Catalase, a Specific Antigen in the Feces of Human Subjects Infected with Helicobacter pylori

Abstract

Recently, we reported the production of three new monoclonal antibodies with high specificity for a Helicobacter pylori antigen suitable for diagnosis of H. pylori infection. The aim of the present study was to identify the antigen recognized by these monoclonal antibodies concerning both H. pylori and the feces of human subjects infected with H. pylori. The cellular antigen was purified from an H. pylori cell extract by immunoaffinity column chromatography with the monoclonal antibody as a ligand. The amino-terminal amino acid sequences (eight residues) of the purified antigen and H. pylori catalase were the same. The molecular weights of native and subunit, specific catalase activity, and UV and visible spectra of the purified antigen were in good agreement with those of H. pylori catalase. The human fecal antigens were purified from two fecal samples of two H. pylori-positive subjects by ammonium sulfate precipitation, CM-Sephadex C₅₀ chromatography, and the same immunoaffinity chromatography used for the H. pylori cellular antigen. The fecal antigens had catalase activity. The amino-terminal amino acid sequences (five residues) of the human fecal antigen and H. pylori catalase were the same. The monoclonal antibodies reacted with the native cellular antigen, but did not react with the denatured antigen, human catalase, and bovine catalase. The results show that the target antigen of the monoclonal antibodies is native H. pylori catalase and that the monoclonal antibodies are able to specifically detect the antigen, which exists in an intact form, retaining the catalase activity in human feces.

Helicobacter pylori infection can be diagnosed by tests requiring endoscopic biopsy of the gastric mucosa (culture, histology, and a rapid urease test) and by noninvasive tests (serology and urea breath test) (20). Recently, two enzyme immunoassays (EIAs) for the direct detection of the fecal H. pylori antigens have been developed. One uses polyclonal rabbit antibody (Premier Platinum HpSA; Meridian Diagnostics, Inc., Cincinnati, Ohio), and the other uses plural kinds of monoclonal antibodies (MAbs) (FemtoLab H. pylori; Connex GmbH, Martinsried, Germany). They have been shown to be reliable tools for noninvasive diagnosis of H. pylori infection (3, 11, 12, 19). However, the lower specificity and considerable lot-to-lot variation of Premier Platinum HpSA have been reported in several papers (4-6, 18), and the H. pylori antigen profile in human feces recognized by the polyclonal antibody or the plural kinds of MAbs remains uncertain.

To develop a diagnostic test for H. pylori infection with higher specificity and reproductivity, we previously reported three new MAbs (21G2, 41A5, and 82B9) that recognize the same fecal H. pylori antigen, partial characterization of the cellular and fecal antigen, and development of a new single-step EIA that uses one kind of MAb, 21G2, for the detection of fecal H. pylori antigen (17). The developed EIA was able to detect 41 H. pylori isolates and fecal samples from seven H. pylori-positive human subjects. Other members of the Helicobacter species, major bacteria in feces, and fecal samples from six H. pylori-negative human subjects gave negative results. The antigen was characterized as follows. (i) The molecular masses of the cellular antigen and the fecal antigen were the same, 260 kDa. (ii) The antigen was labile to sodium dodecyl sulfate (SDS) and heat treatment. (iii) The structure of the antigen was considered to be composed of more than one epitope, such as a homodimer or homotetramer, because the sandwich EIA consisting of one kind of MAb could detect the antigen. For the purpose of contributing to the elucidation of the H. pylori antigen profile in human feces and establishing the basic aspects of the newly developed EIA that uses one kind of MAb, we intended to clarify one of the human fecal antigens originating from H. pylori.

In the present paper, we describe both the identification of the antigen purified by immuoaffinity chromatography with the MAb 21G2 as a ligand and the specificity of the MAbs for the antigen.

MATERIALS AND METHODS

MAbs and human fecal samples.

The MAbs 21G2, 41A5, and 82B9 were produced as described previously (17). The precipitates containing MAbs were obtained by addition of (NH₄)₂SO₄ to mouse ascitic fluid to a concentration of 50% saturation and then centrifugation at 2,000 × g for 20 min. The MAbs were further purified with Affi-Gel according to the procedures indicated by the manufacturer (Bio-Rad Laboratories). Human fecal samples were obtained from two healthy Japanese male subjects (subject A, 56 years old; subject B, 53 years old) and stored at −35°C before use. The subjects were determined to be H. pylori positive by the urea breath test and serology. Consent was obtained from the participants in the study.

Antigen extraction from H. pylori cells.

H. pylori ATCC 43504 was cultured on brain heart infusion (BHI) agar (Difco) plates containing 5% horse blood in a microaerobic environment (Anaero Pack Helico system; Mitsubishi Gas Chemical Co., Inc.) for 4 days at 37°C. The bacterial cells were harvested, washed in phosphate-buffered saline (PBS), suspended in PBS containing 0.5% formalin, and then stored overnight at 4°C. The bacterial cells were washed three times in PBS and disrupted by sonication with a Biom'c model 7250 (output 3, 50% duty cycle for 10 min; Seiko Instruments and Electronics, Ltd.). The soluble fraction containing the antigen was obtained by ultracentifugation at 90,000 × g for 30 min.

Antigen extraction and partial purification from human fecal samples.

Two human fecal samples (165 g from subject A and 92 g from subject B) were used. A fecal sample was suspended in fourfold volumes of PBS, and the supernatant was obtained by centrifugation at 10,000 × g for 30 min and then ultracentrifugation at 90,000 × g for 30 min. (NH₄)₂SO₄ was added to the supernatant to a concentration of 40% saturation, and the mixture was centrifuged at 8,000 × g for 30 min. (NH₄)₂SO₄ was further added to the resultant supernatant to a concentration of 80% saturation, and the mixture was centrifuged at 8,000 × g for 30 min. The precipitate contained the antigen, and it was dissolved in 10 mM potassium phosphate buffer (pH 7.0) and dialyzed against the same buffer. The dialysate was applied to a column of CM-Sephadex C₅₀ (1 by 2.5 cm) equilibrated with the same buffer. The column was washed with the buffer, and the antigen was eluted with PBS. The antigenicity-positive fractions in the eluate were pooled (partially purified antigen).

Affinity chromatography on MAb 21G2 column.

Fifty milligrams of MAb 21G2 was immobilized on 3 g of CNBr-activated Sepharose 4B according to the procedures indicated by the manufacturer (Amersham Pharmacia Biotech). The soluble fraction from H. pylori (5 ml, 4 mg of protein/ml) or the partially purified antigen from human feces was applied to a column of MAb 21G2-immobilized Sepharose 4B (2 by 3 cm) equilibrated with PBS for 2 h at room temperature. The column was washed with PBS, and the antigen was eluted with 0.2 M glycine-HCl buffer (pH 3.0). Each fraction was neutralized with 1 M Tris, and the fractions containing the antigen were pooled and concentrated by dehydration with polyethylene glycol 20000 (Wako Pure Chemical Co.).

Identification of antigen.

The purity and subunit molecular weight of the antigen were estimated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (4 to 20% gradient) (9). A low-molecular-mass-SDS calibration kit (Amersham Pharmacia Biotech) was used for molecular mass standards. Proteins were visualized by staining the gel with silver stain (Kanto Chemicals Co., Inc.). The molecular weight of the native antigen was estimated by Sephacryl-S300RH gel filtration column chromatography (1.5 by 140 cm, 0.1 M phosphate buffer [pH 6.5]) with molecular mass calibration kits from Amersham Pharmacia Biotech. The amino-terminal amino acid sequence was determined by automated Edman sequencing with the HP G1005A protein sequencing system (Hewlett-Packard Company). UV and visible spectra (220 to 600 nm) were recorded by a spectrophotometer (U-3210; Hitachi).

Assays.

Antigenicity detection by a direct sandwich EIA with MAb 21G2 was performed as described previously (17). Briefly, plastic 96-well EIA microtiter plates (Coster) were coated with MAb 21G2 (5 μg of protein/ml in PBS) overnight at 4°C. After blocking nonspecific binding sites with 250 μl of PBS containing 1% skim milk (Difco) (blocking buffer) for 1 h at 4°C, 50 μl of samples and 50 μl of peroxidase-conjugated MAb 21G2 were added to each well of the plates, and the plates were incubated for 1 h at 25°C. After the plates were washed five times with 250 μl of PBS containing 0.05% Tween 20 (washing buffer), 100 μl of substrate solution containing 3, 3′, 5, 5′-tetramethylbenzidine and H₂O₂ (TMB 1 component Microwell peroxidase substrate; BioFX Laboratories) was added. The reaction was stopped after 10 min by adding 50 μl of 1 N H₂SO₄, and the absorbance was measured on a microplate reader (model 550; Bio-Rad Laboratories) at dual wavelengths (450 and 630 nm). Direct sandwich EIAs with the single MAb (41A5 or 82B9) were developed and performed as described above.

Immunoblotting (Western blotting) was performed as described previously (10). For dot blotting, 2 μl of samples was dotted on a nitrocellulose membrane. The samples denatured by heat treatment (100°C, 5 min) in the presence of 1% SDS were also used. The membrane was blocked with 4% (wt/vol) Block Ace (Snow Brand Milk Products Co., Ltd.) for 1 h at room temperature, washed in washing buffer, and incubated in the MAb solution (50 μg/ml in Block Ace). The membrane was then washed in washing buffer, incubated with peroxidase-conjugated antimouse immunoglobulin G (Cappel; dilution, 1:800 in Block Ace) for 1 h at room temperature, and the peroxidase activity was developed in substrate solution (0.6 mM 3, 3′-diaminobenzidine and 4 mM H₂O₂ in PBS).

Catalase activity was determined at 25°C by the spectrophotometric method of Beers and Sizer (2) with a molar absorption coefficient of 43.48 liters/mol/cm at 240 nm (7). Protein concentration was determined with a bicinchoninc acid protein assay reagent (Pierce Chemical Co.) with bovine serum albumin as a standard.

Cross-reactivity with other catalases.

Human erythrocyte catalase and bovine liver catalase were obtained from Sigma and Amersham Pharmacia Biotech, respectively. Cell extract containing catalase of other bacteria was prepared from the following catalase-positive type cultures: H. pylori ATCC 43504, Helicobacter felis ATCC 49179, Helicobacter hepaticus ATCC 51448, Helicobacter mustelae ATCC 43772, Helicobacter cinaedi ATCC 35683, Campylobacter jejuni ATCC 29428, and Escherichia coli ATCC 25922. Helicobacter species and C. jejuni were cultured on BHI agar plates containing 5% horse blood in a microaerobic environment for 4 days. E. coli was cultured aerobically on BHI agar plates for 3 days. All cultures were done at 37°C. Bacterial cells were harvested, washed in PBS, suspended in PBS containing 0.5% formalin, and then incubated overnight at 4°C. The bacterial cells were washed three times in PBS and disrupted by sonication as described above. The cell extract containing catalase was obtained by ultracentrifugation at 90,000 × g for 30 min. Reactivity of the MAbs 21G2, 41A5, and 82B9 with other catalases was tested in the direct sandwich EIAs with the single MAb.

RESULTS

Purification and identification of cellular antigen from H. pylori.

We purified the H. pylori cellular antigen from the soluble fraction in the cytoplasm by immunoaffinity chromatography. An elution profile of the antigen is shown in Fig. 1A. The cellular antigen was eluted in fractions 14 to 19. SDS-PAGE of each fraction showed that the cellular antigen was purified to a single band, and the molecular mass of the antigen was estimated to be 59 kDa (Fig. 1B). The native molecular mass of the purified antigen was estimated to be 260 kDa by gel filtration chromatography (data not shown), suggesting that the antigen consists of a homotetramer structure. The absorption maxima of the purified antigen were 278 and 406 nm (data not shown).

FIG. 1.

Purification of H. pylori cellular antigen by immunoaffinity chromatography with MAb 21G2 as a ligand. (A) Five milliliters of the soluble fraction (20 mg of protein) from H. pylori ATCC 43504 cells was loaded onto an immunoaffinity column. The column was washed with PBS, and then the antigen was eluted with 0.2 M glycine-HCl buffer (pH 3.0). Fractions of 10 ml were collected. The arrow indicates the start of elution. ○, antigenicity (dilution, 1:2,000); •, protein; ▵, catalase activity. (B) Ten microliters of the eluate fractions was mixed with an equal volume of sample buffer. The samples were boiled for 5 min and electrophoresed on an SDS-polyacrylamide gel (4 to 20% gradient), and then the gel was visualized with silver stain. M, marker proteins for molecular weight determination (phosphorylase b, 94,000; bovine serum albumin, 67,000; ovalbumin, 43,000; carbonic anhydrase, 30,000; soybean trypsin inhibitor, 20,100; and α-lactalbumin, 14,400).

We concluded that the purified antigen is native H. pylori catalase from the following facts. (i) The native molecular mass of the antigen was close to that of H. pylori catalase (200 kDa), which is a homotetramer (7). (ii) The molecular mass of the subunit of the antigen (59 kDa) was in good agreement with the value predicted from the amino acid sequence of the H. pylori catalase gene product (58,599 Da) (14) and was also close to the experimentally determined value (50 to 57 kDa) (7, 15, 21). (iii) Eight amino acids from the amino terminus of the purified antigen were determined to be Met-Val-Asn-Lys-Asp-Val-Lys-Gln, which are identical to those in the reported sequence of H. pylori catalase (13, 14, 21). (iv) Catalase activity was eluted parallel with the antigenicity from the immunoaffinity column (Fig. 1A). (v) The absorption maxima of the purified antigen were very similar to that of H. pylori catalase (280 and 405 nm) (7). (vi) The specific catalase activity of the purified antigen (68 mmol of H₂O₂/min/mg of protein) was almost the same as that of the purified catalase (60 ± 3 mmol of H₂O₂/min/mg of protein) (7).

Purification of human fecal antigen from two H. pylori-positive subjects.

By comparing the reactivity of the ultracentrifugal fecal supernatant with that of the H. pylori cell extract in the EIA, as described previously (17), the relative amount of the antigen in feces was estimated to be extremely low (ca. 5 × 10⁻⁵). Therefore, the partial purification by ammonium sulfate precipitation and cation-exchange chromatography on the CM-Sephadex column was performed before the immunoaffinity chromatography. The cation-exchange chromatography was chosen because the isoelectric point of the H. pylori catalase is relatively high (pI = 9.0 to 9.3) (7). The antigenicity was adsorbed on a CM-Sephadex column and eluted with PBS, indicating that the fecal antigen was cationic (data not shown). Elution profiles on the immunoaffinity column are shown in Fig. 2. The antigenicity was detected in the eluate fractions, and both peak fractions of the antigenicity showed catalase activity. SDS-PAGE of the purified human fecal antigen from subject A showed a single band (59 kDa), and five amino acids from the amino terminus of the purified antigen were determined to be Met-Val-Asn-Lys-Asp, which are identical to the reported sequence of H. pylori catalase. The facts described above show the existence of intact catalase originating from H. pylori in fecal samples from human subjects infected with H. pylori.

FIG. 2.

Purification of H. pylori fecal antigen from subject A (A) and subject B (B) by immunoaffinity chromatography with MAb 21G2. Six milliliters of partially purified fecal antigen (2.4 mg of protein and 1 mg of protein for panels A and B, respectively) was loaded onto an immunoaffinity column. The column was washed with PBS, and the antigen was eluted with 0.2 M glycine-HCl buffer (pH 3.0). Fractions of 10 ml were collected. The arrows indicate the start of elution. ○, antigenicity (1:10 dilution for panel A and no dilution for panel B); ▵, catalase activity.

Specificity of the MAbs with the antigen.

To guarantee the high specificity of the MAbs for H. pylori catalase, we examined the reactivity of the MAbs with other catalases. The H. pylori catalase reacted dose dependently in the EIAs that use one kind of MAb—21G2, 41A5, or 82B9 (data not shown)—and the lower limit of the detection for H. pylori catalase was 6 ng of protein/ml. Whereas, no reaction of human erythrocyte catalase and bovine liver catalase was observed up to 100,000 ng/ml in the EIAs. Furthermore, no cross-reactivity with other bacterial catalases was observed in the EIAs (Table 1). Immunoblotting of the H. pylori catalase with the MAbs 21G2, 41A5, and 82B9 showed no detectable band. The MAbs 21G2, 41A5, and 82B9 reacted with native H. pylori catalase, but did not react with denatured H. pylori catalase by dot blotting. The results indicate that the MAbs recognized a conformational epitope consisting of a homotetramer structure, because H. pylori catalase should be converted to a subunit structure in the process of immunoblotting.

TABLE 1.

Reactivity of MAbs in single-step direct sandwich EIA with other bacterial catalases

Origin	Catalase activity (μmol of H2O2/min/ml)a	A450/630b
		21G2	41A5	82B9
H. pylori ATCC 43504	1.6 (10,000)	0.22	0.16	0.14
	6.4 (10,000)	2.0	1.3	1.2
H. hepaticus ATCC 51448	22 (60)	0.014	0.023	0.013
H. felis ATCC 49179	20 (1,200)	0.014	0.032	0.021
H. mustelae ATCC 43772	22 (8,200)	0.014	0.034	0.062
H. cinaedi ATCC 35683	5.4 (12)	0.014	0.037	0.051
C. jejuni ATCC 29428	2.3 (0.81)	0.017	0.031	0.042
E. coli ATCC 25922	20 (120)	0.018	0.033	0.012

^aCell extracts were used as bacterial catalases. Specific activities (in micromoles of H₂O₂ per minute per milligram of cellular protein) are shown in parentheses.

^bData represent mean values in duplicate.

DISCUSSION

We purified H. pylori cellular antigen recognized by MAb 21G2 from the soluble fraction by immunoaffinity chromatography and identified the antigen as H. pylori catalase. Similarly, the human fecal antigens were purified from the two H. pylori-positive human subjects. The amount of the purified fecal antigens was extremely low. Thus, we could not determine the specific catalase activity of the purified fecal antigens precisely. However, from the results obtained and the fact that native molecular masses of the fecal antigen from three positive subjects were the same, 260 kDa (17), we concluded that the existence of intact catalase originated from H. pylori in fecal samples from human subjects infected with H. pylori. The interesting fact that native catalase is excreted in intact form without being denatured in the gut is reported first in the present paper. The catalase activity of H. pylori was significantly greater than that of the related bacterium C. jejuni (7). We also estimated that the H. pylori catalase content was 10% of the total soluble protein from the results of the antigen purification (data not shown). H. pylori catalase activity was positive for aged cultures (the coccoid cells) even up to 160 days (22). H. pylori catalase in whole cells, the supernatant, and soluble enzyme preparations remained active after the exposure to pH ≥3, but lost its activity after exposure to pH 2 (1). However, it is well known that the pH in the stomach could rise to more than 3 while sleeping or during and after eating. When the release of H. pylori happened under such conditions, H pylori should be delivered into the duodenum and finally excreted in the stool, retaining the active form of catalase. Therefore, we consider that the above in vitro studies describing the larger quantity and higher stability of H. pylori catalase are consistent with this conclusion.

Catalase is a ubiquitous enzyme found in eukaryotic and most prokaryotic organisms. Odenbreit et al. reported that H. pylori KatA, a catalase gene (katA) product, is highly homologous to catalases in both prokaryotes and eukaryotes (14). In addition, Klotz et al. reported that the alignment of 74 catalase sequences, including 29 bacterial (including H. pylori), 8 fungal, 7 animal, and 30 plant sequences, revealed a core region of approximately 360 amino acids with significant sequence similarity among all catalases (8). We previously showed no cross-reactivity of MAb 21G2 with other bacteria (H. hepaticus, H. felis, H. mustelae, H. cinaedi, C. jejuni, Bacteroides vulgatus, E. coli, Bifidobacterium breve, and Bifidobacterium infantis) (17). These bacteria, except for B. vulgatus and the two Bifidobacterium species, are catalase positive. In this study, we confirmed the high specificity of the MAbs (21G2, 41A5, and 82B9). The EIAs using the MAbs showed no cross-reactivity with other catalases (human erythrocytes and bovine liver) or bacterial cell extracts containing catalase (H. hepaticus, H. felis, H. mustelae, H. cinaedi, C. jejuni, and E. coli) (Table 1). Newell et al. reported the production of an anti-H. pylori catalase MAb, CP30, which could react with the subunit of the catalase (D. G. Newell, P. Nuijten, A. R. Stacey, and S. L. Hazell, abstract from Microb. Ecol. Health Dis. 4(Suppl.):S120, 1991), and the MAb was used for immunoblotting analysis of the H. pylori catalase in other reports (13, 15). Thus, the target epitope of the MAb CP30 is considered to be on the primary sequence; however, the specificity of the MAb has not been shown. The high specificity of the MAbs (21G2, 41A5, and 82B9) for H. pylori catalase, which may refer to the properties of recognizing a conformational epitope with a homotetramer structure of H. pylori catalase, could be emphasized.

For several pathogenic bacteria, catalase is involved in the defense mechanisms against in vivo killing by polymorphonuclear granulocytes. H. pylori catalase activity was also shown to apparently be essential for the survival of H. pylori at the phagocytes' cell surface (16). H. pylori catalase-deficient mutants occurred spontaneously in vitro (13, 21); however, there has been no report on the isolation of catalase-deficient strains clinically (21). Moreover, a vaccine study indicates that H. pylori catalase is a highly effective antigen, suggesting that it may be essential in vivo (15). The findings suggest that H. pylori catalase is necessary for H. pylori infection. Therefore, the EIA that used the well-characterized MAbs with high specificity for native H. pylori catalase would be a useful diagnostic test for H. pylori infection. A clinical study for the evaluation of the EIA that used the MAb 21G2 is in progress and will be reported elsewhere.