Introduction
Infections with pathogenic food-borne bacteria constitute one of the leading causes of morbidity and mortality in humans. The World Health Organization (WHO) suggests that the human population worldwide suffers from about 4.5 billion incidences of gastroenteritis annually, causing approximately 1.8 million deaths [
1]. Various
Campylobacter species have been identified as the leading enteric bacterial infection worldwide [
2,
3].
Campylobacter jejuni is considered as a classical zoonotic pathogen, as it is found in the normal intestinal flora in many birds and mammals. Since
C. jejuni colonizes various food animals, it can contaminate food products during processing [
4]. After ingestion by a human host, these bacteria use their flagella-driven motility to colonize the epithelial cells of the ileum and colon. Here, they can interfere with normal functions in the intestinal tract, leading to diseases associated with fever, malaise, abdominal pain and watery diarrhoea [
2,
3]. In addition, a minority of infected individuals may develop late complications such as Reiter’s reactive arthritis or Guillain-Barrè and Miller-Fisher syndromes [
5]. There is increasing evidence showing that
C. jejuni disturbs the normal absorptive capacity of the human intestine by damaging epithelial cell functions, either by cell invasion, the production of pathogenicity-associated factors or indirectly by triggering inflammatory responses [
3,
6‐
8].
It has been proposed that transmigration across and invasion into intestinal epithelial cells during infection is a major reason of
C. jejuni-triggered tissue damage [
2‐
4]. Investigation of gut biopsies obtained from infected patients and in vitro infection experiments of intestinal epithelial cells indicated that
C.
jejuni can enter human host cells [
9‐
11].
Campylobacter jejuni expresses various adhesins in the outer-membrane including CadF, FlpA, JlpA and PEB1 [
12‐
15]. For example, in vitro CadF is a well-known bacterial factor that binds to fibronectin, an important extracellular matrix (ECM) protein and bridging factor to integrin receptors [
13,
16,
17]. Maximal bacterial adherence and invasion of INT-407 intestinal epithelial cells is dependent on CadF and is associated with tyrosine phosphorylation of paxillin, a focal adhesion-based scaffolding factor [
18]. The expression of CadF also seems to be required for the stimulation of the small Rho GTPases Rac1 and Cdc42 via fibronectin and integrin member β1, that are required for
C. jejuni host cell entry. The signalling pathways involved in the latter process have been described in detail [
19‐
21]. However, fibronectin and integrin β1 are basolateral receptor molecules and not commonly exposed at apical surfaces in the intestine. It is therefore unclear how
C. jejuni gains access to these receptors during infection.
To access deeper tissues and cause short- or long-term infections in the human body, various pathogenic bacteria, including
Salmonella, Shigella Listeria or
Neisseria, must overcome the epithelial barrier [
22,
23]. These important bacterial pathogens are able to cross polarised intestinal epithelial cells by different mechanisms, known as the paracellular and the transcellular routes. Bacteria using the transcellular route enter host cells at apical surfaces followed by intracellular trafficking and leave these cells at the basolateral surface. In contrast, bacteria specialised on the paracellular route cross the epithelial barrier by passage between neighbouring epithelial cells and overcome the tight junctions and adherens junctions [
24]. In the case of
C. jejuni, the literature is highly controversial. While some groups reported the paracellular route, others described the transcellular model or a mix of both [
25‐
30]. In general, the host factors and bacterial factors involved in the transmigration process of
C. jejuni are still unclear [
31].
We have recently shown that a closely related pathogen,
Helicobacter pylori, secretes a novel bacterial virulence determinant into the culture supernatant, the serine protease HtrA [
32‐
34], which is also present in
C. jejuni[
35‐
37]. HtrA proteins constitute a group of heat shock induced serine proteases that influence the adhesion and invasion properties of different bacterial pathogens. HtrA proteins typically consist of a signal peptide, a trypsin-like serine protease domain and one or two protein interaction (PDZ) domains. In addition, by binding of the PDZ domain in one HtrA molecule to that in other HtrA molecules, HtrA can build-up to highly proteolytic active oligomers that also function as a chaperone [
38]. The HtrA protease domain consists of an active site, called the catalytic triad, which is formed by the conserved amino acid residues histidine, aspartatic acid and serine [
39]. Many bacterial HtrA proteins are suggested to be localized in the periplasm and to be involved in quality control of envelope proteins by degradation of misfolded proteins as well as prevention of formation of aggregates [
40]. Thus, it was surprising to find that HtrA exhibits the capability of extracelluar transport in
H. pylori[
34,
41], where it could cleave host surface molecules. We identified that
H. pylori HtrA directly cleaves the junctional protein and tumor suppressor E-cadherin and fibronectin on the surface of gastric epithelial host cells. HtrA-mediated cleavage of E-cadherin facilitated the loss of the adherence junction complex leading to the disruption of the epithelial barrier function in response to
H. pylori infection [
32] and may also apply to
C. jejuni HtrA [
33]. Here, we present the results from a detailed investigation to determine if
C. jejuni HtrA can cleave both E-cadherin and fibronectin, and whether HtrA protease activity is required for transmigration across polarised epithelial cells. Our findings show that
C. jejuni can effectively cross polarised epithelial cells in an HtrA-protease dependent fashion without affecting TER.
Conclusions
The intestinal mucosa in the human intestine forms a tight barrier, which protects against host invasion by commensals, non-pathogenic microbes residing in the intestinal lumen. Some enteric pathogenic bacteria, such as
Salmonella Shigella, or
Listeria, have specific tissue-invading properties and can physically breach the intestinal mucosal barrier [
43‐
45]. In general, these bacterial pathogens can translocate via a paracellular route or a transcellular route. A well studied example is
Salmonella enterica serovar Typhimurium which can cross the intestinal barrier preferentially by entering M cells, although they can also enter and pass through epithelial cells of the intestinal tract in vivo and in cultured polarized epithelial cells in vitro [
46‐
48]. However, very little is known about
C. jejuni transmigration
. Previous work has revealed that
C.
jejuni can translocate across Caco-2 and other polarized cell monolayers without a concomitant loss in TER [
25,
49‐
51], indicating that
C.
jejuni can cross a given polarised cell monolayer whose integrity, however, remains intact. In contrast, other research groups reported on a time-dependent decrease of TER caused by
C. jejuni infection, while the bacterial factor(s) triggering a reduction in TER were not addressed [
52‐
54]. Thus, there are some conflicting data in the literature and a consensus is yet to be reached among investigators as to the mechanism of translocation.
Our previous data suggested that HtrA chaperone activity plays a major role in
C. jejuni host cell binding, whereas HtrA protease activity mainly affected invasion [
36]. Novel data presented in this work show that HtrA from
C. jejuni can be secreted into the cell culture supernatant during bacterial growth or during infection. In addition, it was shown that
C. jejuni can cross polarised epithelial monolayers very rapidly. The first viable transmigrated wt
C. jejuni CFU were detected after 15–30 min (Figure
6 and data not shown). In contrast,
C. jejuni invasion of different host cell types was commonly observed at much later time points and was obvious between 4–6 hours or later during infection [
18‐
21,
55,
56]. These facts alone already indicate that transmigration of
C. jejuni exclude the transcellular route as a major mechanism in MKN-28 cells, which would of course take much longer time until the first bacteria reach the basolateral compartment. Instead, our findings strongly argue for the paracellular route mainly used by
C. jejuni 81
– 176 and NCTC11168. Moreover, it was found that deletion of
htrA or complementation with a protease-inactive S197A mutant exhibited a strongly reduced transmigration potential, indicating that HtrA’s protease activity indeed plays a role in this process. In addition, all
htrA mutants described here expressed flagella and were highly motile. Thus, we describe here the first
C. jejuni mutants with very high motility, but having very low transmigration and invasion potential, thus behaving like a classical avirulent Δ
flaA/B mutant.
In addition, evidence was presented that recombinant HtrA from C. jejuni can cleave-off in vitro and during infection in vivo the NTF domain from E-cadherin, a major adherens junctional protein, while it leaves the receptor molecule fibronectin uncleaved. Thus, cleavage of E-cadherin may be involved in C. jejuni transmigration. The exact cleavage site(s) in E-cadherin, however, are yet unknown and should be investigated in future studies. In addition, the total amount of cell-based E-cadherin dropped down during the course of infection, but did not lead to a complete cleavage, even at late time points of infection (8 hours). We therefore propose that cleavage of E-cadherin by HtrA during infection is a strictly controlled, temporary and locally restricted process, possibly achieved by surface-exposed and/or secreted HtrA proteins when the bacteria enter the intercellular space. Host cells continuously translate large amounts of E-cadherin proteins, and therefore the host cell machinery could rapidly replace cleaved proteins. This hypothesis could also explain why no significant reduction in TER was observed during infection with C. jejuni, and suggests that these bacteria can close the “door” behind them, which appears as a clever novel infection mechanism during bacterial transmigration across polarised gut epithelial cells.
Methods
Campylobacter strains
The
C. jejuni strains RM1221, ATCC43430, TGH-9011, NCTC11168, 1849, 81–176, 1543/01, 2703/01, ST3046 and F38011 were used in this study. The isogenic mutants 81-176Δ
cadF, 11168Δ
htrA and 11168
htrA S197A were recently described [
33,
35‐
37]. The isogenic F38011Δ
cadF and 81-176Δ
flaA/B mutants were kindly provided by Michael Konkel [
57] and Patricia Guerry [
58]. All
C. jejuni strains were grown on
Campylobacter blood-free selective Agar Base (Oxoid) containing
Campylobacter growth supplement (Oxoid) or on Mueller-Hinton (MH) agar amended with 50 μg/ml kanamycin or 30 μg/ml or chloramphenicol at 37°C under microaerobic conditions (generated by CampyGen, Oxoid) for 48 hours.
Other bacterial species
Salmonella typhimurium strain NCTC12023 was kindly provided by M. Hensel (University Osnabrueck/Germany). Neisseria gonorrhoeae strain 6B10 is a gift of T. Meyer (Max Planck Institute for Infection Biology Berlin/Germany). Shigella flexneri strain 15.4 is a clinical isolate from the Medical School Magdeburg/Germany, and Listeria monocytogenes strain EGD (Serotyp 1/2a) was kindly provided by J. Wehland (HZI Braunschweig/Germany). As control, we used the non-pathogenic Escherichia coli strain Top10 (Invitrogen). Each of these bacteria was grown overnight at 37°C on conventional LB agar plates.
HtrA secretion assays
C. jejuni wild-type and ΔhtrA deletion mutant strains were grown in BHI broth medium for 12 hours to an OD600nm ~1.0. The supernatant and the cell pellets were separated by centrifugation at 4,000 rpm, and the supernatant was further purified from remaining bacterial cells by passage through a 0.21 μm sterile filter. The resulting bacterial pellets and supernatants were analysed by immunoblot and casein zymography analyses. Absence of live bacteria in the supernatant was confirmed by incubation on agar plates showing no growth.
Motility assays
Motility phenotypes of strains were tested in MH media containing 0.4% agar. Bacterial cells were harvested from a 36 h culture on conventional agar plates and resuspended in PBS to obtain an optical density at 600 nm of 0.45 (approximately 1 × 109 CFU/ml). Subsequently, 2 μl of a bacterial suspension of 2 × 108 CFU/ml were stabbed into motility agar. Plates were incubated at 37°C under microaerophilic conditions for 36 h, followed by measuring the diameter of the resulting swarms. The final data were the mean of at least five separate measurements from three experiments.
Host cell lines
Human embryonic intestinal epithelial cells (INT-407, non-polarised), obtained from the American Type Culture Collection (ATCC CCL-6) and polarised MKN-28 cells were grown in RPMI-1640 medium containing L-glutamine and Earle’s salts (Gibco). After reaching about 70% confluency, the cells were washed two times with PBS, and then starved for 12 h before infection.
Infection studies
For the infection experiments, INT-407 cells were seeded to give 4 × 105 CFU in 12-well tissue culture plates. The culture medium was replaced with fresh medium without antibiotics 1 h before infection. Bacteria were suspended in culture medium, added to the cells at a multiplicity of infection (MOI) of 100, and co-incubated with host cells for the indicated periods of time per experiment.
Transepithelial electrical resistance (TER) assay
MKN-28 cells were cultured on 0.33 cm² cell culture inserts with 3 μm pore size (Millipore). The cells were allowed to form confluent monolayers, and then incubated for another 14 days. TER was measured with an Electrical Resistance System (ERS) (Millipore). Maximum resistance indicated that the cells reached maximal polarity. TER was calculated as Ohms x cm² by subtracting fluid resistance and multiplying by the monolayer surface area. Bacteria were suspended in culture medium, added to the cells at a multiplicity of infection (MOI) of 50, and co-incubated with host cells for the indicated periods of time per experiment. The number of CFU was determined by growth on MH or LB plates, respectively.
HtrA expression, purification and E-cadherin cleavage in vitro
Cloning of
H. pylori htrA (
Hp HtrA aa18-aa475) and
C. jejuni htrA (CjHtrA aa17-aa472) was described previously [
33,
34]. Briefly, the genes were amplified from genomic DNA excluding predicted signal peptides. PCR fragments flanked by restriction sites for
Bam HI/
Xma I were cloned into pGEX-6P-1 (GE Healthcare) to generate a GST-fusion protein. The expression and purification protocol was described in detail [
34]. Cleavage assays of purified HtrA with recombinant human full-length E-cadherin (R&D Systems), recombinant human His-tagged N-terminal NTF domain (Sino Biological) or human fibronectin (Calbiochem) were performed as described [
32].
SDS-PAGE and western blot
Cells were lysed [
32], proteins were separated by SDS-PAGE and tested for fibronectin (Santa Cruz) and E-cadherin using polyclonal antibodies recognizing the extracellular domain of E-cadherin (H-108 from Santa Cruz or HECD1 from BD Biosciences) and whole cell lysates were tested for GAPDH. The polyclonal anti-His tag antibody is from Qiagen and the rabbit HtrA antibody was described in [
36,
37]. Bacterial HtrAs were detected by Coomassie staining (BioRad).
Casein zymography
Bacterial lysates, culture supernatants or recombinant HtrA were separated in casein containing gels under non-reducing conditions. Subsequently, gels were renatured in 2.5% Triton-X-100 and equilibrated in developing buffer [
34]. Caseinolytic activity was visualized by staining with 0.5% Coomassie Blue R250.
Field emission scanning electron microscopy (FESEM)
Plate-grown C. jejuni strains were harvested and fixed in a sterile solution containing 5% formaldehyde, 2% glutaraldehyde in cacodylate buffer (0.1 M cacodylate, 0.01 M CaCl2, 0.01 M MgCl2, 0.09 M sucrose, pH 6.9) for 1 hour on ice. The solution was centrifuged and passed through a sterile filter. After several washes with cacodylate buffer and TE buffer (20 mM Tris, 1 mM EDTA, pH 6.9), samples were dehydrated in serial dilutions of acetone (10, 30, 50, 70, 90 and 100%) on ice for 15 min each step. Samples were then allowed to reach room temperature before another change of 100% acetone, after which they were subjected to critical-point drying with liquid CO2 (CPD030, Bal-Tec). Samples were finally covered with ca. 10.0-nm 11 thick gold film by sputter coating (SCD500, Bal-Tec) and examined in a field emission scanning electron microscope (Zeiss DSM 982 Gemini) using an Everhart Thornley SE detector and in-lens detector in a 50:50 ratio at an acceleration voltage of 5.0 kV.
Electron microscopic analysis by negative staining
For negative staining, thin carbon support films were prepared by indirect sublimation of carbon on freshly cleaved mica. Samples were then absorbed to the carbon film and negatively stained with 1% (wt/vol) aqueous uranyl acetate (pH 4.5). After air drying, samples were examined by transmission electron microscopy (TEM) in a Zeiss TEM 910 at an acceleration voltage of 80 kV.
Statistical analysis
All data were evaluated using Student t-test with SigmaStat statistical software (version 2.0). Statistical significance was defined by P ≤ 0.05 (*) and P ≤ 0.005 (**). All error bars shown in figures and those quoted following the +/− signs represent standard deviations.
Acknowledgements
We thank Ina Schleicher (HZI Braunschweig, Germany), Dr. Sabine Brandt (University Magdeburg, Germany) and Dr. Marguerite Clyne (UCD Dublin, Ireland) for technical support. We also thank Drs. Patricia Guerry (Fayetteville State University, USA), Michael Konkel (Pullman University, USA), Michael Hensel (University Osnabrueck, Germany), Thomas Meyer (Max Planck Institute for Infection Biology Berlin, Germany) and Juergen Wehland (HZI Braunschweig, Germany) for providing the indicated pathogens. The work of S.B. is supported through a SFI grant (UCD 09/IN.1/B2609).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MB, BH, MR and NT performed and designed the experiments. KTB, LB, OAO and SW provided crucial materials and advise for the experiments. SB, the senior/corresponding author, supervised the experiments and wrote the manuscript together with SW. All authors read and approved the final manuscript.