Role of a bacterial organic hydroperoxide detoxification system in preventing catalase inactivation

H. pylori cells (WT or ahpC mutant) were grown to late log phase and harvested, from which the cell extract or membrane fraction were prepared. The total amount of lipid hydroperoxides were determined and calculated as the amount (nmol) present in the samples from 10 10 cells. The data are averaged from 3 independent experiments with standard deviation.


Summary
In the gastric pathogen Helicobacter pylori, catalase (KatA) and alkyl hydroperoxide reductase (AhpC) are two highly abundant enzymes that are crucial for oxidative stress resistance and survival of the bacterium in the host. Here we report a previously unidentified connection between the two stress-resistance enzymes. We observed that the catalase in ahpC mutant cells, in comparison to the parent strain, is partially (about 50%) inactivated. The decrease of catalase activity is well correlated with the perturbation of heme environment in catalase as detected by electron paramagnetic resonance spectroscopy (EPR). To understand the reason for this catalase inactivation, we examined the inhibitory effects of hydroperoxides on H.
pylori catalase (either present in cell extracts or added to the purified enzyme) by monitoring the enzyme activity and the EPR signal of catalase. H. pylori catalase is highly resistant to its own substrate, without loss of enzyme activity by treatment with a molar ratio of 1:3000 H 2 O 2 .
However, it is inactivated by lower concentrations of organic hydroperoxides (the substrate of AhpC). Treatment with a molar ratio of 1:400 t-butyl hydroperoxide resulted in about 50% inactivation of catalase. UV-visible absorption spectra indicated that the catalase inactivation by organic hydroperoxides is due to the formation of a catalytically incompetent compound II species. To further support the idea that organic hydroperoxides which accumulate in the ahpC mutant cells are responsible for the inactivation of catalase, we measured the level of lipid peroxidation in ahpC mutant cells compared to wild type cells. The results showed that the total amount of extractable lipid hydroperoxides in the ahpC mutant cells is about 3 times that in the wild type cells. Our findings reveal a novel role of the organic hydroperoxide detoxification system in preventing catalase inactivation. 4 16, 17, 18), demonstrating the importance of these enzymes in oxidative stress resistance and host colonization.
H. pylori express abundant levels of catalase and AhpC proteins (19). The genetic and biochemical characterization of H. pylori catalase and AhpC have been performed in different laboratories (13,20,21,22,23,24). H. pylori catalase is a homotetrameric protein, with each subunit having a molecular mass of 59 kDa. It is a monofunctional catalase without peroxidase activity. A unique property of H. pylori catalase is that it has an isoelectric point (pI) of >9.
Another property of H. pylori catalase distinct from other typical catalases is its stability at very high concentrations of H 2 O 2 (20). H. pylori cells were shown to be resistant to high concentration (~100mM) of H 2 O 2 and this resistance was abolished in katAmutants (24). H. pylori AhpC is a major component of the thioredoxin-dependent peroxiredoxin system (AhpC-Trx-TrxR) that catalyzes reduction of hydroperoxides including H 2 O 2 and organic hydroperoxides (22,25), as well as reduction of peroxynitrite (26). It was extremely difficult to obtain an ahpC knock-out mutant (21,22); eventually the mutant was obtained by screening transformants at very low O 2 (1% partial pressure) condition (13). AhpC mutant cells exhibited severe growth sensitivity to hydroperoxides and to the superoxide-generating agent paraquat (13).
We studied the relationship between these two important antioxidant proteins and discovered that the organic hydroperoxide reductase (AhpC) plays a previously unidentified role in protecting catalase from inactivation by organic hydroperoxides. For the first time, it is shown that loss of AhpC function leads to a significant increase of organic hydroperoxides within the cells, and that these hydroperoxides are potent inhibitors of catalase. With a similar procedure, the ferritin mutant strain (43504 pfr:Kan) was constructed as follows. A 1213-bp fragment containing H. pylori pfr gene was PCR amplified using primers pfrF (5'-TGGCTAGTTTTAAGGGCATG-3') and pfrR (5'-AAGCGCAAAATTTGCAAGCG-3'), and cloned into pGEM-T vector. Subsequently, the 301-bp HindIII fragment within pfr gene was replaced by a kanamycin resistance cassette (Kan).
Cell-free extract and membrane fraction. Plate-grown H. pylori cells were harvested and suspended in phosphate-buffered saline (PBS). The cells were collected by centrifugation (10,000 xg for 10 min), resuspended in PBS, and broken by two passages through a French pressure cell at 18,000 lb/in 2 . Crude extracts were then cleared of unbroken cells by centrifugation at 10,000 xg for 10 min. The supernatant (cell-free extract) was then subject to ultracentrifugation (45,000 xg for 60 min) to obtain the membrane fraction (the pellet).

Protein concentration determination and gel electrophoresis. Protein concentrations
were determined with a bicinchoninic acid protein assay kit (Pierce, Rockford, Ill.). For SDS-PAGE, 5 µg of cell extract was placed into SDS buffer, boiled for 5 min, and applied to a denaturing 12.5% acrylamide gel. With densitomeric measurement of all the protein bands on SDS gel, the portion (%) attributed to a specific protein (KatA or AhpC) was calculated. Based 7 on this percentage, the molecular weight of the protein, and the total protein concentration, we estimated the molar concentration of KatA or AhpC in the cell extract.
Purification of catalase. Native H. pylori catalase was purified following the method similar to that described by Radcliff et al. (28). Briefly, H. pylori cells grown on plates were harvested by suspension in 0.1 M sodium phosphate buffer (pH7.5). After centrifugation, the cell pellet was resuspended in the buffer. Cells were disrupted by 3 cycles through a French pressure cell at 18,000 lb/in 2 , and the lysate was centrifuged at 28,000xg for 10 min to remove the cell debris. The supernatant was collected and subjected to ultracentrifugation at 100,000xg for 45min. The supernatant was applied to a SP Sepharose cation-exchange column (Amersham) that had been equilibrated with 25mM sodium phosphate buffer (pH7.5), and proteins were eluted by the creation of a gradient with 1 M NaCl in 25mM sodium phosphate buffer (pH7.5).
Catalase positive fractions were selected by checking for oxygen-reducing activity in 3% H 2 O 2 .
The pooled catalase positive fractions were further purified by gel filtration chromatography using Sephacryl S-200 column (Amersham) and eluted with 25mM sodium phosphate buffer (pH7.5). The purified catalase was then filter sterilized, stored at 4 o C, and protected from light.
Determination of catalase activity. Quantitative catalase activity of H. pylori cell extract or purified catalase protein was determined following the method described by Hazell et al. (20). Briefly, catalase activity was measured spectrophotometrically at 25 o C by following the decrease in absorbance at 240 nm (ε 240nm = 43.48 M -1 cm -1 ) of 13 mM H 2 O 2 in PBS. All assays were repeated to give 12 rate determinations for the first minute of reaction. One unit was defined as the amount of enzyme that catalyzes the oxidation of 1 µmol H 2 O 2 min -1 under the assay condition. X-band (~9.6 GHz) EPR spectra were recorded on a Bruker ESP-300E EPR spectrometer equipped with an ER-4116 dual-mode cavity and an Oxford Instruments ESR-9 flow cryostat.
The intensity of the EPR signals was normalized to the OD of whole cell samples and the protein absorption band of cell-free extracts and purified samples.

Monitoring KatA and AhpC proteins in H. pylori cell extract by SDS-PAGE.
Both KatA and AhpC are major proteins expressed in H. pylori cells. Hazell et al. (20) reported that catalase accounts for ~1% of the cell's total protein in H. pylori. The proteome analysis of Jungblut et al. (19) showed that AhpC (TsaA) is the third most abundant protein in H. It was observed in Pseudomonas aeruginosa that catalase activity is affected by mutation of a separate gene (30); a bfrA mutant of P. aeruginosa had only 47% the KatA activity of wild type strain, despite possessing wild type expression level of KatA. BfrA, composed of 24 subunits capable of binding 700 iron atoms, is the major iron storage protein in P. aeruginosa (31). The results of Ma et al. (30) suggested that BfrA is required as a source of iron for the heme prothetic group of KatA. H. pylori possess two iron-storage proteins, NapA and Pfr. NapA is a homologue of bacterial DNA-protecting proteins (Dps), and its molecular structure has been determined (32,33). It has a dodecameric structure (12 subunits) capable of binding up to 500 iron atoms (32). Pfr is the major iron storage protein in H. pylori (34). As a typical ferritin, Pfr 11 forms a 24mer structure and binds more than 2000 iron atoms (35). As shown in Fig.2A, H. pylori KatA activity is not significantly affected by loss of either NapA or Pfr. Therefore, the observation that H. pylori KatA activity is affected by loss of AhpC is an unusual phenomenon, unrelated to iron storage, which is different from the modulation of catalase activity by BfrA in P. aeruginosa.
As shown in Fig.1, the resolution of KatA protein band and its density relative to the total protein pattern allows it to be unambiguously monitored. Comparison of the protein profiles from crude extracts of the wild type and various mutant strains (Fig.1) indicated that knock-out of AhpC did not significantly change the level of catalase protein expression. Hence a regulatory role for AhpC in catalase expression was ruled out. Therefore, we hypothesized that the catalase in ahpC mutant cells might have undergone certain structural changes leading to its partial inactivation.
Purified H. pylori AhpC was shown to be able to reduce H 2 O 2 in vitro at a similar rate as its t-butyl hydroperoxide (tBOOH) reducing activity (22). However, the H 2 O 2 -decomposing activity in the katA mutant (AhpC+) cells was undetectable (our result Fig. 2A and ref. 23, 24). Because AhpC requires thioredoxin (Trx) and thioredoxin reductase (TrxR) for its activity, failure to detect the H 2 O 2 -decomposing activity of AhpC could be due to a limited amount of available Trx-TrxR proteins. Therefore, we tested whether the WT cell extract has the activity to reduce organic peroxides. Using tBOOH as a substrate, and without adding reductant, the WT cell extract showed reductase activity, with the specific activity being 15.  Fig.2A,B), despite no change in KatA protein expression level (Fig.1). The intensity of EPR signals was strictly normalized based on the protein concentration of each sample. These results suggested that disruption of AhpC in the cells results in a modification of the catalase heme environment, leading to the partial inactivation of the catalase.

13
At present, the identity of the heme structural perturbation signal corresponding to the g value of 6.8, 5.1, 2.0 resonance is not clear, but the complete loss of these signals in the katAahpC double mutant cells indicated that both resonances arise from catalase, not from other proteins that might have been overexpressed in the ahpC mutant cells. Since similar EPR signals with increased rhombicity have been observed in purified H. pylori catalase samples with formate or azide bound in the distal heme pocket (g ~ 6.6, 5.4, 2.0) (37), the heme environment is likely to be modified by binding of a small molecule in close proximity to the heme iron.

Inactivation of catalase in cell-free extracts by organic hydroperoxide.
The primary function of AhpC is associated with the detoxification of organic hydroperoxides (8,9). The active site of the AhpC enzyme can accommodate virtually any

ROOH, including HOOH (H 2 O 2 ) (22, 38). H. pylori AhpC mutant cells exhibited growth
sensitivity to hydroperoxides and to the superoxide-generating agent paraquat (13). To test whether any of these oxidative agents is responsible for inactivation of catalase, we treated the cell-free extract of wild type H. pylori with various agents followed by measuring catalase activity and monitoring EPR signals (Fig. 3). Treatment with up to 0.5 M H 2 O 2 did not significantly affect catalase activity and did not change the catalase heme EPR signal, indicating H. pylori catalase is highly stable in the presence of a high concentration of its own substrate.
Similarly, treatment with up to 0.1 M paraquat had no effect (data not shown). Since the whole cells of a sodB mutant strain (Fig. 2) also showed no effect on catalase (neither on the enzyme activity nor the EPR signal), it seems that H. pylori catalase is not sensitive to superoxide.
To test whether organic hydroperoxides are responsible for inactivation of catalase, we used t-butyl hydroperoxide (tBOOH), a small molecule of organic hydroperoxide that is soluble in aqueous solution, to treat an H. pylori cell extract. By treatment with various concentrations of 14 tBOOH, we observed progressive inactivation of catalase and a concomitant decrease in the g = 6.4, 5.4, 2.0 resonance and a slight increase in the g = 6.8, 5.1, 2.0 resonance (Fig. 3). For example, after treatment with 150 mM tBOOH, about 70% of catalase activity was lost and the g = 6.4, 5.4, 2.0 resonance was decreased to approximately one third of the intensity in untreated cell-free extract. Treatment of wild-type cell extract with approximately 70 mM tBOOH reduced the catalase activity to one half of the control level, and the EPR signal mimiced that of the whole cells of the ahpC mutant strain (Fig.2B). As a control, the cell-free extract of the ahpC mutants without treatment (not shown) gave rise to an EPR signal similar to that of whole cells of ahpC mutant strain. is stable with 500 mM H 2 O 2 , while it is inhibited 50% with 70 mM tBOOH. These results indicated that an organic hydroperoxide is able to induce inactivation and perturbation of the catalase heme environment in cell-free extracts analogous to those induced by the absence of AhpC in the ahpC mutant cells.

Inactivation of purified H. pylori catalase by organic hydroperoxide.
To closely examine the sensitivity of catalase to hydroperoxides in vitro, H. pylori catalase was purified to near homogeneity (Fig.4) and thus the accumulation of compound II leads to the deactivation of catalase (39). In contrast, H. pylori catalase can withstand very high concentrations of H 2 O 2 (molar ratio = 1:3000) without loss of enzyme activity (Fig. 5). However, purified H. pylori catalase showed sensitivity to organic hydroperoxides, confirming the results observed for cell-free extracts. Treatment with increasing concentrations of tBOOH resulted in progressive decrease in the catalase g = 6.4, 5.4, 2.0 EPR signal and concomitant decrease in the catalase activity (Fig. 5). Under these experimental conditions, pre-treatment of purified catalase with tBOOH in a molar ratio of approximately 1:400 caused the loss of 50% of catalase activity, and an approximate 50% decrease in the catalase EPR signal.
Parallel absorption studies (Fig.6) indicate that addition of a 800-fold excess of tBOOH to purified catalase results in the immediate and near complete one-electron oxidation to yield the Fe(IV) compound II species as evidenced by the shift in the Soret-band maximum from 405 to 430 nm (36). The Fe(IV) compound II species is EPR silent and is not catalytically competent for dismutation of hydrogen peroxide into water and oxygen. Hence the inhibition and loss of catalase EPR signal on addition of tBOOH is attributed to the formation of compound II species.
Compared to the results observed for the whole cells (Fig.2) or cell extracts (Fig.3), g = 6.8, 5.1, 2.0 EPR signals are not evident with the purified catalase after tBOOH treatment. The binding of an organic hydroperoxide in the distal heme pocket is a good candidate for the origin of the g = 6.8, 5.1, 2.0 EPR signal in the cells or cell extracts. But this binding may require one or more additional small molecules that are not present in the purified enzyme sample. We also purified the catalase from H. pylori ahpC mutant cells, and examined its sensitivity to hydroperoxides. The results (not shown) were the same as those for the catalase from the wild type strain.

Higher levels of lipid hydroperoxides in ahpC mutant cells.
Unsaturated fatty acids have been repeatedly found as a constituent of lipids in H. pylori, and the growth of H. pylori displayed sensitivity to addition of unsaturated free fatty acids due to their incorporation into phospholipids and subsequent membrane destruction (40, 41). Thus, under the physiological (oxidative stress) condition, there might be a steady flow of lipid hydroperoxides present within H. pylori cells; this situation would require an abundant organic peroxide reductase activity to remove the damaging organic hydroperoxides.
The results presented above suggested that loss of AhpC function leads to accumulation of organic peroxides, which is responsible for inactivation of catalase. To further support this notion, we measured the levels of lipid hydroperoxides in wild-type and ahpC mutant H. pylori cells ( Table 1). The total amount of lipid hydroperoxides (in the entire cell extract) in ahpC mutant cells was determined to be about 3 times that of wild type cells. The majority of lipid hydroperoxides within the cells was present in the membrane fraction. When considering only the membrane fraction, the extent of lipid peroxidation in ahpC mutant cells was about 4 times that of wild type cells. This result indicated that organic hydroperoxides indeed accumulated to a significantly greater extent in the ahpC mutant cells compared to the wild type cells.

17
The above result also suggests that under physiological conditions the major function of H.
pylori AhpC is to reduce organic hydroperoxides. Knock-out of AhpC (an abundant protein) to create ahpC mutants enabled us to detect a significant increase of lipid hydroperoxides in the cells. To our knowledge, this is the first direct demonstration of lipid peroxidation in bacterial cells. The detoxification of organic hydroperoxides in H. pylori seems quite different from that in E. coli. It is reported that E. coli lacks the polyunsaturated fatty acids necessary for lipid peroxidation (42), and AhpC is not such an abundant protein in E. coli as in H. pylori. In accordance with these, the recent study of Seaver and Imlay (38) suggested that the major physiological substrate of E. coli AhpC is H 2 O 2 rather than organic hydroperoxides.
We showed that in vitro treatment of purified catalase with tBOOH in a molar ratio of approximately 1:400 caused 50% inactivation of catalase (Fig.5). To complement this result, we found that the ahpC mutant cells accumulated much more lipid hydroperoxides than the wild type (Table 1)