|Home | About | Journals | Submit | Contact Us | Français|
Helicobacter pylori bacteria colonize the human stomach where they stimulate a persistent inflammatory response. H. pylori is considered noninvasive; however, lipopolysaccharide (LPS)-enriched outer membrane vesicles (OMV), continuously shed from the surface of this bacterium, are observed within gastric epithelial cells. The mechanism of vesicle uptake is poorly understood, and this study was undertaken to examine the roles of bacterial VacA cytotoxin and LPS in OMV binding and cholesterol and clathrin-mediated endocytosis in vesicle uptake by gastric epithelial cells. OMV association was examined using a fluorescent membrane dye to label OMV, and a comparison was made between the associations of vesicles from a VacA+ strain and OMV from a VacA− isogenic mutant strain. Within 20 min, essentially all associated OMV were intracellular, and vesicle binding appeared to be facilitated by the presence of VacA cytotoxin. Uptake of vesicles from the VacA+ strain was inhibited by H. pylori LPS (58% inhibition with 50 μg/ml LPS), while uptake of OMV from the VacA− mutant strain was less affected (25% inhibition with 50 μg/ml LPS). Vesicle uptake did not require cholesterol. However, uptake of OMV from the VacA− mutant strain was inhibited by a reduction in clathrin-mediated endocytosis (42% with 15 μg/ml chlorpromazine), while uptake of OMV from the VacA+ strain was less affected (25% inhibition with 15 μg/ml chlorpromazine). We conclude that VacA toxin enhances the association of H. pylori OMV with cells and that the presence of the toxin may allow vesicles to exploit more than one pathway of internalization.
Infection with the gastric pathogen Helicobacter pylori results in chronic gastritis (13) and is associated with increased risk for the development of peptic ulcer disease (35), gastric carcinoma (41, 57), and gastric lymphoma (5, 60). H. pylori persistence, in an environment where peristalsis and sloughing of cells are continually occurring, is mediated by a variety of adhesins present on the bacterial surface (14, 21, 36, 40). However, despite the ability to adhere to the gastric epithelium, the majority of organisms remain unattached to surface cells (32), leading to speculation that lipopolysaccharide (LPS)-enriched outer membrane vesicles (OMV) shed by these bacteria (15, 19, 26) contribute to H. pylori pathogenesis via the persistent delivery of bacterial virulence factors (including the vacuolating cytotoxin VacA) and antigens to the gastric mucosa (26, 27). Observations that H. pylori OMV modulate gastric epithelial cell proliferation (22), induce apoptosis (3), stimulate secretion of the proinflammatory cytokine interleukin-8 (22), increase micronucleus formation (8), and are at the luminal surface (15, 26) and within cells of the gastric epithelium (15) support this hypothesis.
OMV shedding by Gram-negative bacteria is well described in the literature (reviewed by Kuehn and Kesty ), yet little is known of the mechanisms of vesicle adherence to and internalization within mammalian host cells. The adherence of enterotoxigenic Escherichia coli (ETEC) OMV to host cells is mediated via a heat-labile enterotoxin (LT) associated with these OMV (31), whereas leukotoxin, associated with OMV shed by Actinobacillus actinomycetemcomitans, is not involved in vesicle binding (12). The internalization of ETEC, Porphyromonas gingivalis, and Pseudomonas aeruginosa OMV has been shown to involve cholesterol-rich lipid rafts (6, 16, 31), and recently, Kaparakis and colleagues (25) reported that uptake of H. pylori OMV is also lipid raft dependent. This is in contrast to the uptake of Shigella flexneri OMV, which occurs via phagocytosis, with the proposed subsequent fusion of OMV with the phagosomal membrane and the release of vesicle contents into the cell cytoplasm (24).
In this study, we sought to examine whether VacA cytotoxin associated with H. pylori OMV was involved in vesicle binding. We also examined the rate of OMV internalization and the involvement of LPS, cholesterol, and clathrin-mediated endocytosis in vesicle uptake by AGS gastric epithelial cells. We report that within 20 min essentially all VacA+ OMV associated with AGS cells were intracellular and that uptake was enhanced by the presence of vesicle-associated cytotoxin. Excess H. pylori LPS reduced vesicle uptake, having a more significant effect on VacA+ OMV than VacA− vesicle uptake. Uptake of both VacA+ and VacA− OMV did not require cholesterol. However, a reduction in clathrin-mediated endocytosis inhibited VacA− OMV uptake and to a lesser extent VacA+ OMV internalization.
The well-characterized H. pylori clinical isolate 60190 (ATCC 49503) was used in this study. H. pylori 60190 is a vacA s1m1 strain (2) producing a cytotoxin that induces extensive vacuolation in cultured epithelial cells (9). H. pylori strain 60190:v1, in which the VacA gene has been disrupted by insertional mutagenesis resulting in both the absence of the 87-kDa protein and vacuolating cytotoxin production (11), was also used.
H. pylori was grown in 2.8% (wt/vol) Brucella broth (Becton Dickinson, Sparks, MD) supplemented with 5% fetal bovine serum (Invitrogen, Auckland, New Zealand) at 37°C under microaerobic conditions with constant rotation (120 rpm). At 72 h of incubation, bacteria were removed by two centrifugations (12,000 × g, 15 min, 4°C), and the final supernatants ultracentrifuged (200,000 × g, 2 h, 4°C) to recover OMV. After three washes in phosphate-buffered saline (PBS), OMV were stored at −20°C until required. Vesicles to be labeled were washed once after recovery and then, to standardize labeling, were suspended in 1 ml PBS/50 mg OMV pellet. OMV were labeled with 1% (vol/vol) 3,3′-dioctadecyloxacarbocyanine perchlorate or 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine perchlorate (Vybrant DiO or Vybrant DiD, respectively; Molecular Probes, Eugene, OR) by incubation for 20 min at 37°C. Free dye was removed by two washes with PBS, and labeled OMV were stored at 4°C for up to 6 weeks. The protein concentration of OMV was determined (37), and preparations were examined by transmission electron microscopy (TEM) to confirm the absence of whole bacteria and flagella by overlaying aliquots of OMV suspension onto Formvar-coated 200-mesh copper grids and negatively staining with 1% ammonium molybdate (pH 7.4).
The AGS human gastric adenocarcinoma cell line (ATCC CRL-1739) was cultured in F-12 nutrient mixture (Ham) (plus l-glutamine) (Invitrogen, Auckland, New Zealand) and supplemented with 10% (vol/vol) fetal bovine serum and 1% (vol/vol) penicillin-streptomycin-glutamine supplement. Cells were cultured at 37°C with 5% CO2.
Cell proliferation was determined using a colorimetric assay that measures the amount of 5-bromo-2′-deoxyuridine (BrdU) incorporated into cellular DNA (Roche Diagnostics, Manneheim, Germany). AGS cells (1 × 104) were cultured overnight in 96-well plates and then incubated for a further 24 h with labeled or unlabeled OMV from strain 60190 (0.05 to 20 μg). The BrdU assay was performed per the kit instructions. Briefly, individual wells were incubated with BrdU labeling solution (10 μM BrdU) for 2 h. After removal of the medium, a FixDenat solution was added at room temperature for 30 min and then replaced with anti-BrdU-peroxidase conjugate (diluted 1:100) for 2 h. After washing, substrate solution was applied, and absorbance read at 370 nm every minute for 30 min with a SpectraMax plate reader (Molecular Devices). The level of BrdU incorporation was calculated from the slope of the linear portion of the graph.
Cells (3 × 105) were cultured overnight and then washed once with PBS, and the medium was replaced prior to the addition of labeled OMV (200 μg OMV protein) for up to 6 h. After incubation, cells were washed to remove unbound OMV and lifted with trypsin/EDTA (Invitrogen). Fluorescence measurements were made using a fluorescence-activated cell sorter (FACS) vantage flow cytometer (Beckman Coulter Cytomics FC 500 MPL; Australia) and CXP software (Beckman Coulter 2005). A total of 10,000 events were collected for each sample. Mean fluorescence intensity (MFI) values of cells incubated in the absence of OMV were subtracted from the values of OMV-treated cells.
To determine the proportion of internalized OMV, extracellular vesicle fluorescence was quenched with trypan blue (0.025% final concentration). To confirm trypan blue quenched DiO fluorescence, cells were incubated with labeled OMV for 4 h, fixed, and permeabilized using a Fix and Perm cell permeabilization kit (Caltag Laboratories, Burlingame, CA) per the manufacturer's instructions. Fluorescence was measured before and after the addition of trypan blue.
AGS cells (3 × 105), cultured overnight on glass coverslips, were incubated with 200 μg DiO-labeled OMV for 5 h. Following fixation (4% paraformaldehyde, 45 min), the cells were washed extensively in PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)] buffer, and coverslips mounted in SlowFade without glycerol (Molecular Probes, Eugene, OR). Where immunofluorescence microscopy was used to visualize intracellular OMV, the cells were blocked with 5% (wt/vol) bovine serum albumin (BSA) for 60 min at room temperature. After washing, the cells were permeabilized with 0.5% (vol/vol) Triton X-100 for 20 min and washed again before being incubated overnight at 4°C with a polyclonal antibody to H. pylori OMV (27). Primary antibody binding was detected using a phycoerythrin-conjugated goat anti-mouse secondary antibody. Cells were examined and imaged using a Leitz Aristoplan microscope (Germany) fitted with a Photometrics KAF1400 CCD camera and QUIPS Smartcapture software, version 1.3 (Vysis Inc., IL). Images were formatted in Adobe Photoshop (CS2).
To investigate the role of LPS in OMV association, DiO-labeled OMV were added to cells at the same time as H. pylori LPS and then cells were incubated for 4 h before fluorescence measurements. The effect of several pharmacological agents (cycloheximide, nystatin, chlorpromazine, and methyl-β-cyclodextrin (MβCD); Sigma, St. Louis, MO) on OMV uptake was examined by pretreating cells with each drug for the following times: 1 h, nystatin and MβCD; 30 min, chlorpromazine; and 15 min, cycloheximide. Cells were then incubated with labeled OMV for 2 to 4 h. Control cells were incubated without inhibitor for the same period of time. To assess drug cytotoxicity, AGS cells were incubated in media alone or with the highest concentration of each inhibitor for 4 h and then stained with 0.75% propidium iodide for 10 min, and fluorescence was measured by flow cytometry. Cytotoxicity was expressed as the percentage of propidium iodide-positive cells. To assess inhibition of clathrin-mediated endocytosis, cells were incubated with DiO-labeled low-density lipoprotein (LDL) (5 to 15 μg) for 4 h.
Extraction of H. pylori LPS was performed using a conventional hot phenol-water treatment (23). Subsequent purification of this crude LPS was performed by using enzymatic treatments with RNase A, DNase II, and proteinase K (all from Sigma) as described previously (38).
AGS cells (3 × 105) were cultured overnight and washed once with PBS prior to incubation for 1 h at 37°C in F12 media containing 5 mg/ml of MβCD. The cholesterol content of cells was measured using the cholesterol CHOD-PAP reagent (Roche Diagnostics, Indianapolis, IN).
Cells (3 × 104) were cultured overnight in 96-well microtiter plates and incubated with OMV (10 to 50 μg OMV protein) for 6 h. Untreated AGS cells were used as controls for background vacuolation. Vacuolation was assessed by staining with neutral red as previously described (10).
Results are the means ± standard errors (SE) of the means. Data were analyzed by t test, Pearson's correlation, one-way analysis of variance (ANOVA), or two-way ANOVA. If the ANOVA P value was <0.05, the ANOVA was followed by Dunnett's posthoc test.
We and others have previously shown that H. pylori OMV are internalized by gastric epithelial cells (8, 15, 44, 45). In this study, we sought to examine the rate and mechanism of OMV internalization using the lipophilic fluorescent membrane dye DiO to monitor OMV localization. The addition of labeled OMV to AGS cells for 24 h induced a 73% decrease in proliferation (27% ± 4% of control) similar to that induced by unlabeled OMV (30% ± 2% of control [mean ± SE; n = 3]), indicating that labeling did not interfere with the interaction of vesicles with AGS cells.
Internalization of OMV was assessed quantitatively by flow cytometry using the cell impermeant dye trypan blue to quench extracellular DiO-vesicle fluorescence. The effectiveness of trypan blue at quenching DiO-OMV fluorescence was confirmed by incubating cells with labeled OMV and comparing the fluorescence of permeabilized and nonpermeabilized cells treated with this dye. Complete loss of DiO fluorescence was observed in permeabilized cells upon addition of trypan blue, while nonpermeabilized cells showed no change in fluorescence (Fig. 1A and B). To examine the rate of OMV internalization, AGS cells were incubated with DiO-labeled OMV for up to 4 h, and fluorescence was measured before (total associated OMV) and after (intracellular OMV) the addition of trypan blue. A linear increase in total associated OMV was observed over the 4 h, with a Pearson's coefficient correlation, r value, of 0.997 (P = 0.003) (Fig. (Fig.1C).1C). At each time point assessed, the majority of the OMV associated with the cells were intracellular (Fig. (Fig.1C).1C). Examination of earlier time points showed that internalization of OMV peaked at 20 min and then plateaued (Fig. (Fig.1D1D).
The steady increase in OMV association with cells over time could be explained by specific vesicle binding that is dependent on upregulation or recycling of receptors after internalization of OMV. To examine whether vesicle binding stimulates de novo synthesis of OMV receptors, AGS cells were pretreated with increasing concentrations of the protein synthesis inhibitor cycloheximide. Inhibition of protein synthesis had no effect on the association of OMV with AGS cells (P = 0.878, ANOVA) (Fig. (Fig.1E1E).
We have previously shown that VacA is localized on the surface of H. pylori OMV (26) and that OMV-associated cytotoxin is biologically active (22). These findings, plus evidence that heat-labile enterotoxin has a role in the internalization of OMV shed from enterotoxigenic E. coli (31), led us to consider that vesicle-associated VacA may also play a role in the uptake of H. pylori OMV. Uptake of OMV from the vacuolating strain 60190 was compared to that of its isogenic mutant 60190:v1 that does not produce a toxin (11). The vacA gene has been disrupted in this mutant by insertion of a kanamycin cassette preventing formation of the cytotoxin (11). Although the proteome of H. pylori OMVs has not been defined, OMV from this mutant strain are assumed to have the same phenotype as that of wild-type OMV, except for the absence of the VacA protein. Examination by fluorescence microscopy after 5 h confirmed a previous observation (8) that vesicles shed from both wild-type and VacA− mutant strains are internalized by AGS cells (Fig. 2A and B). To confirm that the observed fluorescence was OMV and not free dye, we labeled vesicles with DiO and, after incubation with AGS cells, used a polyclonal antibody directed against H. pylori OMV to detect vesicles. Immuno-labeled OMV colocalized with DiO-labeled OMV (Fig. (Fig.2C2C).
While the data obtained by fluorescence microscopy demonstrated that uptake of VacA− mutant OMV can occur, there was no indication of the rate of association of these OMV. To investigate this, AGS cells were incubated with labeled OMV from either strain for up to 6 h and examined at regular time points by flow cytometry. As shown in Fig. Fig.3A,3A, wild-type OMV associated with AGS cells at a higher rate than OMV from the VacA− strain (P < 0.01, at 6 h).
To further examine the role of VacA in OMV binding, vesicles from strain 60190 were labeled with DiO (green) and vesicles from strain 60190:v1 with DiD (red). Cells were incubated with OMV from either strain or a 50:50 mix of both. Coincubation of wild-type OMV with VacA− OMV had no effect on the association of wild-type vesicles with cells (Fig. (Fig.3B).3B). In contrast, the association of OMV from the VacA mutant strain was significantly inhibited (P = 0.002) (Fig. (Fig.3B).3B). This observation that incubation with wild-type OMV inhibited the association of VacA− OMV suggests that the presence of VacA enhances vesicle binding.
LPS has been implicated as a potential H. pylori adhesin (14). AGS cells were coincubated with wild-type or VacA− OMV and increasing concentrations of purified H. pylori LPS. OMV uptake was inhibited in a dose-dependent manner by H. pylori LPS, with the effect most evident with VacA+ OMV (58% ± 2.5% inhibition of wild-type OMV uptake compared with 25% inhibition ± 2.4% of VacA− OMV uptake with 50 μg/ml LPS) (Fig. (Fig.4)4) (P < 0.05, Dunnett's posthoc test). In combination with the data shown in Fig. Fig.3,3, this suggests that VacA is important in facilitating the uptake of OMV by an LPS-inhibitable mechanism, while in the absence of VacA, LPS plays a less significant role.
Depletion of plasma membrane cholesterol, which disrupts lipid rafts (34), has been reported to inhibit entry of purified VacA into host cells (42, 53). To examine whether depletion of plasma membrane cholesterol could likewise reduce the uptake of VacA+ OMV, AGS cells were pretreated with 5 mg/ml MβCD, a cholesterol sequestering agent (47). MβCD treatment resulted in no difference in cholesterol levels per mg of total cell protein (37.74 μg ± 1.87 μg for untreated cells, 33.16 μg ± 2.21 μg for MβCD-treated cells) (means ± SE; n = 3; P = 0.133). This was unexpected, as a reduction in cholesterol of up to ~50% has been reported for similar concentrations of MβCD (42, 53). However, pretreatment of AGS cells with as little as 1 mg/ml MβCD was sufficient to significantly inhibit vacuole formation in response to 60190 OMV (Fig. (Fig.5A),5A), while vesicle uptake was increased by MβCD at all concentrations tested (Fig. (Fig.5B)5B) (P < 0.001 and P < 0.05, respectively, Dunnett's posthoc test). Higher concentrations of MβCD were not used, as, in addition to disrupting lipid rafts, these have been reported to inhibit clathrin-mediated endocytosis (47, 55) and to induce marked changes in cellular morphology in HeLa cells (53). Instead, cells were treated with nystatin, a cholesterol-binding agent (50), which disrupts lipid rafts but does not affect clathrin-mediated endocytosis (49). Pretreatment with nystatin (25 μg/ml) inhibited wild-type OMV-mediated cell vacuolation (Fig. (Fig.5C)5C) (P < 0.05, Dunnett's posthoc test) but had no effect on wild-type or VacA− mutant OMV uptake (Fig. (Fig.5D)5D) (P = 0.114 and P = 0.811, respectively, ANOVA).
Clathrin-mediated endocytosis was explored as a potential pathway for vesicle internalization. This pathway can mediate the constitutive uptake of ligands, such as transferrin or low-density lipoprotein, as well as ligand-triggered receptor uptake of proteins (17). Chlorpromazine, a known inhibitor of clathrin-mediated endocytosis (59), significantly inhibited vacuole formation in AGS cells treated with 60190 OMV when added at a concentration of 15 μg/ml (Dunnett's posthoc test, P < 0.001) (Fig. (Fig.6A)6A) and had a dose-response effect on wild-type OMV uptake (ANOVA, P = 0.011) (Fig. (Fig.6B).6B). Chlorpromazine (15 μg/ml) reduced the uptake of VacA− OMV to a greater extent (42% ± 8.9) (Dunnett's posthoc test, P < 0.05) (Fig. (Fig.6C).6C). A reduction in clathrin-mediated endocytosis was confirmed by measurement of LDL uptake (Dunnett's posthoc test, P < 0.05) (Fig. (Fig.6D).6D). We were not able to completely inhibit clathrin-mediated endocytosis with chlorpromazine, as concentrations of this drug greater than 15 μg/ml were cytotoxic, as assessed by propidium iodide staining (data not shown).
The H. pylori vacuolating cytotoxin is secreted in a soluble form and is also associated with OMV (45). Vesicle-associated VacA has been observed at the surface (26) and within the gastric mucosa (15); however, little is known of the interactions of OMV-associated cytotoxin with host cells. The data presented in this study indicate that VacA is not essential for H. pylori OMV uptake but that the presence of the cytotoxin increases the rate of vesicle association with cells.
Purified VacA is reported to bind to multiple cell surface components, including sphingomyelin (18), glycosphingolipids (46), glycosylphosphatidylinositol (GPI)-anchored protein(s) (43), and receptor-like protein tyrosine phosphatase ß (39), with subsequent clustering of the receptor-bound toxin in cholesterol-rich lipid rafts (39, 42, 53). Horstman et al. speculated that such binding of vesicle-bound bacterial toxins to multiple binding sites acts to bring vesicles into closer proximity with host cells, thus allowing secondary adhesions to increase the “intimacy” between the two (20). In support of this, VacA+ OMV associated with cells quicker than VacA− OMV and inhibited the association of VacA− vesicles, suggesting that the toxin increases the avidity of vesicle binding. In addition, H. pylori LPS inhibited OMV uptake to a greater extent with VacA+ OMV. We speculate that binding of vesicle-associated VacA to cells increases the interaction of OMV LPS with cells, increasing the avidity of the overall vesicle binding reaction. In studies of other bacteria, the aminopeptidase PaAP, one of the major protein constituents of OMV from P. aeruginosa cystic fibrous isolates, was recently shown to increase the rate of association of P. aeruginosa OMV with lung epithelial cells (4). In addition, a heat-labile enterotoxin (LT) associated with ETEC OMV increases the association of E. coli OMV with cells (31). In the absence of LT, E. coli OMV binding is significantly reduced (31).
Uptake of vesicles from several bacterial genera has been shown to be cholesterol dependent (6, 25, 31). In addition, Kaparakis and colleagues (25) recently reported that H. pylori OMV uptake is cholesterol-rich lipid raft dependent. However, in our study, binding of cholesterol with MβCD or nystatin, disruptors of lipid raft formation (25, 49), did not significantly inhibit H. pylori vesicle uptake. The disparity between the Kaparakis study and ours may be a reflection of the uptake of vesicles from different H. pylori strains and/or the use of serum-free cell culture medium (25).
MβCD, although having no significant effect on AGS cell cholesterol levels, significantly decreased vacuolation in cells treated with VacA+ OMV. Low concentrations of MβCD have also been shown to inhibit vacuolation in HeLa cells treated with purified VacA (53), adding weight to evidence that membrane cholesterol is essential for VacA toxigenesis (39, 42). A reduction in vacuolation in the absence of an effect on OMV uptake may be explained by a greater sensitivity of vacuole biogenesis, compared with vesicle uptake, to membrane cholesterol levels.
VacA translocation to lipid rafts following binding to receptor-like protein tyrosine phosphatase ß in nonlipid raft domains on the cell surface is reportedly inhibited by treatment with the same concentration of MβCD (5 mg/ml) used in our study (39). Thus, if lipid rafts are required for VacA+ OMV uptake, we expected to see reduced vesicle internalization. Instead we observed that OMV uptake in MβCD-treated cells was increased. The reason for this is not immediately apparent and requires further investigation.
Disruption of clathrin-mediated endocytosis with chlorpromazine reduced the uptake of OMV. To our knowledge this is the first report demonstrating OMV internalization via clathrin-mediated endocytosis. Internalization of ETEC and P. aeruginosa OMV is unaffected by chlorpromazine, and these OMV rarely colocalize with clathrin (4, 31). Indeed, ETEC OMV frequently colocalize with caveolae (31). In addition, P. gingivalis OMV internalize via a Rac1-regulated, lipid raft-dependent pathway that is independent of caveolin and clathrin (16).
The observation that VacA+ OMV uptake was less inhibited by chlorpromazine than uptake of VacA− OMV suggests that toxin-containing H. pylori OMV may be taken up into gastric epithelial cells via more than one pathway. That nystatin had no significant effect on VacA+ OMV uptake also supports this. However, further investigation is required to confirm this hypothesis. Of note, several bacterial toxins (including anthrax, cholera, and Shiga toxins) are reportedly endocytosed by both clathrin-dependent and clathrin-independent mechanisms (1, 52, 56).
The increase in OMV association over time, in conjunction with the observation that the majority of associated OMV were intracellular 20 min and thereafter, suggests to us that binding is the rate-limiting step in OMV uptake, possibly due to a limited numbers of receptors. The steady increase in fluorescence over time also suggests that, once internalized, OMV persist within AGS cells for some time. This is supported by the observation of intact OMV within cells after >72 h coincubation (54). The association of ETEC and P. aeruginosa OMV with cells has also been shown to increase with time (4, 31). In agreement with Bauman and Kuehn (4), we consider that this is consistent with (i) vesicle binding initiating upregulation of OMV receptors or (ii) receptors recycling back to the plasma membrane after internalization of their cargo (7). We found OMV association was unaffected by cycloheximide, indicating that if specific receptors are involved they are already translated. We have observed that perturbation of the actin cytoskeleton inhibits vesicle association (H. Parker and J. Keenan, unpublished data), suggesting disruption of receptor recycling.
The small size of OMV (50 to 300 nm) makes their enumeration difficult. In this study, OMV were quantified based on protein concentration. While this may vary between strains, it is one of the more commonly used methods for OMV enumeration (16, 24, 25, 48, 58). As an alternative, OMV may be quantified based on the concentration of lipopolysaccharide (51) or toxin (6) in vesicle preparations. However, the amount of LPS and VacA cytotoxin also varies between H. pylori strains (26, 29), as well as within a strain grown under different conditions (28, 30). In addition, use of a VacA− strain in this study precluded OMV enumeration based on cytotoxin concentration. Immunolabeling of VacA+ OMV shows relatively few (<10) toxin molecules per OMV (15, 26, 54). Therefore, we considered that its absence from OMV may not significantly change the overall vesicle protein concentration, thus allowing for reasonable comparison between OMV from wild-type and VacA− OMV.
In summary, we have shown that VacA toxin enhances the association of H. pylori OMV with cells and that VacA− OMV are internalized via clathrin-mediated endocytosis, while VacA+ OMV may be able to exploit more than one pathway of internalization. We propose that constitutive release (26) and uptake (15) of OMV from the surface of these predominantly noninvasive bacteria contributes to H. pylori pathogenesis via the persistent delivery of virulence factors and antigens to gastric epithelial cells.
We thank Tim Cover (Division of Infectious Diseases, Vanderbilt University School of Medicine) for the kind gift of the VacA mutant.
This work was supported by a Bright Futures Doctoral Scholarship to H.P. and funding from the Cancer Society of New Zealand (to J.I.K. and M.B.H.).
Editor: S. R. Blanke
Published ahead of print on 27 September 2010.