INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Actinobacillus pleuropneumoniae is the causative agent of porcine fibrinohemorrhagic necrotizing pleuropneumonia (23). Twelve serotypes of A. pleuropneumoniae based on capsular and lipopolysaccharide (LPS) antigens have been recognized (24). Serotypes 1 and 5 are predominant in Québec, while serotype 2 is dominant in most European countries (22). Several bacterial factors have been suggested as important virulence attributes of A. pleuropneumoniae. The hemolytic and cytotoxic RTX toxins have been shown to be major virulence factors (6). In addition, the polysaccharidic capsule, some outer membrane proteins, and LPS also seem to be involved in virulence (9, 12, 30). We have shown that LPS play a major role in adherence of A. pleuropneumoniae to porcine respiratory tract cells and mucus (1, 2, 13, 25). The LPS are complex molecules composed of three well-defined regions: lipid A; the core region, which is an oligosaccharide containing Kdo; and the O antigen, a polysaccharide chain consisting of repeated units (11).

Selection of various tissues (tropism) by bacteria, virus, and toxins prior to colonization and infection is a well-known phenomenon (15, 16, 21, 32). In the colonization process, recognition of the carbohydrate moiety of glycoproteins and glycosphingolipids is a specific interaction which requires an adhesin (3, 30). A number of pulmonary pathogens, associated with infections in humans, specifically recognized the carbohydrate sequence GalNAc1-4Gal isolated from human lung tissues (19). It was recently shown that the gangliotetraosylceramide GgO₄ (asialo-GM1) glycosphingolipid, expressed by human regenerating respiratory epithelial cells, is recognized by Pseudomonas aeruginosa (5). Another report demonstrated a specific binding of P. aeruginosa LPS to GgO₄ glycosphingolipid on thin-layer chromatograms (TLC) (8).

In this study, putative glycosphingolipid receptors for A. pleuropneumoniae serotype 1 and serotype 2, whole cells as well as extracted LPS, were identified by using a TLC binding assay and various glycosphingolipids of acid and nonacid nature.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Glycosphingolipids. The lipids and glycosphingolipids used in this study (Table 1) were purchased from Calbiochem (La Jolla, Calif.) or Sigma-Aldrich (Oakville, Ontario, Canada).

Bacterial strains and growth conditions. A. pleuropneumoniae reference strain 4074 of serotype 1 (with a semirough LPS profile) and reference strain 4226 of serotype 2 (with a smooth LPS profile) were obtained from the National Veterinary Institute, Uppsala, Sweden. Bacterial strains were cultivated on brain heart infusion agar (Difco Laboratories, Detroit, Mich.) supplemented with 15 µg of NAD per ml. Inoculated agar plates were incubated overnight at 37°C in a 5% CO₂ atmosphere.

Extraction and isolation of LPS. LPS from A. pleuropneumoniae serotypes 1 and 2 was extracted and isolated by the method of Darveau and Hancock (4), with some modifications (27). Briefly, disrupted cells were treated with DNase, RNase, pronase, and sodium dodecyl sulfate and were subjected to MgCl₂ precipitation and high-speed centrifugation. These LPS preparations contained less than 1% protein as determined by a dye-binding assay (Bio-Rad Laboratories, Richmond, Calif.), and no bands were detected after silver staining of sodium dodecyl sulfate-polyacrylamide gels.

LPS hydrolysis. Ten milligrams (dry weight) of LPS was hydrolyzed at 100°C for 2 h in 1 ml of 1% (vol/vol) acetic acid previously saturated with nitrogen. Lipid A (insoluble) and polysaccharides (soluble) were separated by centrifugation at 12,000 × g for 10 min after neutralization with 5 N NaOH (25). The polysaccharidic fraction (also referred to as detoxified LPS) was used in the TLC binding assay.

TLC binding assay. The binding of A. pleuropneumoniae to glycosphingolipids (Table 1) separated on TLC was carried out as described previously (10). TLC were prepared in duplicate, using 4 µg of pure glycosphingolipids. The glycosphingolipids were separated on aluminum-packed silica gel plates (high-performance TLC; Merck, Darmstadt, Germany), using chloroform-methanol-water (60:35:8 by volume) as a solvent system. One TLC plate was treated with anisaldehyde as described previously (31). An identical plate was immersed in 0.3% (wt/vol) polyethylmethacrylate (Plexigum P28; Röm, GmbH, Darmstadt, Germany) in a mixture of diethylether and n-hexane (3:1 by volume) for 1 min prior to the overlay binding assay. The plates were dried at room temperature (RT) and then incubated in a blocking solution containing 2% (wt/vol) bovine serum albumin and 0.1% (wt/vol) Tween 20 in phosphate-buffered saline (PBS; 0.01 M KH₂PO₄, 0.1 M Na₂HPO₄, 1.37 M NaCl, 0.027 M KCl [pH 7.4]) for 2 h at RT. The chromatograms were then overlaid with either a bacterial suspension resuspended in PBS to an A₅₄₀ of 1.8, equivalent to approximately 3 × 10⁹ CFU/ml, or extracted LPS resuspended in PBS (2 mg/ml) and incubated for 2 h with gentle agitation. After five washes in PBS to remove unbound bacteria or extracted LPS, the TLC plates were incubated with either rabbit polyclonal antibodies raised against whole cells of A. pleuropneumoniae serotype 1 or serotype 2 or monoclonal antibodies (MAbs) directed against an epitope in the O chain of LPS of serotype 1 (5.1 G8F10) (20) or serotype 2 (102-G02) (7). After 2 h incubation at RT, the plates were washed three times with PBS, overlaid with goat anti-rabbit immunoglobulin G (heavy plus light chain)-horseradish peroxidase conjugate (Jackson ImmunoResearch Laboratories, Mississauga, Ontario, Canada) or goat anti-mouse immunoglobulin IgG (heavy plus light chain)-horseradish peroxidase conjugate (Bio-Rad Laboratories), and incubated at RT for 1 h. The chromatograms were washed again, and the conjugate-bound peroxidase was developed with 4-chloro-1-naphthol and hydrogen peroxide (Sigma). The plates were then dried at RT in a dark place before the binding activities were evaluated and photographed with an Alpha Imager 2000 (Canberra Packard Canada, Montréal, Québec, Canada). As controls, TLC plates that had not been overlaid with bacterial cells or extracted LPS were incubated with primary and secondary antibodies; these antibodies did not bind directly to the glycosphingolipids.

Inhibition with MAbs. The bacterial suspensions were preincubated with various dilutions of the MAbs against serotype 1 O antigen (5.1 G8F10) or against serotype 2 O antigen (102-G02) at 37°C with mild agitation for 30 min. The prepared mixtures were used in the overlay assay. The immunostaining and development steps were carried out as described under "TLC binding assay."

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Binding of A. pleuropneumoniae whole cells. Binding of A. pleuropneumoniae cells of serotype 1 and serotype 2 to various acid and nonacid glycosphingolipids separated on TLC was evaluated. Both serotypes of A. pleuropneumoniae bound to glucosylceramide (GlcCer), galactosylceramide sulfate (sulfatide; SO₃-3GalCer), lactosylceramide (LacSer; Gal4GlcCer), gangliotriaosylceramide (GgO₃), and GgO₄ (Table 2 and Fig. 1). A. pleuropneumoniae serotypes 1 and 2 bound weakly GalCer, whether the fatty acids were hydroxylated or not. Weak and sporadic binding of serotype 1 strain to sulfatide was detected. Strong binding of both serotypes, manifest by thick and intensely stained bands, to GgO₃ and GgO₄ on TLC plates was observed. No binding was observed for serotype 1 and serotype 2 whole bacteria to any of the gangliosides, globoseries glycosphingolipids, phosphatidylethanolamine (PE), and ceramide (Cer) on TLC.

View larger version (128K): [in this window] [in a new window]   FIG. 1.   TLC after chemical detection with anisaldehyde (A) and immunostained chromatograms obtained by overlay with cells of A. pleuropneumoniae serotype 2 (B). The purified glycosphingolipids were separated on TLC plates by using chloroform-methanol-water (60:35:8 by volume). The lanes contain glycosphingolipids (4 µg) as listed by number in Table 1.

Binding of A. pleuropneumoniae extracted LPS. Since we have previously identified LPS as the major adhesin of A. pleuropneumoniae, we wished to evaluate this molecule in the TLC binding assay. Extracted LPS of serotypes 1 and 2 bound to the same glycosphingolipids recognized by whole cells (Fig. 2B). Sulfatide was bound sporadically by extracted LPS of serotype 1 as did whole cells of serotype 1. No binding was observed by extracted LPS of both serotypes to gangliosides, globoseries glycosphingolipids, PE, and Cer.

View larger version (54K): [in this window] [in a new window]   FIG. 2.   TLC showing separated glycosphingolipids stained with anisaldehyde (A) after overlay with A. pleuropneumoniae serotype 1 extracted LPS (B) or serotype 1 detoxified LPS (C). The following glycosphingolipids (4 µg) were tested: lane 1, GlcCer; lane 2, GalCer type I; lane 3, GalCer type II; lane 4, sulfatide; lane 5, LacCer; lane 6, GgO3; lane 7, GgO4.

Inhibition of binding by MAbs specific for LPS O antigen. The binding of A. pleuropneumoniae to glycosphingolipids on TLC plates was inhibited by O-antigen-specific MAbs (Table 3). After incubation of bacteria with the specific MAbs, both serotypes 1 and 2 failed to recognize GlcCer, GalCer with hydroxy and nonhydroxy fatty acids, sulfatide, and LacCer. Incubation of bacterial strains with MAbs raised against O antigen completely abolished binding to mono- and disaccharide glycosphingolipids, whereas no such effect was shown with GgO₃ and GgO₄ asialogangliosides. As a negative control, bacterial cells were incubated with MAb against the heterologous serotype; no inhibition of binding to glycosphingolipids was detected.

Binding activity and concentration of glycosphingolipid. The relative binding affinity of GlcCer, LacCer, GgO₃, and GgO₄ on TLC plates was evaluated (Fig. 3). Different concentrations of these glycosphingolipids (3.75 ng, 7.5 ng, 15 ng, 30 ng, 60 ng, 125 ng, 250 ng, 500 ng, 1 µg, 2 µg, and 4 µg) were separated on TLC plates. The lowest amounts of GlcCer, LacCer, GgO₃, and GgO₄ which showed visible binding by bacteria were 2 µg, 500 ng, 125 ng, and 125 ng, respectively. The final amount of the glycosphingolipids needed to obtain visible bands, after chemical detection with anisaldehyde, was 250 ng.

View larger version (102K): [in this window] [in a new window]   FIG. 3.   TLC stained with anisaldehyde (A) and immunostained chromatograms obtained by overlay with A. pleuropneumoniae serotype 2 cells (B). Lanes contain the following concentrations of GlcCer (I), LacCer (II), GgO3 (III), or GgO4 (IV); lane 1, 4 µg; lane 2, 2 µg; lane 3, 1 µg; lane 4, 500 ng; lane 5, 250 ng; lane 6, 125 ng; lane 7, 60 ng; lane 8, 30 ng; lane 9, 15 ng; lane 10, 7.5 ng; lane 11, 3.75 ng.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

It has been reported by our laboratory that LPS is involved in the binding of A. pleuropneumoniae to porcine respiratory tract cells (1, 25). Interaction between receptors on target tissue and different pathogens, by means of one or more adhesin(s), is the first step in colonization and infection. The successful selection of the specific tissues by bacteria is dependent, among other factors, on the presence or absence of specific receptor(s) for that organism. The majority of described receptors for microbes on host cells are glycoconjugates, which is explained in part by the abundance of these substances at the cell surfaces (15). An in vitro TLC binding assay, described by Karlsson and Strömberg (18), was used to evaluate the binding profile of A. pleuropneumoniae whole cells and LPS to glycosphingolipids.

A. pleuropneumoniae reference strains of serotypes 1 and 2, as well as their extracted LPS, bound to GlcCer, GalCer, LacCer, sulfatide, GgO₃, and GgO₄ on TLC plates. Interestingly, these glycosphingolipids were bound by whole bacterial cells as well as their extracted LPS. To the best of our knowledge, this is one of the first reports describing the binding of LPS to glycosphingolipids, including sphingolipids with mono- and disaccharide moieties. Gupta et al. (8) reported that P. aeruginosa whole cells and extracted LPS bound the GgO₄ glycosphingolipid on TLC plates, but extracted LPS failed to detect GM1, lactosylceramide, and globoside.

Both serotypes of A. pleuropneumoniae recognized weakly hydroxylated GalCer as well as nonhydroxylated GalCer. It is known that the fatty acid of Cer plays a role on the presentation of the carbohydrates. The sulfatide was also detected by serotype 1 and serotype 2 reference strains, but serotype 1 strains bound this isoreceptor only sporadically. The presence of a sulfate group in sulfatide increased the binding compared to GalCer, which may suggest that the sulfate group is most probably important for this interaction. The sulfatide, isolated from human gastric tissue, was shown to be recognized by Helicobacter pylori, a causative agent of human gastritis (14, 26).

The binding data shown in Fig. 3 indicate that A. pleuropneumoniae recognized less well GlcCer than LacCer glycosphingolipids. The binding of A. pleuropneumoniae to LacCer, which is stronger than that to GalCer of types I and II or GlcCer, clearly indicates that the binding of bacteria to LacCer requires the Gal1-4Glc sequence. Recognition of internal sequences in various glycosphingolipid as well as binding of LacCer was previously reported for Propionibacterium spp., Neisseria gonorrhoeae, and Neisseria subflava whole bacterial cells (17, 28, 29).

To identify the region of LPS involved in binding to glycosphingolipids, detoxified LPS as well as MAbs against LPS O antigen were used. Detoxified LPS (devoid of lipid A) bound to the same glycosphingolipids recognized by extracted LPS, which suggests the involvement of the saccharide moiety of the LPS in the binding. Preincubation of bacteria with the O-antigen specific MAbs resulted in inhibition of binding to mono- and disaccharide glycosphingolipids (GlcCer, GalCer, LacCer, and sulfatide). However, binding to the asialogangliosides was not affected by incubation of bacteria with O-antigen specific MAbs. Therefore, our findings suggest that the polysaccharide moiety (presumably the O antigen) of LPS seems to be involved in binding to GlcCer, GalCer types I and II, sulfatide, and LacCer. The lactose (Gal1-4Glc) is a core structure present in most acid and nonacid glycosphingolipids. Linkage of a neuraminic acid residue abolished binding of A. pleuropneumoniae and its extracted LPS to LacCer, as demonstrated in Fig. 1. Hence, one may speculate that carbohydrate linkage to Gal1-4Glc of LacCer abolishes the binding to this receptor. Furthermore, the large difference in binding affinity between LacCer and the asialogangliosides suggest that the binding to these asialogangliosides is not based solely, if at all, on the interaction with the lactose moiety. The structural difference between GgO₃ and GgO₄ is a Gal1-3 linkage to GalNAc. However, A. pleuropneumoniae bound with comparable strength to the GgO₃ and GgO₄, indicating that the minimal binding epitope sequence appears to be the GalNAc1-4Gal disaccharide. This observation indicates also that the Gal1-3 linkage is not part of the binding epitope of GgO₄. Taken together, our data from binding of LPS and inhibition of binding by specific MAbs against LPS O antigen of serotypes 1 and 2 suggest that the core region of LPS may represent a conserved binding domain of the LPS adhesin which binds the GalNAc1-4Gal sequence found in both GgO₃ and GgO₄ glycosphingolipids. It was reported that many human pulmonary pathogens recognize the saccharide sequence GalNAc1-4Gal in fucosylasialo-GM1, asialo-GM1, and asialo-GM2 glycosphingolipids (19). On the other hand, these pathogens failed to detect GlcCer, GalCer, sulfatide, and LacCer (19), which may indicate that the binding mechanisms developed by A. pleuropneumoniae are advantageous for successful colonization and infection of its host. Inability to bind sialic acid-containing glycosphingolipids by A. pleuropneumoniae or their extracted LPS, despite the presence of the specific GalNAc1-4Gal binding sequence, may correlate with different conformational presentation of the binding epitope in a way not accessible for binding by the LPS adhesin. Human pulmonary pathogens also failed to recognize these gangliosides (19). Interestingly, the terminal part of LPS (O side chain) is involved in binding to mono- and disaccharide glycosphingolipids, which are in nature short and therefore close to the cell membrane and poorly accessible.

Multiple-carbohydrate binding specificities have been observed in various microbial pathogens (15). The fact that A. pleuropneumoniae LPS effectively recognized various saccharide sequences found in different glycosphingolipids probably represents an advantage for this pathogen. However, competitive binding studies should help establish the specificity of the interaction between LPS and various glycosphingolipids. The binding of A. pleuropneumoniae LPS to these putative receptors may eventually be used in the development of an effective multivalent antiadhesion reagent containing carbohydrates representing specific receptor binding sequences to eliminate A. pleuropneumoniae infection. Further studies are needed to determine the presence, location, and functionality of these putative receptors in the pig respiratory tract.