilis low-temperature-induced protein (Anava-YLT2; GenBank accession no. P29712), and E. coli Dps/PexB (Ec-Dps/PexB) (1). (B) Alignment of the N-terminal amino acid sequence of a B. fragilis AhpC homologue (Bf-AhpC) with those of E. coli AhpC (Ec-AhpC; GenBank accession no. D90701), S. typhimurium AhpC (Sty-AhpC; GenBank accession no. P19479), B. subtilis AhpC (Bacsu-AhpC) (17), and E. faecalis AhpC (Efaec-AhpC; GenBank accession no. AF016233). Consensus of at least 50% identical amino acid residues is denoted by black boxes; conserved amino acid substitutions are depicted by grey boxes.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

When mid-log-phase cells of B. fragilis are shifted from anaerobic to aerobic conditions, they are no longer able to maintain cell division but cell viability remains high even after oxygen exposure for up to 3 days. We are interested in understanding the physiological adaptations that occur during this period of reversible growth arrest under aerobic conditions. The initial studies suggested that the ROS-scavenging enzymes superoxide dismutase and catalase were important for aerotolerance in B. fragilis (24). More recently, Rocha et al. (29) have shown that response to oxidative stress in B. fragilis induces the synthesis of at least 28 proteins in the presence of oxygen or hydrogen peroxide. This finding is very interesting since this obligate anaerobe is not able to shift to an aerobic metabolism and maintain growth under aerobic conditions but does respond with the synthesis of a protective mechanism against the toxic effects of ROS. In this study, we describe the isolation and partial characterization of a B. fragilis H₂O₂- and organic peroxide-resistant mutant and identification of constitutively expressed proteins that may be potentially involved in protection against ROS in anaerobic bacteria.

The resistance to high levels of H₂O₂ in mutant strain IB263 was correlated with constitutive catalase activity, and the important role that KatB plays in protection against an exogenous source of H₂O₂ is consistent with this observation (29). In addition, evidence showing that B. fragilis produces homologues of AhpC and Dps, which are known to be involved in resistance to peroxides and oxidative DNA damage, was presented (1, 22, 41). Taken together, the constitutive resistance to hydrogen peroxide and organic peroxide and the overexpression of KatB, AhpC, and Dps in IB263 suggest that these activities are linked to a peroxide resistance response and may be under a common transcriptional regulator. Strong support for this idea was provided by the promoter fusion experiments. The findings in Fig. 4B clearly showed that the wild-type katB promoter was deregulated in IB263 but regulated normally in the parent, which suggests that a trans-acting regulatory mechanism controlling the peroxide response exists in B. fragilis. However, at this point it is not possible to rule out the formal possibility that the deregulation of KatB, AhpC, and Dps resulted from multiple mutations induced by the H₂O₂ enrichment technique. This matter will be clarified in further studies on complementation or isolation of the mutation.

The peroxide response in facultative and aerobic bacteria has been extensively studied and found to be complex and tightly regulated. In E. coli and S. typhimurium, the H₂O₂ redox sensor and transcriptional activator OxyR induces in mid-log-phase cells the synthesis of nine proteins, including KatG, AhpCF, GorA, and Dps (2, 9). A constitutive OxyR mutant confers overexpression of these proteins and resistance to hydrogen peroxide (9). More recently, Zheng et al. (47) have shown that two OxyR conserved cysteine residues specifically sense peroxides, forming disulfide bonds leading to intramolecular conformational changes and activation of OxyR. In B. subtilis, a hydrogen peroxide-resistant mutant overexpresses KatA, AhpC, and MrgA (a metalloregulated Dps homologue), possibly due to a mutation in a transcriptional repressor (17) that may contain a redox-active metal ion cofactor (8). Thus, there are a variety of mechanisms able to sense oxidative stress in cells, and it will be of interest to determine the mode of control in anaerobic organisms.

Dps protein has been found in several aerobic organisms (1, 5, 8, 13, 26, 46), and one of its major functions is to protect DNA against oxidative damage (22). However, Dps may participate in gene regulation in stationary-phase and in peroxide-consuming reactions located at the chromosome (1, 26). Dps protein is related to the ferritin/bacterioferritin family of iron storage proteins, suggesting the possibility of divergence from a common ancestor (5, 26). Moreover, both AhpC and Dps in S. typhimurium (15, 43) and AhpC in Mycobacterium tuberculosis (12) were induced during interactions with macrophages, suggesting that these oxidative stress proteins may participate in survival to the macrophage oxidative burst killing mechanisms. These findings raise several questions as to how anaerobic bacteria have acquired genes involved in detoxification and protection against toxic ROS since they are adapted to live in the absence of molecular oxygen.

Despite the high resistance to hydrogen peroxide in B. fragilis IB263 under anaerobic conditions, there is no corresponding increase in resistance to oxygen exposure. In contrast, it seems that the mutant strain was slightly more sensitive to oxygen than the parent strain. Although the peroxide response may be effective in scavenging and detoxifying peroxides formed during oxygen exposure, it is not clear whether this deregulated peroxide response would eventually affect other protective mechanisms, such as a superoxide response, that might be involved in controlling aerotolerance in this strain. On the other hand, it is possible that other secondary mutations affecting the sensitivity to oxygen are present in IB263. However, evidence supporting the possibility that oxygen toxicity is somehow different from the toxicity exerted by the H₂O₂ is available from previous studies showing that the effect of H₂O₂ on the degradation and repair of DNA differs from the effect of oxygen in irradiated B. fragilis cells (34, 37). B. fragilis Bf-2 was more sensitive to DNA-damaging agents such as far-UV radiation, N-methyl-N'-nitrosoguanidine, ethyl methanesulfonate, acriflavine, and mitomycin C in the presence of oxygen than when treated with H₂O₂. Aerotolerance and resistance to ROS in anaerobic bacteria may be part of a more complex mechanism which still remains to be elucidated. In this regard, we have identified several classes of genes that are differentially induced by various oxidative stresses (unpublished), and this investigation will be the focus of a future report.