Helicobacter Catalase Devoid of Catalytic Activity Protects the Bacterium against Oxidative Stress*♦

Catalase, a conserved and abundant enzyme found in all domains of life, dissipates the oxidant hydrogen peroxide (H2O2). The gastric pathogen Helicobacter pylori undergoes host-mediated oxidant stress exposure, and its catalase contains oxidizable methionine (Met) residues. We hypothesized catalase may play a large stress-combating role independent of its classical catalytic one, namely quenching harmful oxidants through its recyclable Met residues, resulting in oxidant protection to the bacterium. Two Helicobacter mutant strains (katAH56A and katAY339A) containing catalase without enzyme activity but that retain all Met residues were created. These strains were much more resistant to oxidants than a catalase-deletion mutant strain. The quenching ability of the altered versions was shown, whereby oxidant-stressed (HOCl-exposed) Helicobacter retained viability even upon extracellular addition of the inactive versions of catalase, in contrast to cells receiving HOCl alone. The importance of the methionine-mediated quenching to the pathogen residing in the oxidant-rich gastric mucus was studied. In contrast to a catalase-null strain, both site-change mutants proficiently colonized the murine gastric mucosa, suggesting that the amino acid composition-dependent oxidant-quenching role of catalase is more important than the well described H2O2-dissipating catalytic role. Over 100 years after the discovery of catalase, these findings reveal a new non-enzymatic protective mechanism of action for the ubiquitous enzyme.

Catalase was described over 100 years ago (1), and the enzyme's role clearly is to detoxify H 2 O 2 by converting it into H 2 O and O 2 . It is one of he most abundant proteins in cells, and it is present in most organisms, including in both plant and animal cells (2). It is oftentimes a highly expressed protein. For example, in the gastric pathogen Helicobacter pylori, catalase (KatA) levels are estimated to be 4 -5% of the total protein content (3), and the katA gene is one of the most highly expressed genes in H. pylori cells recovered from the human stomach (4). HOCl-mediated oxidation of six identified Met residues in H. pylori KatA leads to methionine sulfoxide formation (Met-O), 2 protein oligomerization, and loss of catalase activity (3). Most organisms, including Helicobacter, possess a peptide repair enzyme, methionine sulfoxide reductase (Msr), that reduces Met-O back to Met in certain oxidation-susceptible Msr-targeted proteins (5,6). This Msr-mediated repair (along with added GroEL/ES) returns most of the catalase activity (3). Although the identified protein targets of repair are few among the total of all organisms, some of these repair targets are themselves stress-combating enzymes.
Purified KatA and Msr enzymes were shown to physically interact (6), and this interaction resulted in Msr-mediated repair of five out of the six oxidized KatA Met residues (3). In addition, H. pylori KatA is ubiquitous, present in both the cytoplasm and in the periplasm and on the cell surface as well as being readily detected (like Msr) extracellularly (7)(8)(9). Taken together, this information (reactivity toward HOCl; presence of multiple Msr-repairable Met residues, enzyme abundance) caused us to investigate the possibility that the catalase primary sequence or composition represents a Met-recycling sink providing oxidant protection per se to this bacterium. This Metmediated quenching role would be separate from the enzyme's catalytic role, and no such whole cell protective mechanism has yet been assigned to any protein. However, the possibility that a Met-rich protein could exist to serve a protective oxidantquenching role (via cyclic turnover of Met residues) was raised 20 years ago (10,11), but results to support this have been unavailable. While our study identifies a new oxidant protective mechanism for catalase, the model organism we use, H. pylori, is an important pathogen. Although this bacterium is known to be the agent of human gastritis, which can develop into peptic ulcer disease (12), factors that allow it to persist in the host need to be identified, as such persistence is responsible for the most severe outcomes of the infection, namely gastric and duodenal cancers (13). In response to H. pylori infection, the host produces a battery of harmful reactive oxygen species (ROS), such as superoxide anion (O 2 . ), hydrogen peroxide (H 2 O 2 ), hydroxyl radical ( ⅐ HO), and hypochlorous acid (HOCl). Indeed, exposure of gastric cells (14) or phagocytes (15) to H. pylori increases host cell ROS production, and H. pylori-infected patients show elevated levels of ROS (16). HOCl is produced in large amounts by neutrophil myeloperoxidases; its concentration can reach up to 5 mM at inflammatory sites (17), and it is 100 times more toxic than H 2 O 2 (18). HOCl primarily targets sulfur-containing amino acids, cysteine (Cys) and methionine (Met), and indeed these residues are the preferred amino acids for oxidation under physiological conditions (19). Met oxidation by HOCl and similar small molecule oxidants (superoxide and hydroxyl radicals) leads to formation of Met-O in many proteins, and further oxidation results in methionine sulfone formation within the peptides, and either oxidized form at Met residues can lead to protein dysfunction (20). Oxidation of Met residues within a protein can "protect" other susceptible residues in that same protein from oxidation (5), but documentation of Met-containing residues within a protein to act as antioxidants in whole cell physiology is unavailable. Here, we address the roles of a catalase protein devoid of catalytic activity to provide oxidative stress protection.

Results and Discussion
To investigate whether catalase can quench oxidants independently of its H 2 O 2 removal activity, we engineered two H. pylori mutants to synthesize only apo-catalase, e.g. devoid of catalase activity. Based on the published crystal structure of H. pylori catalase, two residues are essential for catalysis as follows: the proximal Tyr-339, coordinated to the heme iron, and the distal His-56, essential for the formation of the main reaction intermediate, compound I (21,22). Therefore, two markerless chromosomal katA mutant versions, katA H56A and katA Y339A , were constructed in two H. pylori wild-type (WT) strains, strain 43504 and the mouse colonizing strain X47 (23). Complete deletion of ⌬katA mutants were also constructed in both parental strains. When grown on plates, colonies of ⌬katA mutants, as well as katA H56A and katA Y339A site-directed mutants, displayed strikingly different phenotypes compared with their individual parental strain (43504 or X47). While wild-type colonies were dark brown, the mutant strains were light brown (katA H56A ) or yellow (katA Y339A or ⌬katA) in color, suggesting that the heme b cofactor binding might be either disturbed (katA H56A ) or even lacking (katA Y339A or ⌬katA). Catalase assays revealed both 43504 ⌬katA and X47 ⌬katA mutants, as well as isogenic katA H56A and katA Y339A site-directed mutants, had no detectable catalase activity (Table 1). However, catalase was still synthesized in both katA H56A and katA Y339A strains, albeit to lower levels than WT (Fig. 1, A and  B). No catalase protein was detected in ⌬katA gene deletion strains (Fig. 1, A and B).
To investigate the respective involvement of each residue in the heme-binding ability of the Helicobacter H 2 O 2 -dissipating enzyme (24), we likewise studied specific mutant versions. KatA H56A and KatA Y339A as well as the native catalase KatA WT (control) were expressed as recombinant proteins in Escherichia coli and purified to near homogeneity. Purified proteins were phenotypically different. KatA WT was light brown; KatA H56A was dark brown, and KatA Y339A was colorless ( Fig.  2A). When protein samples were loaded on a gel in a nonreducing buffer without prior heating, both KatA WT and KatA H56A ran as apparent tetramers (apparent molecular mass above 200 kDa) (Fig. 2B), in agreement with previous studies (3,24). By contrast, KatA Y339A was unable to tetramerize. When subjected to reducing conditions and boiling, KatA WT , KatA H56A , and KatA Y339A proteins migrated as monomers with an apparent molecular mass below 60 kDa, in agreement with their calculated mass (58.6 kDa). Catalase activity of the purified KatA H56A was less than 1% of the activity of purified KatA WT , although KatA Y339A had no detectable activity ( Table  1). The three proteins were analyzed by UV-visible scan spectrophotometry (Fig. 2C). A heme b-specific absorption Soret peak centered at ϭ 410 nm was observed for KatA WT ; this peak was shifted for KatA H56A ( ϭ 402 nm) and was essentially absent for KatA Y339A , indicating that the heme b environment is modified in KatA H56A , and it is abolished in KatA Y339A protein. These results were expected based on heme ligands predicted from the crystal structure (21,22).
Previous work demonstrated that HOCl reacts with H. pylori catalase, leading to catalase oligomerization, methionine oxidation, and enzyme inactivation (3). To determine whether purified wild-type or variant catalase could quench HOCl, subsequently protecting H. pylori cells from oxidant-mediated death, non-growing H. pylori 43504 WT cells were incubated either  Results shown are average Ϯ S.D. from two independent batches of purified proteins, and assays were done in triplicate (n ϭ 6 total). b ND means not detectable. c Catalase activity was measured as mol of H 2 O 2 /min/10 9 cells. Results shown are average Ϯ S.D. from three biological replicates, and assays were done in triplicate (n ϭ 9 total). with PBS buffer, or with HOCl, or with HOCl that had been pre-incubated for 15 min with either purified KatA WT , purified KatA H56A , or purified KatA Y339A . In addition, two unrelated protein controls, HypC (high Met content, 6%) and UreE (low Met content, 1%), were included in the study. Exposure to HOCl only (no protein added) killed all the cells, e.g. we could not recover any CFU following this treatment (n ϭ 3) (Fig. 3). In contrast, when HOCl was incubated with purified native or either variant catalase, the final CFU count was in the same range as the PBS only control, indicating that each purified catalase protein, whether active (KatA WT ) or inactive (KatA H56A and KatA Y339A ), can protect H. pylori in vitro against the deadly effect of the oxidant (Fig. 3). When the same experiment was repeated with decreasing concentrations of catalase, cell survival rates decreased accordingly; however, there was no significant difference in protection between the native (WT) and mutant versions of KatA (data not shown). Furthermore, addition of HypC to the HOCl mixture conferred levels of protection similar to catalase, whereas addition of UreE was not protective, suggesting a correlation between Met content and HOCl quenching. This protective effect is apparently independent of Msr and Met recycling, as the Msr repair mixture was not added for this experiment. Nevertheless, it demonstrates the oxidant quenching capacity of catalase that is inherent in its sequence and independent of its catalytic activity. A previous biochemical study almost 30 years ago suggested that pure bovine catalase, at high concentration, and either active or azide-inactivated, could protect bovine ␣1-antiproteinase enzyme against inactivation by HOCl (25); but it was con-cluded the heme ring of catalase likely reacted with HOCl to somehow dissipate the oxidant.
To determine whether holo-or apo-catalase can protect H. pylori cells against HOCl in vivo, disk inhibition assays with HOCl were conducted using wild-type strain 43504 as well as ⌬katA, katA H56A , and katA Y339A mutant strains (Fig. 4). In  Results (CFU/ml recovered after 1 h) represent the mean and standard deviation from three independent challenge experiments, with each serial dilution plated in triplicate. ND, no CFU could be detected (detection limit, 10 2 CFU/ml). addition, because Msr was shown to interact with and repair KatA in H. pylori (3,6), ⌬msr single mutants, as well as ⌬msr ⌬katA, ⌬msr katA H56A , and ⌬msr katA Y339A double mutants were included in this study. The sensitivity (diameter of inhibition) of the ⌬katA mutant was significantly greater than the parental strain (43504) or the mutant strains katA H56A and katA Y339A (p Ͻ 0.01, ANOVA). There was no significant difference between WT, katA H56A , and katA Y339A strains (Fig. 4). These results suggest that catalase, even when inactive and lacking the heme moiety, enables H. pylori cells to combat HOCl. In addition, ⌬msr single mutants were significantly (p Ͻ 0.01) more sensitive to HOCl than the WT strain, in agreement with previously published results (26). Double mutants ⌬katA ⌬msr, katA H56A ⌬msr, and katA Y339A ⌬msr were also more sensitive than WT (p Ͻ 0.01). Mean diameters of inhibition for katA H56A ⌬msr and katA Y339A ⌬msr were significantly greater than those of their respective parental backgrounds, e.g. katA H56A and katA Y339A (p Ͻ 0.01), confirming the importance of Msr in repairing HOCl-oxidized Met-O residues. The ⌬katA ⌬msr double mutant strain was as sensitive to HOCl as the ⌬katA mutant strain, suggesting that absence of catalase protein is the single most important factor contributing to HOCl sensitivity.
The ability of katA H56A and katA Y339A mutant strains to colonize the gastric mucosa of mice was investigated. Each mutant were orally given to mice, and their colonization levels after 3 weeks were compared with those obtained with the wild-type strain (X47) and the ⌬katA catalase negative mutant (Fig. 5). The WT strain was able to colonize 11 mice (n ϭ 14 total), with an average of 1.34 ϫ 10 6 CFU recovered per g of colonized stomach. In contrast, the ⌬katA deletion mutant was detected in only one mouse (n ϭ 10 total), and the colonization load was low (5.5 ϫ 10 3 CFU per g). Interestingly, the site-directed katA H56A catalase mutant was able to colonize 10 mice (n ϭ 12 total), with an average CFU number of 5.7 ϫ 10 5 CFU per g of colonized stomach, although the site-directed katA Y339A catalase (heme-deficient) mutant successfully colonized 5 out of 10 mice, with an average of 1.1 ϫ 10 6 CFU per g of colonized stomach. Colonization counts for both katA H56A and katA Y339A mutants were not significantly different from those obtained for the WT, while ⌬katA deletion mutants had significantly lower counts compared with the other three strains. To rule out an effect of the presence of the KSF cassette on the colonization-deficient phenotype of ⌬katA::KSF mutants, we constructed a markerless X47 ⌬katA deletion mutant strain (see under "Experimental Procedures"). As expected, those mutants had no measurable catalase activity, and catalase could not be detected by immunoblotting (data not shown). In addition, mouse colonization assays revealed those markerless ⌬katA mutants were unable to colonize mice (no CFU detected, n ϭ 4), although the X47 parental strain (WT control) colonized all inoculated mice (n ϭ 4, average colonization of 3.9 ϫ 10 5 CFU per g of stomach), therefore confirming that the lack of  Mouse stomachs were homogenized, and serial dilutions were plated. Data are presented as a scatter plot of numbers of CFU/g of stomach (log 10 scale) as determined by plate counts. Each symbol represents the mean CFU count for one stomach (n ϭ 14 for WT, n ϭ 12 for katA H56A , n ϭ 10 for katA Y339A , and n ϭ 10 for ⌬katA, respectively). Each horizontal bar represents the geometric mean of the colonization load for each group. The ⌬katA mutant (catalase null strain) geometric mean is significantly lower than the wild-type strain (p Ͻ 0.01, Student's t test), or the katA H56A mutant (p Ͻ 0.01), or the katA Y339A mutant (p Ͻ 0.08), while there is no significant difference between average colonization loads (geometric mean) for WT, katA H56A , and katA Y339A strains (p Ͻ 0.01). A dashed horizontal line shows the detection limit, which represents a count below 1.5 ϫ 10 2 CFU/g of stomach.
Novel Role for Helicobacter Catalase NOVEMBER 4, 2016 • VOLUME 291 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 23369 colonization by ⌬katA::KSF was likely due to the absence of the katA gene, rather than the presence of the KSF cassette. Taken together, these results suggest that catalase is critical for combating oxidants in vivo, and this does not depend on H 2 O 2 dismutation catalytic activity.
Results from a previous study suggested that katA was only needed for long-term colonization (24 weeks) in mice, as there was no significant difference (between the wild-type strain and a ⌬katA null mutant) in colonization after 8 days or 12 weeks (27). However, a major difference between that study and ours is the H. pylori parent strain (mouse colonizer) used. SS1 was used as parental strain in the previous study, and this strain is known to preferentially colonize the mouse stomach antrum (27). By contrast, in this study, we used X47, a strain that is unrelated to SS1, and whose primary tropism is the mouse stomach corpus (28). Therefore, differences in oxidative stress levels between the corpus and the antrum, combined with genetic differences between SS1 and X47, are likely to account for the different outcomes in both colonization assays.
At least two independent studies have suggested that catalase is a good vaccine candidate for prevention of gastric ulcers and cancers (29,30). In one such study, 90% of the mice were protected from H. pylori (SS1 strain) colonization by raising anti-KatA antisera (30). This strong protective immunization effect is another indication of the importance of this Met-containing enzyme. It would be of interest to assess whether the Helicobacter catalase could be detected in the mucus (i.e. extracellular to the bacterium) of infected animals, and if so, what downstream affects this catalase may have on the microflora.
More than 100 years after the initial report describing catalase activity (1), we now know its physiological importance in at least one organism is due to an additional mechanism. The finding that inactive catalase was protective to the bacterium was not expected, but it was nevertheless hypothesized based on the known susceptibility of its Met residues to oxidation and to (Msr-mediated) turnover. It is possible this new role can only be documented in Helicobacter, as the bacterium makes a large amount of catalase, and the mouse model is a facile way to assess viability in a natural stress environment. We do not know whether the cyclic turnover (oxidation and repair) of catalase occurs extracellularly or in the host mucous lining, but the oxidant quenching must occur to some extent, and both enzymes (Msr and catalase) are readily detected outside the (Helicobacter) cell (at least in vitro). Assessing the extent of extracellular repair (i.e. Met-O-Met turnover) must await repair assays and mass spectral studies to assess the Met(-O) state of the target using extracellular fractions.
Catalase in many organisms is up-expressed upon oxidative stress exposure, so it is reasonable that the protein would serve a secondary protective role. This secondary role, independent of activity, would be expected to operate even in cells that are iron-or heme-starved, conditions known to occur in host-infected tissue for many pathogens. Additionally, oxidative stress and protein oxidations are correlated with a number of human neurodegenerative diseases, including Parkinson's and Alzheimer's disease (31). Interestingly, the neurological tissues of concern here contain both catalase and methionine sulfoxide reductase.
In addition to catalase, other abundant and/or secreted proteins, especially ones that are repaired by Msr (such as GroEL and AhpC (26)) could also play an unexpected stress-combating role, e.g. quenching of oxidants to confer stress protection. Indeed, any Met-rich protein, and particularly if localized outside the cytoplasm, should be studied via mutagenesis or other approaches for roles in combating or quenching oxidants. Even if these proteins have an already known function, a secondary role as a protector from oxidants is a possibility. Of course, a Met-rich protein without an enzyme function could serve this role; in theory, only stability and MetO-forming ability would be required to be an antioxidant; Met recycling via recognition by Msr would be expected to aid this role. It seems possible this methionine-mediated antioxidant role within proteins may impact the physiology and survival of multiple organisms.

Experimental Procedures
Bacterial Strains and Plasmids-Strains and plasmids are described in supplemental Table S1. E. coli TOP10 (Invitrogen) was used for all cloning experiments, and E. coli BL21(DE3)RIL (Novagen) was used to express recombinant H. pylori KatA proteins. H. pylori wild-type strain 26695 was used as a source of DNA for PCR. All plasmids and polymerase chain reaction (PCR) products were sequenced at the Georgia Genomics Facility, University of Georgia, Athens, GA, and compared with DNA sequences from strain 26695 (32) to ensure that no error had been introduced following PCR amplification, as well as to verify the presence of engineered site-directed mutations within katA.
Construction of H. pylori katA Deletion and Site-directed Mutants-Markerless chromosomal katA H56A and katA Y339A mutants were constructed following a two-step kanamycin-sucrose (selection-counter selection) method, as described previously (33). Briefly, a 1720-bp DNA sequence containing the katA gene (hp0875 in strain 26695) and flanking sequences was amplified by PCR using genomic DNA from H. pylori strain 26695 as template and primers KatA-KpnI and KatA-BamHI (supplemental Table S2). The PCR product was digested with KpnI and BamHI and ligated into similarly digested pBS-KS plasmid to generate pKS-katA. Next, a unique HindIII site located within katA was used to insert a 3-kb Kan R -sacB-P flaA (KSF) cassette previously excised from pKSF-II plasmid (33). Insertion of the KSF cassette is not expected to have any polar effect, because the gene downstream of katA (kapA, hp0874 in strain 26695) has its own promoter. The resulting plasmid, pKS-katA::KSF, was introduced into H. pylori strain 43504, and mutants were selected on BA plates supplemented with kanamycin. Colonies appeared after 3-5 days and were shown to be sucrose-sensitive as well as catalase-negative. Those mutants, ⌬katA::KSF, were used as negative controls (absence of katA) for all experiments. In addition, they were used as parental strains to generate site-directed point mutants following transformation with plasmid pKS-katA H56A or pKS-katA Y339A . Each point mutation was generated using a Splicing by Overlap Extension-PCR method. Briefly, external primers KatA-KpnI and KatA-BamHI were used concomitantly with internal primers designed to introduce a mutation in the open reading frame of katA, resulting in either a H56A or a Y339A substitution (supplemental Table S2). After a final round of amplification with purified PCR products and external primers, each 1720bp-long PCR product (containing the desired mutation) was digested with KpnI and BamHI and ligated into pKS to yield plasmid pKS-katA H56A or pKS-katA Y339A . Each plasmid was introduced separately into the ⌬katA::KSF mutant, and transformants (site-directed mutants) were isolated after 3-5 days on BA supplemented with 5% sucrose. Those sucrose-resistant katA H56A and katA Y339A chromosomal mutants were shown to be kanamycin-sensitive as well as catalase-negative. The same procedure was used to generate ⌬katA::KSF, katA H56A , and katA Y339A mutations in the mouse colonizing strain X47 (33). Finally, to rule out an effect of the presence of the KSF cassette on the colonization-deficient phenotype of ⌬katA::KSF mutants, we constructed a markerless ⌬katA deletion mutant. Briefly, plasmid pKS-katA was digested with HindIII, treated with T4 polymerase, and ligated onto itself, to generate a new plasmid, pKS-katA*. This treatment introduces a frameshift in the katA gene after the Val-55 codon, leading to early translation termination. Transformation of ⌬katA::KSF mutants with plasmid pKS-katA* generated ⌬katA markerless mutants that were sucrose-resistant, kanamycin-sensitive, and catalase-negative. The chromosomal disruption of katA by the KSF cassette, the frameshift of katA in ⌬katA mutants, and the presence of chromosomal mutation katA H56A or katA Y339A in each H. pylori strain were confirmed by DNA sequencing.
Construction of H. pylori ⌬msr Mutants-⌬msr deletion mutants were generated using Splicing by Overlap Extension-PCR. Briefly, primers ⌬msr-cat1 and ⌬msr-cat2 were used to amplify a 412-bp-long DNA sequence located upstream of msr (hp0224 in strain 26695). Primers ⌬msr-cat3 and ⌬msr-cat4 were used to amplify a 411-bp-long sequence located downstream of hp0224. The final amplification step included each purified PCR product, a cat (Cm R ) cassette (0.8 kb) and primers ⌬msr-cat1 and ⌬msr-cat4. The resulting 1620-bp-long PCR product was introduced into either H. pylori wild-type strain 43504 or ⌬katA, katA H56A , or katA Y339A mutant strain, to generate ⌬msr single mutant, or ⌬katA ⌬msr, katA H56A ⌬msr, or katA Y339A ⌬msr double mutants, respectively (supplemental Table S1). H. pylori cells were transferred after 16 h onto BA plates supplemented with 25 g/ml chloramphenicol. Colonies appeared after 3-5 days of incubation. The concomitant chromosomal deletion of msr and the insertion of cat were confirmed by PCR using genomic DNA from each strain and primers ⌬msr-cat1 and ⌬msr-cat4 and by DNA sequencing.

Cloning, Expression, and Purification of Recombinant Native and Variant KatA Proteins-Recombinant native and variant
KatA proteins were expressed in E. coli BL21 (DE3)-RIL (Novagen). Briefly, in three separate PCRs, primers KatA-NdeI and Kat-BamHI were used with three different plasmid templates, e.g. pKS-katA, pKS-katA H56A , or pKS-katA Y339A to amplify a 1540-bp-long DNA sequence containing either the native katA gene, the katA H56A mutated gene, or the katA Y339A mutated gene, respectively. Each PCR product was digested with NdeI and XhoI, gel-purified, and cloned into similarly digested pET21b plasmid, generating plasmids pET-KatA WT , pET-KatA H56A , and pET-KatA Y339A . E. coli BL21-RIL cells transformed with the appropriate pET-KatA plasmid were grown at 37°C in 500 ml of LB supplemented with ampicillin and chloramphenicol until an absorbance of 0.5 at 600 nm was reached. Gene expression was induced by the addition of 0.5 mM isopropyl ␤-D-thiogalactopyranoside followed by incubation (at 37°C) for 3-5 h. Recombinant native and variant catalase proteins were purified as reported previously for the H. pylori native catalase, by using a stepwise dual approach combining anion-exchange and size exclusion chromatography (34). Briefly, cells were centrifuged and washed with 200 ml of 25 mM Na 2 HPO 4 (pH 7.5), 50 mM NaCl buffer (buffer A), and pellets were suspended in 5 ml of buffer A. Protease inhibitor (cOmplete Mini, Roche Applied Science) was added, and cells were disrupted by passage through a French pressure cell at 18,000 pounds/inch 2 , and the lysate was centrifuged at 17,000 ϫ g for 15 min to remove cell debris. The supernatant was then collected and subjected to ultracentrifugation at 100,000 ϫ g for 1 h. The membrane-free supernatant was applied to an SP-Sepharose cation-exchange column (GE Healthcare) that had been equilibrated with buffer A, and the protein was eluted with a linear gradient of 0.05-1 M NaCl in buffer A. Catalase-containing fractions were selected, pooled, concentrated using Amicon Ultra-4 devices with a 10-kDa molecular mass cutoff (Merck Millipore, Billerica, MA) and further purified by size exclusion chromatography using a HiLoad 16/60 Superdex 75 column (GE Healthcare) in buffer A with 0.3 M NaCl. Catalasecontaining fractions were selected, pooled, concentrated, and stored at 4°C protected from light. Upon purification, recombinant native catalase (KatA WT ) was light brown in solution, purified KatA H56A variant was dark brown, and purified KatA Y339A variant was colorless. The protein purity was assessed by SDS-PAGE, and the final protein concentration was determined using the BCA kit (ThermoFisher Pierce).
Catalase Assays-Catalase assays were done spectrophotometrically using whole cells or purified catalase proteins, in phosphate-buffered saline (PBS) containing 15 mM H 2 O 2 , as described previously (7).Briefly, for whole cells assays, cells were washed and resuspended in PBS to a final A 600 of 1.0. Five l of whole cells were mixed with 495 l of PBS containing 15 mM H 2 O 2 , and the initial H 2 O 2 disappearance (decrease in A 240 ) was followed for up to 1 min. Catalase assays with purified proteins were carried out using 0.05 to 0.1 g of purified KatA WT and 1 to 10 g of purified KatA H56A or KatA Y339A .
UV-visible Spectrometry-Measurements were recorded at 22°C on a SpectraMax plus spectrophotometer (Molecular Devices, Sunnyvale, CA) using a final volume of 500 l in a quartz cuvette with a 1-cm path length. Each purified catalase protein (native KatA WT , KatA H56A , or KatA Y339A variant) was diluted to a final concentration of 10 M in size exclusion chromatography buffer (see above). UV-visible spectroscopy scan was run from 250 to 600 nm. Scans were run three times, with three independent batches of purified proteins.
Immunoblotting Experiments-For detection of KatA in crude extracts, cells were grown on BA, harvested, and resuspended in loading buffer. All samples were subjected to SDS-12.5% PAGE using a Mini-Protean II apparatus (Bio-Rad), according to the method of Laemmli (35) and transferred to a nitrocellulose membrane (0.2-m pore size; Bio-Rad). The membrane was blocked by incubation in 20 mM Tris-HCl (pH 7.6), 137 mM NaCl buffer (Tris-buffered saline, TBS) supplemented with 3% gelatin (Mallinckrodt Baker). This was followed by a 1-h incubation along with a 1:1,000 dilution of anti-KatA (rabbit polyclonal) antiserum, in TBS buffer with 0.1% Tween 20 (TTBS), 1% gelatin. When needed as control, anti-UreA (rabbit polyclonal, Santa Cruz Biotechnology, Dallas, TX) antiserum was used with a 1:2,000 dilution. The membrane was washed with TTBS and then incubated for 1 h with the secondary antibody (goat anti-rabbit immunoglobulin G coupled with alkaline phosphatase, Bio-Rad) diluted 1:2,000 in TTBS, 1% gelatin. The membrane was again washed with TTBS buffer. Bound antibodies were detected following addition of the chromogenic reagents nitro blue tetrazolium (0.25 mg/ml) and 5-bromo-4-chloro-3-indolyl phosphate (0.125 mg/ml) (Sigma) in 10 mM Tris-HCl (pH 9.5), 150 mM NaCl.
HOCl Challenge of H. pylori in Presence of Purified Proteins-H. pylori wild-type strain 43504 cells were grown for 24 -36 h on BA plates, harvested, spun down, and suspended in sterile PBS to a final A 600 of 1.1 in PBS buffer. Then 0.9-ml aliquots of this suspension were incubated for 60 min at 37°C (no shaking) with either 0.1 ml of PBS, 0.1 ml of PBS with NaOCl (10 -15% available chlorine; Sigma), or 0.1 ml of PBS/NaOCl that had been previously incubated for 15 min at 37°C in the presence of purified protein Kat WT , Kat H56A , Kat Y339A , HypC, or UreE. Both HypC and UreE were expressed and purified as described previously (26,36). Final protein and NaOCl concentrations were 0.25 and 200 M, respectively, and final A 600 was equal to 1 in each tube (ϳ5 ϫ 10 8 cells per ml). After 60 min at 37°C (no shaking), cells were serially diluted in sterile PBS, and 10 l of each dilution was spotted in triplicate on BA plates. CFU were counted after 3 days of incubation at 37°C under microaerophilic conditions (4% O 2 ). Results (remaining CFU after 1 h) represent means and standard deviations from three independent challenge experiments and serial dilutions spotted in triplicate.
Disk Inhibition Assays-H. pylori wild-type, ⌬katA or ⌬msr single mutant, or ⌬katA ⌬msr double mutant cells were grown on BA for 24 h before being resuspended to a final A 600 of 1 in sterile PBS buffer. Then 0.1 ml of cells were homogenously spread on top of 25-ml standardized BA plates (3-5 replicates per strain). A sterile paper disk (7.5 mm diameter) was placed in the center of each plate, and 10 l of undiluted NaOCl (Sigma, 1.4 M, with NaClO ⑀ 292 ϭ 350 M Ϫ1 cm Ϫ1 ) was added onto the disk. Cells were allowed to grow for 48 h, and the diameter of the inhibition zone was measured. Results shown are means and standard deviations from three to seven independent growth experiments, each with three to five replicates (n ϭ total number of measurements). ANOVA was used to compare diameter means between strains.
Mouse Colonization Experiments-All procedures were approved by the Institutional Animal Care and Use Committee of the University of Georgia. H. pylori X47 (mouse-adapted, parental strain) and X47 ⌬katA, X47 katA H56A , or X47 katA Y339A mutant strains were grown for 24 h or less on BA plates, harvested, and resuspended in sterile PBS buffer (pH 7.4) to a final A 600 of 2. Five-to 6-week-old female C57BL/6NCr mice (NCI, Frederick, MD) were infected via oral gavage with 0.15 ml of bacterial suspension (ϳ1.5 ϫ 10 8 H. pylori cells per mouse). Mice were sacrificed by CO 2 asphyxiation and cervical dislocation 3 weeks post-inoculation. Stomachs were quickly removed, weighed, and gently homogenized in 5 ml of sterile PBS using a Dounce hand homogenizer. Dilutions were made in sterile PBS and plated (0.1 ml) in duplicate on plates supplemented with amphotericin B, bacitracin, nalidixic acid, and vancomycin. Plates were incubated for 5-7 days at 37°C in a 4% O 2 partial pressure atmosphere for colony counting. Data are expressed as CFU recovered per g of mouse stomach. The detection limit of the assay is 150 CFU per g of stomach. Student's t test was used to compare geometrical means of colonization between strains.