DNA sequence conservation and diversity in transposable element IS605 of Helicobacter pylori

BACKGROUND

IS605, a transposable element-like sequence identified in the virulence-associated cag region of Helicobacter pylori reference strain NCTC11638, is unusual in containing two oppositely-oriented open reading frames whose products are homologues of the single transposases of the unrelated elements, IS200 and IS1341.

METHODS

One hundred independent H. pylori isolates from different parts of the world were screened by PCR and dot blot hybridization to determine the presence of IS605. For some positive isolates, southern hybridizations and sequence analyses were done.

RESULTS

Of the 100 isolates, 31 were found to contain sequences related to each ORF with orientation and spacing matching those in canonical IS605-hybridizing sequences. No isolate containing just one ORF and not the other was found. The frequencies of IS605 carriage were independent of geographical origin (U.S. vs. non-U.S.), and of the probable virulence of the isolate (cag status, toxin production or vacA alleles, patient symptoms). Southern blot hybridization of six IS605-containing strains revealed one to nine IS605 copies per genome. Two types of DNA sequence diversity were found: first, a specific 100 bp deletion in two isolates; second, base substitution divergence of 0.4% to 7.5% in pairwise comparisons among the eight isolates characterized, a level of divergence similar to that seen in other H. pylori chromosomal genes.

CONCLUSIONS

Based on these findings, we speculate that IS605 is a relatively ancient component of the H. pylori gene pool that has proliferated in this species by horizontal gene transfer, homologous recombination, and transposition.

Activation of IL-8 gene expression by Helicobacter pylori is regulated by transcription factor nuclear factor-kappa B in gastric epithelial cells

In vivo, gastric infection with Helicobacter pylori leads to substantial production of the inflammatory cytokines IL-1, IL-6, TNF-alpha, and IL-8. H. pylori strains that contain the cag pathogenicity island (cag+) and are associated with ulceration and gastric carcinoma induce greater cytokine production than cag- strains. Expression of these cytokines is often regulated by the transcription factor complex, nuclear factor-kappa B (NF-kappa B) through kappa B-binding elements in the enhancer/promoter regions of their genes. We report that more virulent cag+ H. pylori strains induce increased NF-kappa B-DNA binding activity, which elevates IL-8 expression in AGS gastric epithelial cells. The cag+ H. pylori strains induce significant stimulation of IL-8 promoter-driven reporter activity, while cag- strains do not. Furthermore, mutation of specific genes within the cag island (picA1 and picB) ablates enhanced NF-kappa B activation and IL-8 transcription. Increased IL-8 expression is inhibited by mutation in either the NF-kappa B or NF-IL-6 binding element. The cag+ strains, compared with the cag- strains, induce enhanced nuclear localization of a RelA-containing NF-kappa B binding complex, but no increase in NF-IL-6 binding activity. These studies demonstrate that the ability of different types of H. pylori strains to activate NF-kappa B correlates with their ability to induce IL-8 transcription and indicate a mechanism for the heightened inflammatory response seen in subjects infected with cag+ H. pylori strains.

Characterization of Helicobacter pylori dapE and construction of a conditionally lethal dapE mutant

Helicobacter pylori colonizes the human gastric mucosa and causes gastritis, ulceration, or gastric cancer. A previously uncharacterized region of the H. pylori genome was identified and sequenced. This region includes a putative operon containing three open reading frames termed gidA (1,866 bp), dapE (1,167 bp), and orf2 (753 bp); the gidA and dapE products are highly homologous to other bacterial proteins. In E. coli, dapE encodes N-succinyl-L-diaminopimelic acid desuccinylase, which catalyzes the hydrolysis of N-succinyl-L-diaminopimelic acid to L-diaminopimelic acid (L-DAP) and succinate. When wild-type H. pylori strains were transformed to select for dapE mutagenesis, mutants were present when plates were supplemented with DAP but not with lysine; orf2 mutants were selected without DAP supplementation. Consistent with the finding that GidA is essential in Escherichia coli, we were unable to obtain a gidA mutant in H. pylori despite evidence that insertional mutagenesis had occurred. The positions of gidA, dapE, and orf2 suggest that they form an operon, which was supported by slot blot RNA hybridization and reverse transcriptase PCR studies. The data imply that the H. pylori dapE mutant may be useful as a conditionally lethal vaccine.

Role of vacuolating cytotoxin in gastritis due to Helicobacter pylori in gnotobiotic piglets

To investigate the role of the Helicobacter pylori cytotoxin in the pathogenesis of gastritis, gnotobiotic piglets were colonized with either toxigenic H. pylori or a nontoxigenic isogenic mutant. Only piglets given the toxigenic strain developed toxin-neutralizing antibodies (indicating that toxin is expressed in vivo), but there was no difference in bacterial colonization, epithelial vacuolation, or gastritis between the two groups of piglets.

Structure-function relationships among wild-type variants of Staphylococcus aureus beta-lactamase: importance of amino acids 128 and 216

beta-Lactamases inactivate penicillin and cephalosporin antibiotics by hydrolysis of the beta-lactam ring and are an important mechanism of resistance for many bacterial pathogens. Four wild-type variants of Staphylococcus aureus beta-lactamase, designated A, B, C, and D, have been identified. Although distinguishable kinetically, they differ in primary structure by only a few amino acids. Using the reported sequences of the A, C, and D enzymes along with crystallographic data about the structure of the type A enzyme to identify amino acid differences located close to the active site, we hypothesized that these differences might explain the kinetic heterogeneity of the wild-type beta-lactamases. To test this hypothesis, genes encoding the type A, C, and D beta-lactamases were modified by site-directed mutagenesis, yielding mutant enzymes with single amino acid substitutions. The substitution of asparagine for serine at residue 216 of type A beta-lactamase resulted in a kinetic profile indistinguishable from that of type C beta-lactamase, whereas the substitution of serine for asparagine at the same site in the type C enzyme produced a kinetic type A mutant. Similar bidirectional substitutions identified the threonine-to-alanine difference at residue 128 as being responsible for the kinetic differences between the type A and D enzymes. Neither residue 216 nor 128 has previously been shown to be kinetically important among serine-active-site beta-lactamases.

Effect of growth phase and acid shock on Helicobacter pylori cagA expression

Helicobacter pylori strains possessing cagA are associated with peptic ulceration. To understand the regulation of expression of cagA, picB, associated with interleukin-8 induction, and ureA, encoding the small urease subunit, we created gene fusions of cagA, ureA, and picB of strain 3401, using a promoterless reporter (xylE). Expression of XylE after growth in broth culture revealed that basal levels of expression of cagA and urea in H. pylori were substantially greater than for picB. For cagA expression in stationary-phase cells, brief exposure to acid pH caused a significant increase in xylE expression compared with neutral pH. In contrast, expression of xylE in urea or picB decreased after parallel exposure to acid pH (pH 7 > 6 > 5 > 4), regardless of the growth phase. Expression of the CagA protein varied with growth phase and pH exposure in parallel with the observed transcriptional variation. The concentration of CagA in a cell membrane-enriched fraction after growth at pH 6 was significantly higher than after growth at pH 5 or 7. We conclude that the promoterless reporter xylE is useful for studying the regulation of gene expression in H. pylori and that regulation of CagA production occurs mainly at the transcriptional level.

Effect of Helicobacter pylori on gastric epithelial cell migration and proliferation in vitro: role of VacA and CagA

Helicobacter pylori infection is associated with inflammation of the gastric mucosa and with gastric mucosal damage. In this study, we sought to test the hypothesis that two H. pylori virulence factors (VacA and CagA) impair gastric epithelial cell migration and proliferation, the main processes involved in gastric mucosal healing in vivo. Human gastric epithelial cells (MKN 28) were incubated with undialyzed or dialyzed broth culture filtrates from wild-type H. pylori strains or isogenic mutants defective in production of VacA, CagA, or both products. We found that (i) VacA specifically inhibited cell proliferation without affecting cell migration, (ii) CagA exerted no effect on either cell migration or proliferation, and (iii) undialyzed H. pylori broth culture filtrates inhibited both cell migration and proliferation through a VacA- and CagA-independent mechanism. These findings demonstrate that, in addition to damaging the gastric mucosa, H. pylori products may also impair physiological processes required for mucosal repair.

Helicobacter pylori picB, a homologue of the Bordetella pertussis toxin secretion protein, is required for induction of IL-8 in gastric epithelial cells

Approximately 60% of Helicobacter pylori strains are cagA+ and this genotype is more frequently associated with duodenal ulcer disease. Although most wild-type cagA+ strains are both cytotoxigenic and induce enhanced Interleukin-8 (IL-8) secretion in gastric epithelial cells, isogenic cagA- mutants retain full activity in these assays; thus, cagA appears to be a marker of enhanced virulence. Delineation of the nucleotide sequence of a 4 kb region upstream of cagA allowed the identification of 966 bp (picA) and 2655 bp (picB) open reading frames encoding 36 kDa and 101 kDa polypeptides, respectively. picA and picB constitute an operon in opposite orientation to cagA. The deduced picB product showed significant homology (26% identity and 50% similarity) with the Bordetella pertussis toxin secretion protein (PtlC). Of 55 H. pylori clinical isolates, the picA and picB segment was conserved exclusively in cagA+ strains and present in all isolates from patients with duodenal ulceration, versus 59% of isolates from patients with gastritis alone (P = 0.01). Using gene-replacement techniques, we constructed picA and picB mutant H. pylori strains and demonstrated that the picB gene product is involved in the induction of IL-8 expression in gastric epithelial cells. Further, Northern blot hybridization and RT-PCR data showed that picA and picB are co-transcribed and an insertional mutation in picA ablates picB expression. These studies indicate a role of picA and picB in the induction of an inflammatory response in gastric epithelial cells either directly or by enabling secretion of an unidentified product, and suggest a mechanism for the overrepresentation of strains possessing these genes in patients with peptic ulceration.

Role of the Helicobacter pylori virulence factors vacuolating cytotoxin, CagA, and urease in a mouse model of disease

The pathogenic role of Helicobacter pylori virulence factors has been studied with a mouse model of gastric disease. BALB/c mice were treated orally with different amounts of sonic extracts of cytotoxic H. pylori strains (NCTC 11637, 60190, 84-183, and 87A300 [CagA+/Tox+]). The pathological effects on histological sections of gastric mucosae were assessed and were compared with the effects of treatments with extracts from noncytotoxic strains (G21 and G50 [CagA-/Tox-]) and from strains that express either CagA alone (D931 [CagA+/Tox-]) or the cytotoxin alone (G104 [CagA-/Tox+]). The treatment with extracts from cytotoxic strains induced various epithelial lesions (vacuolation, erosions, and ulcerations), recruitment of inflammatory cells in the lamina propria, and a marked reduction of the mucin layer. Extracts of noncytotoxic strains induced mucin depletion but no other significant pathology. Crude extracts of strain D931, expressing CagA alone, caused only mild infiltration of inflammatory cells, whereas extracts of strain G104, expressing cytotoxin alone, induced extensive epithelial damage but little inflammatory reaction. Loss of the mucin layer was not associated with a cytotoxic phenotype, since this loss was observed in mice treated with crude extracts of all strains. The pathogenic roles of CagA, cytotoxin, and urease were further assessed by using extracts of mutant strains of H. pylori defective in the expression of each of these virulence factors. The results obtained suggest that (i) urease activity does not play a significant role in inducing the observed gastric damage, (ii) cytotoxin has an important role in the induction of gastric epithelial cell lesions but not in eliciting inflammation, and (iii) other components present in strains which carry the cagA gene, but distinct from CagA itself, are involved in eliciting the inflammatory response.

Mosaicism in vacuolating cytotoxin alleles of Helicobacter pylori. Association of specific vacA types with cytotoxin production and peptic ulceration

Approximately 50% of Helicobacter pylori strains produce a cytotoxin, encoded by vacA, that induces vacuolation of eukaryotic cells. Analysis of a clinically isolated tox- strain (Tx30a) indicated secretion of a 93-kDa product from a 3933-base pair vacA open reading frame. Characterization of 59 different H. pylori isolates indicated the existence of three different families of vacA signal sequences (s1a, s1b, and s2) and two different families of middle-region alleles (m1 and m2). All possible combinations of these vacA regions were identified, with the exception of s2/m1 (p < 0.001); this mosaic organization implies that recombination has occurred in vivo between vacA alleles. Type s1/m1 strains produced a higher level of cytotoxin activity in vitro than type s1/m2 strains; none of 19 type s2/m2 strains produced detectable cytotoxin activity. The presence of cagA (cytotoxin-associated gene A) was closely associated with the presence of vacA signal sequence type s1 (p < 0.001). Among patients with past or present peptic ulceration, 21 (91%) of 23 harbored type s1 strains compared with 16 (48%) of 33 patients without peptic ulcers; only 2 (10%) of 19 subjects harboring type s2 strains had past or present peptic ulcers (p < 0.005). Thus, specific vacA genotypes of H. pylori strains are associated with the level of in vitro cytotoxin activity as well as clinical consequences.

Segmental conservation of sapA sequences in type B Campylobacter fetus cells

Campylobacter fetus cells may exist as either of two defined serogroups (type A or B) based on their lipopolysaccharide (LPS) composition. Wild-type strains contain surface array proteins (S-layer proteins) that have partial antigenic cross-reactivity but bind exclusively to LPS from homologous (type A or B) cells. Type A cells possess 8 homologs of sapA, which encodes a 97-kDa S-layer protein; the gene products of these homologs have a conserved N terminus of 184 amino acids. To further explore the structural relationships between the C. fetus S-layer proteins and their encoding genes, we sought to clone and express an S-layer protein from type B strain 84-91. The cloned type B gene (sapB) was similar in structure to the previously cloned type A gene (sapA) and encoded a full-length 936-amino acid (97-kDa) S-layer protein. Sequence analysis of sapB indicated that the conserved N-terminal encoding region in sapA was absent but that the remainder of the ORF (encoding 751 amino acids) was identical to that of sapA in spite of the nonconserved nature of this region among sapA homologs. Noncoding sequences both 300 base pairs 5' and 1000 base pairs 3' to the sapB and sapA ORFs, including the sapA promoter and transcriptional terminator sequences, were essentially identical. Southern analyses revealed that the sapB N-terminal encoding region was conserved in multiple copies in type B strains but was absent in type A strains. Recombinant sapA and sapB products bound to a substantially greater degree to cells of the homologous LPS type compared with the heterologous LPS type, indicating that the conserved sapA- and sapB-encoded N termini are critical for LPS binding specificity. The parallel genetic organization and identity at the nucleotide level in both coding and noncoding regions for sap homologs in types A and B cells indicates the necessity of both homolog conservation and high fidelity DNA replication in the biology of sap diversity.

Serologic detection of infection with cagA+ Helicobacter pylori strains

Approximately 60% of Helicobacter pylori isolates possess the cagA gene and express its 120- to 140-kDa product (CagA). In this study, the cagA gene was detected in H. pylori isolates from 26 (81.3%) of 32 patients with duodenal ulcers (DU), 17 (68.0%) of 25 patients with gastric ulcers, and 23 (59.0%) of 39 patients with nonulcer dyspepsia (NUD). By Western blotting (immunoblotting) with antiserum to CagA, in vitro CagA expression was demonstrated for 95.5% of cagA+ strains compared with 0% of strains lacking cagA. Sera from patients infected with cagA+ strains (n = 66) reacted with recombinant CagA in an enzyme-linked immunosorbent assay to a significantly greater extent than either sera from patients infected with strains lacking cagA (n = 30) or sera from uninfected persons (n = 25) (P < 0.001). A strain lacking cagA was isolated from eight patients who had serum immunoglobulin G antibodies to CagA, which suggests that these patients were infected with multiple strains. Serum immunoglobulin G antibodies to CagA were present in 87.5, 76.0, and 56.4% of patients with DU, gastric ulcers, and NUD, respectively (odds ratio, 5.41; 95% confidence interval, 1.44 to 24.72; P = 0.004 [DU versus NUD]). These data demonstrate an association between infection with cagA+ H. pylori and the presence of duodenal ulceration and indicate that serologic testing is a sensitive method for detecting infection with cagA+ strains.

Interleukin-8 response of gastric epithelial cell lines to Helicobacter pylori stimulation in vitro

Gastric infection with Helicobacter pylori activates a mucosal inflammatory response by mononuclear cells and neutrophils that includes expression of cytokines interleukin-1 beta (IL-1 beta), IL-6, tumor necrosis factor alpha, and IL-8. In this study, we analyzed the IL-8 response of human gastric cancer cell lines (Kato III, AGS, and MKN28) to H. pylori infection in vitro. IL-8 mRNA expression was detected by reverse transcription-PCR amplification of RNA extracted from epithelial cells after incubation with different H. pylori wild-type and mutant strains, and IL-8 secretion was measured by an enzyme-linked immunosorbent assay. Exposure to viable H. pylori induced IL-8 mRNA and protein synthesis in all three gastric cell lines but not in nongastric epithelial cell lines. Heat-killed H. pylori and a crude cytotoxin preparation did not induce significant IL-8 secretion. IL-8 mRNA peaked between 2 and 4 h postinfection, and IL-8 protein production was maximal 24 h postinfection. Exposure of gastric carcinoma cells to other gastrointestinal bacteria, such as Pseudomonas aeruginosa, Campylobacter jejuni, and Escherichia coli, but not Campylobacter fetus, induced IL-8 synthesis. Wild-type strains that expressed the vacuolating cytotoxin (Tox+) and a cytotoxin-associated gene (cagA) product (CagA+) induced significantly more IL-8 than did CagA- Tox- strains. However, there was no decrease in IL-8 induction by isogenic mutants of CagA-, Tox-, or Cag- Tox- strains or by a mutant lacking the urease subunits. These results indicate that exposure to H. pylori and other gram-negative organisms that do not colonize the gastric mucosa induces IL-8 production by gastric carcinoma cells in vitro. Although the CagA+ Tox+ phenotype of H. pylori is associated with enhanced IL-8 production by gastric cell lines, other bacterial constituents are clearly essential.

A lipopolysaccharide-binding domain of the Campylobacter fetus S-layer protein resides within the conserved N terminus of a family of silent and divergent homologs

Campylobacter fetus cells can produce multiple S-layer proteins ranging from 97 to 149 kDa, with a single form predominating in cultured cells. We have cloned, sequenced, and expressed in Escherichia coli a sapA homolog, sapA2, which encodes a full-length 1,109-amino-acid (112-kDa) S-layer protein. Comparison with the two previously cloned sapA homologs has demonstrated two regions of identity, approximately 70 bp before the open reading frame (ORF) and proceeding 550 bp into the ORF and immediately downstream of the ORF. The entire genome contains eight copies of each of these conserved regions. Southern analyses has demonstrated that sapA2 existed as a complete copy within the genome in all strains examined, although Northern (RNA) analysis has demonstrated that sapA2 was not expressed in the C. fetus strain from which it was cloned. Further Southern analyses revealed increasing sapA diversity as probes increasingly 3' within the ORF were used. Pulsed-field gel electrophoresis and then Southern blotting with the conserved N-terminal region of the sapA homologs as a probe showed that these genes were tightly clustered on the chromosome. Deletion mutagenesis revealed that the S-layer protein bound serospecifically to the C. fetus lipopolysaccharide via its conserved N-terminal region. These data indicated that the S-layer proteins shared functional activity in the conserved N terminus but diverged in a semiconservative manner for the remainder of the molecule. Variation in S-layer protein expression may involve rearrangement of complete gene copies from a single large locus containing multiple sapA homologs.

High-frequency S-layer protein variation in Campylobacter fetus revealed by sapA mutagenesis

Campylobacter fetus utilizes paracrystalline surface (S-) layer proteins that confer complement resistance and that undergo antigenic variation to facilitate persistent mucosal colonization in ungulates. C. fetus possesses multiple homologues of sapA, each of which encode full-length S-layer proteins. Disruption of sapA by a gene targeting method (insertion of kanamycin (km) resistance) caused the loss of C. fetus cells bearing full-length S-layer proteins and their replacement by cells bearing a 50 kDa truncated protein that was not exported to the cell surface. After incubation of the mutants with serum, the survival rate was approximately 2 x 10(-2). Immunoblots of survivors showed that phenotypic reversion involving high-level production of full-length (98, 127 or 149 kDa) S-layer proteins had occurred. Revertants were serum resistant but caused approximately 10-fold less bacteraemia in orally challenged mice than did the wild-type strain. Southern hybridizations of the revertants showed rearrangement of sapA homologues and retention of the km marker. These results indicate that there exists high-frequency generation of C. fetus sapA antigenic variants, and that intracellular mechanisms acting at the level of DNA reciprocal recombination play key roles in this phenomenon.

Mutation of the cytotoxin-associated cagA gene does not affect the vacuolating cytotoxin activity of Helicobacter pylori

Helicobacter pylori now is recognized as an etiological agent in chronic superficial gastritis and peptic ulcer disease. Although only about 60% of H. pylori isolates produce an immunodominant 128-kDa antigen (CagA; cytotoxin-associated gene product), virtually all H. pylori-infected patients with duodenal ulceration develop a serologic response to the 128-kDa protein, which suggests an association of this gene with ulceration. The cloned cagA gene from H. pylori 84-183 was disrupted by insertion of a kanamycin resistance gene, and this inactivated cagA construct was introduced into H. pylori 84-183 by electrotransformation. Southern hybridization of kanamycin-resistant H. pylori transformants demonstrated that the wild-type cagA gene had been disrupted by insertion of the kanamycin cassette, and immunoblot analysis showed that the mutant strains no longer produced the 128-kDa CagA protein. Similar results were obtained when the cagA mutation was introduced by natural transformation into H. pylori 60190, a high-level toxin-producing strain. The cagA-negative H. pylori strains showed cytotoxin, urease, and phospholipase C activities, C3 binding and adherence similar to those of the isogenic wild-type strains. These findings demonstrate that the cagA gene product does not affect the vacuolating cytotoxin activity of H. pylori.

Divergence of genetic sequences for the vacuolating cytotoxin among Helicobacter pylori strains

Approximately 50% of Helicobacter pylori isolates produce a cytotoxin in vitro that induces vacuolation of eukaryotic cells. Screening a lambda ZapII library of H. pylori 60190 chromosomal fragments permitted the identification of a 3864-base pair (bp) open reading frame (vacA) that encoded the vacuolating cytotoxin, and a > or = 567-bp upstream gene that was homologous to Escherichia coli cysteinyl-tRNA synthetase. The sequence data suggest that a 33-amino-acid leader sequence and a C-terminal peptide are cleaved from a 139-kDa protoxin to yield the mature 87-kDa cytotoxin. The vacA gene product contains a C-terminal motif that is present in several other bacterial proteins that undergo C-terminal cleavage, including IgA proteases of Haemophilus influenzae and Neisseria gonorrhoeae. Isogenic H. pylori mutants with insertional mutation of the vacA gene lacked vacuolating cytotoxin activity and failed to produce the 87-kDa protein. Southern analysis of naturally occurring tox-H. pylori strains with vacA probes indicated the presence of hybridizing bands, but both Southern analysis and polymerase chain reaction studies suggested that the vacA sequences of tox- strains differed from those of tox+ strains. Sequence analysis of a 1541-bp region of polymerase chain reaction-amplified vacA from tox- strain 87-203 indicated 64.8% amino acid identity with the corresponding region from tox+ strain 60190. Thus, sequence divergence in vacA genes may explain the lack of functionally active cytotoxin production by some H. pylori isolates.

Rearrangement of sapA homologs with conserved and variable regions in Campylobacter fetus

The Campylobacter fetus surface-layer (S-layer) proteins mediate both complement resistance and antigenic variation in mammalian hosts. Wild-type strain 23D possesses the sapA gene, which encodes a 97-kDa S-layer protein, and several sapA homologs are present in both wild-type and mutant strains. Here we report that a cloned silent gene (sapA1) in C. fetus can express a functional full-length S-layer protein in Escherichia coli. Analysis of sapA and sapA1 and partial analysis of sapA2 indicate that a block of approximately 600 bp beginning upstream and continuing into the open reading frames is completely conserved, and then the sequences diverge completely, but immediately downstream of each gene is another conserved 50-bp sequence. Conservation of sapA1 among strains, the presence of a putative Chi (RecBCD recognition) site upstream of sapA, sapA1, and sapA2, and the sequence identities of the sapA genes suggest a system for homologous recombination. Comparison of the wild-type strain (23D) with a phenotypic variant (23D-11) indicates that variation is associated with removal of the divergent region of sapA from the expression locus and exchange with a corresponding region from a sapA homolog. We propose that site-specific reciprocal recombination between sapA homologs leads to expression of divergent S-layer proteins as one of the mechanisms that C. fetus uses for antigenic variation.

Cloning and expression of a high-molecular-mass major antigen of Helicobacter pylori: evidence of linkage to cytotoxin production

A high-molecular-mass (120- to 128-kDa) Helicobacter pylori antigen has been associated with peptic ulcer disease. We created a bank of 40,000 random chromosomal fragments of H. pylori 84-183 by using lambda ZapII. Screening of this bank in Escherichia coli XL1-Blue with absorbed serum from an H. pylori-infected person permitted the isolation and purification of a clone with a 3.5-kb insert. Subcloning of this insert (pMC3) permitted the expression of a recombinant H. pylori protein that had a mass of approximately 96 kDa and that was recognized by the human serum. Sera that were obtained from H. pylori-infected persons and that recognized the native 120- to 128-kDa H. pylori antigen recognized the recombinant 96-kDa pMC3 protein to a significantly greater extent than did sera that did not recognize the native H. pylori antigen. All 19 H. pylori isolates producing the 120- to 128-kDa antigen hybridized with pMC3; none of 13 nonproducers did so (P < 0.001). Because all 15 isolates producing the vacuolating cytotoxin hybridized with pMC3, we called the gene cagA (cytotoxin-associated gene). Sequence analysis of pMC3 identified an open reading frame of 859 amino acids, without a termination codon. Parallel screening of a lambda gt11 library with human serum revealed positive plaques with identical 0.6-kb inserts and sequences matching the sequence of the downstream region of pMC3. To clone the full-length gene, we used the 0.6-kb fragment as a probe and isolated a clone with a 2.7-kb insert from the lambda ZapII genomic library. Nucleotide sequencing of this insert (pYB 2) revealed a 785-bp sequence that overlapped the downstream region of pMC3. Translation of the complete nucleotide sequence of cagA revealed an open reading frame of 1,181 amino acids yielding a protein of 131,517 daltons. There was no significant homology with any previously reported protein sequence. These findings indicate the cloning and characterization of a high-molecular-mass H. pylori antigen potentially associated with virulence and with cytotoxin production.

Characterization of the Campylobacter fetus sapA promoter: evidence that the sapA promoter is deleted in spontaneous mutant strains

Wild-type Campylobacter fetus cells possess S-layer proteins (S+ phenotype), whereas after laboratory passage, spontaneous stable mutants that do not express these proteins (S- phenotype) arise. To determine the molecular mechanisms by which C. fetus changes to the S- phenotype, we studied wild-type strain 23D, from which the sapA gene encoding the 97-kDa S-layer protein has been cloned, and strain 23B, a spontaneous S- mutant. We compared these strains with another pair of strains, LP (S+) and HP (S-). Southern analysis with the cloned sapA gene as a probe indicated that both pairs of strains have multiple sapA homologs. Using gene disruption and replacement techniques, we constructed an isogenic strain of 23D that differed only in sapA expression (strain 23D:401:1). A 6.0-kb HindIII fragment from 23D:401:1 containing 3.4 kb of sapA upstream region then was cloned into pBluescript to produce pBG101. Nucleotide sequence analysis of sapA upstream region revealed a consensus promoter at -121 bp from the translational start site. Primer extension analysis placed a single in vivo transcription initiation site at the -114-bp position of sapA. A DNA probe derived from the sapA promoter region hybridized to a 5.5-kb HindIII fragment of chromosomal DNA from strain 23D but not to DNA from strain 23B. Northern RNA blot analysis showed no sapA mRNA in strain 23B. These data indicate that the lack of S-layer protein expression in spontaneous mutant strains is caused by the deletion of promoter sequences.

The in vitro conversion of chorismate to isochorismate catalyzed by the Escherichia coli entC gene product. Evidence that EntA does not contribute to isochorismate synthase activity

The entC and entA genes, coding for the enzymes isochorismate synthase and 2,3-dihydro-2,3-dihydroxybenzoate dehydrogenase, respectively, were subcloned behind the T7 promoter in the expression plasmid pGEM3Z. Their protein products were overproduced and partially purified for in vitro analysis of the conversion of chorismate to isochorismate. Whereas previous genetic experiments suggested that the EntA enzyme has a role in this conversion, this study clearly indicates that EntC alone catalyzes the reaction. Addition of EntA had no effect on isochorismate synthase activity. As a result, the mutation (previously designated entC401) in strain AN191 was characterized by nucleotide sequence analysis. The lesion is a single base substitution in the entA gene, resulting in a glutamic acid-for-glycine substitution at the penultimate amino acid (residue 247) of the EntA enzyme. The mutant protein was partially purified and shown to be devoid of 2,3-dihydro-2,3-dihydroxybenzoate dehydrogenase activity, whereas the entC gene product from strain AN191 exhibited normal isochorismate synthase function. These results conflict with the earlier characterization of the entC401 mutation in a different genetic background. The data presented herein establish that the EntA protein does not contribute to isochorismate synthase activity and that the mutant strain that led to this suggestion harbors a defective allele of entA rather than entC.