Epithelial Cell Surface Sites Involved in the Polyvalent Adherence of Porphyromonas gingivalis: a Convincing Role for Neuraminic Acid and Glucuronic Acid

Abstract

We investigated the target structures of the epithelial cells responsible for the attachment of Porphyromonas gingivalis by immunocytofluorimetry, enzyme-linked immunosorbent assay, and confocal microscopy. Integrins (β1, β3, and αV) and E-cadherin played no significant role. Carbohydrates (such as α-d-methylglucoside, l-fucose, d- and l-mannose, N-acetylglucosamine, and N-acetylgalactosamine) had little inhibitory effect on bacterial binding. Enzymatic treatments of the epithelial membranes and sugar competition studies showed that N-acetylneuraminic acid and glucuronic acid were involved in binding.

Porphyromonas gingivalis plays an important role in the initiation and progression of periodontal disease. It produces several adherence factors that contribute to bacterial virulence (15). P. gingivalis or P. gingivalis components have been shown to bind to collagen, fibronectin, laminin, vitronectin, and cytokeratin (13, 18, 25), but the ability of the bacteria to bind to glycoaminoglycans has been poorly studied. It has been reported that glycoaminoglycan-binding proteins may mediate the adherence of bacteria to eukaryotic cells (28), making it possible for the bacteria to initiate an infection. The goal of the study was to investigate the potential receptors in mediating the attachment of whole P. gingivalis to oral epithelial cells.

Bacterial cultures and preparation of bacterial cell surface antigens.

P. gingivalis (ATCC 33277) was maintained on blood agar plates and grown in Todd-Hewitt broth supplemented with hemin (10 μg/ml) and vitamin K₁ (1 μg/ml) for 18 h to the exponential phase in an anaerobic chamber at 37°C (1). Bacterial concentrations were determined by measuring the absorbance at 660 nm. For some experiments, the bacteria were labeled with 0.04% fluorescein isothiocyanate (FITC) in phosphate-buffered saline (PBS) for 30 min in the dark at room temperature.

Bacterial antigens were prepared by a mechanical attrition procedure by using sterile glass beads (glass bead extracts [GBE]) (4, 20). Antisera against the surface soluble antigens of P. gingivalis were raised by subcutaneous injection of young rabbits (<2 kg) with 2 mg of proteins every 3 weeks for 3 months.

KB cell cultures and cell membrane preparations.

KB cells (ATCC CCL-17) were grown to confluency in RPMI 1640 medium (BioWhittaker) supplemented with 10% fetal calf serum, 2 mM l-glutamine, and antibiotics (penicillin [100 IU/ml] and streptomycin [100 μg/ml]).

The membrane extract preparation procedure was performed at 4°C. Briefly, the cells were detached by using 0.02% EDTA, resuspended in 10 mM Tris-HCl (pH 7.6) containing 0.5 mM MgCl₂ for 10 min, and then disrupted by using a Dounce homogenizer. NaCl was added to a final concentration of 0.15 M, and the cell suspension was centrifuged at 260 × g for 5 min. A protease inhibitor mixture containing 10 μg of aprotinin/ml, 0.5 μg of leupeptin/ml, 1 mM 4-(2-aminoethyl) benzene sulfonylfluoride (AESF), 1 mM EDTA, and 1.8 mg of iodoacetamide/ml was added to the supernatant. The membrane extracts were pelleted at 150,000 × g for 45 min and solubilized in 50 mM Tris-HCl (pH 7.6) containing 300 mM NaCl, 5 mM MgCl₂, and 1% Nonidet P-40. After solubilization, the extracts were centrifuged at 10,000 × g for 15 min, and the supernatant were used as a source of KB cell membrane antigens. This soluble material was used to raise polyclonal antisera against the surface antigens of epithelial cells by subcutaneous injection of young rabbits (<2 kg) with 2 mg of proteins every 3 weeks for 3 months.

ELISA test.

Some experiments were performed to determine the optimal conditions for bacterial binding to the KB cell membrane extracts. Cell membrane extracts in 50 mM Na₂CO₃-NaHCO₃ buffer (pH 9.6) were incubated in 96-well microtiter plates (0.5 μg/100 μl/well) overnight at 4°C. The wells were washed three times with PBS containing 0.1% bovine serum albumin (BSA) and 0.05% Tween 20 (PBT). After a saturation step with PBS-3% BSA (200 μl/well), GBE suspensions (100 μl in PBS-0.1% BSA) ranging in concentration from 0.156 to 20 μg/ml were added to the wells, and the plates were incubated for 1 h at room temperature. The wells were then rinsed three times with PBT, and anti-P. gingivalis antiserum (100 μl; 1:1,000 dilution in PBT) was allowed to bind to the plates for 1 h at room temperature. After a washing step with PBT, the plates were incubated for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (100 μl, 1:1,000 dilution in PBT) and then rinsed again. The substrate solution containing 0.4 mg o-phenyldiamine (Dako)/ml in 0.05 M phosphate-citrate buffer (pH 5.0)-0.012% H₂O₂ was added, and the A₄₉₂ values were determined after the reaction had been quenched with 3 N HCl. Duplicate assays were performed for each GBE concentration. When the mean A₄₉₂ values were plotted against the bacterial membrane concentrations, the linear slopes of the exponential curves fell between 0.625 and 10 μg/ml for GBE. The amounts of KB surface antigens attached to the plates were assayed by enzyme-linked immunosorbent assay (ELISA). This assay differed from the one described above by the omission of GBE and by the use of the rabbit anti-epithelial cell membrane antisera (1:1,000 dilution in PBT) instead of anti-P. gingivalis antibodies. There were no significant differences in levels of detectable KB cells antigens with or without enzyme pretreatment.

Effects of heparinase, chondroitinase, neuraminidase, and trypsin on GBE binding to KB membrane extracts.

Heparinase type III (from Flavobacterium heparinum), chondroitinase type ABC (from Proteus vulgaris), trypsin type I (from bovine pancreas), and neuraminidase type V (from Clostridium perfringens) were purchased from Sigma.

KB membrane extracts (20 μg/250 μl) were treated for 1 h at 37°C with heparinase III (2.5 to 10 U/ml) in PBS, chondroitinase ABC (2.5 to 10 U/ml) in PBS, trypsin (5 to 120 μg/ml) in PBS, or neuraminidase (2.5 to 10 U/ml) in 50 mM acetate sodium (pH 5.5) containing 5 mM CaCl₂ and 0.1% BSA. ELISA buffer (3.75 ml) was then added to the reaction mixtures in order to assess the GBE-binding capacity of the treated cell membrane extracts.

Treatments with heparinase III (2.5 to 10 U/ml) or chondroitinase ABC (2.5 to 10 U/ml) had no negative effect on bacterial binding to the KB cell membrane antigens (Fig. 1). In fact, the heparinase treatment (10 U/ml) caused a significant 32% increase in GBE binding (P < 0.001), which may be due to conformational changes to epithelial antigens that unmask the receptor sites commonly referred to as cryptitopes. Treatment with 10 U of neuraminidase/ml (Fig. 1) or with 30 μg of (∼300 U/ml) trypsin/ml (data not shown) caused a 80% decrease in the GBE-KB cell membrane interaction (P < 0.001) compared to the results obtained in the absence of enzyme treatment. These results indicate that the GBE can bind to KB cell membrane antigens through close interactions between proteins and sugars.

FIG. 1.

Effect of various enzymes on GBE binding to KB membrane extracts. The KB membrane extracts were pretreated for 1 h at 37°C with heparinase III (x), chondroitinase ABC (▵), neuraminidase (▪), and then allowed to bind to the wells of 96-well microtiter plates (0.5 μg/well). The plates were incubated for 1 h at room temperature and binding was assessed by ELISA as described above. The results are expressed as the means of duplicate tests with three GBE concentrations (1.25, 2.5, and 5 μg/ml).

Role of epithelial glycoprotein adhesins in the P. gingivalis-KB cell interaction.

Many studies have demonstrated the involvement of eukaryotic cell adherence molecules (CAM) in bacterial binding. Isberg et al. (9) have shown that Yersinia species have a high binding affinity for β1 integrin receptors on the epithelial cell surface, whereas Mengaud et al. (16) have shown that the binding of Listeria monocytogenes to E-cadherin epithelial cell receptors results in phagocytosis of the bacteria. β3 integrin receptors are also involved in microbial binding (3, 27). More specifically, Nagakawa et al. (17) have found that anti-α5 β1 integrin antibodies inhibited the adhesion of fimbriae conjugated to microspheres to epithelial cells.

The role of cadherins and the β1, β3, and αV integrins in epithelial cell recognition of P. gingivalis antigenic structures was assessed by flow cytofluorimetry (FACScan; Becton Dickinson). Suspended KB cells (2 × 10⁵) were fixed in 1 ml of 1% paraformaldehyde then blocked for 1 h in PBS-1% BSA. After centrifugation (260 × g for 5 min), they were incubated with 5 μg of anti-pan cadherin (Sigma) or of anti-β1, -β3, or -αV integrin polyclonal antibodies (Chemicon)/ml in PBS-0.1% BSA for 45 min at 4°C. The cells were then rinsed three times with PBS and incubated in PBS-0.1% BSA containing either FITC-conjugated anti-rabbit antibody (diluted 1:20; positive control) for 45 min at 4°C or FITC-labeled bacteria for 30 min at room temperature. The bacteria were used at an infection multiplicity of 100 (i.e., 100 bacteria per epithelial cell). Bacterial adherence to epithelial cells was optimum at this bacterium/epithelial cell ratio (8). Positive controls for inhibition of P. gingivalis binding to KB cells were carried out by pretreatment of the KB cells or the bacteria with the respective immune sera diluted 1:5. The inhibition of fluorescence intensity was of 33 and 35%, respectively (Fig. 2). Negative controls were carried out with nonimmune serum. No change in the intensity of the fluorescence of the KB cells was observed compared to the negative controls, indicating that none of the anti-CAM antibodies had an inhibitory effect on the adherence of P. gingivalis to the KB cells. Cadherins and the β1, β3 and αV integrins thus do not play an important role in mediating P. gingivalis binding to epithelial cells. However, due to the complexity of the receptor-ligand interactions involved in the adherence of P. gingivalis to KB epithelial cells, a direct role for integrin or cadherin in bacterial internalization cannot be ruled out. For instance, it is well known that integrin ligand uptake is regulated from inside the cell (12, 14, 23). Recent works underlined the importance of β1 integrin on epithelial cell invasion by P. gingivalis (30). Since we used cell membranes as well as whole cells fixed with paraformaldehyde, only the binding phenomenon was studied here.

FIG. 2.

Flow cytometry histograms showing inhibition of adherence of P. gingivalis to KB cells. FITC-labeled P. gingivalis (100 bacteria per cell) were allowed to bind to suspended KB cells (2 × 10⁵). (A) The eukaryotic cells were preincubated for 45 min at 4°C with anti-keratinocyte immunserum (red line) or nonimmune serum (blue line) at 1:5 dilution. (B) The FITC-labeled P. gingivalis were preincubated for 1 h at room temperature with anti-P. gingivalis antiserum (red line) or nonimmune serum (blue line) at 1:5 dilution. In both figures, cell autofluorescence is visualized in green.

Identification of KB membrane sugars involved in the GBE-KB cell interaction.

Sugars (N-acetylglucosamine, N-acetylgalactosamine, l-mannose, d-mannose, l-fucose, and α-d-methylglucoside) were purchased from Sigma. Inhibition assays of the GBE-KB cell interaction were performed with GBE (1.25, 2.5, and 5 μg/ml in PBS-0.1% BSA) preincubated for 1 h at room temperature with sugars at various concentrations (from 0 to 200 mM).

Alpha-d-methylglucoside (200 mM) did not affect significantly the binding of GBE to KB cell membrane antigens. At a concentration of 100 mM, other sugars caused no more than an average 26% decrease in GBE binding compared to the basal level. We obtained a binding percentage of 74.0 ± 16.1, 87.5 ± 1.3, and 84.1 ± 5.5 with l-mannose, d-mannose and l-fucose (P < 0.05), respectively, and 81.1 ± 7.0 and 86.7 ± 8.8 with N-acetylglucosamine and N-acetylgalactosamine, respectively (P < 0.001). At a concentration of 200 mM, N-acetylglucosamine and d-mannose slightly decreased GBE binding of 2 to 5% compared to the percentage obtained with 100 mM (P < 0.01). A 31% inhibition of the GBE-epithelial cell membrane antigen interaction was obtained by using 200 mM l-mannose (P < 0.001). The slight inhibitory effect of simple sugars on P. gingivalis binding can be explained by the low affinity of multiple bacterial ligands at individual sites. There may also be more than one P. gingivalis recognition site on the KB cell surface.

A number of bacteria have been shown to interact with mannose-containing domains of phagocytes (2, 11, 21, 26). It has also been reported that type I Escherichia coli fimbriae mediate bacterial binding to mannose residues (5, 22). Several studies have shown that P. gingivalis fimbriae can attach to a wide range of extracellular matrix molecules (18, 24) and cell receptors (10, 19, 25, 29). The competition between sugars (including mannose) and sugar-containing epithelial cell sites for fimbriae would shed some light on the mechanisms involved in host cell-fimbria interactions.

KB cell membrane extracts were also preincubated with concanavalin A, PNA, ECA, SBA, RCA, and UEA lectins (Sigma), which possess the following glycan specificities: α-Man, α-Glc; β-Gal (1→3)GalNAc; β-Gal(1→4)GlcNAc; GalNAc; β-Gal; and α-l-Fuc, respectively. The effect of lectins on GBE binding to KB cell membrane extracts was assessed by adding lectins (6.25 to 200 μg/ml in PBS-0.1% BSA) to the wells after the saturation step. The plates were incubated for 1 h at room temperature and then rinsed three times with PBT before being incubated with GBE suspensions (at 2.5 and 5 μg/ml). None of the lectins interfered with the epithelial cell binding sites (data not shown).

GBE preincubated with different concentrations of N-acetylneuraminate (sodium salt, pH 7) displayed a dose-dependent inhibition of binding to KB membrane extracts (Fig. 3). This result confirmed that sialic residues are involved in the binding process. Interestingly, 200 mM glucuronate (sodium salt, pH 7) also inhibited binding by more than 80% (P < 0.0001) (Fig. 3). Whole-bacteria binding to KB cells was assessed by confocal microscopy by using the protocol of Houalet-Jeanne et al. (7). KB cells were first detached with EDTA 0.02%, resuspended in PBS and then incubated with FITC-labeled P. gingivalis, in the presence or absence of sugars, at an infection multiplicity of 100 for 30 min at room temperature. The addition of 25 mM glucuronate (Fig. 4) or 100 mM neuraminate (data not shown) caused a significant decrease in ATCC 33277 P. gingivalis adherence to KB cells compared to the controls.

FIG. 3.

Inhibition of GBE binding to KB membrane extracts by N-acetylneuraminic acid or glucuronic acid (ELISA). GBE were incubated for 1 h at 37°C with N-acetylneuraminic acid (▪) or with glucuronic acid (□) at the indicated concentrations before being added to wells coated with KB membrane extracts. The results are expressed as the means of duplicate tests with two GBE concentrations (1.25 and 2.5 μg/ml).

FIG. 4.

Confocal photomicrographs. KB cells were incubated with FITC-labeled P. gingivalis at an infection multiplicity of 100 for 30 min at room temperature, in the presence of 25 mM glucuronate (A) or not (B).

Sialic residues and glucuronic acid thus played an important role in P. gingivalis binding to epithelial cells. N-Acetylneuraminic acid and glucuronic acid are both highly polar molecules found in proteoglycans. Neuraminic acid is a structural component of hyaluronan, which has been also shown to mediate Treponema denticola binding to periodontal tissue (6).

The aim of our study was to identify the main eukaryotic structures involved in the P. gingivalis adherence process. We have demonstrated the predominant role played by complex sugars (N-acetylneuraminic acid and glucuronic acid) in bacterial binding to oral epithelial cells. Their involvement in adherence, and thus the subsequent stages of the infectious process, indicates that they may be plausible targets for new anti-infective agents.