The enteric microbiota contributes to the pathogenesis of inflammatory bowel disease, but the pathways involved and bacterial participants may vary in different hosts. We previously reported that some components of the human commensal microbiota, particularly Clostridium perfringens (C. perfringens), have the proteolytic capacity for host matrix degradation and reduce transepithelial resistance. Here, we examined the C. perfringens-derived proteolytic activity against epithelial tight junction proteins using human intestinal epithelial cell lines. We showed that the protein levels of E-cadherin, occludin, and junctional adhesion molecule 1 decrease in colonic cells treated with C. perfringens culture supernatant. E-cadherin ectodomain shedding in C. perfringens-stimulated intestinal epithelial cells was detected with antibodies against the extracellular domain of E-cadherin, and we demonstrate that this process occurs in a time- and dose-dependent manner. In addition, we showed that the filtered sterile culture supernatant of C. perfringens has no cytotoxic activity on the human intestinal cells at the concentrations used in this study. The direct cleavage of E-cadherin by the proteases from the C. perfringens culture supernatant was confirmed by C. perfringens supernatant-induced in vitro degradation of the human recombinant E-cadherin. We conclude that C. perfringens culture supernatant mediates digestion of epithelial cell junctional proteins, which is likely to enable access to the extracellular matrix components by the paracellular pathway.
- colonic epithelial cells
- tight junction proteins
- protein degradation
- Clostridium perfringens
the intestinal epithelium lies at the interface between the intestinal microbiome and the gastrointestinal-associated tissues and represents a functional barrier to foreign antigens and microorganisms. The gut epithelium uses several defense mechanisms against microbes, such as the luminal microbiota, the mucus layer, epithelial integrity, epithelial cell turnover, and immune responses (3, 26). Some bacteria deploy enzymes, such as proteases, to bypass or overcome host defense mechanisms, and the role of the proteases produced by gut microflora in health and inflammation is beginning to be explored (4, 33).
Translocation of bacteria to the subepithelial layer is prevented by cell-cell and cell-basement membrane interactions of the intestinal epithelial cells that are responsible for the epithelial integrity. The tight junction and adherens junction (AJ) proteins involved in epithelial cell-cell interactions consist of transmembrane and cytoplasmic domains that interact with the actin cytoskeleton and regulate epithelial paracellular permeability. Tight junctions are composed of occludin, claudin, junctional adhesion molecule 1 (JAM-1), and zonula occludens proteins that play an important role in the barrier function and the establishment and maintenance of apico-basal polarity in epithelial cells (16, 27, 29). E-cadherin, the major representative of the AJ, is a transmembrane glycoprotein that consists of an extracellular domain with five homologous repeats, a single transmembrane region, and a highly conserved cytoplasmic domain (40). The extracellular domains play important roles in connecting neighboring cells through calcium-dependent homophilic interactions, and the cytoplasmic (intracellular) domain of E-cadherin interacts with the actin cytoskeleton via β-and α-catenins. In addition, cadherins play important roles in cell signaling, proliferation, and differentiation (34, 37).
The tight junctions are disrupted during inflammation resulting in increased paracellular permeability and microbial penetration into intestinal tissues (1, 39). Patients with inflammatory bowel disease (IBD) have an increased intestinal paracellular permeability and defective regulation of tight junction proteins (38, 39).
In a previous study, we identified a number of proteolytic bacteria by screening the bacterial colonies from fecal samples of 31 patients with IBD and 20 healthy controls (24). The proteolytic activity was associated with a diversity of bacteria (Clostridium perfringens, Staphylococcus epidermidis, Enterococcus faecalis, and Bacillus thuringiensis/Bacillus cereus) but was predominantly linked with Clostridium perfringens (C. perfringens) (24).
Host and bacterial proteases have a broad range of substrates spanning from extracellular matrix components to tight junction proteins, cytokines, chemokines, and growth factors (18–20, 23, 41). Several investigators have reported that components of the gut microflora can degrade or delocalize epithelial junction proteins (14, 30, 31, 43). In this study, we examined the C. perfringens-derived proteolytic activity against epithelial tight junction proteins using human intestinal epithelial cell lines.
MATERIALS AND METHODS
C. perfringens and Escherichia coli (E. coli) K12 were grown overnight at 37°C under anaerobic conditions in Tryptone Soya Broth (TSB) containing 0.5 g/l l-cysteine. C. perfringens isolated from a patient with ulcerative colitis (UC8.1) was chosen based on its high proteolytic activity (24). Sterile filtered supernatants from bacterial cultures were obtained by centrifugation of 16-h cell cultures for 10 min at 5,000 revolution/min followed by subsequent passing of the supernatant through a 0.22-μm pore size, 33-mm Millex filter with Durapore (PVDF) membrane (the lowest protein-binding syringe filters available, from Millipore, Billerica, MA).
The human intestinal Caco-2 and C2BBe1 epithelial cell lines (ATCC HTB-37 and ATCC CRL-2102, respectively; ATCC, Manassas, VA) were maintained in Dulbecco's Modified Eagle Medium (DMEM; Gibco, Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum (Sigma-Aldrich, St. Louis, MO), 100 μg/ml penicillin and 100 U/ml streptomycin (Gibco), 100 μM nonessential amino acids (Gibco), and 0.01 mg/ml transferrin (Calbiochem, San Diego, CA). The C2BBe1 cell line is a clonal derivative of human colon adenocarcinoma Caco-2 cells. The colonic cells were allowed to grow to 90–100% confluence, then aliquoted into six-well plates at 5 × 105 cells/well in supplemented DMEM, and incubated for 7 days at 37°C in 5% CO2. For polarized monolayers, C2BBe1 cells were seeded on a Millicell hanging cell culture insert (Millipore) with a 3.0-μm pore size at a density of 2 × 105 cells/insert and cultured for 21 days in 24-well plates until the transepithelial electrical resistance was greater than 300 Ω/cm2 when cells were fully differentiated. Culture medium was replaced every 2 days until the cells reached confluency or complete differentiation. The cells were passaged based on ATCC-suggested protocols using a 0.25% trypsin, 0.53 mM EDTA solution (ATCC). Confluent or polarized intestinal epithelial cells were stimulated with different volumes of C. perfringens cell culture or filtered sterile culture supernatant for various times as described in the respective figure legends. Bacterial growth medium was used as a no-stimulation control.
Resazurin-based cytotoxicity assay.
The cytotoxic effect of C. perfringens-filtered sterile culture supernatant on colonic C2BBe1 cells was evaluated by the resazurin test (Sigma-Aldrich). Intestinal epithelial cells were seeded into six-well plates at 5 × 105 cells/well in supplemented DMEM and incubated for 7 days at 37°C in 5% CO2 before treatment with C. perfringens cell-free supernatant. Following treatment with the bacterial supernatant, the colonic cells were washed with PBS and subsequently incubated with 1 ml of resazurin solution (10% vol/vol, in DMEM) for 5.5 h at 37°C, to allow viable cells to convert resazurin from the oxidized blue form to the reduced pink form (resorufin). After incubation, the resazurin solution was collected, and absorbance was measured at 570 and 600 nm using a SpectraMax M2 microplate reader (Molecular Devices, Sunnyvale, CA). The percentage of resazurin reduction was calculated and compared with the controls per manufacturer's instructions (AbD Serotec, Raleigh, NC). Culture medium (TSB or DMEM) was used as a no-stimulation control, and treatment with EDTA was used to show the cytotoxic effect (decreased cell viability as a decrease in resazurin reduction to resorufin) on colonic cells.
Proteins were extracted from the human intestinal epithelial cells using cell lysis buffer (Cell Signaling Technology, Danvers, MA) supplemented with halt protease and phosphatase inhibitor cocktail (protein research products; Thermo Scientific Pierce, Rockfield, IL), according to the manufacturer's instructions. Protein concentration was determined using Bradford protein assay (Bio-Rad, Hercules, CA) or the Thermo Scientific NanoDrop 1000 spectrophotometer. Total protein (90 μg) of the cell protein extracts was loaded and electrophoresed in running buffer on a 10 or 15% Tris-glycine SDS-polyacrylamide gel after boiling for 5 min at 100°C in Laemmli sample buffer (Bio-Rad). Following the SDS PAGE, proteins were transferred to nitrocellulose membrane at 100 V, 4°C for 60 min. After being blocked in Tris-buffered Saline with 0.1% Tween-20 containing 5% wt/vol nonfat dry milk, the membrane was incubated at 4°C overnight with primary antibodies [at 1:1,000 dilution, against caspase-3 from Cell Signaling Technology; 1:500 dilution, against matrix metalloproteinase (MMP)-7 from R&D Systems, Minneapolis, MN; 1:2,000 dilution, against ADAM-10 and 1:5,000 dilution, against β-actin from Sigma-Aldrich; 1:2,000 dilution, against JAM-1 and 1:100 dilution, against occludin from Abcam, Cambridge, UK; 1:1,000 dilution, against claudin-1 from Invitrogen, Camarillo, CA]. The primary antibodies against the extracellular domain of E-cadherin were used at 1:10,000 dilution (Millipore) and against the cytoplasmic domain at 1:5,000 dilution (BD Biosciences, San Jose, CA).
Immunoreactive bands were detected on membranes after incubation with appropriate anti-rabbit or anti-mouse IRDye 800-conjugated secondary antibodies (at 1:10,000 dilution, from Rockland, supplied by Tebu-Bio, Peterborough, UK) using a LI-COR Odyssey Infrared Imaging Scanner (LI-COR Biosciences, Lincoln, NE).
For caspase-3 detection, an anti-rabbit secondary antibody conjugated to horseradish peroxidase (Cell Signaling Technology), a biotinylated protein ladder (Cell Signaling Technology), an anti-biotin antibody conjugated to horseradish peroxidase (Cell Signaling Technology), and a chemiluminescent detection system were used (SuperSignal West Dura Extended Duration Substrate, Thermo Fisher Scientific).
For protein detection in Caco-2 cell protein extracts, 150 μg total protein was loaded, and, following the SDS PAGE, proteins were transferred to Immobilon-P Membrane, PVDF, 0.45 μm (Millipore) at 100 V, 4°C for 60 min. Immunoreactive bands were detected on membranes using the anti-rabbit Alexa Fluor 488-conjugated (at 1:2,000 dilution, Dianova, Hamburg, Germany), anti-mouse Alexa Fluor 594-conjugated (at 1:2,000 dilution, Abcam, Cambridge, UK) secondary antibodies and the Typhoon FLA 7000 scanner (GE HealthCare Life Sciences, Freiburg, Germany).
For all Western blots, the Precision Plus Protein Dual Color standards protein marker (Bio-Rad) was used, and the densitometric quantitation was performed with ImageJ software. The densitometry of protein abundance was normalized to the loading control (namely β-actin).
In vitro degradation of rhE-cadherin.
Recombinant human E-cadherin Fc chimera (R&D Systems) was incubated with C. perfringens-filtered sterile culture supernatant. When different protease inhibitors were tested, they were incubated first with C. perfringens culture supernatant for 5 min at room temperature before the proteolytic substrate was added and subsequently incubated at 37°C for the times specified in the figure legends. Electrophoresis was carried out using a Mini-PROTEAN Tetra System (Bio-Rad) and 10% SDS-polyacrylamide gels, which were stained with Coomassie Blue, and the protein bands were detected on gels using a LI-COR Odyssey Infrared Imaging Scanner (LI-COR Biosciences).
Culture supernatant of C. perfringens decreases E-cadherin protein levels in human colonic epithelial cells in a time- and dose-dependent manner.
Reduced transepithelial resistance after exposure of rat distal colon to culture supernatants of C. perfringens in Ussing chambers (24) prompted us to investigate the effect of C. perfringens-derived proteolytic activity against epithelial tight junction proteins in human intestinal C2BBe1 epithelial cell line. The selection of this cell line was based on its enterocytic differentiation characteristics and formation of a polarized monolayer with an apical brush border morphologically comparable to that of the human colon (21).
Intestinal epithelial cells grown for 7 or 21 days for confluent or polarized monolayers, respectively, were treated for various times, from 6 to 19 h, with increasing volumes (100 to 600 μl) of C. perfringens cell-free supernatant (Fig. 1).
The full-length E-cadherin (the major adhesion molecule in AJ) protein levels were analyzed by Western blotting using an antibody against either the extracellular (Fig. 1, A–C) or cytoplasmic domain of E-cadherin (Fig. 1D). Western blot analysis revealed a decrease in the full-length E-cadherin protein levels in a time- and dose-dependent manner (Fig. 1A). Remaining extracellular E-cadherin levels reached ∼50% of the levels in control sample, treated with broth (TSB) alone, at high concentrations (500 and 600 μl) of C. perfringens cell-free supernatant after 6-h treatment (Fig. 1A). With an increase in the time of treatment with C. perfringens culture supernatant, ∼50% of the remaining extracellular E-cadherin levels (compared to controls) were reached at intermediate concentrations (300 μl after 10-h treatment and 200 μl after 19-h treatment) of C. perfringens cell-free supernatant (Fig. 1A and Fig. 2A).
The protein levels of the full-length E-cadherin decreased dramatically in confluent and polarized C2BBe1 cells treated for 19 h with either C. perfringens cell culture or cell-free supernatant (Fig. 1B).
We showed a dramatic decrease in the full-length E-cadherin protein levels in another cell line, the Caco-2 human colonic cell line, utilizing similar treatment conditions with the C. perfringens bacterial supernatant (Fig. 1C).
After 19-h treatment, a decrease in the cytoplasmic full-length E-cadherin levels was observed at high concentrations of C. perfringens culture supernatant concomitantly with the generation of an ∼28-kDa COOH-terminal E-cadherin fragment (Fig. 1D). For further experiments, we mainly used confluent C2BBe1 cells and 19-h treatment with the bacterial supernatant.
C. perfringens induces E-cadherin ectodomain shedding in human intestinal epithelial cells.
Following 19-h treatment of colonic C2BBe1 cells with C. perfringens cell-free supernatant, the extracellular full-length E-cadherin protein levels, detected in whole cell lysates of C2BBe1 with an antibody against the extracellular domain, were reduced in a dose-dependent manner (Fig. 2A). No cleavage of intact 120-kDa E-cadherin was detected in untreated cells (DMEM) or samples treated with broth (TSB) alone, and the full-length extracellular E-cadherin levels were similar in these controls (Fig. 2A, inset). To determine whether the reduction in the full-length extracellular E-cadherin levels in C. perfringens-treated samples is the result of ectodomain shedding, we assessed the release of the ∼80-kDa ectodomain of E-cadherin into cell supernatants of C2BBe1 after treatment with C. perfringens cell-free supernatant. The proteins in the cell culture supernatants were precipitated with 10% trichloroacetic acid, and the E-cadherin ectodomain was detected with an antibody against the E-cadherin extracellular domain.
The reduction in the extracellular full-length E-cadherin in whole cell lysates was coupled with the accumulation of the ectodomain in the cell culture supernatant of C2BBe1 cells in a dose-dependent manner (Fig. 2B). Lower levels of full-length E-cadherin were reached at increasing concentrations of C. perfringens culture supernatant with a concomitant significant accumulation of the E-cadherin ectodomain in colonic cell culture supernatants.
No induction of E-cadherin ectodomain shedding was detected in E. coli K12 cell-free supernatant-stimulated colonic cells (Fig. 2B, inset). E-cadherin ectodomain levels were similar in samples treated with broth (TSB) alone or E. coli-stimulated C2BBe1 cells. These data indicate that secreted components of C. perfringens are required for E-cadherin cleavage and release of the soluble E-cadherin ectodomain into the cell culture supernatants.
Culture supernatant of C. perfringens decreases protein levels of occludin and JAM-1 in colonic cells.
Stimulation of C2BBe1 cells for 19 h with increasing volumes (100 to 600 μl) of C. perfringens culture supernatant resulted in a dramatic decrease in protein levels of the tight junctions occludin and JAM-1 but not claudin-1 (Fig. 3). The decrease in the tight junction protein levels occurred in a dose-dependent manner, lower levels being reached at treatments with higher concentrations of C. perfringens cell-free supernatant (Fig. 3). To test whether the tight junction proteins are affected in a similar manner in another cell line, we chose to measure the JAM-1 protein levels in C. perfringens-treated Caco-2 colonic cells under similar conditions and observed a similar decrease as in treated C2BBe1 cells (data not shown).
C. perfringens culture supernatant has no cytotoxic activity on human intestinal epithelial cells.
To determine whether the C. perfringens cell-free supernatant has a cytotoxic effect on the intestinal epithelial cells, we used a resazurin-based cytotoxicity assay. The colonic C2BBe1 cells were treated with increasing volumes (100 to 600 μl) of C. perfringens supernatant in similar conditions to the ones tested for the E-cadherin and tight junction protein levels.
Cell viability as percentage of control was similar in C2BBe1 colonic epithelial cells treated with either broth (TSB) alone or increasing volumes of C. perfringens culture supernatant (Fig. 4A). Percentage reduction of resazurin was similar in C2BBe1-untreated samples (DMEM) and in samples treated with either broth (TSB) alone or increasing volumes of C. perfringens culture supernatant (Fig. 4, B and C). Treatment with EDTA was used to show the cytotoxic effect (decreased cell viability) on C2BBe1 colonic cells.
C. perfringens culture supernatant does not induce or activate the host proteases involved in E-cadherin cleavage.
The release of the ectodomain by C. perfringens treatment could be due to direct cleavage of E-cadherin by the proteases secreted into culture supernatant by C. perfringens or to the activation of host MMPs involved in E-cadherin shedding by C. perfringens-secreted factors.
Several host MMPs including stromelysin-1 (MMP-3), matrilysin (MMP-7) (17), ADAM-10 (12), and meprin-β (9) are involved in E-cadherin shedding, and E-cadherin is also a substrate for presenilin-1/γ-secretase system and caspase-3 (11, 35).
Therefore, we wanted to know whether the secreted factors of C. perfringens activate the host proteases. Using Western blot analysis, we determined the MMP-7 protein levels in untreated C2BBe1 colonic epithelial cells or treated with either broth (TSB) alone or increasing volumes of C. perfringens culture supernatant (Fig. 5A). The pro-MMP-7 levels were similar in untreated and treated samples (data not shown), and no activation of the proenzyme was detected in any sample using either whole cell lysates or cell culture supernatants of C2BBe1 cells (Fig. 5A).
ADAM10 is involved in the proteolytic processing of E-cadherin in vitro and in vivo, and ADAM10 overexpression, inhibitor studies, and RNA interference-mediated downregulation of endogenous ADAM10 demonstrated that ADAM10 represents an E-cadherin sheddase in established and primary epithelial cell lines (12). To test whether the secreted proteases of C. perfringens affect the host ADAM-10 levels, we determine the ADAM-10 protein levels in treated and untreated C2BBe1 colonic epithelial cells, and similar levels were detected in all samples (data not shown).
During apoptosis, in addition to the caspase-3-mediated cleavage of E-cadherin proximal to the transmembrane segment of E-cadherin releasing the cytoplasmic domain into the cytosol, a host metalloproteinase sheds the extracellular domain from the cell surface (35). We wanted to know whether the treatment with C. perfringens cell-free supernatant activates caspase-3 in colonic epithelial cells. Caspase-3 protein levels were similar in treated and untreated C2BBe1 cells, and no caspase-3-cleaved fragments were detected in any sample (Fig. 5B).
In vitro degradation of recombinant human E-cadherin (rhE-cadherin) by C. perfringens culture supernatant and the effects of protease inhibitors.
We established that the protein levels of some host proteases involved in E-cadherin cleavage (MMP-7, ADAM-10, and caspase-3) are similar in untreated and treated colonic cells with C. perfringens cell-free supernatant (Fig. 5). We were next interested in determining whether the secreted proteases of C. perfringens are directly involved in E-cadherin degradation. Therefore, we tested in vitro degradation of human recombinant E-cadherin (rhE-cadherin) by C. perfringens cell-free supernatant. rhE-cadherin was degraded in vitro when incubated with C. perfringens supernatant, and we believe that a secreted metalloprotease of C. perfringens cleaves rhE-cadherin at a specific site, resulting in the generation of a fragment corresponding to the molecular weight of ∼80 kDa of the E-cadherin ectodomain and a smaller fragment of about 35 kDa (Fig. 6, A and B). Efforts to determine the NH2-terminal amino acid sequence of the smaller protein fragment, which will allow us to map the cleavage site of the bacterial metalloprotease, were unsuccessful. Our speculation was based on the results with the metalloprotease inhibitor, EDTA, when the rhE-cadherin was cleaved at different sites (Fig. 6, A and B). This proteolytic activity revealed in the presence of EDTA is most likely the result of a cysteine protease because EDTA is an activator of cysteine proteases and based on previous studies that showed that no proteases other than a metalloprotease (with collagenolytic and gelatinolytic activities) and a cysteine protease (Clp, clostripain-like protease) have been detected in culture filtrates of C. perfringens (10, 13, 28, 36).
We next used cysteine protease inhibitors (antipain and leupeptin) to determine whether the cysteine protease secreted by C. perfringens is involved in rhE-cadherin cleavage. Complete inhibition of the proteolytic activity of C. perfringens supernatant against rhE-cadherin was achieved in the presence of cysteine protease inhibitors (Fig. 6C). Because these are also serine protease inhibitors, the involvement of any serine proteases was eliminated by using PMSF, a serine protease inhibitor, which had no effect on rhE-cadherin degradation by C. perfringens cell-free supernatant (Fig. 6, A and D).
Longer incubation (20 min) of the rhE-cadherin with C. perfringens supernatant in the presence of EDTA resulted in complete degradation of the ∼82-kDa fragment generated after 10-min incubation (Fig. 6C). The ∼80-kDa fragment generated after 10-min incubation with C. perfringens supernatant in the absence of EDTA was not degraded by longer incubation, demonstrating again the distinct cleavage of E-cadherin by the secreted proteases of C. perfringens (Fig. 6C). No cleavage of rhE-cadherin by E. coli K12 cell-free supernatant was detected in vitro (Fig. 6D).
In this study, we explored C. perfringens-derived proteolytic activity against epithelial tight junction proteins. We detected a dramatic decrease in protein levels of the tight junction occludin and JAM-1, but not claudin-1, in intestinal epithelial cells treated with C. perfringens-filtered sterile culture supernatant. The unchanged claudin-1 protein levels could be explained by changes in protein localization after treatment with the bacterial supernatant, making claudin-1 inaccessible to the bacterial proteases, and/or by substrate specificity of the C. perfringens-secreted proteases.
E-cadherin (the major component of epithelial AJ) protein levels detected with an antibody against the extracellular domain decreased in a time- and dose-dependent manner in C. perfringens-treated colonic cells. We concluded that this decrease was the result of E-cadherin ectodomain shedding because the decrease in the full-length extracellular E-cadherin was nicely coupled with the accumulation of the soluble ectodomain in the epithelial cell culture supernatants.
The direct involvement of the C. perfringens-secreted proteases in E-cadherin cleavage was shown by C. perfringens supernatant-induced in vitro degradation of the human recombinant E-cadherin. We suggest an interplay between the previously described C. perfringens-secreted proteases (a metalloprotease with collagenolytic and gelatinolytic activities and a clostripain-like cysteine protease) based on the following findings: 1) in the presence of the metalloprotease inhibitor, EDTA, the initial cleavage was inhibited and the rhE-cadherin was cleaved at different sites; 2) no E-cadherin cleavage was observed in the presence of cysteine protease inhibitors; and 3) PMSF, a serine protease inhibitor, had no effect on the E-cadherin cleavage, alone or with the C. perfringens supernatant. Therefore, we speculate that the activity of the C. perfringens-secreted cysteine protease is necessary for activation of the metalloprotease. Once activated, the metalloprotease sheds the E-cadherin ectodomain and in turn inhibits cleavage of E-cadherin by the cysteine protease. To dissect this crosstalk between the C. perfringens-secreted proteases, we would like to mention the need for using purified proteases and specific protease inhibitors for future in vitro degradation experiments. We are in the process of purifying both the endogenous and recombinant (COOH-terminally 6xHis-tagged) proteases. The pET28a(+) vector was used for construction and overexpression of the His-tagged variants in E. coli BL21(DE3). For purification of the endogenous proteins from the cell culture supernatant, we used ammonium sulfate precipitation and ion exchange chromatography. Our preliminary experiments showed that the collagenase (metalloprotease) binds to the DEAE anion exchange column, but further purification steps are necessary (e.g., gel filtration chromatography).
Several studies showed the involvement of microbial-derived proteases in disruption of the tight junctions and alterations in the intestinal epithelial barrier function. Decreased protein levels of occludin, E-cadherin, and desmoglein-2 were detected in Caco-2 monolayers inoculated with Candida albicans (6). The extracellular domain of E-cadherin that serves as a receptor for Listeria monocytogenes surface protein IntA, Candida albicans invasin Als3, and pneumococcal surface adhesin A (PsaA) of Streptococcus pneumonia, is the target for several bacterial proteases (2, 15, 22). The Bacteroides fragilis metalloprotease toxin, BFT or fragilysin, induces in vitro release of the E-cadherin ectodomain in intestinal epithelial cells and intestinal tissues cultured with purified BFT (25, 44). Helicobacter pylori secretes the HtrA serine protease that has been reported to disrupt epithelial barrier function by inducing ectodomain E-cadherin shedding independently of cellular host proteases (8, 42). In addition, different HtrAs expressed by other Gram-negative gastrointestinal pathogens including enteropathogenic E. coli, Shigella flexneri, and Campylobacter jejuni cleave E-cadherin on host cells (5, 7).
E. faecalis-derived gelatinase triggers the loss of extracellular E-cadherin that leads to the impairment of the intestinal barrier function and contributes to development of chronic intestinal inflammation in mice that are susceptible to intestinal inflammation (32).
Our study brings additional evidence for the involvement of Gram-positive bacteria-derived proteases in degradation of the tight junction proteins. Another aspect that will be further explored is the role of C. perfringens-derived proteases in cell signaling based on the hypothesis that cleavage of E-cadherin will affect the β-catenin levels, cellular localization, and the expression of target genes of the Wnt/β-catenin signaling.
This work was supported by Science Foundation Ireland and, in part, by funding for the Intestinal Proteases: Opportunity for Drug Discovery consortium through Coordination Theme 1 (Health) of the European Union's FP7, Grant agreement number HEALTH-F2-2008-202020.
F. Shanahan is affiliated with a multidepartmental university campus company, Alimentary Health, which investigates host-microbe interactions. The content of this paper was neither influenced nor constrained by that fact. No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: M.P. and F.S. conception and design of research; M.P. performed experiments; M.P. analyzed data; M.P. interpreted results of experiments; M.P. prepared figures; M.P. drafted manuscript; M.P. and F.S. edited and revised manuscript; M.P. and F.S. approved final version of manuscript.
Present affiliation for M. Pruteanu: Genome Biology Unit, European Molecular Biology Laboratory, Meyerhofstrasse 1, Heidelberg 69117, Germany.
The authors thank Barry Kiely, Alimentary Health, Cork, Ireland for providing the C. perfringens UC8.1 strain. We also thank Nassos Typas from EMBL Heidelberg, Germany for use of laboratory facilities.
- Copyright © 2013 the American Physiological Society