• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Feb 17, 2004; 101(7): 1852–1857.
Published online Feb 9, 2004. doi:  10.1073/pnas.0307329101
PMCID: PMC357016

Molecular cloning and characterization of two Helicobacter pylori genes coding for plasminogen-binding proteins


Helicobacter pylori binds a number of host cell proteins, including the plasma protein plasminogen, which is the proenzyme of the serine protease plasmin. Two H. pylori plasminogen-binding proteins have been described; however, no genes were identified. Here we report the use of a phage display library to clone two genes from the H. pylori CCUG 17874 genome that mediate binding to plasminogen. DNA sequence analysis of one of these genes revealed 96.6% homology with H. pylori 26695 HP0508. A subsequent database search revealed that the amino acid sequence of a lysine-rich C-terminal segment of HP0508 is identical to the C terminus of HP0863. Recombinant proteins expressed from HP0508 and HP0863 bound plasminogen specifically and in a lysine-dependent manner. We designate these genes pgbA and pgbB, respectively. These proteins are expressed by a variety of H. pylori strains, have surface-exposed domains, and do not inhibit plasminogen activation. These results indicate that pgbA and pgbB may allow H. pylori to coat its exterior with plasminogen, which subsequently can be activated to plasmin. The surface acquisition of protease activity may enhance the virulence of H. pylori.

Helicobacter pylori is a microaerophilic, Gram-negative bacterium that colonizes the mucosa of the human stomach. One of the most common pathogens of humans (1), this microbe is associated with chronic gastritis, peptic ulcers, gastric cancer, and mucosa-associated lymphoid tissue lymphoma of the stomach. H. pylori expresses a wide range of virulence factors that aid in its pathogenesis by affecting host cells or altering the bacterium's microenvironment: urease (2), superoxide dismutase (3), collagenase (4), vacuolating cytotoxin A (5), and CagA (6). Urease is essential for host colonization whereas CagA is a major determinant of disease severity and outcome.

H. pylori was recently found to bind plasminogen (Pg), and this property may contribute to the virulence of this bacterium. Pg is the precursor of the broad-specificity protease plasmin, and it is the main component of the fibrinolytic system. During wound healing, Pg is proteolytically cleaved into plasmin, a serine protease that cleaves fibrin, the main component of blood clots, and the extracellular matrix molecules fibronectin, laminin, and vitronectin. Pg also activates procollagenases. Because indiscriminate activation of plasmin could contribute to significant tissue damage, the recruitment and activation of Pg must be carefully controlled (7).

Pg-binding activity has been documented for group A, C, and G streptococci (8), Neisseria meningitidis (9), Salmonella enterica (10), Haemophilus influenzae (11), and Borrelia burgdorferi (12). For group A streptococci (13), B. burgdorferi (14), Borrelia crocidurae (15), and Yersinia pestis (16), Pg has been demonstrated to be important for virulence in vivo. It is believed that tissue penetration and dissemination are facilitated by the acquisition of Pg on the bacterial surface and subsequent activation of this surface-associated Pg by a host-derived Pg activator such as tissue Pg activator (tPA) or urokinase Pg activator. Pg binding is used by the influenza A virus to enter host cells (17). The prion protein PrP(Sc) that is associated with transmissible spongiform encephalopathy has also been reported to bind Pg (18), although the physiological significance has been questioned (19).

H. pylori Pg-binding proteins (Pgbs) have been identified (20, 21), but the genes were not identified. This study identifies the H. pylori 26695 genes HP0508 and HP0863 (pgbA and pgbB, respectively) as responsible for H. pylori Pg-binding activity, and the characterization of these gene products suggests that they may be important virulence factors. Although H. pylori is not known to disseminate to tissues outside the stomach and is generally believed to be noninvasive, we hypothesize that the chronic nature of H. pylori infection in conjunction with acquisition of proteolytic activity may contribute to the impaired wound healing and tissue damage characteristic of gastric ulcers.

Materials and Methods

Bacterial Strains and Culture Media. H. pylori strains CCUG 17874 and 26695 (CCUG 41936) were obtained from the Culture Collection at the University of Gothenburg. H. pylori clinical isolates Poland 43, Turkey 33, Iran 23, Sweden 5, and Somalia 3 were obtained from Carina Bengtsson (Karolinska Hospital), and AH244 was obtained from James Fox (Massachusetts Institute of Technology, Cambridge). H. pylori was grown on blood agar plates (4.3% Columbia blood agar base/0.01% dl-tryptophan/6% horse blood) in a humidified microaerobic environment (85% N2/10% CO2/5% O2) at 36°C for 3 days.

Escherichia coli strain TG1 (Stratagene) was used for phage display. TOP10 (Invitrogen) and BL21 Star (Invitrogen) were used for protein expression. E. coli was grown in Luria broth with 100 μg/ml ampicillin or carbenicillin when appropriate.

Construction and Biopanning of the M13 Phage Library. H. pylori CCUG 17874 DNA was sonicated (10 sec on, then 10 sec off over 3–5 min) at maximum power. DNA fragments in the 100- to 2,000-bp range were chosen for constructing the library. In brief, the DNA was ligated into the phagemid pG8SAET (gift of Lars Frykberg, Swedish University of Agricultural Science, Uppsala) and transformed into E. coli TG1. Subsequent infection with the helper phage R408 enabled expression of in-frame protein fusions to the M13 phage major outer protein VIII.

The biopanning procedure was performed as follows. In brief, 200 μl of the phage library was added to Pg coated on microtiter plates (MaxiSorp, Nunc) containing 20 μg of Pg per well and incubated for4hat room temperature (RT). The wells were then washed with PBS, pH 7.5, with 0.05% Tween 20. Phage particles were eluted with 50 mM sodium citrate/140 mM NaCl, pH 5.0, followed by 2 M Tris·HCl, pH 8.0. This eluate was used in a second round of panning to enrich for correct clones.

DNA Sequencing and Protein Sequence Alignment. Plasmids, phagemids, and PCR amplicons were prepared by using the Wizard plasmid purification kit (Promega) or the JETquick PCR purification spin kit (Saven, Malmö, Sweden) and sequenced by using an ABI-PRISM 310 genetic analyzer (Perkin–Elmer/Applied Biosystems). Appropriate H. pylori DNA sequences were retrieved from the National Center for Biotechnology Information/GenBank. DNA sequence alignment was performed by using sequencher 3.1.1 software (Gene Codes, Ann Arbor, MI). Multiple protein sequence alignment was performed by using the program clustalw 1.8 (22). The accession number corresponding to the nearly full-length H. pylori CCUG 17874 pgbA (corresponding to H. pylori 26695 gene HP0508) is AJ550456.

Cloning and Expression of H. pylori pgbA, pgbB, and pgbA3 as His-Tagged Proteins. Gene fragments were cloned by PCR using AmpliTaq Gold (Perkin–Elmer/Applied Biosystems) in 25 μl containing 1× AmpliTaq Gold buffer, 2.5 mM MgCl2, 1 μM dNTP, and 25 pmol of each primer. The following cycle program was used: 94°C for 8 min; 25 cycles at 94°C for 30 sec, 55°C for 30 sec, 72°C for 30 sec; and 72°C for 7 min. pgbA and pgbB were cloned without their signal sequences. pgbA from CCUG 17874 was cloned into pET-100D (Invitrogen) by using primers pET-100D/ADE (5′-CACCGATTTGCAAGTGGGGGAGTTTGG-3′) and 3′-stop/CCUG (5′-CTACTTCTTATTCATTTCTAAAGCTTT-3′). pgbB from 26695 was cloned into pET-100D by using primers pET-100D/5′-0863 (5′-CACCTTGCAAGAAAGTATCGTTTC-3′) and pET-100D/3′-0863stop (5′-TCATTACTTCTTATTCATTTCTAAAGCTTT-3′). pgbA3 from CCUG 17874 was cloned into pET25B+ (Invitrogen) by using primers 5′-Heplg comb (5′-CAATGAAAAAGTCAATGCCAAAG-3′) and 3′-Heplg comb (5′-CTCTTTCTTGATGCTCTCT-3′). All clones were verified by DNA sequence analysis.

Plasmid DNA was transformed into E. coli BL21 Star. Proteins were induced according to the manufacturer's directions (Invitrogen) and purified by using the BugBuster kit and Ni-chelate columns (Novagen) in the presence of 6 M urea.

ELISA Pg Attachment Assays. One microgram of Pg in 100 μlofPBS was coated onto microtiter wells (Nunc) for 1 h at RT or overnight (ON) at 4°C. The remaining sites were blocked by incubating with 200 μl of 1% BSA in PBS for 1 h at RT or ON at 4°C, followed by a 5-min incubation in 200 μl of PBS containing 0.02% sodium azide, and stored dry at 4°C for 2–3 weeks. Attachment of Pgb, expressed as N-terminal His-tag fusion proteins, to Pg-coated wells, was assayed by incubating Pgb in 100 μl of 0.1% BSA in PBS (PBSB) for 1 h, followed by three 5-min washing steps with 200 μl of PBSB. Next, an allophycocyanin-conjugated Ab directed against the N-terminal His-tag (Invitrogen) was incubated for 1 h. Absorbance (405 nm) was read after incubating with 100 μl of 1 M diethanolamine/1 mM MgCl2/1 mg/ml p-nitrophenyl phosphate at 37°C for 30 min.

Inhibition of PgbA or -B attachment by fetuin or Pg was assayed by preincubating various concentrations of fetuin or Pg with 0.5 μM Pgb in PBSB for 1 h at RT before allowing the Pgb to attach to Pg-coated microtiter wells. Inhibition of Pg binding was assayed by adding either ε-aminocaproic acid (EACA) or lysine with the indicated amount of Pgb to Pg-coated wells.

Abs. Rabbit polyclonal antisera (termed Ab 114) were raised against recombinant PgbA3 (Biodesign International, Kennebunkport, ME). The amino acid sequence of PgbA3 is given in Fig. 1. Antisera were adsorbed against soluble proteins from E. coli BL21 Star, the strain from which the protein was purified for immunization. E. coli BL21 Star cells were sonicated, and proteins in the supernatant were coupled to CNBr 4B Sepharose (Pharmacia) according to the manufacturer's instructions. Fifteen microliters of antisera were diluted 1:1,000 in TBS, passed over a 1-ml column of coupled Sepharose, and stored at 4°C for up to 4 months.

Fig. 1.
Multiple protein sequence alignment of PgbA and PgbB from strains 26695 and J99. The sequences from phage display clones pPC18 and pPC21 and their combined sequence, PgbA3, are also shown. Black shading indicates identical residues, and gray shading indicates ...

Western Blots. Whole cell lysates of H. pylori were electrophoretically separated on NuPAGE 10% Bis-Tris gels (Invitrogen) in Mops buffer or 10% Tris·HCl Ready Gels (Bio-Rad) and electroblotted onto a poly(vinylidene difluoride) membrane. The membranes were then incubated in blocking buffer (5% nonfat dried milk in PBS) for 1 h, washed three times with TBS (20 mM Tris·HCl, pH 7.5/140 mM NaCl), and probed with Ab 114 diluted 1:1,000 in TBS containing 0.05% Tween 20 (TBST). After three 5-min washes in TBST, the membranes were incubated for 1 h with allophycocyanin-conjugated goat anti-rabbit Ab (Sigma-Aldrich) diluted 1:30,000 in TBST, followed by three washes in TBST at RT. The blot was developed with Sigma FastTab nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate substrate (Sigma-Aldrich).

Surface Localization of H. pylori PgbA and PgbB. Intact H. pylori G27 (2 × 109 colony-forming units per ml) in PBS/5 mM MgCl (PBSM) were treated with 10 mg/ml proteinase K in a volume of 325 μl at 37°C for 10 min. Phenylmethylsulfonylfluoride in isopropanol was added to 160 μg/ml, and cells were centrifuged, washed twice with PBSM, and resuspended in 100 μl of PBSM. A total of 107 cells were subjected to Western blot with Ab 114.

Inhibition of Pg Binding to H. pylori. Whole H. pylori were washed in TBS, adjusted to an absorbance (600 nm) of 1, and 200 μl were coated on microtiter plates for1hatRTorONat4°C. The wells were then blocked with 1% BSA in PBS for 1 h at RT or ON at 4°C. Pg was labeled with digoxigenin (Roche) according to the manufacturer's instructions, preincubated with PgbA, PgbB, or fetuin, and added to wells in PBSB for 1 h. Digoxigenin-labeled Pg was used at a concentration of 2 nM. Inhibitors were used at the indicated concentrations. After washing with PBSB, wells were incubated for 1 h with allophycocyanin-conjugated anti-digoxigenin Ab (Roche) in PBSB at a 1:3,000 dilution. After washing, the wells were incubated with 1 mg/ml p-nitrophenyl phosphate in 1 M diethanolamine/1 mM MgCl2 at 37°C for 30 min. Absorbance was read at 405 nm.

Plasmin Chromogenic Assay. Microtiter wells were coated with 100 μl of 10 μg/ml PgbA or PgbB in PBS for 1 h at RT. A total of 200 μl of 1% BSA in PBS was used to block remaining sites ON at 4°C. All subsequent washes and incubations were performed in PBS with 1% BSA. Pg at 10 μg/ml was allowed to attach for 1 h, the wells were washed three times for 5 min each, and 0.32 μg/ml tPA (Sigma-Aldrich) was incubated for 1 h at RT. After washing, the wells were incubated with 0.5 mg/ml V-0882 (Sigma-Aldrich), a chromogenic plasmin substrate, at 37°C for 30 min. Plasmin cleavage was detected by reading absorbance at 405 nm.


Identification of Pgbs by Phage Display and Database Search. Previous studies showed that H. pylori CCUG 17874 can bind Pg, an activity attributed to 42-, 57-(20), and 58.9-kDa (21) water-extractable surface proteins. However, the genes encoding these proteins were not identified. To clone the gene(s) responsible, we constructed a phage display library and panned against immobilized Pg. Two clones containing overlapping regions were isolated (pPC18 and pPC21; Fig. 1). The combined pPC18 and pPC21 sequences possessed 97% identity on the DNA level to the 3′ end of H. pylori 26695 HP0508. Henceforth we will call HP0508 pgbA and refer to the combined 3′ gene segment (from pPC18 and pPC21) as pgbA3. Subsequent amino acid sequence analysis revealed 100% identity between the C-terminal 50–54 aa of PgbA and the C-terminal 50–54 aa of the protein encoded by HP0863, which we then designated PgbB (Fig. 1). This common C-terminal domain is lysine-rich. Because PgbA and PgbB are characteristically lysine-rich and Pg-binding is mediated by these lysine residues (8, 23, 24), this finding suggested that the protein encoded by pgbB also binds Pg. Both genes had been annotated as hypothetical proteins (25) and have no homologues in other species.

Comparison of the entire amino acid sequences of PgbA and PgbB showed only 27% identity, but both proteins have similar calculated pKi values (10.3 and 9.6, respectively), similar total lysine content (12.5% and 11.3%, respectively), and nearly identical hydrophobicity profiles with hydrophobic N termini and hydrophilic C termini (data not shown). The N termini are likely to be membrane-bound whereas the C termini, which are identical and encompass most of the binding domain identified by phage panning, are likely to be surface-exposed.

Recombinant PgbA and PgbB Bind Pg. To test the proteins for Pg-binding activity, both pgbA and pgbB were cloned without their signal sequences, expressed as His-tagged fusions, and affinity-purified by Ni-chelate chromatography. In ELISA Pg attachment assays, both proteins exhibited concentration-dependent Pg attachment (Fig. 2). In this assay, PgbA attaches to Pg ≈20% better than PgbB, although both reach saturation at 0.25 μM (Fig. 2). Henceforth, PgbA and PgbB are termed Pgbs.

Fig. 2.
Attachment of PgbA and PgbB to Pg. Various concentrations of PgbA or -B containing His-tagged fusions were incubated with Pg bound to microtiter wells. Attachment was detected by incubation with alkaline phosphatase-conjugated anti-His Ab followed by ...

To determine whether Pg bound specifically to PgbA and PgbB, we examined the ability of Pg and fetuin to inhibit the interaction between Pg and PgbA and PgbB. Fetuin is a sialic acid-containing serum glycoprotein. With Pg as an inhibitor, attachment was decreased by >80% for both PgbA and PgbB, whereas, with fetuin, attachment of PgbB was decreased by only 20% and not at all for PgbA (Fig. 3). These results demonstrate that the interactions are specific and do not depend on sialic acid.

Fig. 3.
Specificity of PgbA and PgbB attachment to Pg. Recombinant PgbA or -B containing His-tagged fusions was preincubated with Pg or fetuin before being allowed to attach to Pg-coated wells. Absorbance was taken at 405 nm in triplicate, and results are representative ...

The Interactions of PgbA and PgbB with Pg Are Lysine-Dependent. Pg has lysine-binding domains, and other Pgbs interact with Pg by means of lysine-rich domains (8, 23, 24). The Pg-binding activity of other Pgbs is characteristically inhibited by the addition of lysine or the lysine analog EACA (8, 23, 24). Because PgbA and PgbB have lysine-rich C-terminal ends, we investigated the lysine dependence of binding to Pg by adding lysine or EACA with PgbA or -B to compete for binding to Pg-coated microtiter wells (Fig. 4). At an inhibitor concentration of 150 mM, the binding of PgbA was inhibited 100% with EACA and 90% with lysine (Fig. 4A), and PgbB protein was inhibited 57% with EACA and 36% with lysine (Fig. 4B), indicating less lysine dependence for the binding of the latter to Pg. For both PgbA and PgbB, EACA was a slightly more effective inhibitor for reasons that are unknown.

Fig. 4.
Inhibition of PgbA (A) and PgbB (B) attachment to Pg with lysine and EACA. Absorbance readings were taken at 405 nm in triplicate, and results are representative of at least three independent experiments. Standard deviations are shown.

Western Blot of Multiple H. pylori Strains Using Ab Raised Against the C Terminus of PgbA. We purified recombinant PgbA3 and raised rabbit polyclonals against it. This Ab, called Ab 114, recognizes the purified recombinant PgbA and PgbB in a Western blot (data not shown). Preimmune serum does not recognize PgbA3 or recombinant PgbA and PgbB. We then used Ab 114 to determine whether PgbA and PgbB are expressed by various H. pylori strains, including laboratory-passaged strains and clinical isolates. The 11 strains tested included the common laboratory strains CCUG 17874, 26695, G27, ATCC 43504, and SS1 and clinical isolates from Poland, Turkey, Iran, Sweden, Somalia, and the U.S. Western blot analysis of these 11 strains detected two bands with apparent molecular masses of 55 and 65 kDa, roughly corresponding to the sizes of PgbA (50 kDa) and PgbB (60 kDa) (Fig. 5). No E. coli proteins were detected by Ab 114 (Fig. 5), and preimmune sera did not recognize any proteins (data not shown). The discrepancy in predicted and apparent protein molecular masses may reflect posttranslational modifications. These data demonstrate that PgbA and PgbB are widely expressed proteins. Attempts to create a deletion mutation in pgbA were not successful, suggesting that it may be essential. We did not attempt to delete pgbB.

Fig. 5.
Expression of PgbA and PgbB by H. pylori strains of diverse geographical origins. Shown is a Western blot of H. pylori whole cell lysates using Ab 114, which recognizes both PgbA and -B. Lanes: 1, Poland 43; 2, Turkey 33; 3, Iran 23; 4, Sweden 5; 5, Somalia ...

C Termini of PgbA and PgbB on H. pylori Are Surface-Localized. Previous studies demonstrated that whole H. pylori CCUG 17874 can bind radiolabeled Pg (20), suggesting that the binding moieties are located on the surface of the bacterium. To determine whether PgbA and PgbB are surface-localized, we treated whole H. pylori with proteinase K to degrade surface-exposed protein domains. Treated and untreated cells were analyzed by Western blot (Fig. 6) using Ab 114, which crossreacts with the C termini of PgbA and PgbB proteins. The antisera showed reduced reactivity against proteinase K-treated H. pylori (Fig. 6), indicating that this antigen is exposed on the surface and was degraded by proteinase K treatment. Preimmune sera were nonreactive (data not shown). These results indicate that the C-terminal region of PgbA and PgbB are exposed on the surface of H. pylori.

Fig. 6.
Surface localization of C termini of PgbA and -B. Shown is a Western blot of proteinase K-treated intact H. pylori CCUG 17874. Treated and untreated whole cell lysates were probed with Ab 114, which recognizes the C termini of both PgbA and -B. Similar ...

PgbA and PgbB Inhibit Attachment of Pg to Intact H. pylori. We wanted to determine whether recombinant PgbA and -B could inhibit the attachment of Pg to intact H. pylori. Digoxigenin-labeled Pg was preincubated with various concentrations of purified PgbA, PgbB, a combination of the two, Pg, or fetuin before being allowed to attach to intact H. pylori CCUG 17874 coated on microtiter wells. Recombinant PgbA and PgbB were able to inhibit Pg attachment to whole H. pylori whereas fetuin could not (Fig. 7). At 0.2 μM, both PgbA and PgbB individually inhibited Pg attachment by 62% and 68%, respectively, whereas the mixture inhibited by 82%. Interestingly, combining both proteins resulted in slightly greater inhibition at any concentration compared with either Pgb alone, suggesting that they act synergistically. Fetuin could not efficiently inhibit Pg attachment at any of the concentrations we used (Fig. 7). These data show that PgbA and PgbB can compete relatively efficiently for Pg compared with their native counterparts on the surface of H. pylori. That these Pgbs could not completely abolish Pg attachment to H. pylori may be due to reduced affinity of the recombinant proteins or the existence of other Pgbs that bind by means of a different site on Pg.

Fig. 7.
PgbA and -B inhibit attachment of Pg to whole H. pylori. Digoxigenin-labeled Pg was preincubated with PgbA or -B (individually or as a mix) or fetuin before being allowed to attach to H. pylori strain CCUG 17874 immobilized onto microtiter wells. For ...

Pg Bound to PgbA or PgbB Can Be Converted into Enzymatically Active Plasmin. It has been demonstrated that Pg bound to the surface of H. pylori can be activated to plasmin (20), and we have confirmed these results independently (data not shown). Although H. pylori is not an invasive organism, we hypothesize that harnessing proteolytic activity could contribute to host tissue damage (as seen in chronic ulcers) and/or nutrient release. For H. pylori to acquire such activity, PgbA and PgbB must adhere to Pg/plasmin without interfering with enzymatic function. Therefore, we asked whether Pg bound to PgbA and -B could be converted to active plasmin. Specifically, we used a chromogenic assay to measure proteolytic plasmin activity after Pg bound to PgbA or -B was incubated with a Pg activator. We demonstrated that Pg bound to either PgbA or -B was capable of being converted to functionally active plasmin (Fig. 8). No substrate cleavage was detected when either tPA or Pg was omitted from the assay or when wells were coated with BSA instead of PgbA or -B (Fig. 8). These results show that binding of PgbA or -B to Pg does not inhibit activation to plasmin and that PgbA and -B alone cannot activate Pg.

Fig. 8.
Pg bound to PgbA or -B can be activated to plasmin. Pg was allowed to attach to PgbA or -B immobilized on microtiter wells. Control wells had no Pg added. Bound Pg was then incubated with or without tPA. Activation to plasmin was assessed by development ...


Bacterial proteases are important virulence factors that can enhance tissue degradation, bacterial dissemination, and/or nutrient release (26). There is growing evidence that bacterial interaction with host Pg is a common method for enhancing virulence (27). There are numerous examples of microorganisms that bind Pg, which can subsequently be activated to plasmin, thereby arming the bacterium with a surface-localized protease that can mediate degradation of extracellular matrix and tissue penetration (for a review, see ref. 7). The role of Pg in bacterial pathogenesis has been confirmed in animal studies with group A streptococci (13), Y. pestis (16), B. burgdorferi (14), and B. crocidurae (15).

Relatively little has been done in the area of Pg interaction with H. pylori, although such binding has been known since 1994 (28) and the significance in other organisms is well established. In this article, we present the characterization of two Pgbs universally expressed by diverse H. pylori strains (Fig. 5). We hypothesize that H. pylori Pg binding contributes to the chronic gastric ulcers commonly associated with infection based on the following rationale. Pg and Pg activators are recruited to sites of tissue injury, e.g., a site of gastric ulceration. Although primarily found in the mucous, H. pylori also adheres to the gastric epithelium and is present at ulcer edges (29). Acquisition of Pg/plasmin could enable the motile bacterium to degrade host components with which it comes into contact, possibly as a mechanism of nutrient acquisition. Several extracellular matrix proteins including laminin, vitronectin, and fibronectin are susceptible to plasmin cleavage. Fibronectin is produced in response to tissue injury and is necessary for wound healing; its loss is correlated with ulcers (30), and its expression is increased during gastric ulcer healing (31). Once an ulcer forms, H. pylori pgbA and pgbB may contribute to its chronicity by obstructing the natural healing process.

pgbA was cloned from a phage display library panned against Pg, and pgbB was identified by a protein blast search of PgbA. Because the Pg panning procedure is a functional screen and the two clones identified had overlapping sequences (Fig. 1), we were not surprised to find that recombinant PgbA and PgbB did in fact bind Pg in vitro (Fig. 2). The recombinant proteins behaved as expected for Pgbs based on studies in other organisms (8, 23, 24). Both exhibited concentration-dependent (Fig. 2), specific (Fig. 3) attachment to Pg that depended on lysine (Fig. 4). Although both PgbA and PgbB have similar total lysine contents (12.5% for PgbA and 11.3% for PgbB), the former binds Pg better (Fig. 2) and with greater lysine dependence (Fig. 4), suggesting structural dissimilarities that obscure the accessibility of some lysines in the latter. Despite their similarities, PgbA and PgbB likely are not functionally redundant proteins.

It is unclear whether pgbA and pgbB encode proteins that were previously described. The apparent molecular masses of native PgbA and -B are 55 and 65 kDa, respectively, by Western blot (Fig. 5), whereas Pantzar et al. (20) identified two proteins in CCUG 17874 (57 and 42 kDa) with Pg-binding activity, and Yarzabal (21) identified a protein of 58.9 kDa in a clinical isolate from Venezuela. Intriguingly, PgbA and PgbB share an identical stretch of 50–54 aa at their C termini (Fig. 1). The rest of the proteins also share substantial identity, although PgbA is lacking several regions (Fig. 1), suggesting that pgbA evolved from a duplication of pgbB followed by multiple deletion events. Perhaps the N-terminal regions diverged such that they bind each other to display multiple binding sites for cooperative binding. The binding affinity then could be fine-tuned by simply up- or down-regulating PgbA and -B expression. The idea that pgbA and pgbB work synergistically is supported by the finding that a mixture of PgbA and PgbB could inhibit H. pylori binding to Pg slightly better than either Pgb alone (Fig. 7). In the higher-concentration range of inhibitor, this was true when comparing equal molarities of total Pgbs (compare 0.2 μM individual Pgb with 0.1 μM mixture) (Fig. 7).

Minimally, the binding domain(s) identified by phage display are localized to the outer surface (Fig. 6), so the sites are available for anchoring Pg to the bacterial surface. If PgbA and PgbB do function in anchoring an active protease, binding of PgbA or -B to Pg/plasmin must not interfere with protease activity. We showed that Pg bound to PgbA or -B can be activated to plasmin (Fig. 8), indicating that surface proteolytic acquisition could occur in presence of host Pg and Pg activator.

Taken together, our results show that pgbA and pgbB encode surface-exposed proteins that mediate binding to Pg such that it can be converted into plasmin in the presence of a Pg activator. Harnessing protease activity could contribute to virulence by perpetuating host tissue damage (as seen in chronic ulcers). The universal expression of both PgbA and PgbB by a variety of strains (Fig. 5) and the conserved C-terminal domain (Fig. 1) make them promising candidates for H. pylori vaccine development, diagnostic technology, or drug targeting.


We thank B. Wiman for the gift of Pg and S. Chiang for critical reading of the manuscript. This project was supported by National Institutes of Health Grant AI026289 (to J.J.M.) and Karolinska Hospital research support (to G.K.).


Abbreviations: Pg, plasminogen; Pgb, Pg-binding protein; tPA, tissue Pg activator; RT, room temperature; EACA, ε-aminocaproic acid; ON, overnight.

Data deposition: The sequence reported in this paper for H. pylori CCUG 17874 pgbA has been deposited in the GenBank database (accession no. AJ550456).


1. Matysiak-Budnik, T. & Megraud, F. (1997) J. Physiol. Pharmacol. 48, Suppl. 4, 3–17. [PubMed]
2. Dunn, B. E., Campbell, G. P., Perez-Perez, G. I. & Blaser, M. J. (1990) J. Biol. Chem. 265, 9464–9469. [PubMed]
3. Seyler, R. W., Jr., Olson, J. W. & Maier, R. J. (2001) Infect. Immun. 69, 4034–4040. [PMC free article] [PubMed]
4. Kavermann, H., Burns, B. P., Angermuller, K., Odenbreit, S., Fischer, W., Melchers, K. & Haas, R. (2003) J. Exp. Med. 197, 813–822. [PMC free article] [PubMed]
5. Gebert, B., Fischer, W., Weiss, E., Hoffmann, R. & Haas, R. (2003) Science 301, 1099–1102. [PubMed]
6. Amieva, M. R., Vogelmann, R., Covacci, A., Tompkins, L. S., Nelson, W. J. & Falkow, S. (2003) Science 300, 1430–1434. [PMC free article] [PubMed]
7. Lahteenmaki, K., Kuusela, P. & Korhonen, T. K. (2001) FEMS Microbiol. Rev. 25, 531–552. [PubMed]
8. Kuusela, P., Ullberg, M., Saksela, O. & Kronvall, G. (1992) Infect. Immun. 60, 196–201. [PMC free article] [PubMed]
9. Ullberg, M., Kuusela, P., Kristiansen, B. E. & Kronvall, G. (1992) J. Infect. Dis. 166, 1329–1334. [PubMed]
10. Lahteenmaki, K., Virkola, R., Pouttu, R., Kuusela, P., Kukkonen, M. & Korhonen, T. K. (1995) Infect. Immun. 63, 3659–3664. [PMC free article] [PubMed]
11. Sjostrom, I., Grondahl, H., Falk, G., Kronvall, G. & Ullberg, M. (1997) Biochim. Biophys. Acta 1324, 182–190. [PubMed]
12. Hu, L. T., Perides, G., Noring, R. & Klempner, M. S. (1995) Infect. Immun. 63, 3491–3496. [PMC free article] [PubMed]
13. Li, Z., Ploplis, V. A., French, E. L. & Boyle, M. D. (1999) J. Infect. Dis. 179, 907–914. [PubMed]
14. Coleman, J. L., Gebbia, J. A., Piesman, J., Degen, J. L., Bugge, T. H. & Benach, J. L. (1997) Cell 89, 1111–1119. [PubMed]
15. Nordstrand, A., Shamaei-Tousi, A., Ny, A. & Bergstrom, S. (2001) Infect. Immun. 69, 5832–5839. [PMC free article] [PubMed]
16. Sodeinde, O. A., Subrahmanyam, Y. V., Stark, K., Quan, T., Bao, Y. & Goguen, J. D. (1992) Science 258, 1004–1007. [PubMed]
17. Goto, H., Wells, K., Takada, A. & Kawaoka, Y. (2001) J. Virol. 75, 9297–9301. [PMC free article] [PubMed]
18. Fischer, M. B., Roeckl, C., Parizek, P., Schwarz, H. P. & Aguzzi, A. (2000) Nature 408, 479–483. [PubMed]
19. Shaked, Y., Engelstein, R. & Gabizon, R. (2002) J. Neurochem. 82, 1–5. [PubMed]
20. Pantzar, M., Ljungh, A. & Wadstrom, T. (1998) Infect. Immun. 66, 4976–4980. [PMC free article] [PubMed]
21. Yarzabal, A. (2000) Braz. J. Biol. Res. 33, 1015–1021. [PubMed]
22. Higgins, D. G., Thompson, J. D. & Gibson, T. J. (1996) Methods Enzymol. 266, 383–402. [PubMed]
23. Berge, A. & Sjobring, U. (1993) J. Biol. Chem. 268, 25417–25424. [PubMed]
24. Kuusela, P. & Saksela, O. (1990) Eur. J. Biochem. 193, 759–765. [PubMed]
25. Tomb, J. F., White, O., Kerlavage, A. R., Clayton, R. A., Sutton, G. G., Fleischmann, R. D., Ketchum, K. A., Klenk, H. P., Gill, S., Dougherty, B. A., et al. (1997) Nature 388, 539–547. [PubMed]
26. Travis, J., Potempa, J. & Maeda, H. (1995) Trends Microbiol. 3, 405–407. [PubMed]
27. Lahteenmaki, K., Kuusela, P. & Korhonen, T. K. (2000) Methods 21, 125–132. [PubMed]
28. Ringner, M., Valkonen, K. H. & Wadstrom, T. (1994) FEMS Immunol. Med. Microbiol. 9, 29–34. [PubMed]
29. Chan, W. Y., Hui, P. K., Chan, J. K., Cheung, P. S., Ng, C. S., Sham, C. H. & Gwi, E. (1991) Histopathology 19, 47–53. [PubMed]
30. Berman, M., Manseau, E., Law, M. & Aiken, D. (1983) Invest. Ophthalmol. Vis. Sci. 24, 1558–1566.
31. Gillessen, A., Shahin, M., Pohle, T., Foerster, E. & Domschke, W. (1995) J. Physiol. Pharmacol. 46, 57–62. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...