Helicobacter pylori (Hp) carries a type IV secretion system encoded by the cag pathogenicity island (cag-PAI), which is used to: (i) translocate the bacterial effector protein CagA into different types of eukaryotic cells; and (ii) induce the synthesis and secretion of chemokines, such as interleukin-8 (IL-8). The cag-PAI in Hp 26695 consists of 27 putative genes, six of which were identified as homologues to the basic type IV secretion system represented by the Agrobacterium tumefaciens virB operon. To define the role and contribution of each of the 27 genes, we applied a precise deletion/insertion mutagenesis procedure to knock out each individual gene without causing polar effects on the expression of downstream genes. Seventeen out of 27 genes were found to be absolutely essential for translocation of CagA into host cells and 14 out of 27 for the ability of Hp fully to induce transcription of IL-8. The products of hp0524 (virD4 homologue), hp0526 and hp0540 are absolutely essential for the translocation of CagA, but not for the induction of IL-8. In contrast, the products of hp0520, hp0521, hp0534, hp0535, hp0536 and hp0543 are not necessary for either translocation of CagA or for IL-8 induction. Our data argue against a translocated IL-8-inducing effector protein encoded by the cag-PAI. We isolated a variant of Hp 26695, which spontaneously switched off its capacity for IL-8 induction and translocation of CagA, but retained the complete cag-PAI. We identified a point mutation in gene hp0532, causing a premature translational stop in the corresponding polypeptide chain, providing a putative explanation for the defect in the type IV secretion system of the spontaneous mutant.
The Gram-negative bacterial pathogen Helicobacter pylori (Hp) is known as a major causative agent for the induction of medically significant gastroduodenal diseases, including chronic gastritis, peptic ulceration, mucosa-associated lymphoid tissue (MALT) lymphoma and adenocarcinoma (Lee et al., 1993). Specific bacterial factors produced by type I Hp strains as well as a specific host susceptibility have been implicated in the outcome of severe forms of disease (El-Omar et al., 2000).
Type I strains of Hp are often associated with a severe disease outcome of Hp infection in patients. In the Hp infection model of Mongolian gerbils, Ogura et al. (2000) demonstrated recently that type I Hp strains with an intact cag pathogenicity island (cag-PAI) induced strong inflammation and ulceration in the stomach, whereas an isogenic mutant with a defect in the type IV secretion system was unable to elicit such a response. Type I strains carry the cag-PAI, a 37 kb sequence consisting of 27 putative genes. The cagA gene (encoding the cytotoxin-associated antigen A) is a marker for this locus, which is absent in type II strains (Censini et al., 1996). It has been shown recently that the genes of the cag-PAI actually encode a functional type IV secretion system, and that Hp delivers the CagA protein into cultured gastric epithelial cells (Segal et al., 1999; Asahi et al., 2000; Backert et al., 2000; Odenbreit et al., 2000; Stein et al., 2000). CagA is immediately phosphorylated at tyrosine residue(s) by an as yet unidentified eukaryotic kinase, and the phosphorylated protein (CagAP-tyr) is recruited to the membrane close to the location at which the bacteria attach (Segal et al., 1999; Odenbreit et al., 2000; Stein et al., 2000).
A direct consequence of CagA translocation is the dephosphorylation of a complex of tyrosine-phosphorylated cellular proteins in the size range 120–130 kDa (p120–130) and 80 kDa (p80) (Odenbreit et al., 2000). CagA is not only translocated into epithelial cells, but also into human granulocytes and macrophage cell lines. Interestingly, CagAP-tyr is processed in phagocytic cells, giving rise to a tyrosine-phosphorylated 35–45 kDa C-terminal fragment of CagA (Odenbreit et al., 2001).
Furthermore, attachment of type I Hp strains to gastric epithelial cells induces the production and secretion of chemokines, such as interleukin-8 (IL-8) and epithelial cell-derived neutrophil-activating protein 78 (ENA-78) (Rieder et al., 2001), but also granulocyte–monocyte colony-stimulating factor (GM-CSF) or tumour necrosis factor α (TNFα) in gastric epithelial cells (Foryst-Ludwig and Naumann, 2000). Chemokine induction requires direct contact of the bacteria with the epithelial cells (Rieder et al., 1997). The mechanism of this chemokine induction is only partially understood, but several genes of the cag-PAI or components of the type IV secretion apparatus seem to be involved in this process by activation of AP-1 (Naumann et al., 1999) and NFκB (Censini et al., 1996; Keates et al., 1997; Münzenmaier et al., 1997; Segal et al., 1997; Li et al., 1999; Covacci and Rappuoli, 2000), leading to the induction of the corresponding cytokine and chemokine genes. It is unknown whether an additional protein needs to be translocated into epithelial cells, or whether the binding of the type IV secretion apparatus to a cellular receptor triggers the signalling cascade leading to IL-8 induction. Here, we describe the controlled gene replacement/insertion mutagenesis of each individual gene of the cag-PAI and its consequences for both translocation of CagA and induction of IL-8.
Precise gene replacement mutagenesis of the complete Hp cag-PAI
The cag-PAI of Hp 26695 carries a total of 27 open reading frames (ORFs) encoding putative proteins that are probably involved in the function of the type IV secretion system or encode further effector proteins apart from CagA translocated by the system (Fig. 1A) (we do not consider the short reading frame hp0533 as a gene, and hp0548 is an additional element not present in all cag-PAIs; thus, we examine 27 ORFs in this study). We intend to dissect the system into functional components involved in the following putative steps: (i) regulators of gene expression of the cag-PAI; (ii) assembly of the secretion apparatus; (iii) energy delivery for the assembly; and (iv) translocation of CagA and possible other effector proteins. This genetic analysis involved systematically performing a complete mutagenesis of the cag-PAI, ORF by ORF. Regions of 3–6 kb of the PAI of Hp 26695 were amplified by polymerase chain reaction (PCR) and cloned into plasmid pBA (Halter et al., 1984), resulting in 10 plasmid clones used as a basis for gene replacement mutagenesis (cag1–cag9) (Fig. 1B). As outlined in Fig. 1C, PCR primers were designed to amplify the flanking sequences of each ORF and the cloning vector. The deleted gene was replaced by joining the PCR fragment to a cat resistance gene cassette (in sense orientation relative to the gene) without a transcriptional terminator sequence to avoid termination of polycistronic transcripts. The last gene of a putative operon or singular genes without an operon structure were mutated by TnMax5 or TnMax5gfp transposon insertion mutagenesis (see Experimental procedures). This resulted in a set of isogenic mutants covering the complete cag-PAI. Owing to the strategy used, there should be no polar effects on the expression of downstream genes in the particular mutants, as demonstrated in the next section.
Avoidance of polar effects with different mutagenesis procedures
Next, we were interested to test whether the precise replacement by the terminatorless cat cassette of a distinct gene in an operon of the cag-PAI would exert polar effects on transcription and translation of downstream genes. Therefore, a set of selected cag-PAI genes was chosen to generate and purify fusion proteins in an Escherichia coli expression system (Pohlner et al., 1993) for the generation of defined antisera (see Experimental procedures). Interestingly, a pair of mutants generated in hp0541 by TnMax transposon insertion mutagenesis and by the gene replacement procedure (Fig. 1A) had completely different consequences for the expression of a downstream gene (Fig. 2). The TnMax5gfp insertion mutagenesis in hp0541 resulted in a complete loss of expression of the downstream gene hp0539, as demonstrated in the immunoblot, whereas the gene replacement mutagenesis caused only a slight reduction in the amount of HP0539 produced (Fig. 2). This indicates that the gene replacement mutagenesis procedure affects the mutated gene only, which allows the assigning of a certain phenotype of a defined cag-PAI mutant to the corresponding gene that has been knocked out or its gene product. Thus, we have generated a set of knock-out mutants in the cag-PAI, which should allow us to define the role of each gene and its putative gene product according to the capacity of the corresponding strain for IL-8 induction and translocation of CagA. Those proteins without any effect on IL-8 induction and/or CagA translocation are candidates for further putative effector proteins translocated into the host cell other than CagA.
Identification of cag-PAI genes essential for delivery of CagA into the epithelial cell
We then tested the capacity of each mutant strain to translocate and phosphorylate CagA into the gastric epithelial cell line AGS and into the phagocytic cell line J774A.1. Therefore, the cell lines were infected with the wild-type (wt) Hp strain, or an individual isogenic mutant strain, and the tyrosine phosphorylation of CagA was detected in a phosphotyrosine immunoblot (Odenbreit et al., 2000). Seventeen of the 26 putative genes were found to be essential for the delivery of CagA into epithelial cells (Fig. 3). Deletion of hp0538, hp0542 and hp0545 caused only a reduction in the degree of CagA phosphorylation. Other genes, such as hp0520, hp0521, hp0534–hp0536 and hp0543, did not affect CagA translocation and tyrosine phosphorylation at all, suggesting that the corresponding gene products are not involved in the assembly of an intact secretion apparatus for the delivery of CagA. Thus, the complete set of genes in the cag-PAI can be assigned into three groups: those that are absolutely essential for CagA translocation, those that have only a partial contribution and those that are not involved at all.
Components of the cag-PAI essential for IL-8 induction in epithelial cells
The type IV secretion system is necessary for the induction of IL-8 in epithelial cells by Hp. The question was now whether the same components are needed as for the translocation of CagA, or whether a different set of gene products would be necessary. The supernatant of AGS cells infected with the isogenic mutant strains or the wt Hp 26695 (Fig. 3) was used to quantify the secretion of IL-8 using a sandwich enzyme-linked immunosorbent assay (ELISA) with two monoclonal anti-IL-8 antibodies (Schulte and Autenrieth, 1998). Because supernatants from the same infections were used as for the determination of phosphorylated CagA, the quantity of IL-8 induced by a particular mutant strain was directly comparable with the amount of tyrosine-phosphorylated CagA. As expected, most of the genes of the cag-PAI essential for CagA translocation were also absolutely necessary for the induction of IL-8. Rather unexpected, however, was that mutations in hp0520, hp0521, hp0524, hp0534–hp0536 and hp0540 did not have an effect on the ability of the corresponding mutants to induce IL-8, indicating that the corresponding proteins are not involved in the process of IL-8 induction in epithelial cells (Fig. 3). A further set of mutants, including genes hp0526, hp0538, hp0542, hp0543 and hp0545, showed an intermediate and rather variable effect on IL-8 induction rates. Mutations in the remaining genes resulted in only background levels of IL-8, demonstrating that the corresponding proteins are absolutely essential for the process of IL-8 induction (Fig. 3).
Influence of defined cag-PAI mutations on the expression or stability of other components of the type IV secretion system
To estimate whether individual mutations in a cag-PAI gene might have an indirect effect on the stability of other proteins of the type IV secretion system, we tested the stability of one protein of the type IV Cag secretion apparatus in the context of all defined knock-out mutants. Antiserum AK270, which specifically recognizes the product of gene hp0532, was chosen to assess the stability of the corresponding protein in the background of each individual cag-PAI mutant (Fig. 4). The gene product of hp0532, a putative homologue of the outer membrane lipoprotein VirB7 of Agrobacterium tumefaciens (Covacci et al., 1999), is produced under in vitro growth conditions of the bacterium (Fig. 4). In addition, the immunoblot shows that most knock-out mutations in genes of the cag-PAI did not have a negative effect on the expression or stability of Hp0532. Interestingly, however, a mutation in hp0537 causes the complete loss of HP0532, which indicates a causal relationship. HP0532 and HP0537 carry a signal sequence, indicating that they are both secreted proteins, but their precise location has not been determined.
Analysis of spontaneous Hp mutants deficient in type IV secretion and IL-8 induction
The complete mutagenesis approach identified genes that, upon inactivation, revealed different effects in terms of CagA translocation and IL-8 induction. Of particular interest, we also observed that natural variants of Hp strain 26695 existed, which were unable to translocate CagA and showed either a completely abolished (26695pass-1) or a strongly reduced (26695pass-2) capacity to induce IL-8 (Fig. 5A and B). In contrast, the original 26695 strain (26695ori), obtained from the ATCC strain collection (ATCC 700392), was fully competent for translocation of CagA and induction of IL-8 (Figs 3 and 5A and B). Interestingly, the VirB7 homologue HP0532 was absent in Hp 26695pass-1 compared with 26695pass-2 and 26695ori(Fig. 5A). PCR amplification verified the presence of the hp0532 gene with the correct size in all three Hp chromosomes (data not shown). DNA sequencing revealed a point mutation in the hp0532 gene of 26695pass-1 at position 672 of the gene, resulting in a premature termination of the protein (Fig. 5C). As the deletion of hp0532 also caused abrogation of CagA translocation and IL-8 induction, this mutation could explain the defect in Hp 26695pass-1. The mutation in 26695pass-2 has yet to be identified.
The cag-PAI of Hp encodes a type IV secretion system that is involved in the translocation of the bacterial protein CagA into eukaryotic cells and in the induction of IL-8 secretion in gastric epithelial cells. Here, we performed a systematic mutagenesis approach on the complete cag-PAI. We replaced individual genes with a terminatorless cat cassette or inactivated genes by TnMax5 transposon insertion. The procedure was performed in such a way that no or only minor polar effects were exerted on the downstream genes. Each individual mutant that lacks a single protein component of the putative type IV secretion apparatus was tested for its capacity to: (i) translocate the effector protein CagA into epithelial cells; and (ii) induce the production of IL-8 in gastric epithelial cells.
As demonstrated by transposon insertion mutagenesis versus gene replacement mutagenesis, we often face the problem of generating downstream polar effects using the first method (Fig. 2). We have minimized this problem by using gene replacement mutagenesis. As demonstrated for gene hp0541, a TnMax transposon insertion caused a complete loss of the product of the downstream gene hp0539. The replacement mutagenesis resulted in a slight reduction in the amount of the corresponding protein, but did not abolish its expression. This reduction seems not to be critical for the biogenesis of the secretion system, as the hp0540 mutant does not have a negative effect on IL-8 induction, whereas the hp0539 mutant is strongly reduced in IL-8 induction. A possible polar effect of the hp0540 mutant on HP0539 would result in a decrease in IL-8 induction in the hp0540 mutant, which is not the case.
It appears to be very important to avoid such polar effects during cag-PAI mutagenesis as, otherwise, a certain defect cannot be attributed to a single protein in the secretion system. Similar polar effects might also explain the differences seen between our data concerning translocation of CagA and induction of IL-8 and those generated by transposon insertion mutagenesis with mini-Tn3-Km (Suerbaum et al., 1993), as summarized in a recent review by Covacci and Rappuoli (2000). We generally find more mutants that have only partial effects (hp0545, hp0543, hp0542) or no effect at all (hp0540, hp0534, hp0521) on IL-8 secretion or CagA translocation, whereas the corresponding transposon mutants were classified as negative. With the exception of hp0534, we can explain those mutants as having a polar effect on essential genes located downstream in the same operon (Fig. 3).
A further question we addressed was whether the elimination of a single protein by gene mutagenesis might have effects on the stability of other proteins located in the secretion apparatus. This effect was demonstrated for protein HP0532, which was found to be stable in all mutant strains, except in mutant hp0537(Fig. 4). Thus, we would assume that HP0537 might act as a specific chaperone for HP0532. Alternatively, it might be involved as a regulator for the expression of the hp0532 gene or might be part of a protein complex that is unstable in the absence of HP0537.
At present, two major hypotheses about the mechanism of induction of IL-8 synthesis in epithelial cells by the type IV secretion system are conceivable: (i) the effector protein translocation hypothesis; and (ii) the receptor stimulation hypothesis (Fig. 6). The first hypothesis is based on the translocation of an additional effector protein different from CagA, which might be encoded by the cag-PAI, into epithelial cells inducing a signalling cascade via one of two pathways known to be involved in IL-8 induction. These pathways involve Rho-GTPases, PAK, MKK4 and JNK, or PAK, NIK and IκB kinases, resulting in the activation of the transcription factors AP-1 and NFκB, leading to the induction of chemokine gene transcription (Naumann et al., 1999). Our alternative hypothesis depends on the binding of the surface-located intact type IV secretion apparatus to a cellular receptor, which would be activated to transmit a signal into the cell to the same pathway mentioned above. Both hypotheses would explain why many genes from the cag-PAI are essential for IL-8 induction as, in both cases, a more or less intact secretion apparatus must be assembled on the bacterial surface, either for binding to the receptor or for translocation of an effector protein to transmit the signal (Fig. 6).
On the basis of these hypotheses, our mutagenesis data suggest that the cag-PAI does not encode a translocated protein responsible for the induction of IL-8. Otherwise, we would expect at least one mutant strain defective in IL-8 induction, but competent for translocation of CagA as an indicator of an intact secretion apparatus. However, such a mutant was not found in our analysis. Thus, we assume that, if a translocated factor is responsible for the induction of IL-8, it might be encoded outside the cag-PAI. Alternatively, we cannot exclude the possibility that several translocated effector proteins have redundant functions and that the mutagenesis of one gene would not abolish the function of the other genes. On the basis of our data, we favour the second model at present, which relies on receptor binding of the type IV secretion apparatus for signal transduction. This hypothesis is also supported by the finding that the VirD4 homologue HP0524 is not required for IL-8 induction. In A. tumefaciens, the latter protein is strictly involved in the translocation of any protein component, but it is dispensable for pilus formation. Assuming a similar situation for H. pylori, this would mean that no protein translocation would be associated with IL-8 induction. In the case of Yersinia spp., the invasin protein triggers IL-8 induction by binding to the β1 integrin receptor on epithelial cells, which leads to the induction of IL-8 (Schulte et al., 2000). A similar mechanism might exist for Hp(Fig. 6). The assembled type IV secretion apparatus might bind to such a receptor, which would signal via the stress response pathway to the nucleus resulting in IL-8 induction.
Only three proteins, the VirD4 homologue HP0524, HP0526 and HP0540, are essential for CagA translocation, but not for partial or complete IL-8 induction. These proteins are probably not structural components of a functional secretion apparatus in general and might be specific for certain steps in the translocation of CagA. In contrast, proteins HP0538, HP0542, HP0543 and HP0545 might be accessory proteins that stabilize the secretion apparatus and are helpful for both CagA translocation and IL-8 induction (Fig. 6). In contrast to the mutants in groups 1 and 2, these latter mutants show a considerable variation from one experiment to the other, as seen by the high standard deviation of the corresponding IL-8 induction values (Fig. 3).
Several laboratories have reported recently about the structure of the cag-PAI in independent clinical isolates of Hp (Jenks et al., 1998; Maeda et al., 1999; Audibert et al., 2001; Occhialini et al., 2001), and found various configurations ranging from the deletion of single genes within the PAI, groups of genes or the complete cag-PAI. In certain investigations, the authors attempted to correlate the presence or absence of genes with the disease status of the patients sampled. The presence of a complete or a partial cag-PAI was determined by PCR amplification or Southern hybridization. We observe from our analysis with Hp 26695 and the laboratory-passaged derivatives that a complete cag-PAI does not necessarily correlate with a functional secretion system. We identified a point mutation in gene hp0532, which apparently occurred after several passages of the strain in vivo. Although we do not know whether the identified mutation is the only one in the 26695pass-1 strain, it is apparently sufficient to switch off the secretion system by impairing CagA translocation and IL-8 secretion. Assuming that such spontaneous mutations might also occur in vivo during infection, we consider it to be difficult, if not impossible, to correlate the presence of an intact, partially deleted or completely deleted cag-PAI of an isolated strain with the gastroduodenal disease of the patient. Gastric disease is the result of years and decades of a Hp infection accompanied by totally uncontrolled multiple on/off switching events and partial or complete deletions within the cag-PAI.
The VirD4 homologue HP0524 was recently proposed to participate in the control of inflammation by downregulation of IL-8 synthesis in epithelial cells (Crabtree et al., 1999). These authors found a lower IL-8 induction induced by the wt strain compared with the hp0524 mutant strain, and they speculate that the VirD4 homologue might have a potential to downregulate the IL-8-inducing capacity of the cag-PAI. From our mutagenesis data for hp0524, we do not see any effect of the VirD4 homologue on IL-8 synthesis (Fig. 3), but a strong effect on the translocation of CagA. Considering the existence of spontaneous mutants with a partial or complete downregulation of IL-8 (26695pass-1, 26695pass-2), in our view, the data from Crabtree et al. (1999) could be explained by a spontaneous mutation, comparable with our 26695pass-2 mutation, which might have occurred in the wild-type strain, but not in the virD4 mutant strain. A genetic complementation of the mutant with the wild-type virD4 gene might be useful to clarify this situation in future.
Bacterial strains and culture conditions
Hp strains were grown on GC agar plates (Difco) supplemented with horse serum (8%), vancomycin (10 mg l−1), trimethoprim (5 mg l−1) and nystatin (1 mg l−1) (serum plates) and incubated for 2–3 days in a microaerobic atmosphere (85% N2, 10% CO2, 5% O2) at 37°C. Hp mutant strains were grown on serum plates supplemented with chloramphenicol (6 mg l−1) or kanamycin (8 mg l−1). E. coli strains HB101 (Boyer and Roulland-Dussoix, 1969) and DH5α (BRL) were grown on Luria–Bertani (LB) agar plates or in LB liquid medium (Sambrook et al., 1989) supplemented with ampicillin (100 mg l−1), chloramphenicol (30 mg l−1) or tetracycline (15 mg l−1), as appropriate.
DNA manipulations and plasmid constructions
Cloning and DNA analysis procedures were performed according to the methods of Sambrook et al. (1989). Isolation of Hp chromosomal DNA was performed as described by Leying et al. (1992) or using the QIAamp tissue kit (Qiagen). Plasmid DNA was purified from E. coli strains by the boiling procedure, and electroporation-competent E. coli cells were prepared according to the protocol recommended for the gene pulser (Bio-Rad).
Transposon shuttle mutagenesis
The cag-PAI genes were cloned by PCR amplification of the corresponding reading frames from chromosomal DNA of Hp 26695 using primers described in Table 1. The PCR fragments obtained were cloned into the vector pBA. For random insertional mutagenesis, transposon TnMax5 (Kahrs et al., 1995) or TnMax5gfp, a derivative of TnMax5 carrying the gfpmut2 (Cormack et al., 1996) gene close to the right inverted repeat, were used. Plasmids pCag1–pCag9 were transformed into E. coli E181 (Odenbreit et al., 1996) harbouring a pTnMax plasmid, and high-frequency transposition of the minitransposon was induced by growing the bacteria overnight on LB agar containing tetracycline and chloramphenicol to select for the maintenance of both plasmids and 100 µM IPTG. After transposition, total plasmids were transferred into E. coli E145 via transformation. Single transformants were analysed, and the insertion sites of the minitransposon were mapped by restriction digests and DNA sequencing. The interrupted gene was inserted into the chromosome of Hp 26695 by natural transformation of the plasmid after homologous recombination. The correct insertion of the transposon in the chromosome was verified by PCR using primers flanking the corresponding gene (Table 1) or Southern hybridizations.
Table 1. Oligonucleotides used for plasmid constructions and verification of chromosomal deletion mutations in Hp.
Restriction endonuclease cleavage sites GGATCC (BamHI), CTCGAG (XhoI) and CTGCAG (PstI) are underlined.
Gene replacement mutagenesis
Gene replacement mutagenesis was performed by a PCR approach presented schematically in Fig. 1C. Plasmids pCag1–pCag9 were used for inverse PCR amplification of the complete plasmid excluding a single gene using oligonucleotide primers specified in Table 1. The corresponding PCR fragments were ligated with the terminatorless cat cassette and transformed into E. coli DH5α. The clones obtained were tested by restriction digests and used for transformation of Hp 26695. Transformants in Hp were checked for the correct insertion of the cassette into the cag-PAI by PCR or Southern hybridization (data not shown).
Construction of plasmids used for the cag-PAI mutagenesis
Plasmids pCag1–pCag8 were constructed by PCR amplification from 26695 chromosomal DNA with primers cag1left and cag1right to cag8left and cag8right. The PCR fragments obtained were digested with XhoI–BamHI endonucleases and cloned into vector pBA digested with BglII–SalI. Plasmid pJP16 (genes hp0534–hp0536) was generated by cloning of a PCR fragment obtained from chromosomal DNA of Hp 26695 with primers cag6left and cag6right. The PCR fragment was digested with XhoI–BamHI endonuclease and ligated into pMin1 (BglII–XhoI) (Kahrs et al., 1995). Plasmid pSO168 was constructed by introducing an EcoRI deletion into plasmid pCag8A and subsequent religation. Plasmid pSO169 was generated by insertion of pJP16::TnMax5gfp into pJP16 with a subsequent EcoRI deletion. The precise deletion of genes hp0520–hp0523, hp0525, hp0526, hp0528–hp0532, hp0537, hp0538 and hp0540–hp0544 was performed by an inverse PCR approach with primers 520f and 520r to 544f and 544r using pCag1–pCag8A as template, as illustrated in Fig. 1B and C. The amplified PCR fragments were purified, digested with XhoI–BamHI endonucleases and ligated with a cat cassette isolated from plasmid pWS135 as a SalI–BamHI fragment. Plasmids with a deletion in genes hp0524 and hp0527 (pδ524, pδ527) were generated by HindIII digestion of plasmid pCag2 and pCag3, respectively, and ligation with the isolated HindIII fragment of the cat cassette from plasmid pTnMax5. These plasmids carry a cat cassette with a terminator and do not represent precise deletions of the complete orf, but both represent genes at the end of an operon, which should not exert polar effects on downstream genes. Plasmids for precise deletion of hp0545 and hp0546 (pδ545, pδ546) were obtained by inverse PCR with primers 545f and 545r or 546f and 546r, respectively, from pSO168. The fragments were digested with XhoI–BamHI restriction endonucleases and ligated with the cat cassette from pWS135 (SalI–BamHI). The genes p534-Tn, p535-Tn, p536-Tn, p539-Tn and p541-Tn were mutated by TnMax5 transposon insertion mutagenesis of the corresponding pCag plasmids, as described in the section Transposon shuttle mutagenesis.
Southern blotting and hybridizations with DNA fragments were performed using the ECL labelling and detection system according to the manufacturer's protocol (Amersham). For hybridization, 0.5 M NaCl was used; the washing buffer contained 0.5× SSPE (180 mM NaCl, 10 mM sodium phosphate, pH 7.5, and 1 mM EDTA), 6 M urea and 0.4% SDS at 42°C. Filters were washed with 2× SSPE, 0.05% SDS at 42°C.
Natural transformation of Hp
Natural transformation of Hp strains was performed with plasmid or chromosomal DNA according to the procedure described by Haas et al. (1993b). Bacteria were harvested from serum plates and suspended to an OD550 of 0.1 in brain–heart infusion (BHI) medium containing 10% fetal calf serum (FCS). DNA was added, and incubation was extended for 4–6 h under microaerophilic conditions (10% CO2, 37°C) before the suspension was plated on selective serum plates.
Generation of fusion proteins
The genes hp0532 and hp0539 were PCR amplified from chromosomal DNA of strain J99 (Alm et al., 1999), using oligonucleotides JP49/50 and JP51/52 (Table 1). The obtained PCR fragments were digested with PstI and XhoI and ligated into pEV40a (Pohlner et al., 1993) digested with SalI and PstI. The expression of fusion proteins was performed as described previously (Pohlner et al., 1993).
SDS–PAGE, immunoblots and antisera
For detection of gene expression in Hp wild-type and isogenic cag-PAI mutant strains, the bacteria were collected from serum plates and suspended in 300 µl of sample solution (Laemmli, 1970). Boiled aliquots were subjected to SDS–PAGE using a minigel apparatus (Bio-Rad) and blotted onto nitrocellulose membranes at 1 mA cm−2 using a semi-dry blot system (Biotec Fischer). The filters were blocked with 3% bovine serum albumin (BSA) in TBS (50 mM Tris-HCl, pH 7.5, 150 mM NaCl) and incubated with antisera AK270 and AK271 (1:3000 dilution). Alkaline phosphatase-coupled protein A was used to visualize bound antibody.
The authors wish to thank B. P. Burns for critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) (HA2697/1-4) (GK303/2) and the SFB576 (B1-Haas).