Abstract
We have reported that Neisseria gonorrhoeae is extremely resistant to reactive nitrogen species (RNS) including peroxynitrite (PN). Recent literature suggests that catalase can provide protection against commercial preparations of PN. Though wild-type gonococci were shown to be highly resistant to 2 mM PN, Neisseria meningitidis and a gonococcal katA mutant were both shown to be extremely sensitive to 2 mM PN. Analysis of translational fusions to lacZ of the catalase promoters from N. gonorrhoeae and N. meningitidis demonstrated that basal katA expression from gonococci is 80-fold higher than in meningococci, though meningococcal katA retains a greater capacity to be activated by OxyR. This activation capacity was shown to be due to a single base pair difference in the −10 transcription element between the two kat promoters. PN resistance was initially shown to be associated with increasing catalase expression; however, commercial preparations of PN were later revealed to contain higher levels of contaminating hydrogen peroxide (H2O2) than expected. Removal of H2O2 from PN preparations with manganese dioxide markedly reduced PN toxicity in a gonococcal katA mutant. Simultaneous treatment with non-lethal concentrations of PN and H2O2 was highly lethal, indicating that these agents act synergistically. When treatment was separated by 5 min, high levels of bacterial killing occurred only when PN was added first. Our results suggest that killing of N. gonorrhoeae ΔkatA by commercial PN preparations is likely due to H2O2, that H2O2 is more toxic in the presence of PN, and that PN, on its own, may not be as toxic as previously believed.
-
Edited by: J. Moir
Introduction
As an obligate human pathogen, Neisseria gonorrhoeae is well adapted for growth on a variety of mucosal surfaces (Edwards & Apicella, 2004). During the course of infection, gonococci are expected to face an onslaught of oxidative and nitrosative stresses (Bogdan et al., 2000; Hampton et al., 1998). Previous research has shown that multiple gonococcal gene products aid in survival under stressful conditions (Potter et al., 2009a, b; Seib et al., 2004, 2006; Stohl et al., 2005), and that hydrogen peroxide (H2O2), superoxide (
The concurrent presence of both
PN-mediated killing has been shown in a number of bacterial species (Alam et al., 2006; Dyet & Moir, 2006; Kuwahara et al., 2000; Yu et al., 1999; Zhu et al., 1992). Analysis of PN-mediated killing in bacteria can be complicated due to the extremely short half-life of PN in neutral aqueous solution (~1 s), as well as variation in media components that may react with PN to form different reactive species (Goldstein & Merényi, 2008; Schmidt et al., 1998). The use of molecular generators to generate PN through the production of
Methods
Bacterial strains and growth conditions.
All gonococcal mutant strains were derived from laboratory strain F62 (Table 1). N. gonorrhoeae and Neisseria meningitidis strain MC58 were grown on Difco GC medium base (Becton Dickinson) plates with 1 % Kellogg’s supplement (GCK) (Kellogg et al., 1963), in a 5 % CO2 incubator at 37 °C. Broth cultures were grown in GCP broth [proteose peptone no. 3 (Difco, 15 g), soluble starch (1 g), KH2PO4 (4 g), K2HPO4 (1 g), NaCl (5 g) per litre of distilled H2O] supplemented with 1 % Kellogg’s, and 0.042 % sodium bicarbonate (GCK broth), with shaking at 250 r.p.m. Escherichia coli DH10B was grown on plates using either GCK agar or LB agar [Bacto tryptone (Difco, 10 g), yeast extract (Difco, 5 g), NaCl (10 g), Bacto agar (Difco, 15 g) per litre). E. coli broth cultures were grown in GCK or LB broth.
Apr, Cmr, Ermr and Knr, ampicillin, chloramphenicol, erythromycin and kanamycin resistance, respectively.
Chemicals and reagents.
PN (Calbiochem), isoamyl nitrite (Acros), and H2O2 (Acros) were used in these studies. PN was also synthesized as described elsewhere (Uppu, 2006), with the following exception; H2O2 removal was accomplished by stirring PN solutions in the presence of granular MnO2 followed by filtration, rather than running PN solutions through a MnO2-packed column.
Survival counts following treatment with H2O2 and PN.
Broth cultures of N. gonorrhoeae were grown to an OD600 of 0.4–0.6 and diluted back to OD600 0.2 in GCK broth. For simultaneous PN/H2O2 treatment, 1 ml of this culture was added to a culture tube containing PN (final concentration 2.0 mM) and H2O2 (final concentration 0.2 mM). For sequential PN/H2O2 treatment, either PN (2 mM) or H2O2 (0.2 mM) was added to culture tubes before the addition of 1 ml gonococcal culture. The other reactive species was added 5 min later. The tubes were incubated at 37 °C for 45 min, and the bacteria were diluted and plated to determine viability. Survival was measured by dividing the final c.f.u. by the initial c.f.u. To treat cultures with reactive species in the absence of bicarbonate, cultures were filtered through 0.2 µm pore-size, 47 mm diameter Nucleopore polycarbonate Track-Etch membranes (Whatman) and rinsed with two volumes of fresh medium (without bicarbonate). The filter was then added to a tube containing fresh medium and the cells were resuspended. This culture was then diluted back to OD600 0.2 and treated as described above.
Gonococcal transformation.
A light suspension of type I cells (Kellogg et al., 1963) was prepared in 1 ml GCK broth containing 0.042 % NaHCO3 and 10 mM MgCl2. Purified plasmid DNA or ligation mixture was added, and 100 µl of the suspension was plated onto two GCK plates that were incubated right side up for between 6 and 9 h at 37 °C. Cells were then harvested from the plates and streaked on GCK plates containing the appropriate antibiotic for selection of clones. The katA mutation in strain F62 was created by transformation of F62 with genomic DNA from a katA mutant strain derived from gonococcal strain FA1090 (Soler-García & Jerse, 2004).
PCR and cloning.
Genomic DNA from gonococcal strain F62, meningococcal strain MC58 and E. coli strain DH10B was isolated for use as a PCR template. Promoter sequences for lacZ fusions, kat gene cassettes and chromosomal regions for insertional inactivation of genes were amplified with iProof High-Fidelity DNA polymerase (Bio-Rad). Clones were screened by PCR for the presence and orientation of the insert using AmpliTaq (Applied Biosystems). Primer sequences used for all constructs are available from the authors upon request.
Construction of lacZ fusions.
Translational lacZ fusions were constructed with pLES94 (Silver & Clark, 1995), using genomic DNA from F62 or MC58 as the template. PCR fragments and pLES94 were cut with BamHI. The digested insert and plasmid were ligated and cloned into E. coli DH10B. Transformants were selected on LB medium plates containing chloramphenicol (25 µg ml−1) and X-Gal (40 µg ml−1; Invitrogen). For site-specific mutagenesis of the gonococcal katA −10 element, splice overlap extension PCR was used. A compensatory base pair change in nucleotide −8 was incorporated into internal primers to change the −8 nucleotide from A to G. Plasmids were checked for the presence and orientation of the insert by PCR, and those plasmids that contained an insert in the correct orientation were used to transform F62. Colony PCR was performed on chloramphenicol-resistant colonies to confirm the presence of the reporter construct. PCR products were also sequenced to ensure that the appropriate fusion had been made.
β-Galactosidase assays.
Gene reporter activity was determined by β-galactosidase assays from wild-type and ΔoxyR strains grown in the presence or absence of 2 mM H2O2. Gonococcal cells were grown in GCK broth to OD600 0.5–0.7 and incubated with H2O2 (2 mM) for 1 h when appropriate. Cells were pelleted and resuspended in Z buffer (Miller, 1972), lysed with chloroform and 0.1 % SDS, and assayed as described by Miller (1972). Activity was reported in Miller units and the results were reported as the mean of at least six assays performed in duplicate from each day that the cultures were grown.
Construction of inducible gonococcal and E. coli kat strains.
Gonococcal katA, and E. coli katE and katG gene cassettes, which included the wild-type ribosome-binding sites, were PCR-amplified and treated with T4 polynucleotide kinase (NEB) to add phosphates to the ends of the PCR products. These fragments were ligated to PmeI-cut gonococcal complementation vector pGCC4 (Mehr & Seifert, 1998), and the ligation mix was used to transform chemically competent E. coli strain DH10B. Clones were selected on LB plates containing kanamycin (Calbiochem) at 50 µg ml−1. Restriction mapping was used to determine the presence and proper orientation of inserts. The plasmid was isolated from E. coli clones and used to transform gonococcal strain F62. Gonococcal clones harbouring chromosomal integrations of the pGCC4-based constructs were selected on erythromycin at 2 µg ml−1.
Catalase assays.
Catalase activity in whole cells of N. gonorrhoeae and N. meningitidis was measured by reduction in H2O2 concentration, using an Apollo 4000 free radical analyser with an ISO-HPO-100 hydrogen peroxide sensor (World Precision Instruments). A standard curve for conversion of pA to nM H2O2 was generated as per the manufacturer’s instructions. All standard curves and enzyme reactions were performed in temperature-controlled reaction vessels (World Precision Instruments) at a constant temperature of 30 °C. Cells were grown aerobically and suspended in GCK broth for 10 min, after which H2O2 was added to a final concentration of 100 µM. The H2O2 concentration was measured continuously, and data were exported into Microsoft Excel to determine reaction rates. The absorbance of the bacterial suspension was determined in a Beckman DU 640 spectrophotometer. Specific activity is expressed as units (nmoles H2O2 decomposed min−1) per OD600 unit.
Results
Catalase provides protection against PN preparations
When exposed to a 2 mM concentration of PN obtained from Calbiochem, N. gonorrhoeae strain F62 exhibited a high level of resistance compared with N. meningitidis strain MC58 (~1.0 vs ~6.5 logs of killing, respectively; Fig. 1). Recent reports have suggested a role for catalase in the catalytic breakdown of PN, and high catalase activity is a hallmark feature of N. gonorrhoeae strains (Bisaillon et al., 1985; Soler-García & Jerse, 2004). Mid-exponential cultures of N. gonorrhoeae were shown to contain 25-fold higher catalase activity than those of N. meningitidis (Fig. 1). We therefore reasoned that the high basal catalase activity in gonococci may be responsible for the difference in PN resistance between these closely related species.
Survival following 2 mM PN exposure and catalase activity of N. gonorrhoeae and N. meningitidis strains. All cultures were grown to exponential phase and diluted to OD600 0.2. One millilitre was exposed to 2 mM PN and grown out for 45 min before plating to determine viability (a), and 2 ml was used to determine catalase activity (b). For strains incubated with IPTG, cultures were exposed to 0.2 mM IPTG for 90 min before determination of survival and catalase activity. Strains: F62 (parental), ΔkatA (ΔkatA), katA-/+ (ΔkatA strain containing a chromosomal insertion of a Plac-inducible katA allele), MC58 (meningococcal strain). Results are representative of at least three independent determinations; error bars, sd.
A gonococcal katA mutant was constructed and tested for PN resistance. The katA mutant was almost completely killed by 2 mM PN (Fig. 1). PN resistance could be restored by IPTG induction of a cloned katA gene under the control of the Plac promoter in a ΔkatA mutant. These data demonstrate that KatA protects N. gonorrhoeae against killing by commercial PN preparations, and that the susceptibility of N. meningitidis to such preparations may be due to its lower basal catalase activity.
Differences in the neisserial katA upstream region are responsible for the difference in basal catalase activity and OxyR induction capacity
Despite the high sequence similarity between the gonococcal and meningococcal OxyR proteins (only two amino acid substitutions), as well as a high level of sequence conservation in the katA upstream region (Fig. 2), previous reports have demonstrated that a large difference in OxyR-mediated regulation of the katA gene in these two species exists (Ieva et al., 2008; Tseng et al., 2003). In N. meningitidis, katA expression has been shown to be repressed by reduced OxyR, and to be activated by H2O2-induced OxyR oxidation, with OxyR remaining DNA-bound regardless of its oxidation state (Ieva et al., 2008). In N. gonorrhoeae, it has been suggested that OxyR acts primarily as a repressor of katA expression, as the catalase activity of a ΔoxyR mutant is greatly increased compared with that of wild-type gonococci (Tseng et al., 2003).
Schematic representation of the gonococcal and meningococcal katA upstream regions. The OxyR binding site, promoter elements and transcription start site are based on those described for meningococcal katA (Ieva et al., 2008).
To determine the basis for the difference in basal catalase activity and katA regulation between these two pathogenic species, translational promoter–lacZ fusions to the gonococcal F62 and meningococcal MC58 katA promoters were constructed and analysed in an F62 background (Table 2). The strain expressing PkatF62 : : lacZ displayed higher β-galactosidase activity than the strain expressing PkatMC58 : : lacZ, consistent with a higher basal catalase activity in gonococci (Fig. 1, Table 2). Though β-galactosidase activity was extremely high in a ΔoxyR strain expressing PkatF62 : : lacZ compared with a ΔoxyR strain expressing PkatMC58 : : lacZ, both promoters appeared to be equally repressed by OxyR (16.1-fold vs 15.4-fold, respectively; Table 2). The largest difference in expression patterns between the two promoters was observed upon treatment with 2 mM H2O2. Whereas the strain expressing PkatMC58 : : lacZ was induced 43.1-fold over the basal promoter activity, the strain expressing PkatF62 : : lacZ was induced only 2.2-fold. These data are consistent with previous reports that demonstrate N. meningitidis to have intermediate catalase activity in a ΔoxyR mutant, and to have high catalase activity in the presence of H2O2 as a result of OxyR-mediated activation (Ieva et al., 2008). These data also demonstrate that in N. gonorrhoeae, the induction capacity of katA in response to H2O2 is much lower, which is likely the reason that OxyR has been reported to act primarily as a repressor (Tseng et al., 2003). The fact that these lacZ fusion strains were all analysed in gonococcal strain F62 suggests that the difference in katA expression and regulation between these two species is due to differences in the promoter sequence rather than differences in the OxyR protein itself.
Data represent the mean±sd in Miller units of 12 determinations.
Though very few sequence differences exist between the gonococcal and meningococcal katA upstream regions, we reasoned that the single base pair difference in the −10 transcription element could be responsible for the observed differences in basal catalase activity (Fig. 2, Table 2). While the N. gonorrhoeae −10 element contains a perfect match to the E. coli consensus, 5′-TATAAT-3′, the N. meningitidis −10 element contains a G at nucleotide −8 (5′-TATAGT-3′; Fig. 2). In order to determine the effect of nucleotide −8 on katA expression, site-directed mutagenesis was used to change this nucleotide from A to G in PkatF62 : : lacZ. Basal expression of PkatF62 : : lacZ A→G was comparable with that of PkatMC58 : : lacZ (Table 2), suggesting that the lower basal catalase activity observed in N. meningitidis is primarily due to this single nucleotide difference. Surprisingly, this nucleotide was also responsible for OxyR-induction capacity, as PkatF62 : : lacZ A→G, like PkatMC58 : : lacZ, was found to be highly induced in the presence of 2 mM H2O2 (36.7-fold; Table 2). Interestingly, these data demonstrate that the large difference in the pattern of katA expression between these two species can primarily be explained by a single nucleotide difference.
Induction of catalase activity confers protection against commercial PN preparations in a dose-dependent manner
At this point, it appeared that katA had been responsible for the detoxification of the Calbiochem PN preparation, and thus the observed resistance to PN-mediated killing. To determine whether this was a novel function of katA or a general role for catalase proteins, gonococcal katA, as well as E. coli catalase katE, and catalase-peroxidase katG, were placed under the control of an IPTG-inducible Plac promoter, and expression was induced in gonococcal strain F62 ΔkatA to test the ability of these catalases to protect against PN-mediated killing (Fig. 3). E. coli KatE and KatG both had catalase activity when expressed in gonococci, though the activity was not as high as the level observed from induction of gonococcal KatA (Fig. 3a). This was likely due to differences in catalase transcription and/or translation, and not to gross differences in specific activity between the enzymes, as the quantity of these catalase proteins at maximal induction levels (1 mM IPTG), as determined by Coomassie staining, appeared to correlate with the observed catalase activity of the culture (data not shown).
Catalase activity and survival of inducible katA, katE and katG strains of F62 ΔkatA. Exponential phase cultures were diluted to OD600 0.2 and exposed to the indicated concentration of IPTG. After 90 min, the cultures were diluted back to OD600 0.2. (a) An aliquot of the culture was used for measuring catalase activity. Data are representative of three separate experiments. (b) The remainder of the culture was exposed to 2 mM PN and grown out for 45 min before determination of viability. Results are plotted as survival versus catalase activity. Data points are individual values determined from three repetitions of the experiment. Symbols used: gonococcal katA (open and closed diamonds), E. coli katE (open and closed circles), E. coli katG (open and closed triangles).
Catalase activity in gonococcal strain F62 ΔkatA harbouring a katA, katE or katG gene under the control of Plac was induced to varying degrees using different concentrations of IPTG, and survival was determined following exposure to 2 mM Calbiochem PN (Fig. 3b). Increasing catalase activity was shown to correlate with an increase in survival against PN challenge in a dose-dependent manner. These data show that catalase activity protects cells from killing by the PN solution, regardless of the particular catalase protein from which the activity is derived.
Calbiochem PN solution is contaminated with H2O2
At this point, these data were consistent with catalase playing a role in protection against PN-mediated killing. Unfortunately, there was an approximately 6 month hiatus in Calbiochem’s production of the PN solutions that we required to complete our studies. However, the method used by Calbiochem to synthesize PN was relatively simple, and could be completed in less than an hour. Briefly, PN synthesis was performed by the addition of isoamyl nitrite and H2O2 in a basic homogeneous solvent system, followed by the removal of unreacted H2O2 by treatment with granular manganese dioxide (MnO2) (Uppu, 2006). PN production was easy to observe, as PN solutions have a characteristic yellow colour, and PN concentration is easy to quantify at its peak absorbance of 302 nm (ϵ = 1670 M−1 cm−1) (Hughes & Nicklin, 1970).
Unlike the Calbiochem PN solution, PN solutions that were prepared in our laboratory failed to cause significant killing of the gonococcal ΔkatA strain at concentrations up to 2 mM (Fig. 4a). However, when the dose–response curve was repeated in the presence of a sublethal concentration of H2O2, killing increased with PN concentration, reaching levels observed with the commercial preparation. When dose–response curves for H2O2 were performed in the presence and absence of PN (Fig. 4b), PN caused a sharp threshold drop in survival. N. meningitidis was also found to survive challenge to PN solutions that we prepared (data not shown). These data suggest that what had previously been reported to be PN-mediated killing (Barth et al., 2009) of the sensitive N. meningitidis may have actually been due to the combinatorial effect of both PN and H2O2, and that these compounds act synergistically to cause cell killing.
Dose–response of individual and combined PN/H2O2. All cultures were grown to exponential phase and diluted to OD600 0.2 in GCK (0.042 % NaHCO3). (a) Cultures were exposed to a range of PN concentrations (0–1.75 mM), with (squares) or without (diamonds) addition of 0.2 mM H2O2. (b) Cultures were exposed to a range of H2O2 concentrations (0–0.3 mM), with (squares) or without (diamonds) immediate previous addition of 2.0 mM PN. Results are representative of one of three independent trials, all of which generated a similarly shaped curve.
Effects of bicarbonate on PN/H2O2-mediated killing
Gonococci require carbon dioxide for growth, which is supplied in broth culture in the form of sodium bicarbonate. However, carbon dioxide is known to react with both PN and H2O2 (Augusto et al., 2002; Richardson et al., 2000). To determine the effect of carbon dioxide on PN/H2O2-mediated cell killing, the gonococcal ΔkatA strain was subjected to PN or H2O2 treatment in the presence (Fig. 5a) or absence of bicarbonate (Fig. 5b). In the absence of bicarbonate, cells displayed increased sensitivity to the individual addition of PN or H2O2 (~1 log and ~2 logs of killing, respectively). There was relatively little increase in sensitivity to H2O2 and PN when these agents were added in combination (~5 logs of killing, compared with ~6.5 with bicarbonate, within experimental error). This suggests that while bicarbonate decreases sensitivity to these reactive species individually, it has no effect on combinatorial PN/H2O2 killing. In turn, this would also suggest that it is unlikely that the bicarbonate radical plays a role in synergistic PN/H2O2 killing.
Survival of F62 ΔkatA after exposure to PN in the presence or absence of sodium bicarbonate. (a) Cultures were grown to exponential phase in GCK (0.042 % NaHCO3), filtered, diluted to OD600 0.2 in fresh medium containing 0.042 % NaHCO3, and treated with H2O2 and/or PN. (b) Cultures were grown to exponential phase in GCK (0.042 % NaHCO3), filtered, diluted to OD600 0.2 in fresh medium lacking NaHCO3, and treated with H2O2 and/or PN. All cultures were allowed to grow out for 45 min before plating to determine viability. Abbreviations: NA (no addition), PN (2 mM PN), H (0.2 mM H2O2), PN+H (simultaneous 0.2 mM H2O2/2 mM PN treatment). Results are representative of three independent determinations; error bars, sd.
Effect of timing on PN/H2O2-mediated killing
We next wished to determine whether there was an effect on PN/H2O2-mediated killing when these two agents were added sequentially rather than simultaneously. When PN was added 5 min before H2O2 addition, we observed approximately 3 logs of cell killing (Fig. 6). When PN was added 5 min after H2O2 addition, we observed only an approximate single log of killing. These data show that the order of addition of these agents can influence the PN/H2O2-mediated killing effect, which suggests that PN can influence the sensitivity of the organism to H2O2.
Survival of F62 Δkat following sequential exposure to PN and H2O2. All cultures were grown to exponential phase and diluted to OD600 0.2 in GCK (0.042 % NaHCO3). Abbreviations: NA (no addition), H→PN (0.2 mM H2O2 addition followed by 2 mM PN addition 5 min later), PN→H (2 mM PN addition followed by 0.2 mM H2O2 addition 5 min later), PN+H (simultaneous 0.2 mM H2O2/2 mM PN addition). All cultures were allowed to grow out for 45 min before plating to determine viability. Results are representative of three independent determinations; error bars, sd.
Discussion
We have previously reported that N. gonorrhoeae is highly resistant to the toxic effects of PN (Barth et al., 2009). This resistance is seen whether the PN is generated by an NO donor plus a superoxide generator or by direct addition of PN. The gonococcus is not sensitive to NO and grows aerobically in its presence using the NO as a terminal electron acceptor. We were unable to identify a basis for the high level of PN resistance, as mutations in the gonococcal genes orthologous to known RNS resistance genes had no effect on the inability of PN to kill N. gonorrhoeae.
The effects of PN have been studied in many different organisms under various conditions. The mechanism of toxicity of PN seems to differ between organisms, showing the importance of studying PN in individual microbial species/strains. In E. coli, PN has been shown to be more toxic than NO. H2O2, either by pre-treatment, or with simultaneous addition, enhances this toxic effect (Brunelli et al., 1995). Previous experiments with E. coli and PN showed that metal ion chelators did not affect toxicity, nor did the addition of hydroxyl radical scavengers. Addition of DMSO did enhance killing by PN, for the hypothesized reason of increased NO2 production (Zhu et al., 1992). A study examining differences in sensitivity to RNS and ROS between E. coli and N. meningitidis showed that in a xanthine/xanthine oxidase paired system, addition of NO increased toxicity to E. coli, while decreasing toxicity in N. meningitidis (Dyet & Moir, 2006). Studies involving mycobacteria have had similar results. A study of various strains of mycobacteria, including Mycobacterium tuberculosis, showed that while all species were susceptible to NO and NO2, virulent strains showed greater resistance to PN than avirulent strains (Yu et al., 1999). These studies emphasize the ways in which ROS and RNS can have diverse effects on different species. One important reaction for mediating the toxicity of PN is its rapid reaction with CO2. The effects of CO2 on PN toxicity in Trypanosoma cruzi demonstrate that the rapid reaction of CO2 with PN decreases its effective toxicity based on the distance that the PN needs to diffuse (Alvarez et al., 2004). The CO2 reduces the ability of PN to reach its target, and therefore to exert a toxic effect. Another study showed the protective effect of CO2 on PN toxicity, by comparing Helicobacter pylori cultures exposed to PN with and without urea (Kuwahara et al., 2000). Urease breaks down urea into CO2 and NH3, and when urea was added to a culture that produced urease, the bacteria no longer showed sensitivity to PN. Urea added to bacteria which did not produce urease did not demonstrate this protective effect.
Earlier reports have provided evidence that in some organisms, catalase contains PN reductase ability (McLean et al., 2010; Wengenack et al., 1999). Previously, we demonstrated resistance to PN in N. gonorrhoeae, while its close relative, N. meningitidis, was extremely sensitive (Barth et al., 2009). As the difference in catalase expression is one of the major distinctions between these two organisms (Seib et al., 2004), we reasoned that catalase could be involved in PN resistance. We were able to show that the gonococcal catalase, KatA, can protect cells against PN, as increasing catalase activity led to an increase in cell survival against PN challenge in a dose-dependent manner.
At first, these data suggested that KatA had a directly protective role against PN toxicity. When there was an interruption of our ability to obtain commercially prepared PN, we began to synthesize it in our laboratory. Through treatment with MnO2, we removed excess H2O2, a process that should have been performed during commercial preparation as well (Uppu & Pryor, 1996). After performing this treatment, we no longer saw substantial amounts of cell killing, though toxicity could be restored by the addition of sublethal levels of H2O2. This suggests that the large degree of killing observed previously was most likely due to H2O2 contamination, and that the protective effect of katA expression observed in this study was due to the reduction of H2O2.
Even small amounts of H2O2 demonstrated synergistic killing of a katA mutant in combination with PN. Killing was observed with H2O2 at 0.2 mM and PN at 2.0 mM, concentrations that did not result in substantial killing individually. Researchers must be careful when using commercial preparations of PN, as these data suggest that even small concentrations of H2O2 may result in drastic changes in cell survival when combined with PN. It is important to note that the product insert included with Calbiochem PN lists contaminating levels of both nitrite and nitrate, yet makes no mention of residual H2O2 contamination.
H2O2 can freely diffuse across membranes, but does not have a directly toxic effect; instead, it participates in a variety of intracellular reactions (Bienert et al., 2006). In wild-type strains of N. gonorrhoeae, the major reaction will be with catalase (2H2O2→2H2O+O2). H2O2 also reacts with thiol groups within proteins, forming disulfide bridges (Zheng et al., 1998). This reaction can be reversed through the activity of disulfide reductases (Prinz et al., 1997). The major source of H2O2 toxicity is related to its ability to react with iron to form hydroxyl radicals (
Earlier studies have shown a relationship between killing by ROS and exposure to RNS. One experiment studied the effect of acidified nitrite ion addition on H2O2 killing in various micro-organisms, including bacteria, fungi and amoebae (Heaselgrave et al., 2010). They found that acidified nitrite substantially increased the killing effect of H2O2 (~4 logs). This was hypothesized to be due to formation of PN ions in the solution, increasing the toxicity of H2O2 through an unknown mechanism. A second study demonstrated increased H2O2 killing of E. coli by the addition of nitrite ion in a lactate buffer (Kono et al., 1994). This was hypothesized to be due to the formation of PN from nitrite in the presence of H2O2, which then breaks down into
We hypothesize that in the presence of CO2, PN readily decomposes into
At this time we are still unable to explain the mechanism of resistance to PN-mediated killing in gonococci when treated solely with this RNS. Although we are unable to rule out the possibility that KatA has PN reductase activity, our data suggest that this protein is not required for PN resistance. However, gonococci may also encode an as-yet-unidentified PN reductase. As gonococci encode a truncated denitrification pathway (Barth et al., 2009), it is possible that basal expression of denitrification genes may aid in removing toxic RNS or perhaps even prevent their formation. It is also possible that gonococci encode novel repair/detoxification pathways that are able to provide protection against RNS and which are not encoded by more sensitive organisms like E. coli (Brunelli et al., 1995). It is also possible that gonococci have repair mechanisms analogous to those of E. coli, but with greater efficiency or higher cellular concentration. Indeed, a high intracellular glutathione concentration is a hallmark feature of this organism (Archibald & Duong, 1986). As an obligate human pathogen that has become extremely well adapted to its host, it may also be the case that the gonococcus simply does not have any target(s) of PN-mediated killing or has modified target(s), so that it is no longer affected by PN.
ACKNOWLEGEMENTS
This study was supported by the National Institute of Allergy and Infectious Disease (Public Health Service grant R21 AI 080912). In addition, we thank Ann Jerse, Uniformed Services University, for her generous gift of a katA mutant in gonococcal strain FA1090.