OUP user menu

A role for the bacterial outer membrane in the pathogenesis of Helicobacter pylori infection

Jacqueline Keenan , Tony Day , Stephanie Neal , Bramwell Cook , Guillermo Perez-Perez , Randall Allardyce , Philip Bagshaw
DOI: http://dx.doi.org/10.1111/j.1574-6968.2000.tb08905.x 259-264 First published online: 1 January 2000


Helicobacter pylori infection in humans is associated with diverse of clinical outcomes which are partly attributed to bacterial strain differences. Secreted bacterial products are thought to be involved in the pathogenesis caused by this non-invasive bacterium. Electron microscopy of gastric biopsies from infected individuals revealed blebbing of the H. pylori outer membrane, similar to the process of outer membrane vesicle shedding which occurs when the bacterium is grown in broth. Porins, a class of proinflammatory proteins, were observed in the outer membrane vesicles. The VacA cytotoxin, which is produced by 50–60% of H. pylori strains and associated with increased pathogenesis of infection, was also found to be vesicle-associated and biologically active. This supports the hypothesis that these vesicles represent a vehicle for the delivery of damaging bacterial products to the gastric mucosa.

  • Outer membrane
  • Pathogenesis
  • Helicobacter

1 Introduction

Helicobacter pylori has emerged over the last decade as probably the most common chronic infection affecting humans, estimated to involve more than half the people in the world. Most H. pylori infections are subclinical, causing only mild gastric inflammation and mucosal damage. However, a minority of people develop clinically relevant pathology as a consequence of the infection. Several factors, including H. pylori strains differences are attributed with causing this diversity in clinical outcome [1].

Since H. pylori is generally considered non-invasive [2] it is likely that secreted bacterial products play a role in the pathogenesis of this infection. Supporting this hypothesis is a recent study identifying six antigens (including urease) which are released from H. pylori during growth in vitro [3]. The urease enzyme is also detected in the antral lamina propria of H. pylori-infected individuals [4].

Helicobacter spp. shed outer membrane as vesicles when grown in broth culture [5]. This phenomenon is not unusual and is especially prominent in Gram-negative bacteria [6]. We hypothesized that the release of similar vesicles in vivo could contribute to the pathogenesis of H. pylori infection by acting as a vehicle for chronic the delivery of damaging moieties to the host's mucosas, thereby contributing to the ensuing, persistent inflammation which is characteristic of this infection. To investigate this hypothesis, we examined gastric biopsies from H. pylori-infected individuals by electron microscopy (EM) for the presence of bacterial outer membrane vesicle shedding. We then characterized membrane vesicles (harvested from H. pylori type strains in vitro in order to understand the impact these vesicles might have on the host mucosa.

2 Materials and methods

2.1 Preparation of biopsies for EM

Antral biopsies from six randomly selected CLO-test positive individuals were fixed in 2.5% glutaraldehyde (in PIPES buffer) for 30 min at RT, followed by three 5-min washes in PIPES buffer. The samples were post-fixed in osmium tetroxide and block stained with uranyl acetate maleate before being dehydrated and embedded in Spurr's resin.

2.2 Bacterial strains

Three well characterized H. pylori type-strains were used to characterize the membrane vesicles shed by these bacteria. Two of these strains, H. pylori 60190 and 84-183 are known to produce both vaculating cytotoxin (VacA) and the cytotoxin-associated protein (CagA) [7,8] The third strain, H. pylori Tx-30a, fails to produce detectable cytotoxin activity in vitro [7] and lacks cagA[8]. Each strain was individually cultured in 25-ml volumes of 2.8% (w/v) brucella broth base (Difco, Detroit, MI, USA), supplemented with 5% fetal calf serum (Gibco BRL, New Zealand). Cultures were incubated at 37°C in a microaerobic environment (10% hydrogen, 10% carbon dioxide and 80% nitrogen) with constant rotation (120 rpm).

2.3 Outer membrane vesicles

Whole cells were removed from 48–72 h broth cultures by centrifugation (10 000×g, 15 min) and washed three times with PBS. Outer membrane vesicles were recovered from the spent culture supernatants by ultracentrifugation (100 000×g, 2 h, 4°C) and washed twice in PBS [5]. The absence of whole cells and flagella in each membrane vesicle preparation was confirmed by EM after samples were overlaid onto carbon-colloidin coated mesh grids and negatively stained with 1% aqueous phosphotungstic acid (pH 7.0). The protein concentration of each fraction was assayed [9]. All samples were stored at −20°C until required.

2.4 Electrophoresis and dual silver staining

The whole cell and outer membrane vesicle fractions were separated by discontinuous SDS-PAGE under reducing conditions on 12.5% acrylamide gels and fixed overnight (40% ethanol, 5% acetic acid). Dual silver staining [10], which permits simple protein and carbohydrate characterization of bacterial fractions, was used to compare the protein and LPS profiles of the whole cell and membrane vesicle fractions. Following fixation, one of two duplicate gels was gently agitated in 0.7% periodic acid (in fixative) for 10 min to permit oxidation of hexoses to aldehydes. It is this modification of the conventional silver stain which allows the visualization of carbohydrate as well as protein moieties. Thereafter, both gels were silver stained using a conventional method [10].

2.5 Assay for urease activity

The urease activity of the H. pylori whole cell and membrane vesicle fractions was assessed by the enzymatic hydrolysis of urea in a quantitative spectrophotometric assay [11]. Briefly, a dilution of sample was added to cuvettes (1-cm light path length) containing 3 ml of a reaction mixture of 31 mM Tris-HCl (pH 8), 810 μM oxoglutarate, 240 μM NADH and 10 mM urea. The reaction was started by adding 96 U of glutamate dehydrogenase and the reduction of NADH was followed spectrophotometrically for 10 min at 37°C, with a standard wavelength of 340 nm in a dual beam spectrophotometer. One unit of urease activity was defined as that amount capable of hydrolyzing 1 μM urea per min. A urease negative mutant of H. pylori 60190 provided a control for this experiment.

2.6 Immunological characterization of H. pylori membrane vesicles

SDS-PAGE-separated H. pylori membrane vesicle components were electrophoretically transferred from 12.5% acrylamide gels to 0.45 μm nitrocellulose membrane and probed with antisera raised in rabbit to well characterized H. pylori antigens (urease B subunit, 54-kDa heat shock protein (HspB), CagA protein and VacA cytotoxin). The membranes were incubated with primary antibody for 1 h at RT and bound rabbit antibodies detected with an alkaline phosphatase-conjugated secondary antibody (Sigma). Reactive bands were visualized using 5-bromo-4-chloro-3-indolylphosphate as the alkaline phosphatase substrate and nitroblue tetrazolium as the color development reagent [12]. Whole cells provided the control for these experiments.

2.7 Assay for HEp-2 cell vacuolation

Washed membrane vesicles from H. pylori 60190 were incubated with 104 HEp-2 cells (a human larynx epidermoid carcinoma cell line with HeLa markers) in 96-well microtiter plates [13]. Visible vacuolation was assessed by light microscopy; a cytotoxic effect was ascribed to those wells in which >50% of cells were vacuolated. Outer membrane vesicles produced by an isogenic vacA−ve mutant strain of H. pylori 60190 [14] were used as a control for this experiment.

2.8 Ultrastructural immunolocalization of VacA

Antibody to VacA was used to immunolabel the outer membrane of H. pylori. Washed whole cells (H. pylori 60190) were suspended in low salt phosphate buffer (LSP) and overlaid onto carbon-colloidin coated copper grids. Following blocking (30 min in 0.1% BSA in LSP), the grids were incubated in rabbit anti-VacA antisera for 1 h at RT. Unbound primary antibody was removed by three 2-min LSP washes before the grids were incubated in goat anti-rabbit IgG (whole molecule) conjugated to 10-nm gold particles (Sigma). Following three more 2-min washes in LSP, the grids were air-dried and examined by EM. Ultrathin sections of H. pylori+ve gastric biopsies (see above) were overlaid onto carbon-coated nickel grids (300-mesh) and immunolabelled, according to the method of Sommi et al. [15]. Non-specific binding sites were blocked with 1% non-immune goat serum, NGS, diluted in 0.1% BSA in LSP for 1 h at RT. The grids were then incubated overnight in primary antibody (diluted in blocking buffer). After three 10-min LSP-BSA washes, bound primary antibodies were detected using a gold-labelled secondary antibody (see above). Unbound secondary antibody was removed with three more 15-min washes in LSP and the grids stained with lead before being examined by EM.

3 Results and discussion

EM revealed H. pylori outer membrane blebbing in every biopsy examined whether the bacterial outer membrane was adjacent to the gastric epithelium or not (Fig. 1a). Outer membrane appendages have also been described on Neisseria meningitidis in the cerebrospinal fluid of an infant [16] and lends weight to our hypothesis that membrane vesicles contribute to the pathogenesis of H. pylori infection in humans by providing a vehicle for the delivery of toxins and proinflammatory products to the gastric mucosa.


Electron micrograph. a: H. pylori in situ in a gastric biopsy from a patient with antral gastritis. b: Whole cells. c: Outer membrane vesicles recovered following the growth of H. pylori 60190 in broth. Arrows indicate blebbing of the H. pylori outer membrane. (Bar=100 nm)

Similar blebbing of the outer membrane (Fig. 1b) is a feature when H. pylori is cultured in broth [5]. Ultracentrifugation of the spent supernatant results in a pellet of outer membrane vesicles (Fig. 1c). Dual silver staining, which revealed few inter-strain differences between either the whole cells or membrane vesicle protein profiles (Fig. 2a), highlighted the increased ratio of carbohydrate to protein in the membrane vesicles (Fig. 2b). This confirms that the outer membrane of H. pylori strains contain almost all the LPS of the cell [17] and also highlights the marked heterogeneity in LPS content between the strains [18]. H. pylori LPS, however, is known to have low biological activity in mediating macrophage activation [19], inducing mitogenicity in mouse spleen cells, pyrogenicity in rabbits and toxic lethality in galactosamine-sensitized mice [20] when compared to other Gram-negative bacteria.


SDS-PAGE (12.5% acrylamide) separated H. pylori cells and outer membrane vesicles (2 μg protein/lane). a: Conventional silver stain for protein. b: Modified silver stain for protein/carbohydrate. Lane (1) H. pylori Tx-30a; (2) H. pylori 60190; (3) H. pylori 84-183. Molecular mass standards (kDa) are indicated (left).

The H. pylori outer membrane has been shown to contain porin proteins [21]. These ubiquitous proteins in Gram-negative bacteria form water-filled channels across the membrane and play a role in selective permeability[17]. Unusually resistant to heating in strong detergents, porins exhibit a different Mr during SDS-PAGE, depending upon the temperature of solubilization [17,21] Several heat-modifiable proteins were observed in the membrane vesicles when electrophoresed (Fig. 3). This observation shows the potential of membrane vesicles to deliver these proinflammatory and cytokine-inducing proteins [22] to the gastric mucosa.


Silver stained SDS-PAGE (12.5%) of outer membrane vesicles (2 μg protein/lane) from three H. pylori type strains heated at 100°C for 10 min (lane 1) or unheated (lane 2). a: H. pylori Tx-30a. b: H. pylori 60190. c: H. pylori 84-183. Molecular mass standards (kDa) are indicated (left).

No measurable urease activity was detected in the membrane vesicles of the three H. pylori type strains tested (Table 1). Furthermore, immunoblotting with specific antisera failed to identify either the urease B subunit or its associated chaperonin (HspB) in the vesicles (results not shown). This suggests that the demonstration of urease enzyme subunits in sarcosyl-insoluble outer membrane preparations [23] may reflect inner membrane contamination in these samples.

View this table:

Urease activity of H. pylori whole cell and membrane vesicle fractionsa

StrainWhole cellsVesicles
60190 (urease −ve)0.03<0.01
  • aAll values mg−1 protein

Fifty to sixty percent of H. pylori strains express VacA, an extracellular cytotoxin [7]. The majority of these strains also produce an immunodominant, surface associated cytotoxin-associated (CagA) protein [8]. Numerous studies now show that H. pylori strains which express these products are associated with increased pathogenesis of infection. Immunoblotting identified the 87-kDa VacA protein in the outer membrane (Fig. 4a) and vacuolation of HEp-2 cells showed the membrane-associated cytotoxin to be biologically active (Fig. 4b). This finding correlates with the high levels of biologically active heat-labile enterotoxin (LT) found in shed outer membrane vesicles from enterotoxigenic E. coli[24]. EM of immunolabelled H. pylori confirmed that much of the cytotoxin is membrane-associated during growth of H. pylori in vitro (Fig. 5a) and in vivo (Fig. 5b). In contrast, CagA was not present in the membrane vesicle fraction (results not shown).


a: Immunoblot analysis of rabbit anti-VacA reactivity to H. pylori cells (lanes 1–3) and outer membrane vesicles (lanes 4–6) (2 μg protein/lane). Lanes (1,4) H. pylori Tx-30a; (2,5) H. pylori 60190; (3,6) H. pylori 84-183. Molecular mass markers (kDa) are indicated (left). b: HEp-2 cells incubated for 3 h with washed outer membrane vesicles from H. pylori 60190 (top) and H. pylori 60190 isogenic vacA−ve mutant (bottom). Visible vacuolation occurred in the majority of cells in response to the vesicle-associated toxin.


Immunolabelling of H. pylori a: in vitro b: in vivo, following incubation with a polyclonal rabbit antibody against VacA. Primary antibody binding was detected by a gold-labelled goat anti-rabbit IgG antibody. Membrane-bound (solid arrow) and soluble (broken arrow) forms of H. pylori VacA are indicated.

In conclusion, we found that outer membrane vesicles are shed from the surface of infective H. pylori. The presence of porins and biologically active VacA cytotoxin in these vesicles strongly suggests an offensive role for these structures in the pathogenesis of H. pylori infection. We are currently testing whether the variable release and/or adherence of these vesicles in vivo relates to the differing patterns of H. pylori-associated disease.


We thank Dr. T.L. Cover (Department of Medicine, Vanderbilt University School of Medicine) for providing the H. pylori 60190 vacA−ve mutant. This research was generously supported by the Canterbury Medical Research Foundation who made J.K. the recipient of a Canterbury Medical Researcher Interchange Award.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
View Abstract