These authors contributed equally to this work.
Dynamics of the Cag-type IV secretion system of Helicobacter pylori as studied by bacterial co-infections
Article first published online: 8 AUG 2013
© 2013 John Wiley & Sons Ltd
Volume 15, Issue 11, pages 1924–1937, November 2013
How to Cite
Jiménez-Soto, L. F., Clausen, S., Sprenger, A., Ertl, C. and Haas, R. (2013), Dynamics of the Cag-type IV secretion system of Helicobacter pylori as studied by bacterial co-infections. Cellular Microbiology, 15: 1924–1937. doi: 10.1111/cmi.12166
- Issue published online: 10 OCT 2013
- Article first published online: 8 AUG 2013
- Accepted manuscript online: 11 JUL 2013 07:24AM EST
- Manuscript Accepted: 26 JUN 2013
- Manuscript Revised: 21 JUN 2013
- Manuscript Received: 26 OCT 2012
- DFG. Grant Numbers: HA2697/15-1, HA2697/16-1
- Münchner Medizinische Wochenschrift (MMW)
- Top of page
- Experimental procedures
- Supporting Information
Many pathogenic Gram-negative bacteria possess type IV secretion systems (T4SS) to inject effector proteins directly into host cells to modulate cellular processes to their benefit. The human bacterial pathogen Helicobacter pylori, a major aetiological agent in the development of chronic gastritis, duodenal ulcer and gastric carcinoma, harbours the cag-T4SS to inject the cytotoxin associated Antigen (CagA) into gastric epithelial cells. This results in deregulation of major signalling cascades, actin-cytoskeletal rearrangements and eventually gastric cancer. We show here that a pre-infection with live H. pylori has a dose-dependent negative effect on the CagA translocation efficiency of a later infecting strain. This effect of the ‘first’ strain was independent of any of its T4SS, the vacuolating cytotoxin (VacA) or flagella. Other bacterial pathogens, e.g. pathogenic Escherichia coli, Campylobacter jejuni, Staphylococcus aureus, or commensal bacteria, such as lactobacilli, were unable to interfere with H. pylori's CagA translocation capacity in the same way. This interference was independent of the β1 integrin receptor availability for H. pylori, but certain H. pylori outer membrane proteins, such as HopI, HopQ or AlpAB, were essential for the effect. We suggest that the specific interference mechanism induced by H. pylori represents a cellularresponse to restrict and control CagA translocation into a host cell to control the cellular damage.
- Top of page
- Experimental procedures
- Supporting Information
The highly motile, microaerophilic, Gram-negative bacterium Helicobacter pylori is the main causative agent of chronic gastritis, gastric and duodenal ulcers, gastric adenocarcinoma and MALT lymphoma (Blaser and Atherton, 2004). H. pylori is found in the gastric epithelium of about 50% of the human population. Bacterial adherence is thought to play an important role for triggering colonization, but also to closely regulate the functional interplay of the bacteria with the host epithelium.
In the Western world about 50–70% of clinical H. pylori isolates carry the so-called cag-pathogenicity island (cag-PAI) in their genome (Censini et al., 1996; Kusters et al., 2006). The cag-PAI encodes the cag-type IV secretion system (cag-T4SS), which is used by H. pylori to inject the effector protein CagA into eukaryotic cells (Odenbreit et al., 2000; Alvarez-Martinez and Christie, 2009; Fischer, 2011). The presence of this genomic island is an important marker for the virulence of the corresponding strains and allows classification into the more pathogenic type I and the less pathogenic type II H. pylori strains. In this context we understand type I strains as those carrying a complete cag-PAI enabling them to translocate CagA into gastric AGS cells or to induce interleukin-8 (IL-8) secretion in host cells. Type II strains are those that either do not carry a cag-PAI at all (empty site PCR results), or carry a defective cag-T4SS, which does not allow for translocation of CagA or IL-8 induction in AGS cells. β1 integrin heterodimers on the surface of eukaryotic cells serve as a receptor for the cag-T4SS. For the three cag-PAI-encoded proteins CagI, CagL and CagY, as well as for the effector protein CagA itself, a direct interaction with the β1 integrin receptor was shown (Kwok et al., 2007; Jimenez-Soto et al., 2009; Kaplan-Türköz et al., 2012). The precise mechanism for CagA translocation is, however, still unclear. After translocation, CagA is initially phosphorylated by cellular Src and later by Abl kinases at the Glu–Pro–Ile–Tyr–Ala (EPIYA) motif, which is variably repeated in the C-terminal region of CagA (Mueller et al., 2012).
Tyrosine-phosphorylated CagA specifically interacts with and aberrantly activates Src homology 2-containing protein tyrosine phosphatase-2 (SHP2) (Higashi et al., 2001). CagA-activated SHP2 deregulates cell proliferation by stimulating Erk MAP kinase via Ras-dependent and -independent pathways (Higashi et al., 2004). Furthermore, CagA binds to the kinase domains of all partitioning-defective 1/microtubule affinity-regulating kinase (PAR1/MARK) isoforms and inhibits their kinase activity and thereby causes cell junction and polarity defects (Saadat et al., 2007).
Considering such a massive deregulation of cellular pathways by injected CagA, it has to be assumed that either the bacteria take measures to limit CagA injection, or the host cell protects itself against an overload with CagA. We reasoned that the presence and availability of a receptor for the cag-T4SS, the engagement of which is essential for CagA translocation, could be the bottleneck and limiting factor for CagA translocation. To study this question, the dynamics of H. pylori CagA translocation were analysed by competition assays. We show in co-infection experiments using initially two translocation-competent H. pylori wild type strains that pre-infection of AGS cells with a first infecting H. pylori strain can strongly reduce or block the ability of a later infecting strain to inject CagA. We further show that this interference mechanism is independent of any functional T4SS and flagella. To determine if the cellular response seen is exclusive to H. pylori strains, we tested other bacteria (pathogenic and non-pathogenic) such as Escherichia coli (UPEC, EPEC), Campylobacter jejuni and Staphylococcus aureus, or different commensal bacteria, such as lactobacilli. They all were unable to interfere with CagA translocation upon pre-infection experiments, suggesting that the interference mechanism is specific for H. pylori.
- Top of page
- Experimental procedures
- Supporting Information
Design of infection experiments to study the dynamics of H. pylori CagA translocation
A H. pylori infection usually persists over years or decades in the stomach of an individual, but it is unclear how the activity of the cag-T4SS is regulated and how the host cell protects itself against repeated CagA injection by H. pylori. To get some insights into this mechanism, we studied the dynamics of H. pylori CagA translocation in the AGS cell culture model. Therefore, the effect of a pre-infection of AGS cells by H. pylori, on the capability of a later infecting H. pylori strain to translocate CagA via the cag-T4SS was analysed. Co-infections of the cag-PAI-positive H. pylori wild-type (WT) strain P12 were performed together with a second H. pylori strain (‘test’ bacterium) able to translocate CagA (Fig. 1A) to observe the effect on the AGS cells caused by an oversaturation of the infection system. AGS cells in culture are usually in different stages of the cell cycle, which results in a variable surface receptor availability between individual cells. By synchronization (see Experimental procedures) they are brought to the same phase, which results in a co-ordinated expression of cell surface receptors (e.g. β1 integrin receptors). Either synchronized AGS cells were infected with a test bacterium and P12 WT at the same time (t = 0; co-infection), or the test bacterium was infected first (t = 0) and H. pylori P12 WT 1 h later (t = 60) (pre-infection). The cells were harvested 3 h later (t = 180 or t = 240 for co- or pre-infection experiments respectively) and processed for immunoblotting to determine the CagA translocation efficiency of P12 WT.
A primary infecting H. pylori strain reduces the CagA translocation efficiency of a later interacting strain
In order to distinguish the CagA phosphorylation band of the second strain from the CagA band of the P12 WT strain in the immunoblot, H. pylori strain P217 was used as test bacterium, which carries a significantly larger CagA as compared with P12 (170 kDa versus 135 kDa respectively). The P12 CagA (CagAP12) carries three EPIYA motifs, whereas the P217 CagA (CagAP217) consists of six motifs (data not shown), resulting in a stronger tyrosine-phosphorylation of the latter CagA protein upon translocation.
Upon simultaneous infection with H. pylori strains P12 and P217, both CagA phosphorylation bands were strongly reduced in their intensity, as compared with the single infection experiments (Fig. 1B). However, after a 1 h pre-incubation of strain P217, the phosphorylation intensity of CagAP12 (t = 60) was significantly reduced (Fig. 1A and B). But also the CagA translocation of P217 was reduced, both in co-infection (t = 0) and in pre-incubation experiments (t = 60) (Fig. 1A and B).
To determine whether only fully virulent type I H. pylori strains show this effect, we next studied in co-infection experiments different type II H. pylori strains, which carry a defective cag-PAI and produce a non-vacuolating cytotoxin protein (VacA). The presence of different type II H. pylori strains (X47 WT, Tx30a WT) concomitantly with a P12 WT reduced the CagAP12 translocation efficiency of the latter one (Fig. 1C). This effect was clearly enhanced when the type II strain infected already 1 h before the P12 WT strain (t = 60). Both independent H. pylori type II strains, X47 and Tx30a, significantly reduced CagAP12 translocation of strain P12. The intensity of CagA translocation reduction of type II strains was thus comparable to type I strains, suggesting that a complete and functional cag-PAI was not necessary for this effect. Since type II H. pylori strains usually also lack functional VacA (Cover et al., 1994), the data also suggested that a functional VacA is not necessary for this effect. To corroborate these data, we also studied CagAP12 translocation of WT in the presence of increasing amounts of purified activated VacA of H. pylori strain 60190. Neither the VacA addition concomitantly (t = 0) nor 1 h before infection (t = 60) showed any effect on CagA translocation (Fig. S1).
The reduction in CagA translocation of the secondary infecting H. pylori strain is independent of any H. pylori T4SS and flagella of the primary interacting strain
To study the influence of other H. pylori secretion systems and flagella, we used defined congenic deletion mutants defective in all four distinct type IV secretion systems (ΔT4SS) found in H. pylori, the cag-PAI, the comB system, the TFS3 and the TFS4 (Fischer et al., 2010). In addition, the mutant strain also carried a deletion of its flaA and flaB flagellin genes. Thus, the mutant strain, named P12ΔT4SSΔflaAB, did not possess any T4SS nor did it produce active flagella.
AGS cells were co-infected with P12ΔT4SSΔflaAB and P12 WT. The concomitant infection (t = 0) resulted in a reduced translocation and phosphorylation of CagAP12. The inhibition of CagAP12 translocation was even enhanced when P12ΔT4SSΔflaAB infected 1 h in advance to P12 WT (Fig. 2A). Quantification of CagAP12 translocation for the simultaneous co-infections (t = 0) showed a reduction of CagA translocation to 50%, whereas for the subsequent infection (t = 60) a reduction of CagA translocation to 25% of the original level was observed (Fig. 2A).
Live bacteria reduce CagA translocation in a dose-dependent manner, whereas proteinaceous components of bacterial lysates show only a minor effect
To analyse how the bacterial dosage of the first strain might affect the reduction of CagAP12 translocation in the co-infection experiments, we varied the moi (multiplicity of infection) of P12ΔT4SSΔflaAB in co-infection experiments. P12 WT was used constantly at an moi of 60. For simultaneous infection (t = 0), a reduction of CagA phosphorylation was only seen at an moi of 60, whereas in pre-incubation experiments (t = 1 h) a significant effect was already seen at an moi of 30 (Fig. 2B). Quantification of CagA translocation by densitometry of the signal of tyrosine phosphorylation revealed that a direct proportional effect between the moi used and the reduction of CagA translocation was apparent, especially under pre-incubation conditions (Fig. 2C).
Next we tested whether for blocking of CagA translocation live bacteria are necessary, or whether a lysate of P12ΔT4SSΔflaAB may have the same effect. Therefore, P12ΔT4SSΔflaAB was inactivated by either sonication or heating. The ultrasound treatment reduced the number of live bacteria, whereas the heating steps denatured bacterial proteins, whereas other components, such as LPS, would still be stable. Live P12ΔT4SSΔflaAB bacteria showed a small reduction of CagA translocation for the simultaneous infection (t = 0), and a strong reduction in pre-infection conditions (t = 60) (Fig. 3A and B). For the ultrasound-treated P12ΔT4SSΔflaAB bacteria, in pre-infections (t = 60) a significant reduction of CagA translocation was observed, but this effect disappeared in co-infection experiments (t = 0). In the case of heat-inactivated P12ΔT4SSΔflaAB bacteria the normal reduction of CagA phosphorylation disappeared at both time points (t = 0 and t = 60) (Fig. 3A and B). These results argue against thermally stable components of H. pylori P12ΔT4SSΔflaAB, such as LPS, as a cause for the inhibition of CagA translocation, but heat-sensitive factors, such as bacterial proteins, might be involved. In conclusion, the data show that live bacteria have the most drastic effect to reduce CagA translocation in pre-infection experiments. Sonication does apparently not destroy the activity under pre-infection conditions.
Neither a general reduction of bacterial adherence of the secondary strain nor a secreted factor is responsible for CagA translocation inhibition
Helicobacter pylori produces and secretes a set of proteins and enzymes, among them a rapid urease, the vacuolating cytotoxin (VacA) or the proteolytic enzyme HtrA. To study whether a soluble secreted factor of H. pylori might be responsible for the reduction of CagA translocation, AGS cells were treated with bacteria-free supernatants of H. pylori P12ΔT4SSΔflaAB before infection with P12 WT at time points t = 0 or t = 1 h. Pre-treatment with these supernatants did not show any significant effect on CagA translocation (Fig. 3C).
A possible cause for the observed reduced CagAP12 WT translocation during co-infection could be a competition for binding sites on the surface of AGS cells. To rule out such a scenario, we analysed the binding of P12 WT bacteria under conditions of pre-incubation with P12ΔT4SSΔflaAB. For adhesion experiments strain P12 producing GFP from the pHel12 plasmid (P12[pHel12mccA::gfp]), shortly termed P12-GFP, was used. AGS cells were infected with P12ΔT4SSΔflaAB and P12-GFP at time point t = 0. In addition, P12ΔT4SSΔflaAB was used for pre-infection (t = 60) experiments with P12-GFP. After 30 min binding of P12-GFP to AGS cells, unbound bacteria were washed; cells were fixed and measured for GFP fluorescence by flow cytometry. The analysis showed that binding of P12-GFP was not influenced by pre-incubation, or by simultaneous incubation of P12 WT with P12ΔT4SSΔflaAB (Fig. 4A), which might argue against a direct competition for physical binding to the host cell as a cause for the effect.
The availability of β1 integrins on the surface of AGS cells is unaffected by pre-incubation with flagella- and T4SS-deficient H. pylori strains
Next we were interested whether pre-infection of P12ΔT4SSΔflaAB would have an effect on AGS β1 integrin localization on the cell surface, since a reduced β1 Integrin surface availability could explain the reduction in CagAP12 WT translocation. Quantification of β1 integrin of infected versus non-infected cells by flow cytometry revealed no difference in the expression and surface localization of β1 Integrin during a window of 110 min (Fig. 4B), which was significantly longer than the 60 min pre-infection experiments. These data suggest that a reduction in β1 integrin availability for strain P12 WT, e.g. via the induction of integrin internalization, is not the reason for the observed CagA translocation defect.
β1 integrin heterodimers can be activated from outside of the cell by the addition of manganese ions (Mn2+) (Mould et al., 2002), and activated β1 integrin receptors are more efficient for CagA translocation (Jimenez-Soto et al., 2009). Thus, β1 integrins might be in an inactive state after pre-incubation of AGS cells with P12ΔT4SSΔflaAB and therefore not accessible for the T4SS. We therefore tested whether the addition of Mn2+ to the infection would be able to restore the CagAP12 WT translocation defect in co-infection experiments with P12ΔT4SSΔflaAB. The activation of β1 integrin was successful, as seen by a stronger CagA translocation with a P12 WT single infection, as compared with the control without MnCl2 (Fig. 4C). Although treatment of AGS cells with MnCl2 was effective, it was not able to restore the CagA translocation in the pre-infected cells (Fig. 4C). In conclusion, these data suggest that the integrin availability has not been altered, but we cannot exclude that its functional status might be compromised, e.g. by interference with inside-out or outside-in signalling mediated by cytoplasmic tail integrin-binding proteins as a result of the pre-infection, reducing therefore the amount of CagA translocation observed.
Specific H. pylori outer membrane proteins are involved in host cell modulation to block CagA translocation
The cellular changes caused by the pre-infection with H. pylori strains required a direct contact of proteins on the bacteria. We decided to look for the protein(s) on the H. pylori surface and chose to evaluate the role of H. pylori outer membrane proteins (OMPs). H. pylori OMPs might mediate the first contact with AGS cells, making them good candidates to modulate the early cell response for CagA translocation. Accordingly, a set of H. pylori 26695 congenic mutants in outer membrane proteins were used in pre-infection experiments to analyse their ability to interfere with CagAP217 translocation. Some of these omp mutant strains have been generated in our lab before and were already analysed for their effect on CagA translocation (Odenbreit et al., 2002), others were generated during this project (see Table 1). We asked in these experiments whether pre- or co-incubation of AGS cells with a congenic omp gene mutant might have a significant effect on CagA translocation efficiency, as compared with pre- or co-incubation with the corresponding WT strain. In pre-infection experiments, a significantly higher CagA translocation efficiency was recovered when 26695 mutant strains without production of the outer membrane proteins AlpAB, HopI, HopQ or BabA were used, but not for any other mutant omp gene tested (Fig. 5A and C). In co-infection experiments, also mutant strains without AlpAB, HopI, HopQ, and in addition also HofF, HorL and AlpA showed a significant recovery of the CagA phosphorylation in comparison with the 26695 wild type strain (Fig. 5B and C). Interestingly, most of these OMP proteins have been described as bacterial adhesins. The generation and genetic complementation of a P12ΔhopQ strain with a pHel2-based shuttle vector (P12ΔhopQ[phopQ]) resulted in a complete restoration of the competition phenotype (Fig. 5D).
|Parental strain||Mutated gene||Gene name||CagA translocation blocking (> 20% after pre-incubation)||Reference|
|26695 wt||–||–||+||Tomb et al. (1997)|
|26695||hp0252||hopF||+||Odenbreit et al. (2002)|
|26695||hp0254||hopG||+||Odenbreit et al. (2002)|
|26695||hp0472||horE||+||Odenbreit et al. (2002)|
|26695||hp0638||hopH||+||Odenbreit et al. (2002)|
|26695||hp0788||hofF||+||Odenbreit et al. (2002)|
|26695||hp0896||hopT, (babB)||+||Odenbreit et al. (2002)|
|26695||hp0912||hopC, (alpA)||−||Odenbreit et al. (1999)|
|26695||hp1156||hopI||−||Odenbreit et al. (2002)|
|26695||hp1177||hopQ||−||Odenbreit et al. (2002)|
|26695||hp1243||hopS, (babA)||+||Odenbreit et al. (2002)|
|26695||hp1395||horL||+||Odenbreit et al. (2002)|
In conclusion, these data suggest that a pre- or co-infecting H. pylori strain makes use of a set of specific outer membrane proteins, such as AlpAB, HopI and HopQ, in order to communicate with the host cell to restrict or shut down the CagA translocation efficiency or capacity.
Unrelated bacteria do not interfere with H. pylori CagA translocation, but β1 integrin-interacting bacteria compete with H. pylori for integrin availability
An obvious question was now whether other bacteria besides H. pylori would be able to interfere with CagAP12 WT translocation in co-infection and pre-infection experiments, or whether this effect is specific for H. pylori. Therefore a set of different commensal and pathogenic bacteria were tested, including pathogenic and non-pathogenic Gram-negative E. coli, as well as two commensal Gram-positive species of lactobacilli. S. aureus was used due to its interaction with integrins (Agerer et al., 2003), whereas C. jejuni was chosen due to its close phylogenetic relation to H. pylori.
Neither the enteropathogenic (EPEC) nor the non-pathogenic E. coli DH5α lab strain were able to reduce CagAP12 WT translocation in a similar way as observed in H. pylori pre-infection experiments (Fig. 6A), demonstrating the specificity of the inhibitory mechanism. For the commensal lactobacilli we chose Lactobacillus acidophilus and L. johnsonii as candidates for co-infection experiments, but again none of these strains was able to reduce CagAP12 WT translocation (Fig. 6A).
Campylobacter jejuni is genetically closely related to H. pylori and is able to bind β1 integrin via its outer membrane protein CadF and fibronectin, whereas the uropathogenic E. coli (UPEC) strain is able to bind β1 integrin directly (Eto et al., 2007). Interestingly, co-infection but not pre-infection of UPEC or C. jejuni with H. pylori resulted in a reduction of CagAP12 WT translocation (Fig. 6A and B). The pre-infection of AGS cells with S. aureus and P12 WT also resulted in a slightly reduced CagA translocation rate. However, co-infections (t = 0) of S. aureus and P12 WT resulted in a strong reduction in CagA translocation, comparable to the effect of H. pylori P12ΔT4SSΔflaAB (Fig. 6B). Both bacterial pathogens were subsequently grown on agar plates supplemented with cholesterol instead of blood (cholesterol plates) (Jimenez-Soto et al., 2012). Under these growth conditions the bacteria did not get access to extracellular matrix proteins, such as fibronectin. This resulted in the complete loss of their ability to interfere with CagA translocation in co-infection experiments for S. aureus, while for C. jejuni there was a slight but significant increment of CagA translocation without the serum proteins. However, in serum free conditions only C. jejuni retained a rather weak ability of affecting the CagA translocation during co-infection experiments (Fig. 6B). The data suggest that S. aureus and C. jejuni seem to compete with the H. pylori's cag-T4SS for the binding of β1 integrin. The activity is mainly seen upon co-infection conditions (t = 0), which are more competitive condition as the pre-infection situation. Thus, these mechanisms by C. jejuni, UPEC and S. aureus appear to be mainly via competition and therefore different from the interference mechanism used by H. pylori, which are most pronounced under pre-infecting conditions and are independent of a functional T4SS and its integrin interaction.
- Top of page
- Experimental procedures
- Supporting Information
Helicobacter pylori is an important bacterial pathogen colonizing the gastric mucosa of humans. The cag-T4SS and translocation of the effector protein CagA into gastric epithelial cells is an important virulence factor of H. pylori to cause gastric disease. The basic mechanism of CagA translocation into human cells is still not well understood. Especially the question if there is a modulation of the level of CagA injection into a host cell, which theoretically could be determined either by the bacteria or by the host cell. In this study we approached this question by analysing the dynamics of CagA translocation into target cells over time, applying competitive conditions for the translocation-competent bacteria. Co- and pre-infections with different strains and mutants of H. pylori unveiled for the first time that a pre-infection of gastric epithelial cells with any H. pylori strain strongly reduces the capability of a later infecting type I strain to translocate CagA. This reduction in CagA translocation is dependent neither on components of the T4SS of the original infecting strain, nor on the H. pylori cytotoxin VacA. Furthermore, intact bacteria, rather than secreted factors of H. pylori show the most prominent effect. Thus, our data are best explained by a direct contact between the primary infecting strain and the host cell, necessary to signal to the host cell the down-modulating activity. However, the activity was found to be sensitive to heat modification, therefore heat-stable components, such as LPS or other oligosaccharides could be ruled out. Surface appendages of H. pylori, such as flagella or any one of the four T4SSs present in H. pylori P12 (Fischer et al., 2010) could be excluded, since the P12ΔT4SSΔflaAB strain, which lacks all these surface appendages, was fully competent to induce the CagA translocation down-modulating activity. Experiments using a GSK-tagged CagA (Hohlfeld et al., 2006) instead of WT CagA in the tester strain gave similar results as the corresponding tyrosine phosphorylation assays, suggesting that a reduction in CagA translocation, rather than an inhibition of c-Src or c-Abl kinase, might be the reason for the reduced tyrosine phosphorylation of CagA in pre- or co-infection experiments (data not shown).
We find a stronger effect upon pre-incubation, rather than co-incubation of the primary versus secondary H. pylori strain. This might indicate that the host cell needs a certain time to respond to a signal of the primary H. pylori strain for down-modulation of the ‘competence’ for CagA translocation and that this down-modulation might involve the activation capability of β1 Integrin. However, since CagA translocation is a rather fast event (within 5–15 min after an experimental infection, L. Jiménez-Soto, unpubl. data), the reprogramming of the host cell has to be relatively quick. So far the nature of this signal and how the host cell responds to it is still unclear and will be subject for further investigation.
The cag-T4SS interacts with β1 integrin by several of its components, the pilus-associated proteins CagI, CagL, CagY and the effector protein CagA (Kwok et al., 2007; Jimenez-Soto et al., 2009). We therefore asked whether pre-incubation of the primary H. pylori strain could result in the unavailability of β1 integrin receptors for the secondary infecting strain. However, quantification of the cell surface-located β1 integrin pool of infected versus non-infected cells by flow cytometry revealed no difference in the surface exposition or availability of β1 integrin receptors (Fig. 4B). Furthermore, also the activation of the β1 integrin heterodimer by manganese was not able to restore the defect in CagA translocation for the secondary strain.
Other bacterial surface-associated components, which are possible candidates to interact with a putative host cell receptor, are the H. pylori outer membrane proteins. H. pylori strains possess a large family of outer membrane proteins which, according to conserved N- or C-terminal sequences, have been allocated to different families, the Hop (helicobacter outer membrane proteins), the Hor (Hop-related) proteins, the Hof proteins (H. pylori outer membrane protein family) and the Hom family (for H. pylori outer membrane proteins) (Alm et al., 2000). Some of these OMPs of the hop gene family have been characterized as bacterial adhesins, e.g. HopC (also designated as AlpA), BabA, SabA or HopZ. For BabA and SabA the corresponding receptors, the blood group antigen Lewisb and sialyl-LewisX are well described (Ilver et al., 1998; Mahdavi et al., 2002), but for all other adhesins no receptors on the host cell have been identified.
The H. pylori strain 26695 and its corresponding congenic mutants in defined omp genes were tested for their ability to interfere with CagA translocation of a secondary infecting strain (P217). Three out of 12 different OMPs showed a significant effect on CagA translocation under both, pre- as well as co-infection conditions (HopI, HopQ and AlpAB). Under co-infection conditions we cannot exclude that also competition between the two infecting bacterial strains for membrane receptors might play a role, which might explain some of the weak effects on CagA translocation for HofF and HorL outer membrane proteins (Fig. 5B). AlpAB is known to act as adhesin of H. pylori to gastric epithelial cells (Odenbreit et al., 1999; Loh et al., 2008), whereas HopQ and HopI have not been published as bacterial adhesins. The hopQ gene is known to exhibit a high level of genetic diversity, and two families of hopQ alleles exist. The type I hopQ genes are associated with cag-PAI-positive H. pylori strains and the type II alleles are mainly found in cag-PAI-defective strains (Cao and Cover, 2002). Interestingly, only a small set of outer membrane proteins are involved in modulating CagA translocation efficiency. BabA, which mediates binding of H. pylori to the Lewisb receptor, also augments T4SS-dependent pathogenicity of H. pylori. It triggers the production of pro-inflammatory cytokines (Ishijima et al., 2011). Interestingly BabA shows a weak apparent effect on the inhibition of CagA translocation in pre-infections, but not under co-infection conditions (Fig. 5A and B).
We are not aware of any bacterial T4SS/host cell interaction system for which a similar modulation of the T4SS activity has been reported. However, for the plant pathogenic bacteria Pseudomonas syringae, which uses a type III secretion system (T3SS) to inject effector proteins into plant cells, a restriction for effector protein injection has been described (Crabill et al., 2010). Generally plants perceive microorganisms by recognizing microbial molecules known as pathogen-associated molecular patterns (PAMPs). These molecules usually induce a PAMP-triggered immunity (PTI). Alternatively, plants can recognize pathogen effectors directly, inducing effector-triggered immunity (ETI). Interestingly, the P. syringae T3SS can be restricted in its ability to inject a T3SS effector injection reporter molecule (adenylate cyclase, CyaA) into tobacco (Nicotiana tabacum) cells (Crabill et al., 2010). This effect is mediated by a direct or indirect restriction of T3 effector injection and certain T3 effectors can relieve this restriction by suppressing PTI (Crabill et al., 2010).
To better understand the specificity and the mechanism behind this phenomenon, we used non-Helicobacter bacteria for co- or pre-infection to test their effect on CagA translocation of H. pylori. It was attempted to apply a broad-spectrum of diverse bacteria, such as Gram-positive and Gram-negative species, pathogenic and non-pathogenic ones. These experiments should clarify whether the reduction of CagA translocation is H. pylori-specific, or whether it could also be induced by other bacteria. From the set of bacteria used, S. aureus, C. jejuni and uropathogenic E. coli (UPEC) revealed a similar effect as H. pylori. All these pathogens are able to interact with β1 integrin. UPEC interacts directly with β1 integrin (Eto et al., 2007), whereas S. aureus and C. jejuni use serum proteins as bridging molecules for this interaction. To evaluate the relevance of serum proteins for competition with CagA translocation, S. aureus and C. jejuni were grown in serum-free conditions. Whereas S. aureus lost its capacity to interfere, C. jejuni still can reduce CagA translocation, but only in co-infection conditions (Fig. 6B). This suggests that a competition for integrin binding between these pathogens and H. pylori via its cag-T4SS might be the reason for the reduction in CagA translocation. Thus, the shut-down of CagA translocation by subsequent encounters of the bacteria with the same host cell seems to be a specific mechanism used by H. pylori via some of its outer membrane proteins.
These data are consistent with a recent publication showing that in co-infections of EPEC and H. pylori, EPEC inhibits H. pylori-induced AGS cell elongation, but has no effect on CagA translocation and phosphorylation (Brandt et al., 2009). Theoretically it might be possible that by pre- or co-infections certain cellular binding sites for bacterial membrane proteins will be occupied, which might not influence bacterial adhesion in general, but interfere with CagA translocation.
When we compare our in vitro experiments with the situation in vivo, we have to consider that the stomach is not a sterile environment. Although we generally used in the in vitro pre- or co-infection experiments for technical reasons different H. pylori strains producing CagA of different size (P12, P217), we assume that the effect we describe here is also relevant for the infection of the human gastric mucosa with a single H. pylori strain. After having injected CagA into a defined host cell, this cell should be protected against injection of CagA by the same or another individual bacterium of the same strain. However, the human stomach can also be exposed to multiple infections with different H. pylori strains at the same time (Taylor et al., 1995), or to different other bacterial species, as shown by microbiome studies (Bik et al., 2006; Maldonado-Contreras et al., 2011). A recent study in The Gambia reported about a high prevalence of H. pylori in dyspeptic patients with many isolates belonging to the putatively more virulent cagA+, vacAs1 and vacAm1 genotypes. However, the study also showed a significantly lower disease burden in Gambians infected with a mixture of cag-positive and cag-negative strains, as compared with those containing only cag-positive, or only cag-negative strains. The authors concluded from their data that harbouring both cag-positive and cag-negative strains might protect against stomach disease. (Secka et al., 2011). Whether a similar phenomenon as described by our study, e.g. the inhibition of CagA translocation, might be the basis for their in vivo findings will be explored in future studies.
Microbiome studies of the human stomach revealed the presence of other Helicobacteraceae, as well as of Firmicutes, Actinobacteracea, Spirochaetes, Fusobacteria and other bacterial phyla (Bik et al., 2006). Although these studies cannot determine how transient or constant the observed microflora in the stomach really is, they give us an idea about the complexity of the microbial environment in the human stomach. During our studies, we have observed that even short exposure of other bacterial species to the AGS cells can cause reduction in the translocation of CagA. Our pre-infection/co-infection experimental set-up shows unexpected reactions of the target cell not seen before in mono-bacterial experimental infections in vitro. However, since such a situation appears to be rather common for the natural environment of the stomach, it might be possible that such changes in the target cells caused by multiple bacteria/host cell interactions interfering with CagA translocation might be very common during natural infections.
- Top of page
- Experimental procedures
- Supporting Information
Bacterial strains, plasmid constructs and growth conditions
The H. pylori strains P12 was isolated from a duodenal ulcer patient. P217 is a clinical isolate obtained from the Ludwig-Maximilian-University Munich. Strain 26695, G27 and P1 are H. pylori lab strain. These strains are typical type I strains with a functional T4SS encoded by the cag-PAI. H. pylori strain Tx30a and X47 are typical type II strains, which have deletions in the cag-PAI and no functional vacuolating VacA (Atherton et al., 1995). E. coli strains DH5α is a lab strain and the EPEC and UPEC strains are clinical isolates from LMU Munich. C. jejuni C31 is a clinical isolate from the University of Virginia (Guerrant et al., 1987), and S. aureus (ATCC29213) was obtained from the American Type Culture Collection. The lactobacillus strains L. johnsonii NCC-1680 and L. acidophilus were obtained from Nestlé (Lausanne). See also Table 2 for the strains. H. pylori was grown on GC agar plates supplemented with horse serum and trimethoprim (5 mg ml−1) as described (Jimenez-Soto et al., 2012). All antibiotics were obtained from Sigma-Aldrich Chemie (Deisenhofen, Germany). Incubation of the bacteria was performed at 37°C for 1 day in an anaerobic incubator containing a gas mix of 5% O2, 10% CO2 and 85% N2 (Oxoid, Wesel, Germany). S. aureus and E. coli were cultured on Columbia agar plates containing 5% sheep blood. Lactobacilli were grown on Col-SB plates at 37°C in 10% CO2.
|DH5α||Escherichia coli F-Φ80d lacZ ΔM15 Δ(lacZYA-argF) U169 deoR recA1 endA1 hsdR17||Invitrogen|
|EPEC||Enteropathogenic E. coli, clinical isolate #5563||This study|
|UPEC||Uropathogenic E. coli, clinical isolate #6011||This study|
|Staphylococcus aureus||ATCC 29 213 wound Isolate||American Type Culture Collection (ATCC®)|
|Campylobacter jejuni||C 31, Clinical Isolate, University of Virginia||Guerrant et al. (1987)|
|Lactobacillus johnsonii||EW-114 NCC-1680||Nestlé, Lausanne|
|Lactobacillus acidophilus||EW-118 NCC-12||Nestlé, Lausanne|
|P12 WT||pHPP12||Clinical isolate||Schmitt and Haas (1994)|
|P12ΔT4SSΔflaAB||pHPP12||P12Δcag-PAI; ΔTfs-3; ΔTfs-4; ΔcomB; ΔflaA; ΔflaB||This study|
|P12ΔvacA||pHPP12||P12ΔvacA||Gebert et al. (2003)|
|P12 GFP||pHel4-GFP||P12 wt||This study|
|P217||Clinical isolate of H. pylori||Jimenez-Soto et al. (2009)|
|26695||H. pylori wt||Tomb et al. (1997)|
|Tx30a||Clinical isolate (Atherton et al., 1995)||Atherton et al. (1995)|
|X47||mouse-adapted strain of H. pylori||Handt et al. (1995)|
Genetic complementation of H. pylori
Genetic complementation of deletion mutant P12ΔhopQ was performed by using a derivative of the E. coli–H. pylori shuttle plasmid pHel3, designated pIB6 (I. Barwig, L. Holsten and R. Haas, unpublished). For cloning the hopQ gene allele I, chromosomal DNA was amplified using the forward primer CE68 5′-GATCGTCGACATGAAAAAAACGAAAAAAAC-3′ containing a SalI restriction site (italics) and the reverse primer CE69 5′-GATCAGATCTTTTAATACGCGAACACATAA-3′ containing a BglII restriction site (italics). A third primer, CE67, was designed containing an N-terminal HA-tag (bold). Using forward primer CE67 5′-GATCGTCGACATGTACCCATACGATGTTCCAGATTACGCTA-TGAAAAAAACGAAAAAAAC-3′ and reverse primer CE69 a HA-tagged hopQ gene was generated. The resulting PCR fragments were digested with SalI and BglII and incorporated into the corresponding sites of pIB6, obtaining plasmids pCE4 and pCE4HA respectively. The plasmid was introduced into H. pylori by natural transformation.
Cell culture, synchronization and infection and co-infection assays
For infection experiments AGS cells (ATCC CRL 1739a human gastric adenocarcinoma epithelial cell line) were grown in six-well plates containing RPMI 1640 medium (Gibco BRL) supplemented with 10% heat-inactivated fetal bovine serum (FBS, Gibco, Germany) for 2–3 days to reach a confluence of about 80%, corresponding to approx. 1 × 106 cells per well. For synchronization AGS cells were locked in their cell cycle (G0-phase) by removal of serum and cultured for 8–14 h in RPMI medium without FCS at 37°C, 5% CO2. To start the cell cycle of all cells at the same time, 30 min before start of the infection experiment complete medium with serum was added. If not further specified, infections were performed with an moi of 60 for 3–4 h at 37°C and 5% CO2
Quantification of the adherence of P12 WT GFP by flow cytometry (FACS)
AGS cells were infected as described using a GFP-expressing P12 strain for 30 min the cells were washed three times with PBS and fixed with paraformaldehyde (PFA) using a final concentration (w/v) of 2.5%. After 1 h incubation at 4°C in the dark the cells were transferred into an 1,5 ml tube and measured by flow cytometry (BD FACS CantoII).
SDS-PAGE, antibodies, tyrosine phosphorylation assay and Western blot analysis
Cellular fractions or whole-cell pellets with attached bacteria were mixed with equal amounts of 2× SDS-PAGE buffer and boiled for 10 min. Proteins were separated by SDS-PAGE on 6–8% polyacrylamide gels and blotted onto PVDF membranes (Immobilon-P, Millipore). Before addition of the antibodies, membranes were blocked in TBS-T [150 mM NaCl, 25 mM Tris-HCl pH 7.5, 0.075% (v/v) Tween-20] with 3% BSA 1 h or 2 h at room temperature or overnight. Phosphorylated CagA and the other proteins were detected by incubation of the membranes with mouse monoclonal α-phosphotyrosine antibody 4G10 (Upstate Millipore, Schwalbach, Germany), LM534 (Chemicon), Actin-HRP (GenScript, USA), a rabbit polyclonal α-CagA antibody (AK257). As secondary antibody, horseradish peroxidase-conjugated α-mouse, α-rabbit (Sigma-Aldrich, St. Louis, MO, USA) or α-rat IgGAlexa488 (Molecular Probes, Germany) were used. Antibody detection was performed with the ECL Plus chemoluminescence Western blot kit system for immunostaining (Millipore, Schwalbach, Germany) and the GelDoc imaging (Bio-Rad). Quantification of chemoluminescence signals was done using the ImageJ and the ImageLab 4.1 software.
Quantification of β1 integrin surface localization status of AGS cells by flow cytometry
In AGS cells β1 integrin receptors are transferred to the cellular surface in a cell cycle-dependent manner. To analyse whether the infection by H. pylori might interfere with the β1 integrin surface localization, the level of infected and non-infected cells were compared at regular time points within the cell cycle. AGS cells were grown in six well plates to 80% confluence and synchronized. Synchronization was terminated by adding 1 ml serum-containing medium. Some wells were infected with H. pylori at an moi of 60. At different time points wells were fixed by adding 100 μl of 10× paraformaldehyde (PFA) (t = 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 and 110 min). The cells were washed three times in PBS and primary anti-β1 integrin antibody (AIIB2) was added. After 1 h incubation at 4°C in the dark, cells were washed again three times with PBS. The Alexa488-conjugated anti-rat secondary antibody was incubated for 1 h with the cells and samples were washed two times with PBS and used for flow cytometry.
All data are the result of at least three independent experiments and were evaluated using Student's t-test with GraphPad Prism 5 statistical software. P-values < 0.05 were considered as statistically significant.
- Top of page
- Experimental procedures
- Supporting Information
The authors are grateful to Wolfgang Fischer for critical reading of the manuscript and for helpful discussions. This work was supported by research grants of the DFG (HA2697/15-1 and HA2697/16-1) to R.H. and from the Münchner Medizinische Wochenschrift (MMW) to L.J.-S.
- Top of page
- Experimental procedures
- Supporting Information
- 2003) Integrin-mediated invasion of Staphylococcus aureus into human cells requires Src family protein-tyrosine kinases. J Biol Chem 278: 42524–42531. , , , and (
- 2000) Comparative genomics of Helicobacter pylori: analysis of the outer membrane protein families. Infect Immun 68: 4155–4168. , , , , , and (
- 2009) Biological diversity of prokaryotic type IV secretion systems. Microbiol Mol Biol Rev 73: 775–808. , and (
- 1995) Mosaicism in vacuolating cytotoxin alleles of Helicobacter pylori. J Biol Chem 270: 17771–17777. , , , , , and (
- 2006) Molecular analysis of the bacterial microbiota in the human stomach. Proc Natl Acad Sci USA 103: 732–737. , , , , , , et al. (
- 2004) Helicobacter pylori persistence: biology and disease. J Clin Invest 113: 321–333. , and (
- 2009) Dual infection system identifies a crucial role for PKA-mediated serine phosphorylation of the EPEC-Tir-injected effector protein in regulating Rac1 function. Cell Microbiol 11: 1254–1271. , , , , and (
- 2002) Two different families of hopQ alleles in Helicobacter pylori. J Clin Microbiol 40: 4504–4511. , and (
- 1996) Cag, a pathogenicity island of Helicobacter pylori, encodes type I-specific and disease-associated virulence factors. Proc Natl Acad Sci USA 93: 14648–14653. , , , , , , et al. (
- 1994) Divergence of genetic sequences for the vacuolating cytotoxin among Helicobacter pylori strains. J Biol Chem 269: 10566–10573. , , , , and (
- 2010) Plant immunity directly or indirectly restricts the injection of type III effectors by the Pseudomonas syringae type III secretion system. Plant Physiol 154: 233–244. , , , , and (
- 2007) Integrin-mediated host cell invasion by type 1-piliated uropathogenic Escherichia coli. PLoS Pathog 3: e100. , , , and (
- 2011) Assembly and molecular mode of action of the Helicobacter pylori Cag type IV secretion apparatus. FEBS J 278: 1203–1212. (
- 2010) Strain-specific genes of Helicobacter pylori: genome evolution driven by a novel type IV secretion system and genomic island transfer. Nucleic Acids Res 38: 6089–6101. , , , , , , et al. (
- 2003) Helicobacter pylori vacuolating cytotoxin inhibits T lymphocyte activation. Science 301: 1099–1102. , , , , and (
- 1987) Production of a unique cytotoxin by Campylobacter jejuni. Infect Immun 55: 2526–2530. , , , , , and (
- 1995) Characterization of feline Helicobacter pylori strains and associated gastritis in a colony of domestic cats. J Clin Microbiol 33: 2280–2289. , , , , , , et al. (
- 2001) SHP-2 tyrosine phosphatase as an intracellular target of Helicobacter pylori CagA protein. Science 295: 683–686. , , , , , , et al. (
- 2004) Helicobacter pylori CagA induces Ras-independent morphogenetic response through SHP-2 recruitment and activation. J Biol Chem 279: 17205–17216. , , , , , , et al. (
- 2006) A C-terminal translocation signal is necessary, but not sufficient for type IV secretion of the Helicobacter pylori CagA protein. Mol Microbiol 59: 1624–1637. , , , , , and (
- 1998) Helicobacter pylori adhesin binding fucosylated histo-blood group antigens revealed by retagging. Science 279: 373–377. , , , , , , et al. (
- 2011) BabA-mediated adherence is a potentiator of the Helicobacter pylori Type IV secretion system activity. J Biol Chem 286: 25256–25264. , , , , , , et al. (
- 2009) Helicobacter pylori type IV secretion apparatus exploits beta1 integrin in a novel RGD-independent manner. PLoS Pathog 5: e1000684. , , , , , , et al. (
- 2012) Effects of cholesterol on Helicobacter pylori growth and virulence properties in vitro. Helicobacter 17: 133–139. , , , , , and (
- 2012) Structural insights into Helicobacter pylori oncoprotein CagA interaction with beta1 integrin. Proc Natl Acad Sci USA 109: 14640–14645. , , , , , , et al. (
- 2006) Pathogenesis of Helicobacter pylori infection. Clin Microbiol Rev 19: 449–490. , , and (
- 2007) Helicobacter exploits integrin for type IV secretion and kinase activation. Nature 449: 862–866. , , , , , , et al. (
- 2008) Helicobacter pylori HopQ outer membrane protein attenuates bacterial adherence to gastric epithelial cells. FEMS Microbiol Lett 289: 53–58. , , , , and (
- 2002) Helicobacter pylori SabA adhesin in persistent infection and chronic inflammation. Science 297: 573–578. , , , , , , et al. (
- 2011) Structure of the human gastric bacterial community in relation to Helicobacter pylori status. ISME J 5: 574–579. , , , , , , et al. (
- 2002) Integrin activation involves a conformational change in the alpha 1 helix of the beta subunit A-domain. J Biol Chem 277: 19800–19805. , , , , , , et al. (
- 2012) c-Src and c-Abl kinases control hierarchic phosphorylation and function of the CagA effector protein in Western and East Asian Helicobacter pylori strains. J Clin Invest 122: 1553–1566. , , , , , , et al. (
- 1999) Genetic and functional characterisation of the alpAB gene locus essential for adhesion of Helicobacter pylori to human gastric tissue. Mol Microbiol 31: 1537–1548. , , , , and (
- 2000) Translocation of Helicobacter pylori CagA into gastric epithelial cells by type IV secretion. Science 287: 1497–1500. , , , , , and (
- 2002) CagA tyrosine phosphorylation and interleukin-8 induction by Helicobacter pylori are independent from alpAB, HopZ and bab group outer membrane proteins. Int J Med Microbiol 292: 257–266. , , , and (
- 2007) Helicobacter pylori CagA targets PAR1/MARK kinase to disrupt epithelial cell polarity. Nature 447: 330–333. , , , , , , et al. (
- 1994) Genetic analysis of the Helicobacter pylori vacuolating cytotoxin: structural similarities with the IgA protease type of exported protein. Mol Microbiol 12: 307–319. , and (
- 2011) Mixed infection with cagA-positive and cagA-negative strains of Helicobacter pylori lowers disease burden in The Gambia. PLoS ONE 6: e27954. , , , , , , et al. (
- 1995) Long-term colonization with single and multiple strains of Helicobacter pylori assessed by DNA fingerprinting. J Clin Microbiol 33: 918–923. , , , , , , et al. (
- 1997) The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 388: 539–547. , , , , , , et al. (
- Top of page
- Experimental procedures
- Supporting Information
Fig. S1. Pre-treatment of AGS cells with VacA and its effect on the ability of the P12 wild-type strain to translocate CagA. AGS cells were treated with activated VacA at a concentration of 10, 100 or 500 ng ml−1 and where infected either at the beginning of the experiment (t = 0), or after 1 h (t = 1 h) with the P12 wild-type strain. Non-infected cells were used as controls (lane 1) as well as cells which were treated with P12 WT only at time point 0 (lane 2) or 1 h (lane 3). After SDS-PAGE (6% SDS gel) and Western blot the tyrosine-phosphorylated CagA (CagA-Y-P) was detected using the specific antibody 4G10. As a loading control non-phosphorylated CagA was detected using polyclonal rabbit antiserum AK257.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.