Conservation, Localization and Expression of Hopz, a Protein Involved in Adhesion of Helicobacter Pylori

From a sarkosyl-insoluble outer membrane fraction prepared from the Helicobacter pylori strain ATCC 43504, 19 proteins could be sequenced N-terminally by Edman degradation. Oligonucleotides were deduced and used for screening of a genomic library. From the isolated genes, five code for different members of a H.pylori outer membrane protein (Hop) family. Among these, the hopZ gene was characterized in more detail. It encodes a protein which was shown to be located at the bacterial surface by immunofluorescence studies. Sequence analysis of the hopZ gene from 15 different H.pylori strains revealed the existence of two alleles and the possible regulation of hopZ expression by slipped-strand mis-pairing within a CT dinucleotide repeat motif located in the signal-peptide coding region. Among the different strains, the influence of this region on the expression of HopZ was analyzed on a translational level by western blot analysis of bacterial extracts and immunofluorescence studies on intact cells. The protein is expressed only in those strains in which the number of the CT dinucleotide repeats allow for an open reading frame encoding the complete protein. Addionally the function of HopZ was investigated in an adhesion assay. The wild-type strain ATCC 43504 adhered to human gastric epithel cells whereas a knockout mutant strain showed significantly reduced binding to the cells.


INTRODUCTION
Helicobacter pylori is a Gram-negative, microaerophilic, spiral bacterium which colonizes the mucosa of the human stomach.The bacterium is the main cause of chronic active gastritis and peptic ulcer, in particular duodenal ulcer, and plays a role in the development of gastric carcinoma.
Its helical shape and motility, due to four to six flagellae, enables the bacterium to migrate through the gastric mucus in order to reach the boundary layer between the mucus and the mucosa.The bacterium adheres to the endothelial cells of the stomach which protects it from the extreme acidity of the gastric lumen and displacement from the stomach by forces such as those generated by peristalsis and gastric emptying.
A consequence of chronic colonization of the mucosa can be an inflammatory granulocytic and subsequently monocytic infiltration of the epithelium which in turn, by mediators of inflammation, contributes to tissue destruction.Infection stimulates both a local and a systemic humoral immune response which are, however, unable to eliminate the pathogen effectively.Immunization is the conventional way of preventing infectious diseases.Therefore it is important to examine this option with regard to controlling a H.pylori infection.
Bacterial virulence factors or structures accessible to the immune system are candidates for the development of subunit vaccines.It is generally assumed that antigens of this nature are present in the outer membrane of the bacterium.Proteins of 48, 49, 50 and 67 kDa (1) and 31 kDa (2), called HopA, HopB, HopC, HopD and HopE, respectively, have been shown to function as porins in H.pylori. Recently, Tomb et al. (3) identified these proteins among an outer membrane protein family consisting of 32 members.
Three further members of this family were identified as adhesin BabA (4), AlpA and AlpB (5).BabA was demonstrated to bind to the blood group antigen LewisB (4).There are two alleles: babA 1 and babA 2 , differing in the presence of a 10 bp repeat motif which results in the elimination of the start codon and the lack of Le b antigen-binding activity.Furthermore, the existence of repetitive sequences in the signalsequence encoding regions of five further OMP genes with the potential for slipped-strand mispairing and recombination which in turn may cause genotype variation has been described (3).Here we report on the isolation of one of these proteins, HopZ, and describe its localization, its genetic heterogeneity within different H.pylori strains as well as its putative function as an adhesin.

N-terminal amino acid sequence analysis of proteins
The outer membrane protein fraction was separated by SDS-PAGE using 8, 10 or 16% acrylamide depending on the molecular weight of the proteins to be separated.The bands were electrophoretically transferred to Immobilon membranes (7) with a trans blot cell (Bio-Rad, Hercules, CA).To complete the transfer of the proteins, 0.005% SDS was added to the cathode buffer.Membranes were stained with amido black to visualize the protein bands which were then excised.Amino acid sequencing was performed with a sequenator model 477 A (Applied Biosystems, Foster City, CA).Phenylthiohydantoin derivatives were separated with an on-line analyzer model 120A (Applied Biosystems).Data were handled using the control data analysis module 900 (Applied Biosystems).

Construction of an H.pylori genomic library
Helicobacter pylori genomic DNA of strain ATCC 43504 was prepared using the DNAzol ™ reagent (Gibco BRL, Rockville, MD).Genomic DNA was partially digested with the restriction enzyme Sau3AI and separated by agarose gel electrophoresis.DNA fragments between 0.5 and 5.0 kb were electroeluted (7).The Sau3AI treated DNA was cloned into the λZAP vector (Predigested ZAP Express ™ BamHI/CIAP Vector Cloning Kit, Stratagene, La Jolla, CA).

Screening of the H.pylori genomic library and isolation of the complete genes
From the N-terminal amino acid sequences oligonucleotides were deduced using the species-specific codon usage of H.pylori, which had been determined from 19 known H.pylori genes (8).The oligonucleotides were labeled with digoxigenin (DIG) using the DIG Oligonucleotide 3'-End Labeling Kit (Boehringer Mannheim, Germany) and employed for screening of the λZAP H.pylori genomic library under standard conditions, but using a hybridization temperature of 32°C (9).Inserts showing hybridization were excised and sequenced by the dideoxynucleotide chain termination method (10) using the DNA sequencer 373A (Applied Biosystems).From the boundaries of the Sau3AI fragments encoding specific amino acid sequences identical to those determined from the proteins by Edman degradation oligonucleotides were deduced, DIG labeled and used for screening of the ZAP Express ™ library in order to determine the complete sequences of the isolated genes.

Sequence analysis
Sequence analysis including the alignments of the sequences of the different strains was performed using the WISCONSIN PACKAGE, UNIX, v.8, Genetics Computer Group.The genomic DNA sequence from the H.pylori strain 26695 is accessible via http://www.tigr.org/tdb/mdb/hpdb/hpdb.html .The DNA sequences of the hopZ genes of all other H.pylori strains analyzed are accessible under the accession numbers Y18980-Y18993.

Expression and purification of HopZ and HopZ/aa126-281
The hopZ gene fragments encoding the putative mature proteins of the H.pylori strains ATCC 43504 and 5060d(A) as well as the subfragment HopZ/aa126-281 of strain ATCC 43504 were amplified from genomic DNA by PCR following standard procedures and inserted into the vector pQE30 (Qiagen) to obtain plasmids pQE30/hopZ, pQE30/hopZ5060d(A) and pQE30/hopZ/aa126-281, respectively.The plasmids were transformed in Escherichia coli XL1-Blue cells (Stratagene).After induction of the protein expression by IPTG, cells were disrupted and the proteins were purified from the insoluble fraction using nickel-chelate affinity chromatography according to standard protocols (Qiagen).

Western blot analysis on whole bacterial extracts
The recombinant proteins representing the mature HopZ of strain ATCC 43504 as well as the subfragment HopZ/aa126-281 were used to raise antibodies in rabbits following standard procedures.Antibodies directed against E.coli components were removed according to the method of Gruber and Zingales (11).Helicobacter pylori cells of the strains indicated above were grown for 48 h in Müller-Hinton broth containing 0.1% β-cyclodextrin, harvested by centrifugation and dissolved in Tris-glycin SDS sample buffer (Novex, San Diego, CA).Dilutions of the antisera were used for western blot analysis (12) on these H.pylori extracts employing the vectastain ABC kit (Serva, Heidelberg, Germany).Escherichia coli extracts containing the HopZ protein from strain 5060d(A) were used as positive control.

Immunofluorescence studies
Immunofluorescence studies on intact bacterial H.pylori cells, grown as described above, were performed essentially as described earlier (13).For the assay, 2 × 10 8 H.pylori cells were incubated with antisera raised against the three recombinant proteins, respectively, at a dilution of 1:160 using preimmune sera as a control.

Construction of a knockout mutant strain
In order to disrupt the hopZ gene the kanamycin resistance cassette of pILL600 ( 14) was cloned into the single PstI site of the gene using plasmid pQE30/hopZ and plasmid pQE30/ hopZkancas was obtained.This plasmid was used to naturally transform H.pylori strain ATCC 43504 cells according to the procedure described earlier (15).Cells were incubated for 24 h under microaerophilic conditions and then transferred to Columbia agar plates containing 5% horse blood and 20 µg/ml kanamycin.After 48 h, single colonies were visible and after three more passages genomic DNA was prepared as described above.The insertion of the kanamycin resistance cassette was confirmed by PCR using different combinations of primers.

Isolation and characterization of the hopZ gene from H.pylori
The preparation of a Sarkosyl ® -insoluble membrane fraction of H.pylori strain ATCC 43504, its separation using SDS-PAGE and subsequent N-terminal sequencing of the bands allowed the identification of 19 proteins (data not shown).Oligo-nucleotides were deduced and used to screen a λZAP library carrying genomic DNA fragments of H.pylori.
One of the sequences was identical to the N-terminus of HopC, a member of the H.pylori outer membrane family (1).The gene encoding HopC could not be isolated from the screening approach because its nucleotide sequence differs from that of the oligonucleotide deduced by the N-terminal amino acid sequence obtained.However, this oligonucleotide hybridized with two different DNA fragments coding for sequences homologous to HopC.One was shown to contain two homologous genes arranged in tandem; these were termed hopX and hopY; the gene on the second fragment, which was termed hopZ, was selected for further analysis.
The homology of the deduced N-terminal amino acid sequence of HopZ to the corresponding sequences of the other members of the OMP family as well as its hydrophobicity suggested the presence of a signal sequence of 18 amino acid residues cleaved between an alanine and a glutamic acid residue.The hopZ gene from strain ATCC 43504 encodes a mature protein of 681 amino acid residues with a calculated molecular weight of 74.2 kDa and an isoelectric point of 8.4.

Allelic variation of hopZ
The hopZ genes of another 13 H.pylori strains were PCRamplified and sequenced in order to analyze the variability of the HopZ protein.The deduced amino acid sequences are shown in Figure 1.Comparison of the amino acid sequences revealed two allelic variants of the gene.Strains ATCC 43504, 25392, 25144, 1811a and 5060d(A) carry allele I whereas pa1, pa2, ATCC 51110, ATCC 49503, G50, Tx30A, 342, CCUG 17874, 326/01 and 26695 contain allele II.The most prominent difference between both alleles is a 20 amino acid region present only in allele I.At the boundaries of this region, a significant variability is found between both alleles.The amino acid homology within these variable regions encoded by the different alleles is only in the range of 50%.In contrast, the Nterminal 100 amino acids and the C-terminal 400 amino acid residues are conserved between HopZ of all the different strains with 98-100% identity.These regions correspond to the one N-terminal and the seven C-terminal domains which are common to all members of the OMP family (3).Therefore we conclude that these sequences are conserved not only between the different proteins of the OMP family but also between different strains.

Localization of HopZ at the surface of H.pylori
The complete hopZ gene as well as the subfragment encoding amino acid residues 126-281 of strain ATCC 43504, representing the variable region, were expressed in E.coli and the histidinetagged proteins were purified by nickel-chelate chromatography and used for the preparation of antisera in rabbits.
Western blot analysis of the total cellular proteins of H.pylori cultures using antisera raised against the two different proteins revealed a protein band, which corresponds to the expected molecular weight (Fig. 2, lane 1).Only a slight crossreaction with other proteins of the Hop family was observedalso using antisera raised against the entire protein (data not shown)-suggesting that the major epitopes are located in the region specific for HopZ.A hopZ knockout mutant strain, ATCC 43504∆hopZ, constructed by insertion of the kanamycin resistance cassette of Campylobacter coli into the hopZ gene of H.pylori strain ATCC 43504, was used as a control.Indeed the antisera did not recognize HopZ in the mutant strain (lane 2).
Immunofluorescence studies were performed to investigate whether HopZ is exposed on the surface of H.pylori. Figure 3 shows the immunofluorescence of H.pylori strain ATCC 43504 probed with antisera raised against the total membrane fraction (A), as well as against the recombinant HopZ protein (B) and its fragment HopZ/aa126-281 (C).The antisera raised against the complete HopZ protein and the region of amino acids 126-281 showed similar intensities and percentages of immunofluorescence positive cells (>90%) as the antiserum raised against the total membrane fraction.The mutant strain ATCC 43504∆hopZ which was used as a negative control, did not show any fluorescence.
The immunofluorescence studies clearly demonstrate the surface exposition of HopZ.Particularly segments of the highly variable region spanning amino acids 126-281 of this outer membrane protein are located at the cell surface.

Genotypic variation of hopZ and consequences for its expression
Tomb et al. (3) described dinucleotide repeats located in the region encoding the signal sequences of five members of the OMP family.These repetitive sequences were assumed to generate phenotypic variation by a slipped-strand mispairing mechanism (17).Also for the hopZ gene (HP0009) of strain 26695, a repeat of 11 CT dinucleotides, causing a frameshift in the signal-sequence coding region, was identified.In contrast, the hopZ gene of strain ATCC 43504 contains only 7 CT dinucleotide repeats in frame with the residual gene.The sequencing of the hopZ gene of 13 more H.pylori strains allowed a detailed analysis of the gene status within different strains.
The DNA sequence encoding the signal peptide of the HopZ protein contains stretches of CT dinucleotide repeats which differ in their number between the 15 different H.pylori strains analyzed (Table 1).The presence of 7 or 10 CT dinucleotides in the signal-peptide coding region allows the expression of an intact hopZ translation product.A deletion or insertion of one or two dinucleotides in this repeat region leads to a frameshift mutation within the hopZ gene.Therefore strains carrying 6 (5060d(A), ATCC 49503, 25144, 25392, CCUG 17874), 9 (ATCC 51110), 11 (26695) or 12 (326/01) CT dinucleotide repeats are not expected to express an intact HopZ protein (Table 1).To verify the role of the slipped-strand mispairing mechanism in the regulation of expression of the hopZ gene, western blot analysis of bacterial extracts as well as immunofluorescence studies of different H.pylori strains were performed using the antiserum raised against the specific subfragment HopZ/ aa126-281 of strain ATCC 43504.Western blot analysis showed that the antiserum raised against a subfragment of the allele I strain ATCC 43504 reacts with HopZ protein from the allele I and II type (Fig. 2).The HopZ-specific protein band was detected only in the allele I strains ATCC 43504 and 1811a and the allele II strains Tx30A, pa1, pa2, G50 and 342.In contrast, no HopZ-specific protein band was found for the mutant strain ATCC 43504∆hopZ, for the allele I strains 25392, 25144 and 5060d(A) and for the allele II strains ATCC 49503 and CCUG 17874.The allele II strains 326/01 and ATCC 51110 show a weak reaction with a protein of ~70 kDa which is not assumed to be HopZ.
To confirm the results of the western blot analysis, immunofluorescence studies were performed (Fig. 3).No fluorescence could be detected for the strains 326/01 and ATCC 51110 as well as for the other strains which show a frameshift mutation in the signal-sequence coding region.Immunofluorescence indicating HopZ expression was demonstrated only for those strains found to be clearly positive in western blot analysis and exhibiting no frameshift mutation.Therefore the slippedstrand mispairing mechanism in the signal-sequence coding region is suggested to be responsible to switch on or off the expression of HopZ.RT-PCR analysis showed the presence of hopZ-specific mRNA in all 14 H.pylori strains analyzed (data not shown).

Function of the protein HopZ
The adherence of the wild-type strain ATCC 43504 and the mutant strain ATCC 43504∆hopZ to human gastric epithelial cells was compared.Furthermore the H.pylori strains 342 and 326/01 were tested.The wildtype strain strongly adheres to the epithelial cells whereas only a few cells of the mutant strain show adhesion.Additionally strain 342 expressing hopZ is able to adhere to the epithelial cells whereas strain 326/01 lacking HopZ adheres to a lower extent (Fig. 4).These data demonstrate that the lack of HopZ results in a significant reduction of adherence.Therefore we conclude that HopZ is involved in adhesion.

DISCUSSION
A further member of the OMP family of H.pylori, the protein HopZ, was identified and characterized with respect to its conservation, localization, expression and function.
Hancock et al. (18) described an amino acid identity value of >90% between the members of the OMP family of two different H.pylori strains.Indeed, the proteins HopV, HopW, HopX and HopY of the H.pylori strain ATCC 43504 are highly similar to proteins predicted from the sequenced genome of strain 26695 (3) encoded by the open reading frames HP1469, HP1501, HP0252 and HP0254, respectively, with identity values of 96.3, 94.5, 97.2 and 98.4%.We could confirm the high conservation of these proteins in 13 further strains (manuscript in preparation).In contrast, the HopZ protein analyzed here shows much more variability: e.g. the identity between the protein of strain ATCC 43504 and that of strain 26695 is only 79.1%.The sequence analysis of another 13 different strains shows that HopZ is encoded by two alleles with prominent differences in the region located between the 100 N-terminal and the 400 Cterminal highly conserved amino acids.The corresponding region was also described to be variable between the different members of the OMP family of strain 26695 (3).The HopZ protein is the first example within the OMP family for which a significant variability was found between different H.pylori strains.
Immunofluorescence studies using antisera raised against the total HopZ protein as well as against the HopZ-specific region of amino acids 126-281 demonstrated the presence of HopZ on the surface of intact H.pylori cells.Thus HopZ is indeed an outer membrane protein and especially the variable region, amino acids 126-281, is exposed on the cell surface.
The expression of the hopZ gene is regulated on the level of translation by a CT dinucleotide repeat motif of different length in different strains and localized at the 5' end of the coding region.An analogous control mechanism was described for a gene of Haemophilus influenzae encoding a lipopolysaccharide (LPS) synthetase involved in the phenotypic variation of surface LPS (17).This gene can carry a different number of tandem repeat units at its 5' end, which is the result of a slipped-strand mispairing mechanism.Depending on the number of repeats, the gene can either be translated in its full length or the translation product is truncated.Tomb et al. (3) described eight genes from H.pylori which contain dinucleotide repeats in their signalsequence coding regions.Genotypic variation caused by a slipped-strand mispairing mechanism within these repeats was postulated to switch on or off the expression of these genes; however, it has not been demonstrated.Appelmelk et al. (19) showed that two H.pylori genes encoding fucosyltransferases can be switched on or off by slipped-strand mispairing within C-tracts in their 5' coding regions.
Slipped-strand mispairing events within the repetitive sequences of the hopZ gene of H.pylori strains would allow the OMP expression to be readily switched on or off.Western blot analysis on H.pylori extracts from different strains as well as immunfluorescence studies demonstrated that HopZ is expressed only in those strains in which the hopZ gene is not interrupted in its reading frame, although hopZ-specific mRNA could be detected also in those strains which do not express the HopZ protein.Therefore the expression of the hopZ gene seems to be regulated on the translational level.However, the fact that the bacteria are able to abandon hopZ expression revealed that this protein is not essential for H.pylori at least in vitro.
Recently, five porins have been characterized among the members of the OMP family of H.pylori (1,2).However, we have evidence that HopZ is not able to form pores in bilayer experiments, suggesting that HopZ does not function as a porin (data not shown).A further member of the H.pylori OMP family, BabA, was characterized as an adhesin, binding to the histo-blood group antigen Lewis B (4).Moreover the adhesins AlpA and AlpB have been identified (5).We have shown that HopZ is also involved in adhesion of H.pylori.This was demonstrated by comparing the adherence of the wild-type strain ATCC 43504 and the mutant strain ATCC 43504∆hopZ to human epithelial cells.The data suggest that HopZ functions as a major adhesin because knocking out hopZ gene results in significant decrease of adhesive bacteria.Furthermore, strain 342 expressing hopZ adhered to the epithelial cells whereas strain 326/01 lacking HopZ showed reduction in the number of adhesive bacteria.However, some bacteria of the mutant strain as well as strain 326/01 are able to bind the epithelial cells.This adhesion could be mediated by other adhesins of H.pylori. Thus HopZ is the fourth protein of the OMP family which is involved in adhesion of H.pylori to gastric endothelial cells.

Figure 1 .
Figure 1.(Opposite and above) Comparison of amino acid sequences of HopZ of 15 different H.pylori strains.The putative signal peptides are indicated in italics.Asterisks denote the stop codons; note that the hopZ genes of the strains 25392, 25144, 5060d(A), ATCC 51110, ATCC 49503, CCUG 17874, 326/01 and 26695 do not encode an in frame start methionine.Amino acid positions of the mature proteins are indicated on the right side of the figure.Amino acid residues shown to be conserved between the 15 different strains of H.pylori are shown in white letters on black background.The amino acid residues which differ from the consensus sequence are given in black letters.The positions of the amino residues 126-281, which were expressed in E.coli, are indicated by arrows.

Table 1 .
Signal-sequence coding region of the hopZ gene of different H.pylori strains Nucleotide and deduced amino acid sequences of the signal peptide-coding region of the hopZ gene of 15 different H.pylori strains.Asterisks denote the stop codons of the hopZ gene of the strains whose translational start at the designated methionine residue leads to a truncated product.The numbers of the CT dinucleotide repeats, the gene status and the allele type are given on the right.