Correlation between cag Pathogenicity Island Composition and ...

2 downloads 0 Views 440KB Size Report
Oct 7, 2002 - immunosorbent assay in duplicate, using IL-8 Eli-pair (Diaclone, Besançon, ...... Segal, E. D., J. Cha, J. Lo, S. Falkow, and L. S. Tompkins. 1999.
INFECTION AND IMMUNITY, Nov. 2003, p. 6573–6581 0019-9567/03/$08.00⫹0 DOI: 10.1128/IAI.71.11.6573–6581.2003 Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Vol. 71, No. 11

Correlation between cag Pathogenicity Island Composition and Helicobacter pylori-Associated Gastroduodenal Disease Christina Nilsson,1,2 Anna Sille´n,1 Lena Eriksson,1 Mona-Lisa Strand,1 Helena Enroth,3 Staffan Normark,2 Per Falk,4 and Lars Engstrand1,2* Swedish Institute for Infectious Disease Control, Solna,1 Microbiology and Tumorbiology Center2 and Department of Medicine,4 Karolinska Institute, Stockholm, and Clinical Microbiology Laboratory, Capio Diagnostik AB, Sko ¨vde,3 Sweden Received 7 October 2002/Returned for modification 20 January 2003/Accepted 26 June 2003

Helicobacter pylori infection is associated with a variety of outcomes ranging from seemingly asymptomatic coexistence to peptic ulcer disease and gastric cancer. The cag pathogenicity island (PAI) contains genes associated with a more aggressive phenotype and has been suggested to be a determinant of severe disease outcome. The cagA gene has served as a marker for the cag PAI. However, the presence of this single gene does not necessarily indicate the presence of a complete set of cag PAI genes. We have analyzed the composition of the cag PAI in 66 clinical isolates obtained from patients with duodenal ulcer, gastric cancer, and nonulcer dyspepsia. Hybridization of DNA to microarrays containing all the genes of the cag PAI showed that 76 and 9% of the strains contained all or none of the cag PAI genes, respectively. Partial deletions of the cag PAI were found in 10 isolates (15%), of which 3 were cagA negative. The ability to induce interleukin-8 (IL-8) production in AGS cells was correlated to the presence of a complete cag PAI. Strains carrying only parts of the island induced IL-8 at levels significantly lower than those induced by cag PAI-positive isolates. The presence of an intact cag PAI correlates with development of more severe pathology, and such strains were found more frequently in patients with severe gastroduodenal disease (odds ratio, 5.13; 95% confidence interval, 1.5 to 17.4). Partial deletions of the cag PAI appear to be sufficient to render the organism less pathogenic. Colonization of the gastric mucosa by Helicobacter pylori normally occurs in childhood and persists for decades unless it is treated. In most cases, the bacteria cause an asymptomatic chronic inflammation in the antrum of the stomach. However, in a small proportion of individuals, the infection develops into more severe disease, such as peptic ulcer disease and noncardia gastric adenocarcinoma. Both bacterial and host factors have been suggested to play a role in directing the infection toward disease development (9, 16, 22, 23). The CagA (cytotoxin-associated gene A) protein was originally identified as an immunodominant antigen that was found to be associated with the production of VacA (vacuolating cytotoxin A) (11, 50). Since strains carrying these markers were shown to be isolated more frequently from patients with gastroduodenal pathology than from individuals with nonulcer dyspepsia, such strains have been considered to be more virulent. Consequently, H. pylori strains have been divided into two groups. Type I strains carry the cagA gene, express a functional VacA, and are associated with a more aggressive course of infection. In contrast, type II strains lack the cagA gene, carry a nontoxic form of VacA, and are regarded as less virulent (51). The cagA gene was later found to be part of a pathogenicity island (PAI), the cag PAI, a horizontally transferred 40-kb gene fragment containing 27 genes (1, 9). A group of genes within the cag PAI exhibit high sequence homology with type IV secretion systems of other bacteria, the best characterized being the vir genes of Agrobacterium tumefaciens in-

volved in T-DNA transfer to plant host cells (10, 13). CagA has been shown to be secreted via this type IV secretion machinery and translocated into host cells, where it subsequently becomes phosphorylated (3, 5, 37, 42, 47). This is analogous to enteropathogenic Escherichia coli, which translocates the bacterial protein Tir via type III secretion into host epithelial cells, where after phosphorylation it serves as a receptor for the intimin adhesin (29). The tyrosine kinase c-Src has been demonstrated to mediate the phosphorylation of CagA (44), which is a requirement for subsequent interaction with SHP-2 tyrosine phosphatase at the cell membrane, leading to activation of the phosphatase and cellular morphological changes in cultured cells (24). Unphosphorylated CagA that interacts with Grb-2 and activates the Ras/MEK/ERK pathway induces similar cellular phenotypes and cell proliferation (34). Signaling events induced by CagA that alter cellular functions might explain why cag PAI-positive isolates are associated with severe gastro-duodenal pathology. Some H. pylori strains contain only parts of the cag PAI and not the whole set of 27 genes (7, 9, 41). Strains with partially deleted cag PAIs may be cagA positive or negative, and even though they represent an intermediate form, they would be scored as type I or type II strains based on this feature. An important consequence of the interaction between H. pylori and host epithelial cells is the elevated production and secretion of cytokines, such as interleukin-8 (IL-8). Upon contact with host cells, H. pylori induces a signaling cascade involving Ca2⫹-calmodulin and extracellular signal-regulated kinase (ERK) that leads to the activation of the transcriptional regulator NF-␬B, which activates IL-8 production (36). Several of the genes within the cag PAI, including the vir homologues,

* Corresponding author. Mailing address: Swedish Institute for Infectious Disease Control, 171 82 Solna, Sweden. Phone: 46 8 457 24 15. Fax: 46 8 30 17 97. E-mail: [email protected]. 6573

6574

NILSSON ET AL.

have been shown to be required for the stimulation of IL-8 production in host epithelial cell lines (9, 21, 32, 43, 45), indicating a role for the type IV secretion machinery in this inflammatory response. However, the CagA protein per se and its translocation and phosphorylation in the host cells are not involved in this process. The mechanism for signaling between the type IV secretion system and the host, leading to IL-8 secretion, has not yet been described. In the present study, we have analyzed 66 H. pylori isolates from patients with gastric cancer, duodenal ulcer, and nonulcer dyspepsia. We used DNA microarrays and PCR to determine the presence or absence of genes constituting the cag PAI. Our results show that isolates carrying only parts of the cag PAI, whether cagA positive or negative, are less likely to be associated with severe gastroduodenal pathology. Therefore, we suggest that classification of strains into type I and type II should be based on analysis of the entire cag PAI rather than the cagA gene alone. MATERIALS AND METHODS Patients and H. pylori strains. Biopsy samples were obtained from patients included in a hospital-based matched case control study performed at eight hospitals in Sweden that has been described previously (18). Single-colony isolates of H. pylori from 27 patients with gastric cancer, 13 patients with duodenal ulcer, and 26 patients with nonulcer dyspepsia were included in the study. Gastric cancer and duodenal ulcer were designated gastroduodenal disease with severe pathology. The study was approved by the ethics committee of the Medical Faculty, Uppsala University (Uppsala, Sweden). Bacteria were grown on Columbia II agar base BBL plates (Becton-Dickinson, Cockeysville, Md.) containing 8.5% horse blood and 10% horse serum (GC agar plates). The cultures were incubated at 37°C under microaerophilic conditions (5% O2, 10% CO2, 85% N2). From the secondary culture, a single H. pylori colony from each patient was isolated and further analyzed. DNA for PCR and microarray analyses was prepared using the DNeasy tissue kit (Qiagen GmbH, Hilden, Germany). PCR analyses. The 66 H. pylori isolates were typed for cagA status by PCR using two different primer pairs, cagAZ-1 (TTGGAAACCACCTTTTGTATT AGC) and cagAZ-2 (GTGCCTGCTAGTTTGTCAGCG) (6) and cagAsbra forward (ATGATGGCGTGATGTTTGT) and reverse (TTTTCAAGGTCGCTTT TTGC). The isolates were further analyzed by an empty-site PCR to determine whether the whole cag PAI was deleted. A PCR primer pair corresponding to the 5⬘ and 3⬘ flanking regions of the cag PAI, UpCag forward (ACTTTCACGCCC TTTCCCTCC) and DownCag reverse (TTGCATGCGTTATTATTTCAC), was used, and a fragment of 593 bp was amplified when the whole cag PAI was absent (6). The PCR conditions were set to prevent amplification of large DNA segments. Thus, when the major part of the cag PAI was present, no PCR product was obtained. The PCRs were performed by standard procedures using DyNAzyme Taq polymerase and buffer (Finnzymes, Espoo, Finland) with annealing temperatures of 50, 62, and 58°C for cagAZ, cagAsbra, and the empty-site PCR, respectively. As controls, we used isolates with known cag PAI status (6). For verification of the microarray data, PCR on all the cag genes was performed on a subset of the strains with the primers used for the amplification of DNA from strain 26695 that were spotted on the microarray chip (see below and Table 1). DNA fingerprinting analysis was performed by arbitrarily primed PCR, as described elsewhere (19). Microarrays. The 16S rRNA gene was chosen as the positive control for the microarray system; lambda DNA and water were used as negative controls. Primers were designed using The Institute for Genomic Research sequence from strain 26695 (49) and OLIGO software version 5.0 (National Biosciences, Inc., Plymouth, Minn.). The oligonucleotide primers used for PCR were purchased from Genset (Paris, France) (Table 1). Amplification by PCR was performed using standard procedures with DyNAzyme Taq polymerase. The amplified segments representing the different genes were spotted in triplicate onto poly-Llysine-coated PolyPrep slides (Sigma, St. Louis, Mo.) using a GMS 417 arrayer (Genetic MicroSystems Inc., Woburn, Mass.). DNA was UV cross-linked, and the slides were heat snapped, rinsed in 0.1% sodium dodecyl sulfate (SDS) and water, denatured in water at 95°C for 3 min, and cool snapped in ice-cold ethanol before hybridization. Ten micrograms of genomic DNA from each strain was

INFECT. IMMUN. AluI digested, purified using a QIAquick PCR purification kit (Qiagen), and eluted in 50 ␮l of water. After an initial denaturation at 100°C for 7 min, the DNA was randomly labeled with Cy3-dCTP in a mixture of 100 ␮M dATP, 100 ␮M dGTP, 100 ␮M dTTP, 20 ␮M dCTP, 20 ␮M Cy3-dCTP (Amersham Pharmacia Biotech, Uppsala, Sweden), 1⫻ hexanucleotide mix (Boehringer Mannheim, Mannheim, Germany), and 5 U of Klenow enzyme (Boehringer Mannheim). The labeling was performed at 37°C for 16 h, after which the reaction was stopped with 2 ␮l of 0.2 M EDTA. The labeled sample was purified using Quick Spin columns (Boehringer Mannheim) and ethanol precipitated before resuspension in 15 ␮l of hybridization solution (50% formamide, 5⫻ Denhardt’s solution [Sigma], 0.5% SDS, 5 mM potassium phosphate, and 5⫻ SSC [1⫻ SSC is 0.15 M NaCl plus 0.015 M sodium citrate]). The labeled DNA probes were denatured in hybridization solution for 3 min prior to application to the arrays. Hybridization was performed overnight at 42°C in a humid chamber. Posthybridization washing was performed at room temperature for 10 min in 0.2⫻ SSC and 0.1% SDS, followed by two times for 10 min each time in 0.2⫻ SSC. The arrays were dried by centrifugation for 5 min at 40 ⫻ g before being scanned with a GMS 418 array scanner (Genetic MicroSystems Inc.). DNA sequencing. Four of the 10 isolates with a partially deleted cag PAI (Ca34, Ca58, 60:3, and 61:4) were further subjected to DNA sequencing to identify the mechanism behind the deletion. The DNA fragments containing the deletion end points were amplified by PCR using forward and reverse primers from the genes flanking the deletion. The same primers were used for the cycle-sequencing reactions with BigDye Terminator version 3.1 (Applied Biosystems, Foster City, Calif.). The sequencing analysis was performed on an ABI Prism 3100 genetic analyzer. Measurements of IL-8 production in AGS cells. AGS cells (ATCC CRL-1739) were routinely maintained in Nutrient Mixture Ham’s F-12 (Life Technologies, Paisley, United Kingdom) supplemented with 10% fetal bovine serum (Life Technologies) and cultured at 37°C in 95% air and 5% CO2. Cells were seeded in 24-well tissue culture plates and cultured overnight to confluence. H. pylori strains, grown on GC agar plates for 24 h, were harvested and washed in cell culture medium. Following a 6-h coincubation of AGS cells and 108 CFU of each H. pylori strain, the cell culture supernatant was collected and stored at ⫺20°C. The concentration of IL-8 in the supernatant was analyzed by enzyme-linked immunosorbent assay in duplicate, using IL-8 Eli-pair (Diaclone, Besanc¸on, France) according to the manufacturer’s manual. Statistical analyses. Unconditional logistic regression was used for univariate analyses of the relationship between the cag PAI genes and severe disease. Odds ratios and 95% confidence intervals were computed from the model parameters and their standard errors.

RESULTS PCR analyses of the cag PAI. Virulent H. pylori strains associated with disease often carry the cag PAI. We analyzed the compositions of the cag PAIs in 66 clinical isolates from a Swedish population. PCR was performed on all isolates with primers detecting the cagA gene and primers complementary to the flanking regions of the cag PAI, which generate a fragment when the PAI is absent. According to these analyses, 83% (55 of 66) of the strains were cagA positive. In the cag PAI empty-site PCR, 9% (6 of 66) of the strains amplified a fragment, indicating a complete lack of the cag PAI. The remaining 8% (5 of 66) of the isolates were negative in the above-mentioned PCR analyses, suggesting that they contain some of the cag PAI genes but not cagA. Microarray analyses of the whole cag PAI. In order to analyze the composition of the whole cag PAI, we developed a microarray chip containing PCR products from all 27 genes from the cag PAI printed in triplicate on glass slides. Genomic DNAs from the 66 clinical isolates were labeled with Cy3 and hybridized to the chip. Based on the microarray analyses, we found that 76% (50 of 66) of the strains carried an intact cag PAI and that the entire cag PAI was deleted in 9% (6 of 66) of the strains from this Swedish population. Interestingly, isolates

H. PYLORI cag PAI AND DISEASE OUTCOME

VOL. 71, 2003

6575

TABLE 1. Oligonucleotide primers used in this studya ORF name (homologue or reference)

HP0520 HP0521 HP0522 HP0523 HP0524 (virD4) HP0525 (virB11) HP0526 HP0527 (virB10) HP0528 (virB9) HP0529 HP0530 HP0531 HP0532 HP0533 ⫹ 0534 HP0535 HP0536 HP0537 HP0538 HP0539 HP0540 HP0541 HP0542 HP0543 HP0544 (virB4) HP0545 HP0546 HP0547 CagAZ (6) cagAsbra 16S rRNA (17) cag PAI empty site (6)

Primer sequencesb

f: GTGGTGCGATAAAAATCTT r: GGTCATGCAGTGAGTGATAA f: CTAAAGATTGAGGCAAAAC r: AAGTAATGCAGCTGTAAAAA f: ACAGCTGCATTACTTACCTT r: CTTCTTTTTGGCACTCTTA f: AAGGACAAGCTCAAATACGAC r: GCCTGTGGGTTCAAGTGA f: CGCAAAGGCTGAAGTGTA r: TATCAACCACCACAAACAGT f: CGCTATTAACCCTATGAAACT r: AAGAGCTAAATTGATAACCCA f: ACTCAATCTGTCTTCAGTCAT r: ACAGGCTCTTCTAGAGGATA f: TAGAAGAGCCTGTTGTTGGT r: AAAGATGTCGCAGAAACTG f: CAAAATGAATTGATACAAGG r: AGTTGGTGGATCTATGTTTA f: AAGACAGAAACAGCCAACA r: GTCTCTTTTATCGCCTATGC f: AACAATGGGCAAATCAATAG r: TTTTACAACCAACAAACCAAT f: GATAAGACAATCACAACTCAG r: GATTAAAACACTTGCTCTCA f: GTGCAGAATGAAACTAATGA r: CTAATTGATCGCCTTACTTAT f: GGGCGTTAGCTTTACTTGAT r: TGAATTATGACCGCTCTGAA f: CTCTGCGTGAAAGATGAAAA r: GCGATTGGTGGCTACTTAT f: ATGCTAATGTTTGGAAATAA r: TAGATTAGGGTGGTTTTGTA f: AAAATTGTGGTAAAATGTGTT r: AAGAATAAAGAGCGACAGAA f: TTGAATAGGAAAGGAGACACT r: TGATTTTTCTGAGACGATAGG f: GAGAAACAAATCCCAAAGGTT r: AGAAACGACAGCAAGAAACA f: TGCCATCAAAACAGACAAG r: TTAAGGGGCAAAAAGATAGC f: TGTGCCATTCAGAGACAAAC r: GCACCAACATACTCAGATTCA f: GAAATTTCCGATAAGAATCCT r: TTAATGCGCTAGAATAGTGCT f: AGCACTATTCTAGCGCATTA r: AGATAAATATCAGGCCTTGAG f: CCAAAATCGTCAAGACATCA r: AAACAAATGCAACAATGGAG f: GCGCTTTTGAACCTCTTG r: GGCAGGCGGTATTATCTATT f: TTTTTGGCTTCTTTATCGTC r: GGTTGGGACAAGTAGGTTTA f: GTGTAGCGGTTTTTAGCAGT r: CCTTGTATGTCGGTGGTAG 1: TTGGAAACCACCTTTTGTATTAGC 2: GTGCCTGCTAGTTTGTCAGCG f: ATGATGGCGTGATGTTTGT r: TTTTCAAGGTCGCTTTTTGC f: TGGCAATCAGCGTCAGGTAATG r: GCTAAGAGATCAGCCTATGTCC f: ACTTTCACGCCCTTTCCCTCC r: TTGCATGCGTTATTATTTCAC

PCR product size (bp)

Annealing temp (°C)

637

55

414

55

624

58

725

63

1,485

58

1,132

58

690

50

979

55

962

55

808

55

899

55

797

58

980

58

866

58

522

58

492

55

1,250

55

1,027

58

809

58

1,274

55

1,251

55

551

60

886

58

994

58

772

60

620

58

836

60

400

58

823

62

520

58

593

50

a For construction of microarrays, DNA fragments were amplified from strain 26695. A 55-␮l reaction mixture contained 100 ng of genomic DNA, 25 to 35 pmol of primer, 100 ␮M deoxynucleoside triphosphates, PCR buffer, and 0.4 U of DyNAzyme Taq polymerase (Finnzymes). The PCRs were carried out for 35 cycles (95°C, 1 min; 50 to 63°C, 1 min; 72°C, 1 min). b 5⬘—3⬘. f, forward; r, reverse.

6576

NILSSON ET AL.

INFECT. IMMUN.

FIG. 1. Schematic representation of the cag PAI. The 27 ORFs are indicated by arrows in the transcriptional direction and are numbered according to the nomenclature of The Institute for Genomic Research strain 26695 (49). The solid arrows represent genes with sequence homology to type VI secretion systems, with the name of the corresponding vir homologue in A. tumefaciens underneath. The genotypes of the 10 isolates with an intermediate cag PAI are presented as solid and dotted lines, indicating the absence and presence of the genes, respectively. Isolates Ca34, Ca58, 60:3, and 61:4 were subjected to DNA sequencing for determination of the deletion end points. The presence or absence of CagA antibodies in the corresponding sera is indicated by ⫹ or ⫺, respectively. ND, not defined.

with partially deleted cag PAIs, which form an intermediate genotype, were found in 15% (10 of 66) of the strains. To verify the microarray data, we performed PCR analyses of all of the cag PAI genes on DNAs from all 10 cag PAIintermediate strains, two cag PAI-negative strains, and one cag PAI-positive strain. DNAs from strains 67:20 and 67:21 were used as negative and positive controls, respectively (6). A comparison of the PCR analyses and the microarray data shows that 7% of the results differed in the two assays, which indicates that neither of the methods gives a 100% reliable result. This is comparable to error rates observed in other microarray experiments (41). Most differences were false-negative signals in the PCR, which can be explained by improper annealing of the primers due to sequence variation. Furthermore, several of the genes adjacent to the deletion were detected with the microarray but failed to be amplified by PCR. This would reflect the fact that the deletion occurred between the forward and reverse primers. A few genes that were amplified in the PCR assay were not detected in the microarray analyses. Based on both the PCR and microarray data, we estimated the genotypes of the cag PAI-intermediate strains as shown in Fig. 1. All the strains that were genotyped as cag PAI negative in the microarray assay were also identified as negative in the emptysite PCR, showing the reliability of both these methods. In four of the cag PAI-intermediate isolates (strains Ca34, Ca58, 60:3, and 61:4), the sequence neighboring the deletion site was analyzed. PCR fragments were amplified with primers specific for the genes flanking the deleted DNA segment and subjected to sequencing. In all four strains, shorter repeated sequences were found on each side of the deletion (Fig. 2). Correlation between cag PAI status and clinical outcome. The H. pylori isolates analyzed in this study were recovered from patients with different clinical pictures. They were divided into two groups: gastroduodenal disease with severe pathology, including gastric cancer and duodenal ulcer patients, and nonulcer dyspeptic patients with less severe disease. In the group of strains isolated from patients with severe gastro-duodenal pathology, cag PAI-positive strains were found in 88% (35 of 40) of the isolates. The corresponding number for nonulcer

dyspeptic patients was 58% (15 of 26). The entire cag PAI was deleted in 5% (2 of 40) of the cases with severe pathology and in 15% (4 of 26) of the patients with nonulcer dyspepsia. Isolates with partially deleted cag PAIs were found in 8% (3 of 40) of the patients with severe pathology and in 27% (7 of 26) of nonulcer dyspeptic patients (Table 2). Statistical analysis revealed that colonization with H. pylori strains possessing an intact cag PAI was associated with a five-times-increased risk for diagnosis of severe gastroduodenal disease. No significant difference in the risk of developing severe disease was observed between patients infected with cag PAI-negative strains and strains with the intermediate genotype. When patients with cancer or duodenal ulcer were compared, there was no difference between the cag PAI genotypes in these groups. Taken together, these data suggest that H. pylori strains have to carry intact cag PAIs in order to be associated with the development of severe disease. Strains that have internal deletions in the cag PAI have reduced virulence and could be compared to cag PAI-negative strains. The results from the statistical analysis of differences in the cag PAI genetic structure and the severity of gastroduodenal disease are presented in Table 2. Clonal diversity of H. pylori within one host. To analyze the clonal diversity of the cag PAI among single H. pylori cells coexisting in one stomach, microarray analyses were performed on 10 additional colonies from each biopsy specimen

TABLE 2. Statistical analysis of disease outcome linked to presence of cag PAI genes No. of strains cag PAI genotype

Gastric cancer and duodenal ulcer (n ⫽ 40)

Nonulcer dyspepsia (n ⫽ 26)

OR (95% CI)c

Positive Negative and intermediate

35 5a

15 11b

5.13 (1.5–17.4) 1 (reference)

a b c

Two negative and three cag PAI-intermediate strains. Four negative and seven cag PAI-intermediate strains. OR, odds ratio; CI, confidence interval.

VOL. 71, 2003

H. PYLORI cag PAI AND DISEASE OUTCOME

6577

FIG. 2. DNA sequences showing the deletion endpoints in four intermediate strains. PCR fragments of the flanking regions covering the deletion were amplified and sequenced. The sequence obtained in each strain is shown in boldface. For the deleted part, the adjacent corresponding sequence from strain 26695 is shown within brackets. Repeated sequences that were found on both sides of the deletion are shown in color.

from five patients and compared to the original isolates. The selected samples were obtained from patients with cancer of the cardia (Ca57), duodenal ulcer (39:3), and atrophic gastritis (61:4 and 72:3) and with normal endoscopy findings (9:1). The strains initially isolated from the first two patients did not contain any of the cag PAI genes, whereas genotyping of strains from the last three patients had identified strains carrying partially deleted cag PAIs. All 10 of the analyzed colonies from patients Ca57 and 39:3 were negative for the whole cag PAI and thus showed a genotype identical to that of the original isolates. Likewise, all colonies in patient 61:4 had the same genotype and carried a cag PAI with an internal deletion similar to the one identified in the initially analyzed isolate. Interestingly, a mixed population of H. pylori strains carrying different cag PAI genotypes colonized patients 9:1 and 72:3. In both patients, most colonies were of the same genotype as the originally analyzed isolate, but 1 out of 10 had a cag PAI pattern that differed. A complete setup of cag PAI genes was found in one colony isolated from patient 72:3, whereas no genes of the cag PAI were detected in one colony from patient 9:1. However, no fragment was amplified with the empty-site PCR in this colony, indicating that the island was not deleted at the site of the 31-bp repeats flanking the PAI (6). Sequencing from genomic DNA with the primers used for the emptysite PCR (UpCag forward and DownCag reverse) showed that this clone contains the beginning of HP0520 and the downstream sequence of cagA, referred to as HP0548. Genomic fingerprinting was performed on all the isolated colonies from the five selected patients, using arbitrarily primed PCR. All clones within the same stomach showed identical patterns, whereas there were pronounced interpatient variations (data not shown). This result provides an interesting indication of the dynamics of clonal expansion within a microbial population. Correlation between circulating CagA antibodies and cag status. In a previous study, the presence of circulating antibodies against CagA in sera isolated from the patients included in this study was analyzed (18). Sera were not available from three patients colonized with cag PAI-positive strains, one patient with a cag PAI-negative strain, and one patient with a cag PAI-intermediate strain. We found that all but 1 of the 47 patients who had detectable CagA antibodies carried H. pylori strains containing all the cag PAI genes. All five patients with cag PAI-negative strains had not produced any CagA antibodies. Patients colonized with cag PAI-intermediate strains were either CagA antibody positive or negative. Only two of the six strains in the group that carried the cagA gene were isolated

from patients with CagA antibodies. The other four patients, together with the three patients carrying strains without cagA, did not have any detectable antibodies against CagA (Fig. 1). Correlation between cag PAI composition and ability to stimulate IL-8. Since some of the genes in the cag PAI are required for the induction of IL-8 production in epithelial cells, we analyzed a subset of the genetically characterized strains for the ability to induce IL-8 secretion in cultured AGS cells. The strains analyzed included all cag PAI-negative (n ⫽ 6) and -intermediate (n ⫽ 10) isolates and 12 randomly selected isolates out of the 50 containing the whole set of cag PAI genes. With these genetically different clinical isolates, we could confirm the importance of cag PAI genes in the elevated production of IL-8 seen in epithelial cells. Cultured AGS cells coincubated with cag PAI-positive isolates always secreted more IL-8 than cag PAI-intermediate and -negative isolates. Interestingly, strains with internal deletions in their cag PAIs often induced moderate levels of IL-8, significantly different from both cag PAI-positive and -negative isolates (Fig. 3). DISCUSSION In this study, we used microarrays to screen for cag PAI genes in clinical isolates of H. pylori. The results show that cagA-positive isolates do not necessarily have to be cag PAI positive, and vice versa. A partially deleted cag PAI represented a genotype more common than a complete deletion (10 of 66 and 6 of 66, respectively). This approach provides a more accurate view of the genotype and allows a better prediction of the pathogenic traits of the organism than the traditional division of strains into type I and type II based on the presence of the cagA gene alone (51). Censini et al. first identified strains with partially deleted cag PAIs. The molecular mechanism of these genetic rearrangements was explained by incorporation of an insertion element, IS605, in the cag PAI (9). Recently, the composition of the cag PAI in clinical H. pylori isolates has been studied in different populations by various methods, including PCR, Southern blotting, dot blotting, and long-distance PCR (4, 25, 27, 31, 33, 38, 39, 46). Strains with intermediate genotypes, lacking parts of the cag PAI, were found in 0 to 14% of the cases. Some studies have identified a correlation between an intact cag PAI and the development of disease (25, 27, 31, 33), as we have shown in this study, whereas others could not find such a relationship (4, 39). Interestingly, studies have failed to identify IS605 elements within the cag PAI, and no correlation between IS605 elements and intermediate forms of the cag

6578

NILSSON ET AL.

INFECT. IMMUN.

FIG. 3. (A) Induction of IL-8 secretion in AGS cells incubated with the indicated strains for 6 h. The blue, red, and yellow bars represent cag PAI-positive, -intermediate, and -negative strains, respectively. Samples were run in triplicate, and the results are presented as the averages of three to six independent experiments. The error bars show standard errors. (B) The analyzed strains were grouped according to cag PAI genotype, and the mean values with standard deviations of each group are shown. Students t test was used for statistical analysis.

PAI have been found (27, 33, 38, 46). These findings contradict the theory that insertions of IS605 in the cag PAI are the cause of modified PAIs. In this study, we sequenced the deletion end points in four strains with intermediate cag PAI genotypes but were unable to detect the presence of IS605 associated with the deletions in these strains. Instead, we found homologous sequences on each side of the deletion. A probable mechanism for the establishment of these internal deletions within the cag PAI would be that the short repeated sequences serve as homologies enabling slipped-strand mispairing and consequently excision of the enclosed DNA fragment (2). The techniques used to determine the composition of the cag PAI in the aforementioned studies analyze only a subset of the cag PAI genes. Even though they were chosen to cover different parts of the PAI, these results do not give a conclusive picture of the cag PAI structure. Some studies base the selection of target genes for analysis on linkage between certain genes (25, 38, 39). Our results indicate that deletions can occur in all parts of the cag PAI, and such linkage between genes cannot be assumed. Furthermore, due to variations in the sequences of genes, primers used in PCR can fail to amplify a product although the gene is present. This is observed particularly in PCR of the cagA gene, where extensive sequence variations have been reported (15, 38). In the present study, all cag genes were analyzed using hybridization to microarrays and PCR to validate intermediate forms, which gives a more complete view of the composition of the PAI. With this technique, we identified 10 isolates that carried a partially deleted cag PAI. All strains comprised a unique composition of cag genes, often with major deletions in the middle region of the PAI. Our study also shows that H. pylori isolates obtained from patients with severe gastroduodenal disease are more likely to carry a complete cag PAI. This pattern was found in 35 out of 40 cases, a frequency that is statistically significant. Strains that lacked the entire cag PAI, or at least a major portion of it, were

more frequently found in patients with nonulcer dyspepsia than in patients with severe gastroduodenal disease. With a more complete picture of the virulence characteristics of the isolates, it is easier to identify genes and gene products that drive the infection toward persistence or disease. In the gastric niche of the human stomach, genetic and phenotypic diversity within the resident H. pylori population may exist, reflecting the dynamics of the host-microbial cross talk. Genetic changes within the genome of the colonizing strain, accumulated during the years of infection, may give rise to microdiversity between clones. Alternatively, subpopulations of H. pylori can arise from two or more unrelated sources due to multiple infections of the host. We decided to analyze the compositions of the cag PAIs in serial H. pylori clones within the same stomach. Ten additional colonies were isolated from five patients and analyzed with respect to the cag PAI genotype. Most colonies showed a genotype identical to that of the originally analyzed strain. Thus, we considered that the single-colony isolates in the present study were representative of the H. pylori population in these stomachs. However, in two patients, we found that 1 out of 10 colonies carried a cag PAI with a different genotype than the other isolates. All the analyzed subclones from the same stomach showed similar DNA fingerprints, indicating that they all shared a common ancestor. The finding that two of the patients harbored different subtypes in their gastric environments indicates that genetic rearrangements or DNA exchange has occurred in the ancestral strain, which has evolved into two genetically different subtypes. In the case of patient 72:3, the ancestral H. pylori strain was most likely cag PAI positive, and during the time of infection the middle part of the PAI was excised, resulting in a subpopulation of bacteria carrying a partially deleted cag PAI. Similarly, in patient 9:1, the clone with a bigger deletion most probably evolved from the other genotype, which contained the first three genes of the cag PAI. However, other mechanisms for generating these genotypes may be possible, for

VOL. 71, 2003

instance, by DNA transfer from transient colonizing bacteria (30, 48). Our results indicate that the gastric environments of the two patients contain a mixture of clones. Most likely, there is a dynamic relationship between different H. pylori subclones. A combined effect of host response mechanisms and environmental factors may determine which subclones predominate in a given niche at any given time point (8, 26, 28). In this model, it should be expected that subclones that lack virulence genes predominate in a case of nonulcer dyspepsia. But even in these cases, there are seeder populations of more aggressive clones that may expand if the conditions change (12, 13, 35). A clinical observation that may illustrate this event could be the recurrence of peptic ulcer disease. The presence of subclones with various genotypes that allow the bacteria to be more or less pathogenic maximizes the opportunities for survival, persistence, and spread. At any given time, there is an expansion of the subtype that best adapts to the host. Given the facts that H. pylori infection is chronic and that the organism maintains its niche in a continuously changing habitat, this capacity could be of major importance. Disease occurs when this subtle balance is disturbed, for instance, due to changes in the environment or in host mucosal barrier functions that render the host more susceptible to virulent subtypes. Several studies have shown that the detection of CagA antibodies in serum samples does not always correlate with the presence or absence of the cagA gene in the corresponding H. pylori isolate (14, 19). This observation can be explained by three alternative mechanisms. First, a mixed population of both cagA-positive and cagA-negative strains is present in the gastric environment, although only cagA-negative clones have been isolated and analyzed (19, 20). Second, the patient could have been colonized by a cagA-positive strain at an earlier stage of the infection. The isolated strain has subsequently lost the cag PAI, or cag-negative isolates in a mixed infection could have clonally expanded and obliterated the cag-positive strains. Third, the strain carries and expresses the cagA gene but other genes in the cag PAI are deleted or nonfunctional. If the components of the type IV secretion machinery are absent, no secretion of the CagA antigen will occur and antibodies may not be produced. In our study, we found a correlation between CagA antibodies in serum and the presence of a complete cag PAI in the corresponding H. pylori strain. However, in patients colonized with strains intermediate in their cag PAI genotypes, there was a more complex pattern of circulating CagA antibodies. Four patients who were colonized with cagA-positive strains with an intermediate cag PAI genotype did not have any detectable CagA antibodies. These strains all contain large deletions in the cag PAI, including several of the genes required for the type IV secretion apparatus, and the bacteria will therefore not be able to secrete CagA. Since the protein may not be presented to the immune system, this could explain the lack of antibodies against CagA in sera from these patients. However, strains Ca58 and 72:3, which also have large deletions in the cag PAI, were isolated from CagA antibody-positive patients. In these cases, the antibody response might be due to strains other than the ones studied colonizing the host in the past or present. In the case of 72:3, we have shown that there were subclones present at the time of biopsy sampling that contained a whole set of cag PAI genes. After the attachment of virulent H. pylori strains to epithelial

H. PYLORI cag PAI AND DISEASE OUTCOME

6579

cells, production and secretion of IL-8 is stimulated. Direct contact between bacteria and host cells is required (40), and several of the genes in the cag PAI have been shown to play a role in this inflammatory response (9, 21, 32, 43, 45). Recently, the cag PAI genes in strain 26695 were systematically mutagenized, and the abilities of the mutants to induce IL-8 secretion in cultured AGS cells were investigated (21). The study showed that 14 out of the 27 cag genes were required for IL-8 induction. These genes and three additional genes were also essential for translocation of CagA into the host cells. No gene that was involved in IL-8 induction alone was found, indicating that no effector protein encoded by the cag PAI is causing this inflammatory response. In our study, strains harboring all the cag PAI genes were strongly associated with the ability to induce high levels of IL-8 secretion in AGS cells. Strains lacking some or all of the cag PAI genes could invoke only low to moderate IL-8 levels. Since about half of the cag PAI genes are required for IL-8 induction and these genes are scattered over the whole PAI, most strains with partial deletions in the PAI would have lost their ability to induce IL-8. Unexpectedly, strain 1:1 induced comparably low levels of IL-8. This strain contains all the cag PAI genes except cagA, which in isogenic-mutant experiments has been shown not to be involved in IL-8 stimulation. Therefore, the 1:1 strain would be expected to induce as high levels of IL-8 as cag PAI-positive isolates. However, interstrain variations cannot be interpreted in the same fashion as comparisons of isogenic wild-type and mutant strains. Despite the presence of certain cag PAI genes, other genes in the genome might influence the process. Furthermore, point mutations in cag genes that cannot be identified by the microarray technique can influence the type IV secretion apparatus (21). The present findings indicate that intact cag PAIs are associated with the development of gastric pathology. We have chosen to specifically study the genes of the cag PAI, which have been postulated to be associated with virulence. However, with the rapid development of microarray technology, it is now possible to study whole microbial genomes and subsets thereof. The complete genome microarray chip of H. pylori now available provides the possibility to perform more extensive studies of this bacterium that may identify parts of the genome more frequently found in virulent strains (7, 28). It is reasonable to assume that more refined and simplified versions of these techniques will be used as diagnostic tools in the future. Information on microbial genetics and genomics will allow more precise identification of virulent strains. In addition, expression profiling of clinical samples and defined model systems will allow us to better define risk factors in the host and to identify individuals and populations more susceptible to developing H. pylori-associated gastroduodenal diseases. Taken together, the new molecular tools available for studies of pathogenesis will provide the means for designing more rational schemes for treatment and prevention. ACKNOWLEDGMENTS We express our thanks to Sandra Hjalmarsson and Susanne Nilsson for technical assistance. The arrayer and scanner used for the microarray experiments were kindly provided by the Center for Genomics and Bioinformatics at the Karolinska Institute. Grant support was received from the Swedish Foundation for Strategic Research (the Infection and Vaccinology Program), the Swedish

6580

NILSSON ET AL.

Medical Research Council, the Swedish Cancer Foundation, and the Swedish Society of Medicine. REFERENCES 1. Akopyants, N. S., S. W. Clifton, D. Kersulyte, J. E. Crabtree, B. E. Youree, C. A. Reece, N. O. Bukanov, E. S. Drazek, B. A. Roe, and D. E. Berg. 1998. Analyses of the cag pathogenicity island of Helicobacter pylori. Mol. Microbiol. 28:37–53. 2. Albertini, A. M., M. Hofer, M. P. Calos, and J. H. Miller. 1982. On the formation of spontaneous deletions: the importance of short sequence homologies in the generation of large deletions. Cell 29:319–328. 3. Asahi, M., T. Azuma, S. Ito, Y. Ito, H. Suto, Y. Nagai, M. Tsubokawa, Y. Tohyama, S. Maeda, M. Omata, T. Suzuki, and C. Sasakawa. 2000. Helicobacter pylori CagA protein can be tyrosine phosphorylated in gastric epithelial cells. J. Exp. Med. 191:593–602. 4. Audibert, C., C. Burucoa, B. Janvier, and J. L. Fauchere. 2001. Implication of the structure of the Helicobacter pylori cag pathogenicity island in induction of interleukin-8 secretion. Infect. Immun. 69:1625–1629. 5. Backert, S., E. Ziska, V. Brinkmann, U. Zimny-Arndt, A. Fauconnier, P. R. Jungblut, M. Naumann, and T. F. Meyer. 2000. Translocation of the Helicobacter pylori CagA protein in gastric epithelial cells by a type IV secretion apparatus. Cell Microbiol. 2:155–164. 6. Bjo ¨rkholm, B., A. Lundin, A. Sille´n, K. Guillemin, N. Salama, C. Rubio, J. I. Gordon, P. Falk, and L. Engstrand. 2001. Comparison of genetic divergence and fitness between two subclones of Helicobacter pylori. Infect. Immun. 69:7832–7838. 7. Bjo ¨rkholm, B. M., J. L. Guruge, J. D. Oh, A. J. Syder, N. Salama, K. Guillemin, S. Falkow, C. Nilsson, P. G. Falk, L. Engstrand, and J. I. Gordon. 2002. Colonization of germ-free transgenic mice with genotyped Helicobacter pylori strains from a case-control study of gastric cancer reveals a correlation between host responses and HsdS components of type I restriction-modification systems. J. Biol. Chem. 277:34191–34197. 8. Blaser, M. J. 2002. The genetic gymnastics of our indigenous microbes. N. Engl. J. Med. 346:2083–2085. 9. Censini, S., C. Lange, Z. Xiang, J. E. Crabtree, P. Ghiara, M. Borodovsky, R. Rappuoli, and A. Covacci. 1996. cag, a pathogenicity island of Helicobacter pylori, encodes type I-specific and disease-associated virulence factors. Proc. Natl. Acad. Sci. USA 93:14648–14653. 10. Christie, P. J. 2001. Type IV secretion: intercellular transfer of macromolecules by systems ancestrally related to conjugation machines. Mol. Microbiol. 40:294–305. 11. Covacci, A., S. Censini, M. Bugnoli, R. Petracca, D. Burroni, G. Macchia, A. Massone, E. Papini, Z. Xiang, N. Figura, and R. Rappuoli. 1993. Molecular characterization of the 128-kDa immunodominant antigen of Helicobacter pylori associated with cytotoxicity and duodenal ulcer. Proc. Natl. Acad. Sci. USA 90:5791–5795. 12. Covacci, A., and R. Rappuoli. 1998. Helicobacter pylori: molecular evolution of a bacterial quasi-species. Curr. Opin. Microbiol. 1:96–102. 13. Covacci, A., J. L. Telford, G. Del Giudice, J. Parsonnet, and R. Rappuoli. 1999. Helicobacter pylori virulence and genetic geography. Science 284:1328– 1333. 14. Cover, T. L., Y. Glupczynski, A. P. Lage, A. Burette, M. K. Tummuru, G. I. Perez-Perez, and M. J. Blaser. 1995. Serologic detection of infection with cagA⫹ Helicobacter pylori strains. J. Clin. Microbiol. 33:1496–1500. 15. Dong, Q., M. O’Sullivan, W. Hall, C. Herra, C. Kean, C. O’Morain, and M. Buckley. 2002. Identification of a new segment involved in cagA 3⬘ region variation of Helicobacter pylori. FEMS Immunol. Med. Microbiol. 33:51–55. 16. El-Omar, E. M., M. Carrington, W. H. Chow, K. E. McColl, J. H. Bream, H. A. Young, J. Herrera, J. Lissowska, C. C. Yuan, N. Rothman, G. Lanyon, M. Martin, J. F. Fraumeni, Jr., and C. S. Rabkin. 2000. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 404: 398–402. 17. Engstrand, L., A. M. Nguyen, D. Y. Graham, and F. A. el-Zaatari. 1992. Reverse transcription and polymerase chain reaction amplification of rRNA for detection of Helicobacter species. J. Clin. Microbiol. 30:2295–2301. 18. Enroth, H., W. Kraaz, L. Engstrand, O. Nyre´n, and T. Rohan. 2000. Helicobacter pylori strain types and risk of gastric cancer: a case-control study. Cancer Epidemiol. Biomarkers Prev. 9:981–985. 19. Enroth, H., O. Nyre´n, and L. Engstrand. 1999. One stomach—one strain: does Helicobacter pylori strain variation influence disease outcome? Dig. Dis. Sci. 44:102–107. 20. Figura, N., C. Vindigni, A. Covacci, L. Presenti, D. Burroni, R. Vernillo, T. Banducci, F. Roviello, D. Marrelli, M. Biscontri, S. Kristodhullu, C. Gennari, and D. Vaira. 1998. cagA positive and negative Helicobacter pylori strains are simultaneously present in the stomach of most patients with non-ulcer dyspepsia: relevance to histological damage. Gut 42:772–778. 21. Fischer, W., J. Puls, R. Buhrdorf, B. Gebert, S. Odenbreit, and R. Haas. 2001. Systematic mutagenesis of the Helicobacter pylori cag pathogenicity island: essential genes for CagA translocation in host cells and induction of interleukin-8. Mol. Microbiol. 42:1337–1348. 22. Fox, J. G., P. Beck, C. A. Dangler, M. T. Whary, T. C. Wang, H. N. Shi, and C. Nagler-Anderson. 2000. Concurrent enteric helminth infection modulates

INFECT. IMMUN.

23. 24. 25.

26.

27.

28. 29. 30. 31. 32.

33. 34. 35. 36.

37. 38.

39.

40. 41. 42.

43. 44. 45.

46.

inflammation and gastric immune responses and reduces helicobacter-induced gastric atrophy. Nat. Med. 6:536–542. Guruge, J. L., P. G. Falk, R. G. Lorenz, M. Dans, H. P. Wirth, M. J. Blaser, D. E. Berg, and J. I. Gordon. 1998. Epithelial attachment alters the outcome of Helicobacter pylori infection. Proc. Natl. Acad. Sci. USA 95:3925–3930. Higashi, H., R. Tsutsumi, S. Muto, T. Sugiyama, T. Azuma, M. Asaka, and M. Hatakeyama. 2002. SHP-2 tyrosine phosphatase as an intracellular target of Helicobacter pylori CagA protein. Science 295:683–686. Ikenoue, T., S. Maeda, K. Ogura, M. Akanuma, Y. Mitsuno, Y. Imai, H. Yoshida, Y. Shiratori, and M. Omata. 2001. Determination of Helicobacter pylori virulence by simple gene analysis of the cag pathogenicity island. Clin. Diagn. Lab. Immunol. 8:181–186. Israel, D. A., N. Salama, U. Krishna, U. M. Rieger, J. C. Atherton, S. Falkow, and R. M. Peek, Jr. 2001. Helicobacter pylori genetic diversity within the gastric niche of a single human host. Proc. Natl. Acad. Sci. USA 98:14625– 14630. Jenks, P. J., F. Megraud, and A. Labigne. 1998. Clinical outcome after infection with Helicobacter pylori does not appear to be reliably predicted by the presence of any of the genes of the cag pathogenicity island. Gut 43: 752–758. Joyce, E. A., K. Chan, N. R. Salama, and S. Falkow. 2002. Redefining bacterial populations: a post-genomic reformation. Nat. Rev. Genet. 3:462– 473. Kenny, B., R. DeVinney, M. Stein, D. J. Reinscheid, E. A. Frey, and B. B. Finlay. 1997. Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells. Cell 91:511–520. Kersulyte, D., H. Chalkauskas, and D. E. Berg. 1999. Emergence of recombinant strains of Helicobacter pylori during human infection. Mol. Microbiol. 31:31–43. Kidd, M., A. J. Lastovica, J. C. Atherton, and J. A. Louw. 2001. Conservation of the cag pathogenicity island is associated with vacA alleles and gastroduodenal disease in South African Helicobacter pylori isolates. Gut 49:11–17. Li, S. D., D. Kersulyte, I. J. Lindley, B. Neelam, D. E. Berg, and J. E. Crabtree. 1999. Multiple genes in the left half of the cag pathogenicity island of Helicobacter pylori are required for tyrosine kinase-dependent transcription of interleukin-8 in gastric epithelial cells. Infect. Immun. 67:3893–3899. Maeda, S., H. Yoshida, T. Ikenoue, K. Ogura, F. Kanai, N. Kato, Y. Shiratori, and M. Omata. 1999. Structure of cag pathogenicity island in Japanese Helicobacter pylori isolates. Gut 44:336–341. Mimuro, H., T. Suzuki, J. Tanaka, M. Asahi, R. Haas, and C. Sasakawa. 2002. Grb2 is a key mediator of Helicobacter pylori CagA protein activities. Mol. Cell 10:745–755. Montecucco, C., and R. Rappuoli. 2001. Living dangerously: how Helicobacter pylori survives in the human stomach. Nat. Rev. Mol. Cell Biol. 2:457–466. Nozawa, Y., K. Nishihara, R. M. Peek, M. Nakano, T. Uji, H. Ajioka, N. Matsuura, and H. Miyake. 2002. Identification of a signaling cascade for interleukin-8 production by Helicobacter pylori in human gastric epithelial cells. Biochem. Pharmacol. 64:21–30. Odenbreit, S., J. Puls, B. Sedlmaier, E. Gerland, W. Fischer, and R. Haas. 2000. Translocation of Helicobacter pylori CagA into gastric epithelial cells by type IV secretion. Science 287:1497–1500. Owen, R. J., T. M. Peters, R. Varea, E. L. Teare, and S. Saverymuttu. 2001. Molecular epidemiology of Helicobacter pylori in England: prevalence of cag pathogenicity island markers and IS605 presence in relation to patient age and severity of gastric disease. FEMS Immunol. Med. Microbiol. 30:65–71. Peters, T. M., R. J. Owen, E. Slater, R. Varea, E. L. Teare, and S. Saverymuttu. 2001. Genetic diversity in the Helicobacter pylori cag pathogenicity island and effect on expression of anti-CagA serum antibody in UK patients with dyspepsia. J. Clin. Pathol. 54:219–223. Rieder, G., R. A. Hatz, A. P. Moran, A. Walz, M. Stolte, and G. Enders. 1997. Role of adherence in interleukin-8 induction in Helicobacter pylori-associated gastritis. Infect. Immun. 65:3622–3630. Salama, N., K. Guillemin, T. K. McDaniel, G. Sherlock, L. Tompkins, and S. Falkow. 2000. A whole-genome microarray reveals genetic diversity among Helicobacter pylori strains. Proc. Natl. Acad. Sci. USA 97:14668–14673. Segal, E. D., J. Cha, J. Lo, S. Falkow, and L. S. Tompkins. 1999. Altered states: involvement of phosphorylated CagA in the induction of host cellular growth changes by Helicobacter pylori. Proc. Natl. Acad. Sci. USA 96:14559– 14564. Segal, E. D., C. Lange, A. Covacci, L. S. Tompkins, and S. Falkow. 1997. Induction of host signal transduction pathways by Helicobacter pylori. Proc. Natl. Acad. Sci. USA 94:7595–7599. Selbach, M., S. Moese, C. R. Hauck, T. F. Meyer, and S. Backert. 2002. Src is the kinase of the Helicobacter pylori CagA protein in vitro and in vivo. J. Biol. Chem. 277:6775–6778. Selbach, M., S. Moese, T. F. Meyer, and S. Backert. 2002. Functional analysis of the Helicobacter pylori cag pathogenicity island reveals both VirD4-CagAdependent and VirD4-CagA-independent mechanisms. Infect. Immun. 70: 665–671. Slater, E., R. J. Owen, M. Williams, and R. E. Pounder. 1999. Conservation of the cag pathogenicity island of Helicobacter pylori: associations with vac-

VOL. 71, 2003 uolating cytotoxin allele and IS605 diversity. Gastroenterology 117:1308– 1315. 47. Stein, M., R. Rappuoli, and A. Covacci. 2000. Tyrosine phosphorylation of the Helicobacter pylori CagA antigen after cag-driven host cell translocation. Proc. Natl. Acad. Sci. USA 97:1263–1268. 48. Suerbaum, S. 2000. Genetic variability within Helicobacter pylori. Int. J. Med. Microbiol. 290:175–181. 49. Tomb, J. F., O. White, A. R. Kerlavage, R. A. Clayton, G. G. Sutton, R. D. Fleischmann, K. A. Ketchum, H. P. Klenk, S. Gill, B. A. Dougherty, K. Nelson, J. Quackenbush, L. Zhou, E. F. Kirkness, S. Peterson, B. Loftus, D. Richardson, R. Dodson, H. G. Khalak, A. Glodek, K. McKenney, L. M. Fitzgerald, N. Lee, M. D. Adams, E. K. Hickey, D. E. Berg, J. M. Gocayne, T. R. Utterback, J. D. Peterson, J. M. Kelley, M. D. Cotton, J. M. Weidman,

Editor: V. J. DiRita

H. PYLORI cag PAI AND DISEASE OUTCOME

6581

C. Fujii, C. Bowman, L. Watthey, E. Wallin, W. S. Hayes, M. Borodovsky, P. D. Karp, H. O. Smith, C. M. Fraser, and J. C. Venter. 1997. The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 388: 539–547. 50. Tummuru, M. K., T. L. Cover, and M. J. Blaser. 1993. Cloning and expression of a high-molecular-mass major antigen of Helicobacter pylori: evidence of linkage to cytotoxin production. Infect. Immun. 61:1799–1809. 51. Xiang, Z., S. Censini, P. F. Bayeli, J. L. Telford, N. Figura, R. Rappuoli, and A. Covacci. 1995. 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. Infect. Immun. 63:94–98.