Received 2 August 1994/Returned for modification 12 September 1994/Accepted 8 October 1994

Transcription

1 INFECTION AND IMMUNITY, Jan. 1995, p Vol. 63, No /95/$ Copyright 1995, American Society for Microbiology Analysis of Expression of CagA and VacA Virulence Factors in 43 Strains of Helicobacter pylori Reveals that Clinical Isolates Can Be Divided into Two Major Types and that CagA Is Not Necessary for Expression of the Vacuolating Cytotoxin ZHAOYING XIANG, 1 STEFANO CENSINI, 1 PIETRO F. BAYELI, 2 JOHN L. TELFORD, 1 NATALE FIGURA, 2 RINO RAPPUOLI, 1 * AND ANTONELLO COVACCI 1 Immunobiological Research Institute Siena 1 and Institute of Gastroenterology, University of Siena, Siena, Italy Received 2 August 1994/Returned for modification 12 September 1994/Accepted 8 October 1994 Colonization of the mucosa of the stomach and the duodenum by Helicobacter pylori is the major cause of acute and chronic gastroduodenal pathologies in humans. Duodenal ulcer formation strongly correlates with the expression of an antigen (CagA) that is usually coexpressed with the vacuolating cytotoxin (VacA), a protein that causes ulceration in the stomach of mice. However, the relationship between these two virulence factors is unknown. To define whether CagA and VacA are coexpressed in all clinical isolates and their relationships, we collected 43 clinical isolates of H. pylori and studied their genetic and phenotypic properties. Based on this analysis, most of the strains could be classified into two major types. Type I bacteria had the gene coding for CagA and expressed the CagA protein and the vacuolating cytotoxin. Type II bacteria did not have the gene coding for CagA and did not express either the CagA protein or the vacuolating cytotoxin. Type I and type II bacteria represented 56 and 16%, respectively, of the 43 clinical isolates, while the remaining 28% had an intermediate phenotype, expressing CagA independently of VacA or vice versa. This finding shows that although it is present in most cytotoxic strains, CagA is not necessary for the expression of the vacuolating cytotoxin. Helicobacter pylori is a gram-negative, S-shaped, microaerophilic bacterium that infects more than 50% of the human population. It has recently been identified as the cause of acute and chronic active gastritis, peptic ulcer disease, and atrophic gastritis. The bacterium is also suspected to be involved in the genesis of gastric adenocarcinoma and MALT lymphoma (2). Infection usually occurs early in life and, unless eradicated by specific antibiotic treatment, persists forever (1). All clinical isolates of H. pylori produce a number of virulence factors that are essential for the colonization of the stomach and survival in this hostile environment. The best-known virulence factors are the urease, which is supposed to play an important role in the neutralization of gastric acid secretion (9); the flagella, which are essential for swimming through the mucous layer (10, 19); the superoxide dismutase (17); and several molecules that are involved in specific adhesion to the superficial epithelial cells of the stomach (3, 11). In addition to the above factors that are produced by all strains, some factors are produced by only a subset of clinical isolates: these are the cytotoxin-associated protein (CagA) (4, 21) and the vacuolating cytotoxin (VacA) (6, 14, 16, 20). Preliminary studies on a small number of bacterial strains have shown that usually the strains that produce the cytotoxin also produce CagA, while the strains that do not produce the cytotoxin have the cytotoxin gene but lack the gene coding for CagA (4). Serological studies have shown that CagA-producing strains are associated with duodenal ulcer. Antibody titers against CagA correlate with severity of disease, suggesting a role for the CagA protein and the coexpressed cytotoxin in pathogenesis (8, 23, 24). These observations have * Corresponding author. Mailing address: IRIS, Via Fiorentina 1, Siena, Italy. Phone: (39) Fax: (39) recently been confirmed by the finding that purified cytotoxin alone is able to induce ulceration in the stomach of mice (20). In this study we have collected 43 clinical isolates of H. pylori and systematically studied the presence of the CagA and VacA genes and proteins in order to find out whether the two virulence factors are coexpressed in all strains and what the relationship is between the two antigens. The results showed that most of the clinical isolates can be divided into two broad families, which either coexpress or do not coexpress CagA and VacA. However, the isolation of some strains in which the expression of the two factors is dissociated showed that CagA is not necessary for VacA expression and that clinical isolates which express only one of the two virulence factors can be found. MATERIALS AND METHODS Bacterial strains. The 43 H. pylori strains used in this study are listed in Table 1. H. pylori isolates were obtained from patients undergoing upper gastrointestinal examination. The diagnosis obtained by endoscopy and histology was recorded for most of the patients from whom the strains were isolated. Culture conditions. Following primary isolation, H. pylori strains were grown on Columbia agar with 5% horse blood, 0.2% cyclodextrin (12), and Dent or Skirrow s antibiotic supplement (Oxoid, Basingstoke, United Kingdom) in a microaerophilic atmosphere for 3 days at 37 C, then subcultured in an incubator witha5%co 2 atmosphere, and frozen at 70 C. Most of the strains were frozen four to six passages after primary isolation. Subsequent analyses were performed on strains derived from the frozen stocks. Helicobacter felis was cultured on blood agar base no. 2 (Oxoid) supplemented with 5% (vol/vol) horse blood and an antibiotic supplement consisting of 10 g of vancomycin, 3 g of polymyxin B, 5 g of trimethoprim, and 2.5 g of amphotericin B per ml. Bacteria were cultured on freshly prepared agar plates and incubated, lid uppermost, under microaerobic conditions at 37 C for 2 to 3 days (13). Water extracts. Colonies were harvested in distilled water. The concentration of soluble fraction was measured by the A 600. The surface proteins were separated from bacterial cells by vortexing the cells at room temperature. Cells were centrifuged at 10,000 g for 10 min. The supernatant was stored at 20 C and 94

2 VOL. 63, 1995 MOLECULAR CHARACTERIZATION OF H. PYLORI ISOLATES 95 TABLE 1. Genetic and phenotypic properties of 43 isolates of H. pylori and H. felis a Type and strain no. (designation) CagA Vacuolating cytotoxin caga vaca php8.1 CagA size (kda) VacA activity Type I 1 (CCUG 17874) (G39) (60190) (G11) (G20) (G27) (G29) (G32) (G33) (G46) (G56) (G65) (G89) (G103) (G106) (G109) (G120) (932) (933) (Ba158) (Ba167) (Ba185) (Ba212) (Ba211) 136 a 25 (931) (Ba99) (Ba137) (Ba179) (Ba194) 132 b 30 (Ba182) 31 (G12) c 32 (G25) 33 (G204) d 34 (G104) 35 (Tx30a) 36 (Ba142) Type II 37 (Ba82) 38 (925) 39 (G21) 40 (G50) 41 (G198) 42 (2U ) 43 (2U ) 44 (H. felis CS1) a The strains whose designations start with agorabawere isolated from the hospital of Grosseto and from the hospital of Siena (Italy), respectively. Four strains came from England (those whose designation starts with 9), two were from France (2U and 2U ), and one was from Texas (Tx30a). The type strains include CCUG 17874, 60190, and H. felis CS1. All strains except 44 were H. pylori. b For caga and vaca:, gene present;, gene absent. For php8.1:, hybridization to 8.1-kb plasmid probe;, no hybridization. c For CagA and VacA:, protein present;, protein absent. subsequently used for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), Western blot (immunoblot), and vacuolization assays (5). Whole-cell extracts. To obtain cell lysates, bacterial cells were washed once in phosphate-buffered saline (PBS) at ph 7.4 and then centrifuged. The pellet was stored at 20 C and resuspended in SDS-PAGE sample buffer. Chromosomal DNA isolation. Bacteria were harvested, suspended with lysozyme at 100 g/ml in 0.1 M NaCl 10 mm Tris-HCl 1 mm EDTA (ph 8), and incubated at 37 C for 15 min. SDS was added (1%, wt/vol), and the mixture was heated at 65 C. Proteinase K (25 g/ml) was added before incubation at 50 C for 2 h. DNA was purified by CsCl gradient centrifugation with ethidium bromide, precipitated with 77% (vol/vol) ethanol, recovered with a sealed glass capillary, and resuspended in 1 ml of distilled water (15). Plasmid DNA preparation. A 3-day-old culture was harvested from plates and washed in 0.1 M NaCl 10 mm Tris-HCl 1 mm EDTA (ph 8). Bacteria were centrifuged at 10,000 g for 10 min and resuspended in 200 l of ice-cold solution 1 (50 mm glucose, 25 mm Tris-HCl [ph 8], 10 mm EDTA); 10 l of lysozyme (100 mg/ml) was added, and the mixture was incubated at room temperature for 20 min. Then, 400 l of freshly made solution 2 (0.2 N NaOH, 1% SDS) was added, and the mixture was placed on ice for 5 min before the addition of ice-cold solution 3 (15). Cellular debris was centrifuged at 4 C for 15 min. The 800 l of supernatant was extracted once with phenol-chloroform and then once with chloroform. After extraction, the DNA solution was mixed with absolute ethanol at 20 C for 1 h, washed with 75% ethanol, and dried. The DNA pellet was suspended in 50 l of distilled water. DNA was further purified by ethidium bromide high-salt extraction as described by Stemmer (18). Cytotoxicity test on HeLa cells. HeLa cells were cultured in plastic flasks in Dulbecco s modified Eagle s medium (DMEM) containing 25 mm HEPES buffer (N-2-hydroxyethylpiperazine-N -2-ethanesulfonic acid, ph 7.2), 4% fetal bovine serum, and 2 mm glutamine. Cells were maintained at 37 C ina5%co 2 atmosphere. At 24 h before toxin addition, cells were suspended with trypsin- EDTA and seeded in 96-well titration plates in DMEM so that the concentration of cells was 10 4 cells per well. The supernatant prepared by water extracts from different strains was diluted twofold from 1:2 to 1:64 and incubated with HeLa cells for 24 h (5). Vacuolization was assessed by light microscopy. Vacuolization of more than 50% of cells was defined as a positive result. DNA probes. Radioactive probes generated by PCR amplification were used in hybridization experiments. Oligonucleotides were synthesized on an Applied Biosystems synthesizer by the automated phosphoramidite coupling method and purified by standard protocols. The H. pylori caga primers DZ3 (5 -AGTAAG GAGAAACAATGA) and R009 (5 -AATAAGCCTTAGAGTCTTTTTGGAA ATC) were derived from the sequenced caga gene, giving an amplified product of 1,350 bp. The primers F6 (5 -GCTTCTCTTACCACCAATGC) and R20 (5 -TGTCAGGGTTGTTCACCATG) from the H. pylori vaca gene were used to amplify a 1,160-bp DNA fragment. All PCR mixtures contained 50 mm KCl, 10 mm Tris, 200 mm deoxynucleoside triphosphates, 30 pmol of each primer, 0.1 g of bovine serum albumin, 2.5 U of Amplitaq (Perkin-Elmer, Norwalk, Conn.), and 10 ng of H. pylori chromosomal DNA (strain CCUG 17874). Amplifications were performed on a PCR system 9600 thermocycler (Perkin-Elmer) with the following cycling profiles: for caga, 94 C for 45 s, 50 C for 45 s, and 72 C for 45 s for 35 cycles and then extension at 72 C for 10 min; for vaca, 94 C for 1 min, 58 C for 1 min, and 72 C for 1 min for 25 cycles and extension at 72 C for 10 min. The amplified PCR products were purified by electrophoresis on a 1.4% agarose gel. The appropriate fragments were cut out of the gel, placed into a m Spin-X tube (Costar, Cambridge, Mass.), and then centrifuged at 10,000 g for 30 min. The DNA was extracted with 1 volume of phenol-chloroform and then precipitated with 2.5 volumes of absolute ethanol and 1/10 volume of 3 M sodium acetate (ph 5.2) at 80 C for 1 h. The DNA was centrifuged at 4 C for 15 min, and the pellet was resuspended in 20 l of distilled water. The php8.1 probe was derived from the cloned plasmid (3a). The phh361 probe was obtained by HindIII digestion of the cloned vaca gene (nucleotide positions 3215 to 3515) and then treated by the same methods as PCR products. The isolated DNA fragments and plasmid DNA which were used as probes were radioactively labelled by the random priming method (7). Southern hybridization. The genomic DNA and plasmid DNA of bacterial strains were digested with restriction endonucleases, run on a 0.8% agarose gel, and transferred to nylon membranes by standard protocols (15). The membrane filters were then washed three times in 4 SSC 0.2% SDS (1 SSC is 0.15 M NaCl plus M sodium citrate) at 65 C for 30 min and exposed to X- Omat-AR X-ray film (Eastman Kodak, Rochester, N.Y.) with Cronex Lighting- Plus intensification screens (Dupont De Nemours, Wilmington, Del.) for 24 h at 80 C or analyzed in a Phosphorimager 425 with ImageQuant Software (Molecular Dynamics, Sunnyvale, Calif.). SDS-PAGE and Western blot analysis of proteins. Lysates of whole bacterial cells from each strain were separated by SDS-PAGE. The cells were suspended in electrophoresis sample buffer (0.25 M Tris [ph 6.8], 2% SDS, 10% glycerol, 100 mm dithiothreitol, 0.2% bromophenol blue), heated at 100 C for 5 min, and separated in a 10% gel. Electrophoresis was performed at 30 ma for 6 h. The gel was transferred to a nitrocellulose filter (0.45- m pore size; Schleicher & Schuell) at 200 ma for 2 h. The filter was blocked with 3% skim milk in PBS, incubated with a 1:8,000 dilution of antibodies in 3% defatted milk 0.1% Triton- PBS, and shaken constantly at room temperature overnight. After washing, the filters were incubated with a 1:2,500 dilution of horseradish peroxidase-conjugated goat anti-mouse or anti-rabbit immunoglobulin G (Sigma, St. Louis, Mo.) and then washed with solution containing milk, 0.5% Triton, and PBS. The proteins on the filters were incubated with enhanced chemiluminescence detection reagents for 1 min, exposed to X-Omat-AR X-ray films (Eastman Kodak) for 30 min, and developed with an X-Omat M20 processor (Eastman Kodak).

3 96 XIANG ET AL. INFECT. IMMUN. FIG. 1. (A and C) Southern blots of chromosomal DNA from 13 strains cut with HindIII. (A) Restriction fragments were hybridized with the caga probe. The caga gene is present in CCUG 17874, G106, Ba158, 60190, Ba137, Ba182, and G204 (lanes 1 to 7) and absent in G104, Tx30a, G50, 2U, Ba142, and H. felis CS1 (lanes 8 to 13). (C) Restriction fragments were hybridized with the vaca probe. All H. pylori strains contained the vaca gene (lanes 1 to 12). The bands detected by the probe showed a small but visible size heterogeneity that was not investigated further. The two high-molecular-weight bands seen in strains Ba158 and Ba142 are due to partial digestion of DNA. H. felis was negative for vaca gene sequences. Numbers at left are sizes (in kilobases). (B and D) Western blot analysis of CagA and VacA expression in 13 strains. (B) Whole-cell lysates from 13 strains probed with antiserum against CagA protein. CagA protein was expressed in CCUG 17874, G106, Ba158, 60190, and Ba137 (lanes 1 to 5) but not in Ba182, G204, G104, Tx30a, G50, 2U, Ba142, or H. felis (lanes 6 to 13). (D) Whole-cell lysates from 13 strains probed with antiserum against VacA protein. A positive band was shown for CCUG (lane 1), G106 (lane 2), Ba158 (lane 3), (lane 4), Ba182 (lane 6), G104 (lane 8), and Tx30a (lane 9) but Ba137 (lane 5), G204 (lane 7), G50 (lane 10), 2U (lane 11), Ba142 (lane 12), and H. felis (lane 13) were negative. Sizes are shown in kilodaltons. The antibodies used in this study were a polyclonal mouse antiserum against the purified CagA protein and two polyclonal rabbit antisera raised against the amino-terminal and carboxy-terminal portions of the VacA protein expressed as recombinant fragments in Escherichia coli (20). RESULTS Strain collection. Forty-three strains of H. pylori were collected and investigated. Most of the strains (34) were isolated in central Italy from patients undergoing upper gastrointestinal endoscopy examination. They were derived either from the hospital of Grosseto (those designated with G) or from the hospital of Siena (those designated with Ba). Four strains came from England (those designated with 9), two were from France (2U and 2U ), and one was from Texas (Tx30a). In addition to the above strains, we used as a reference well-known laboratory strains, such as CCUG 17874, 60190, and H. felis CS1. Endoscopic and histological test results for the patients from whom the strains were isolated were available for most of the strains. A list of the strains used and their origin is given in Table 1. Presence of the caga gene and protein. The presence of the gene coding for CagA was investigated in all clinical isolates by Southern blotting. HindIII-digested chromosomal DNA was probed with a 32 P-labelled DNA fragment internal to caga. The presence of the protein was detected by Western blotting with bacterial water extracts, probed with a polyclonal serum specific for CagA. Typical examples of the results obtained with Southern and Western blotting are reported in Fig. 1A and B, respectively. The overall results are reported in Table 1. Thirty-three (76.7%) of the 43 clinical isolates had a DNA fragment hybridizing to the caga gene (strains 1 to 33 in Table 1), while 10 (23.3%) did not show hybridizing sequences (strains 34 to 43 in Table 1). Most of the strains having the caga gene showed in Western blotting a band detected by the CagA antiserum (Fig. 1A and B) that varied in size in different strains, ranging from 128 to 140 kda. The size variability of the CagA protein is well visualized in Fig. 2, where selected strains with CagA proteins of different sizes were run next to each other ina4to16%gradient SDS-PAGE gel. The size of CagA in each individual strain is reported in Table 1. Four of the 33 strains having the caga gene did not produce the CagA protein (strains 30 to 33 in Table 1). The reason for the absence of the protein is not clear from the present data; however, one of these strains (Ba182) may have a rearranged caga gene, since the size of the hybridizing band was larger than in the other strains (see Fig. 1A and B). H. felis was negative for the caga gene and protein. Presence of the vacuolating toxin gene and protein and in vitro activity. The vaca gene, tested by Southern blotting with

4 VOL. 63, 1995 MOLECULAR CHARACTERIZATION OF H. PYLORI ISOLATES 97 DISCUSSION FIG. 2. Western blot showing the size variability of CagA protein in different strains. Whole-cell extracts from H. pylori strains were probed with a mouse serum raised against the CagA antigen. two independent probes internal to the coding sequence, was detected in all H. pylori isolates tested and was absent in H. felis (Fig. 1C and Table 1). Both vaca probes detected a prominent band similar in size in all strains tested by Southern blots of HindIII-digested chromosomal DNA (Fig. 1C), indicating that the overall structure of the gene is conserved. However, sequence heterogeneity could be easily detected when PCR was used to amplify regions internal to the gene from different strains (data not shown). The presence of the VacA protein was tested by Western blotting with antiserum against the recombinant fragments A and C, which recognize the amino-terminal and the carboxy-terminal regions of the molecule, respectively. Both sera recognized a band of approximately 94 kda that varied slightly in size in different strains (Fig. 1D). Only 28 (65.1%) of the strains (Fig. 1D and Table 1) showed a band recognized by the anti-vaca antibodies. Vacuolating activity on HeLa cells correlated well with the presence of the VacA protein in Western blotting with two exceptions: strains G12 (strain 31) and Tx30a (strain 35) showed a band in the Western blot but no activity on HeLa cells, and strain Ba142 (strain 36) showed activity on HeLa cells in the absence of a detectable VacA protein by Western blotting. However, transcription of the vaca gene in strain Ba142 was confirmed by Northern (RNA) slot blot (data not shown). Presence of plasmids. Plasmid DNA in the different strains was extracted by the alkaline lysis method and detected in agarose gels. Thirteen of the 43 strains showed plasmid DNA (Table 1). The number of plasmid bands in the gel varied from one to four, and the size of the supercoiled plasmid varied from 4 to 23 kb. In spite of the size heterogeneity observed in agarose gels, all strains containing the plasmid hybridized to an 8.1-kb probe derived from an 8,108-bp plasmid isolated and cloned from strain CCUG (3a). This suggests that all strains shown to contain a plasmid by our method contain a rearranged form of the same plasmid. However, we cannot exclude the presence of other plasmids unrelated to the one isolated from CCUG H. felis also showed a band of plasmid DNA; however, this plasmid did not hybridize to the plasmid derived from CCUG We have collected and analyzed at the molecular level 43 strains of H. pylori and one strain of H. felis. Focusing on markers that are known to vary among different clinical isolates, such as CagA, VacA, and the presence of plasmid DNA, we confirmed in a systematic way the findings reported in many preliminary observations, which can be summarized as follows. H. pylori clinical isolates can be divided into two broad categories that we would have named type I and type II. Type I bacteria have the gene coding for caga and produce the CagA protein and the vacuolating cytotoxin. According to this definition, strains 1 to 24 in Table 1 belong to this category. Type II bacteria do not have the gene coding for CagA and do not produce CagA or the cytotoxin. According to this definition, strains 37 to 43 in Table 1 belong to this category. The observation that purified cytotoxin causes gastric tissue damage and ulceration (20) and the studies showing that duodenal ulcer is always associated with infection with caga-producing type I bacteria (4) suggest that only type I bacteria are associated with severe disease. The difference and the relationships between type I and type II bacteria are not well understood. In addition to the absence of the caga gene and the inability to express the vaca gene, type II bacteria are more fastidious in their growth requirements than type I bacteria; colony formation in agar plates requires 5 days instead of 2 or 3 days (25). At the molecular level, we have observed that type II bacteria have a deletion that covers several kilobases of chromosomal DNA upstream from the caga gene (4a). This region may code for additional virulence factors and regulatory elements necessary for the expression of VacA. Since the deletion is common to all type II bacteria analyzed, the CagA-negative phenotype necessarily includes the loss of phenotypic properties that are expressed in type I bacteria. These properties may be unrelated to CagA expression and can be collectively designated the CagA-positive phenotype. Although 31 of 43 strains (72%) could be classified as either type I or type II, some of the strains showed intermediate phenotypes that could be subdivided into four additional groups, named a to d in Table 1. The a group includes strains 25 to 29, which have the gene coding for CagA and produce CagA but do not produce VacA. The b group includes strains 30 and 31, which, although having the gene coding for CagA, do not produce the CagA protein but do produce the cytotoxin. The c group includes strains 32 and 33, which have the gene coding for CagA but do not express either CagA or the cytotoxin. Group d includes strains 34 to 36, which do not have the gene coding for CagA but produce the cytotoxin. We prefer to consider the strains in groups a to d as subgroups of type I bacteria because they either produce or have the gene for one of the virulence factors that correlate with disease and because they grow fast, as do type I bacteria. The existence of intermediate strains with dissociated expression of CagA and VacA indicates that those two genes, which are coexpressed in most of the strains, do not depend on one another for expression. Genetic disruption of the caga gene confirmed this in a definitive way (22, 25). From these observations, we conclude that an understanding of the linked expression of CagA and VacA must await characterization of the genetic differences between type I and type II bacteria, such as the deletion upstream of the caga gene. The strains described here can also be used in animal models to define the role of different isolates and of individual virulence factors in the establishment of infection and disease. Attempts to correlate clinical disease with infection with type I

5 98 XIANG ET AL. INFECT. IMMUN. or type II bacteria (shown in Table 1) did not show the strong correlation between infection with type I strains and duodenal ulcer that we expected from the serological data that we had obtained previously (23, 24). In fact, some patients from whom a type II strain was isolated also had duodenal ulcer. However, in most of the cases, these patients had antibodies to the CagA protein, indicating that, although only a type II strain was isolated from them, they had been infected by multiple strains. These data indicate that multiple colonies should be isolated from each patient, possibly from different regions of the stomach, and characterized for virulence factor expression. It is also possible that all strains contain the caga gene at some stage and that it is lost from some strains with time. This could explain the antibody response in patients with duodenal ulcer disease whose infecting strains do not have the gene at this point in time. The results reported in this paper provide the rationale for classification of clinical isolates into categories that are defined at the molecular level. This classification could be used for the design of prospective clinical studies to address the prevalence and the role in disease of different strains. ACKNOWLEDGMENTS Zhaoying Xiang is a recipient of the Fellowship Award from UNIDO/ICGEB. We thank Daniela Burroni and Mariangela Dell orco for providing the probe for vaca and the anti-vaca antiserum. We are grateful to M. J. Blaser, A. Labigne, A. Lee, D. Dwarakanath, and F. Mégraud for providing bacterial strains. We thank Giorgio Corsi for the artwork and Catherine Mallia for secretarial assistance. REFERENCES 1. Blaser, M. J Helicobacter pylori and the pathogenesis of gastroduodenal inflammation. J. Infect. Dis. 161: Blaser, M. J Helicobacter pylori: microbiology of a slow bacterial infection. Trends Microbiol. 1: Boren, T., P. Falk, K. A. Roth, G. Larson, and S. Normark Attachment of Helicobacter pylori to human gastric epithelium mediated by blood group antigens. Science 262: a.Censini et al. Unpublished data. 4. Covacci, A., S. Censini, M. Bugnoli, R. Petracca, D. Burroni, G. Macchia, A. Massone, E. Papini, Z. Y. Xiang, N. Figura, and R. Rappuoli Molecular characterization of the 128-kDa immunodominant antigen of Helicobacter pylori associated with cytotoxicity and duodenal ulcer. Proc. Natl. Acad. Sci. USA 90: a.Covacci, A., et al. Unpublished data. 5. Cover, T. L., C. P. Dooley, and M. J. Blaser Characterization of and human serologic response to proteins in Helicobacter pylori broth culture supernatants with vacuolizing cytotoxin activity. Infect. Immun. 58: Cover, T. L., M. K. R. Tummuru, P. Cao, S. Thompson, and M. J. Blaser Divergence of genetic sequences for the vacuolating cytotoxin among Helicobacter pylori strains. J. Biol. Chem. 269: Feinberg, A. P., and B. Vogelstein A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132: Figura, N., M. Bugnoli, M. G. Cusi, A. M. Pucci, P. Lusini, S. Quaranta, A. Barbieri, A. Rossolini, A. L. Di Tommaso, T. De Magistris, R. Rappuoli, L. Marri, R. A. Musmanno, M. Russi, M. Guarna, and M. Losi Pathogenic mechanisms of Helicobacter pylori: production of cytotoxin, p In P. Malfertheiner, and H. Ditschuneit (ed.), H. pylori: gastritis & peptic ulcer. Springer-Verlag, Berlin. 9. Labigne, A., V. Cussac, and P. Courcoux Shuttle cloning and nucleotide sequences of Helicobacter pylori genes responsible for urease activity. J. Bacteriol. 173: Leying, H., S. Suerbaum, G. Geis, and R. Haas Cloning and genetic characterization of a Helicobacter pylori flagellin gene. Mol. Microbiol. 6: Lingwood, C. A Helicobacter pylori: receptors and adhesins, p In C. S. Goodwin and B. W. Worsley (ed.), Helicobacter pylori: biology and clinical practice. CRC Press, Boca Raton, Fla. 12. Olivieri, R., M. Bugnoli, D. Armellini, S. Bianciardi, R. Rappuoli, P. F. Bayeli, L. Abte, E. Esposito, L. De Gregorio, J. Aziz, C. Basagni, and N. Figura Growth of Helicobacter pylori in media containing cyclodextrins. J. Clin. Microbiol. 31: Paster, B. J., A. Lee, J. G. Fox, F. E. Dewhirst, L. A. Tordoff, G. J. Fraser, J. L. O Rourke, N. S. Taylor, and R. Ferrero Phylogeny of Helicobacter felis sp. nov., Helicobacter mustelae, and related bacteria. Int. J. Syst. Bacteriol. 41: Phadnis, S. H., D. Ilver, L. Janzon, S. Normark, and T. U. Westblom Pathological significance and molecular characterization of the vacuolating toxin gene of Helicobacter pylori. Infect. Immun. 62: Sambrook, J., E. F. Fritsch, and T. Maniatis Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 16. Schmitt, W., and R. Haas Genetic analysis of the Helicobacter pylori vacuolating cytotoxin: structural similarities with the IgA protease type of exported protein. Mol. Microbiol. 12: Spiegelhalder, C., B. Gerstenecker, A. Kersten, E. Schiltz, and M. Kist Purification of Helicobacter pylori superoxide dismutase and cloning and sequencing of the gene. Infect. Immun. 61: Stemmer, W. P. C A 20-minute ethidium bromide/high-salt extraction protocol for plasmid DNA. Biotechniques 10: Suerbaum, S., C. Josenhans, and A. Labigne Cloning and genetic characterization of the Helicobacter pylori and Helicobacter mustelae flab flagellin genes and construction of H. pylori flaa- and flab-negative mutants by electroporation-mediated allelic exchange. J. Bacteriol. 175: Telford, J. L., P. Ghiara, M. Dell Orco, M. Comanducci, D. Burroni, M. Bugnoli, M. F. Tecce, S. Censini, A. Covacci, Z. Xiang, E. Papini, C. Montecucco, L. Parente, and R. Rappuoli Gene structure of the Helicobacter pylori cytotoxin and evidence of its key role in gastric disease. J. Exp. Med. 179: Tummuru, M. K. R., T. L. Cover, and M. J. Blaser Cloning and expression of a high-molecular-mass major antigen of Helicobacter pylori: evidence of linkage to cytotoxin production. Infect. Immun. 61: Tummuru, M. K. R., T. L. Cover, and M. J. Blaser Mutation of the cytotoxin-associated caga gene does not affect the vacuolating cytotoxin activity of Helicobacter pylori. Infect. Immun. 62: Xiang, Z., M. Bugnoli, A. Ponzetto, A. Morgando, N. Figura, A. Covacci, R. Petracca, C. Pennatini, S. Censini, D. Armellini, and R. Rappuoli Detection in an enzyme immunoassay of an immune response to a recombinant fragment of the 128-kilodalton protein (CagA) of Helicobacter pylori. Eur. J. Clin. Microbiol. Infect. Dis. 12: Xiang, Z., M. Bugnoli, R. Rappuoli, A. Covacci, A. Ponzetto, and J. E. Crabtree Helicobacter pylori: host responses in peptic ulceration. Lancet 341: Xiang, Z., and A. Covacci. Unpublished data.