INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Helicobacter pylori is a gram-negative bacterium that colonizes the human stomach and causes chronic gastritis and peptic ulceration. Furthermore, colonization with this organism is associated with the development of gastric neoplasms. More than half of the H. pylori strains contain a pathogenicity island, the cag region, whose presence has a marked influence on the virulence of the organism.

Gene transfer between H. pylori strains is extremely common (24) and can generate novel subtypes during colonization with multiple strains (15, 19). The genetic recombination between H. pylori strains includes changes in important virulence markers such as the cag status (15). Therefore, horizontal gene transfer and uptake of foreign DNA play an important role in virulence and host adaptation of H. pylori. Horizontal gene transfer can occur via conjugation, transduction, or transformation. Most H. pylori strains are naturally competent for transformation with linear DNA (27; P. Nedenskov, G. Bukholm, and K. Bovre, Letter, J. Infect. Dis. 161:365-366, 1990) as well as with plasmids (31). In order to get insight into the characteristics of natural transformation in H. pylori, it is necessary to understand the mechanisms involved and their regulation.

When Tomb et al. (25) published the first genomic sequence of H. pylori, based on sequence homologies, a number of potential competence genes could be recognized. However, no integral DNA uptake system was identified (25). At present, a role in transformation has been described for only two loci: the comB operon (13) and dprA (3, 22). To identify other components involved in competence or its regulation, we screened a mutant library for competence-deficient mutants. This resulted in the identification of a novel H. pylori competence gene, comH. Unlike comB and dprA, comH does not belong to a known family of orthologous genes.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and culture media. Bacterial strains and plasmids are listed in Table 1. H. pylori, Helicobacter mustelae (kindly provided by T. Ó Cróinín, Our Lady's Hospital for Sick Children, Crumlin, Ireland), and Helicobacter acinonychis (kindly provided by A. Bart, Academic Medical Center, Amsterdam, The Netherlands) were grown under microaerobic conditions on Dent plates (9) (Dent supplement; Oxoid) supplemented with 40 mg of 2,3,5-triphenyltetrazolium chloride (Sigma Chemical Co., St. Louis, Mo.) per liter. When appropriate, antibiotics were added in the following concentrations: kanamycin, 20 mg/liter (Sigma); chloramphenicol, 15 mg/liter (Serva, Heidelberg, Germany); clarithromycin, 2 mg/liter (Abbott Laboratories Ltd., Queensborough, United Kingdom). Helicobacter felis strains were grown as described by Cattoli et al. (7). Escherichia coli ER1793 (14) and DH5 (Clontech, Palo Alto, Calif.) were cultured in Luria-Bertani broth, with 30 mg of kanamycin or 30 mg of chloramphenicol/liter if appropriate. Plasmid pHel2 is an E. coli-H. pylori shuttle vector that carries the cat_GC chloramphenicol resistance gene (12). The pBC3 suicide vector was derived from the pBC SK+/ plasmid (Stratagene, La Jolla, Calif.) by ligation of the aphA-3 kanamycin resistance cassette (26) into its unique SmaI site (4).

DNA manipulation. Southern blotting and recombinant DNA techniques were performed according to standard protocols (20) unless stated otherwise. Plasmids were isolated with the QIAprep spin miniprep kit (Qiagen GmbH, Hilden, Germany). Restriction enzymes used in this study were obtained from New England Biolabs Inc. (Beverly, Mass.).

Transformation. Natural transformation of H. pylori was performed essentially as described by Wang et al. (31). In brief, 24 h after inoculation bacteria were harvested from their plate and transferred as thick patches onto a fresh plate; after 5 h approximately 1 µg of DNA was added to the patches. After 20 h of incubation, the bacteria were suspended in 120 µl of phosphate-buffered saline and 100-µl portions of appropriate dilutions were spread on selective plates. To calculate a transformation frequency, appropriate dilutions (10⁶ and 10⁸) were plated on nonselective plates. After incubation for 5 days, the colonies were counted.

1

Construction and screening of the library. The construction of the H. pylori mutant library has been described before (4). Individual mutants of the library were inoculated as patches on kanamycin-agar. After 24 h of growth, the patches were covered with 10 µl of a 25-ng/µl chromosomal DNA solution that confers clarithromycin resistance due to an A-2142-to-G mutation in the 23S ribosomal DNA (8). After another 24 h of growth, the patches were transferred to plates containing clarithromycin.

Plasmid rescue, sequencing, and sequence analysis. For plasmid rescue chromosomal DNA of the mutants was isolated and restricted with HindIII (a unique HindIII restriction site is present on pBC3 between the aphA-3 kanamycin and chloramphenicol resistance cassettes; see Fig. 1), self-ligated into circularized HindIII fragments, and transformed into E. coli with selection on chloramphenicol. For the determination of the other point of insertion, the circularized HindIII fragments were used as a template in an inverse PCR with primers that face outward on the aphA-3 cassette: AphA3-R and Kana-L (Table 2). PCR products were cloned in the pGEM-T Easy vector (Promega, Madison, Wis.).

Construction of site-directed mutants in ORF HP1527. A fragment of open reading frame (ORF) HP1527 was amplified from H. pylori strain 1061 with primers HP1527for43 and HP1527rev1156 (Table 2) and cloned into pGEM-T Easy. This HP1527 fragment contains a HindIII site at base 753 of the ORF that was used for restriction and subsequent ligation with the aphA-3-containing HindIII fragment of pJMK30 (29). The resulting HP1527::aphA-3-containing pGEM-T was transformed into E. coli DH5 to obtain pSACHA-1 and pSACHA-2. The orientations of aphA-3 in pSACHA-1 (same direction as the HP1527 ORF) and pSACHA-2 (opposite direction) were determined by PCR with combinations of primers aphA3-R or Kana-L (forward) and HP1527for43 or HP1527rev1156 (reverse) (Table 2) and by sequencing the amplimers. pSACHA-1 and pSACHA-2 were used to create HP1527 mutants in strain 1061 and 26695 by natural transformation (Table 1).

Construction of the rdxA vector. The 5' part of the rdxA gene was amplified by PCR with the primers rdxAIXbaI and rdxAISacI (Table 2), and the resulting amplimers were purified. This 5' fragment of rdxA was cloned into the phagemid pBC-SK using the XbaI and SacI restriction sites that were introduced by PCR, which resulted in the vector pBC-rdxAI. Subsequently the 3' part of rdxA was amplified by PCR with the primers rdxAIIXhoI and rdxAIIKpnI (Table 2), and the resulting amplimers were purified. The introduction of this 3' fragment of rdxA into pBC-rdxAI, with the aid of the XhoI and KpnI restriction sites that were introduced by PCR, gave rise to plasmid pRdxA (Fig. 2).

Complementation analysis with the rdxA vector system. The complete gene HP1527 of strain 1061 was amplified by PCR with primers on the flanking genes: HP1526rev62 and HP1529for1323 (Table 2, Fig. 1). This amplimer was cloned in pGEM-T Easy and ligated into the EcoRI site of the rdxA vector to obtain pRDXA-1527. After being cloned into E. coli DH5, pRDXA-1527 was transformed into H. pylori strain 26695; this gave rise to the metronidazole-resistant (Mtz^r) mutant HpC-1527. HpC-1527 was transformed with pSACHA-1. The resulting kanamycin-resistant (Mtz^r Km^r) colonies were tested for the location of the aphA-3 cassette with a set of PCRs; Apha-L or Apha-3R was used as a forward primer, and reverse primers on the rdxA ORF (MetroF) and the HP1529 ORF (HP1529for1323) were chosen (Table 2, Fig. 3).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Screening of the library. Approximately 1,250 mutants from a random H. pylori 1061 library were screened for transformation deficiency. Each mutant was inoculated as a small patch and, after 24 h, overlaid with chromosomal DNA that confers clarithromycin resistance. After another 24 h, the patches were transferred to selective plates. In this crude but easy-to-perform screening method 1,200 mutants formed one or more Cla^r colonies and thus proved to be competent. The remaining 50 mutants were subjected to natural transformation by the method of Wang et al. (31). In this test, 3 of the 50 were completely transformation deficient and were selected for further examination.

Plasmid rescue, sequencing, and sequence analysis. The library that was used for this screening was created by chromosomal insertion of pBC3 suicide plasmids which contain a random fragment of H. pylori. This random fragment of DNA recombines into the H. pylori chromosome by a single homologous crossover event, which leads to insertion of the complete vector and to a duplication of the DNA fragment of the H. pylori chromosome. Thus, each mutant contains one copy of this fragment on each side of the integrated vector. Because of this duplication, the backbone of the pBC3 vector that interrupts the chromosome has a different insertion point on each side. To determine the first point of insertion, plasmid rescue was performed by restriction with HindIII and religation of the chromosomal DNA, which restores a pBC3-based Cam^r plasmid that contains one flanking sequence of the H. pylori chromosome (Fig. 1). No suitable restriction endonuclease site was available to obtain a rescue plasmid that contains the other flanking sequence. Therefore, the circularized HindIII fragments were used as a template in a reverse PCR with primers that face outward on the aphA-3 resistance cassette. Thus, the chromosomal DNA flanking the aphA-3 cassette was amplified.

View larger version (9K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the genomic region of HP1527 with the location of the pBC3 insertion. Depicted are relevant regions of pBC3 and the HindIII site used for plasmid rescue, the duplicated chromosomal region (gray), and the primers used for complementation of HP1527 (Table 2). The figure is not drawn to scale. aphA-3, kanamycin resistance cassette; Cm, chloramphenicol resistance cassette; colE1, origin of replication. a, HP1528 is present in strain 26695 only.

Construction of site-directed mutants in ORF HP1527 and transformation. To prove that the transformation deficiency of the random mutants was not caused by an unrelated event elsewhere in the genome, site-directed mutants were constructed in strain 1061 by insertion of an aphA-3 cassette in ORF HP1527. First, a fragment of ORF HP1527 of strain 1061 was amplified by PCR and cloned in the pGEM-T Easy vector. Sequence analysis revealed a HindIII restriction site in this DNA fragment. This site was used to insert the aphA-3 cassette, which codes for kanamycin resistance, and the resulting constructs were named pSACHA-1 and pSACHA-2. PCRs with combinations of primers aphA3-R or Kana-L (forward) and HP1527for43 or HP1527rev1156 (reverse) and the sequencing of the amplimers showed that pSACHA-1 has the aphA-3 gene inserted in the same direction as the HP1527 reading frame and that pSACHA-2 has the gene inserted in the opposite direction. pSACHA-1 and pSACHA-2 were used to create mutants in strain 1061. In addition, a SACHA-1 mutant was made in strain 26695 to confirm that the phenotype caused by disruption of HP1527 is similar in an unrelated strain. Disruption of ORF HP1527 in each mutant was confirmed by Southern blotting (results not shown).

6

9

Construction of the rdxA vector. For the complementation of HP1527 we developed a replacement vector that would allow for the ectopic integration of DNA into the chromosome of H. pylori. As a target for replacement we used gene rdxA. Disruption of rdxA causes metronidazole resistance in H. pylori (11). Thus, insertion of any DNA fragment into rdxA will give rise to metronidazole-resistant colonies, and the DNA serves as its own resistance marker to select for the successful integration into rdxA. For the construction of the rdxA vector, two fragments of the rdxA gene were amplified by PCR. With the aid of the restriction sites that were introduced during PCR, the 5' fragment of the gene was cloned into the first two restriction sites of the multiple cloning site of the phagemid pBC SK and the 3' fragment was cloned into the last two restriction sites. This resulted in pRdxA (Fig. 2), a plasmid containing the 5' and 3' parts of the rdxA gene flanking the remainder of the multiple cloning site, which allows for the introduction of a DNA fragment. Sequencing pRdxA with the M13 forward and M13 reverse primers, located just outside the multiple cloning site, confirmed the correctness of the inserts. Transformation of pRdxA into H. pylori yielded metronidazole-resistant colonies, indicating that introduction of the multiple cloning site of pRdxA disrupted the rdxA gene (data not shown).

View larger version (27K): [in this window] [in a new window]   FIG. 2.   Map of vector pRdxA. The multiple cloning site (MCS) allows for cloning between two fragments of the rdxA gene (rdxAI and rdxAII). Relevant restriction sites are indicated.

Complementation analysis with the rdxA vector system. Complementation of ORF HP1527 was performed to confirm that the competence-deficient phenotype of the mutants was caused by disruption of ORF HP1527 and not by a polar effect on surrounding genes. Gene HP1527 was amplified by PCR with primers located on the flanking genes. Strain 26695 contains a small ORF (HP1528) that overlaps the putative promoter region of HP1527. In order to obtain gene HP1527 only and to avoid problems due to this overlap in strain 26695, gene HP1527 was amplified from strain 1061, which lacks ORF HP1528. The amplimer was cloned in the rdxA vector to yield pRdxA-1527 (Fig. 3, left). Because disruption of ORF HP1527 eliminates competence, we performed the complementation of HP1527 as follows. First, pRdxA-1527 was transformed into wild-type H. pylori strain 26695. An Mtz^r mutant of 26695 with a second intact HP1527 inserted into the rdxA gene, directed opposite to the rdxA reading frame, was identified by PCR and called HpC-1527. Next, HpC-1527 was transformed with pSACHA-1, which yielded Mtz^r Km^r transformants with an interruption of either the original or the additional HP1527 ORF. The location of the aphA-3 insertion was identified with a set of PCRs that demonstrate the presence or absence of the aphA-3 cassette, both at the original location and in the rdxA gene, as shown in Fig. 3. A mutant with the aphA-3 insertion in the original HP1527 was called HpC-SACHA. We then tested the wild-type 26695 and its derivative HpC-1527, which contains two intact HP1527 genes, and the mutant with the complemented genotype, HpC-SACHA, for their capabilities to transform to clarithromycin resistance. The transformation frequency of HpC-SACHA was identical to the frequency of the parental strain (Table 4). The duplication of gene HP1527 in mutant HpC-1527 had no marked effect on the transformation frequency.

View larger version (59K): [in this window] [in a new window]   FIG. 3.   (A) Schematic representation of complementation mutant HpC-SACHA showing the rdxA region with the HP1527 insert (left) and the HP1527 region with the aphA-3 insert (right). Indicated are the primers used to construct HpC-1527 and to confirm the site of aphA-3 insertion. Primers 1 (HP1526rev62) and 2 (HP1529for1323) were used to amplify HP1527 from strain 1061 (which lacks ORF HP1528; see text). This HP1527 copy was inserted into the rdxA gene of strain 26695. The presence of an uninterrupted HP1527 gene in the rdxA gene was confirmed by a PCR with primers 3 (MetroF) and 2. Primers 4 (HP1529for1110) and Kana-L confirmed the aphA-3 insertion in the original HP1527 gene, while primer 3 did not yield a product with aphA3-R. (B) PCR confirming the HpC-SACHA genotype in the left panel. Lane M, DNA size markers; lane 1, primers 4 and Kana-L confirmed the aphA-3 insertion in the original HP1527 gene (band at 1,359 bp); lane 2, the presence of an uninterrupted HP1527 gene in the rdxA gene was confirmed by a PCR with primers 2 and 3 (2,410 bp); lane 3, primer 3 did not yield a product with aphA3-R, confirming the absence of the aphA-3 insertion in the complementing gene copy; lane 4, control for lane 1 with HpC-1527 as the template; lane 5, control for lane 2, with the parental strain as the template.

Distribution of HP1527 in the genus Helicobacter. The nine wild-type H. pylori strains from Table 1 were tested for the presence of comH on a Southern blot probed with a comH fragment (bases 62 to 1156) of strain 1061, and comH was demonstrated in all of them (data not shown). These nine strains included both highly competent strains and strains with relatively low competence such as SS1. The same comH fragment also hybridized to three strains of H. acinonychis (Table 1, Fig. 4), a species that is closely related to H. pylori and that is also naturally transformable (results not shown). However, Southern blotting experiments did not demonstrate sequences homologous to comH in two other Helicobacter species, H. felis (two strains; Table 1) and H. mustelae (five strains; Table 1).

View larger version (83K): [in this window] [in a new window]   FIG. 4.   H. acinonychis strains probed with an internal comH fragment. Lane 1, strain Sheeba; lane 2, strain India; lane 3, reference strain ATCC 12686.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

To identify elements of the transformation system in H. pylori, we screened a random-insertion library for loss of competence. We identified a mutant in ORF HP1527 that was incapable of natural transformation. Site-directed mutants in this ORF showed the same phenotype. Complementation of HP1527 in trans completely restored competence, which indicates that the mutation itself rather than a polar effect on surrounding genes causes the transformation deficiency. These results demonstrate that HP1527, an ORF with a heretofore-unknown function, is essential for natural transformation of H. pylori. Although the gene might have additional functions, based on the data presented in this paper a gene name in accordance with current nomenclature for competence genes would be appropriate for HP1527. Because the gene has no orthologs (see below), we decided to name HP1527 comH, which is, to our knowledge, the first available com designation in the alphabet. The comH gene is present in all tested H. pylori strains, not only in highly competent strains but also in less-efficient transformers, and interruption of this gene completely obliterated the natural-transformation competence for both chromosomal DNA and plasmids. comH mutants appear to have a normal growth rate and survival, which suggests that comH has no additional household functions.

The organization of the chromosomal region around comH differs among strains 26695, J99, and 1061. In all three strains, the H. pylori exoA homolog (HP1526) is located downstream of comH, separated from comH by a tRNA gene that lies in the opposite direction. In the 26695 sequence, upstream of comH, are the putative ORF HP1528 and the dnaA homolog (HP1529). The small putative ORF HP1528 is absent from 1061 and from the J99 sequence, which indicates that it is not required for competence in H. pylori. Because of the large and variable intergenic region between comH and the dnaA gene, as well as the putative function of dnaA in chromosomal replication, cotranscription of comH in an operon with this gene is unlikely. Comparison of the ORF comH (strain 26695) with the corresponding ORF of strain J99, JHP1416, revealed an amino acid identity of 93% (95% similarity), which is in line with the variation between other genes in H. pylori. Sequence analysis with the SignalP program (18) showed that comH encodes a product with a presumed transmembrane domain, corresponding to an N-terminal signal peptide for secretion, with a cleavage site between amino acid residues 19 and 20. The putative exported mature protein consists of 460 amino acid residues (52.4 kDa) and has an isoelectric point of 6.35. We did not find sequences that may function as DNA-binding sites in comH.

In Southern blotting experiments with an H. pylori comH probe, a comH homolog was detected in all H. pylori and H. acinonychis strains but not in H. mustelae and H. felis. Likewise, the latter two species did not hybridize with the H. pylori comB operon in earlier experiments by Hofreuter et al. (13). Because H. mustelae is also naturally transformable (data not shown), these results suggest that transformation genes are not conserved among all naturally transformable Helicobacter spp. Database sequence similarity searches did not reveal a significant homology of ORF comH to any genomic sequence available in GenBank, including naturally transformable species such as Bacillus subtilis, Haemophilus influenzae, and C. jejuni. This indicates that part of the H. pylori transformation system is evolutionarily distinct from the systems known from other species. The previously identified comB and dprA genes, however, have orthologs in other competent bacterial species and even in conjugational plasmids (3, 13, 22).

H. pylori contains many ORFs without an obvious ortholog, and the screening of a random library is therefore a powerful method for the identification of gene function. Although a previous screening of this library revealed 8 unique mutants (4), the present identification of three identical clones indicates that the 1,250 insertion mutants are not all independent. In addition to this, not all pBC3 insertions inactivate a gene, and only one restriction enzyme was used to create the random fragments for mutagenesis. The present set of mutants is therefore not a comprehensive library of the 1,500 H. pylori genes. Indeed, none of the known transformation genes (comB operon, recA, and dprA) were identified in our screening. It is therefore possible that other unrevealed competence genes are present.

In this paper we also describe a new complementation strategy for H. pylori based on the rdxA vector. This complementation system has obvious advantages over plasmid-based complementation. It produces a stable, single-copy insertion. Furthermore, the rdxA vector allows for introduction of DNA into H. pylori with an absolute minimum of changes in the genome: it avoids the unknown effects of using additional resistance markers that are unnatural to H. pylori and does not introduce remnants of the vector other than a short polylinker sequence. Many clinical isolates of H. pylori are metronidazole resistant, which indicates that disruption of rdxA does not have a significant effect on the viability of H. pylori. A practical advantage of the lack of an additional resistance marker is the reduced length of the DNA fragment that has to be internalized, which enhances the transformation frequency in less-competent strains.

The lack of orthologs makes it difficult to speculate on the role of comH in the process of transformation. In general, natural transformation can be divided into the following steps: development of a competent state, DNA binding, DNA uptake, and genomic integration (23). The putative N-terminal secretion signal of the comH product suggests that the protein is either anchored in the cytoplasmic membrane or exported to the periplasm and points to a role in the DNA-binding or DNA uptake process, although a function in the development of a competent state cannot be excluded. The results of electroporation experiments that demonstrate a normal recombination in comH mutants imply that comH is not involved in the recombination that follows uptake of chromosomal fragments. This is in accordance with the finding that comH mutants are incapable of plasmid uptake, since RecA-deficient H. pylori mutants are still capable of transformation with self-replicating plasmids but not with chromosomal markers (21).

It has become clear from published genomic sequences that H. pylori contains relatively few operonic loci. Whereas the H. pylori comB competence genes appear to form a small operon, dprA and comH do not. Organization of competence genes in larger loci and operon structures in other naturally transformable bacteria has been described. In B. subtilis, the expression of natural transformation competence is a highly regulated process: competence genes are controlled by a complex signal transduction network that senses environmental changes, and competence is expressed only under specific circumstances (10). In contrast, H. pylori can be transformed under standard culture conditions. The lack of operonic organization of competence genes in H. pylori could therefore well reflect a relatively loose regulation of competence, as in Neisseria spp. (5). The evidence for extensive horizontal gene transfer between H. pylori strains and the conserved nature of comH and other transformation genes stress the importance of natural transformation for this organism.