Vol. 60, No. 4
INFECTION AND IMMUNITY, Apr. 1992, p. 1434-1440 0019-9567/92/041434-07$02.00/0 Copyright X 1992, American Society for Microbiology
Identification of the Mucin-Binding Adhesin of Pseudomonas cepacia Isolated from Patients with Cystic Fibrosis S. U. SAJJAN AND J. F. FORSTNER*
Research Institute, Department of Biochemistry, The Hospital for Sick Children, University of Toronto, Toronto, Ontario M5G 1X8, Canada Received 30 October 1991/Accepted 30 January 1992
In previous experiments, we have shown that isolates of Pseudomonas cepacia from sputa of patients with cystic fibrosis (CF), particularly those with severe lung infection, exhibited specific binding to purified respiratory or intestinal mucins (U. Sajjan, M. Corey, M. Karmali, and J. Forstner, J. Clin. Invest. 89:648-656, 1992). The present report describes the identification of the adhesin as a protein located on fimbriae of mucin-binding P. cepacia. From a total of 53 isolates available (from 22 patients with CF), we used three mucin-binding and three non-mucin-binding isolates for our experiments. Sodium dodecyl sulfatepolyacrylamide gel electrophoresis of crude P. cepacia homogenates was performed, the separated proteins were blotted onto nitrocellulose and overlaid with purified mucin, and mucin-binding components were detected with an antimucin antibody and then a second-antibody-alkaline phosphatase conjugate system. Only mucin-binding isolates exhibited a positively stained band at an Mr of 22,000. The 22-kDa protein was purified, and a polyclonal antibody specific for it was developed in rabbits. By electron microscopy and immunogold labelling, both the antibody and mucin (separately) were localized to pili present over the entire surface of the bacterial cells. Non-mucin-binding isolates did not have (or had very few) pili and did not stain with either mucin or the antibody to the 22-kDa protein. The purified 22-kDa protein and its antibody were each able to inhibit piliated P. cepacia binding to mucin. The amino acid composition of the 22-kDa protein was dissimilar to those of the major pilin proteins ofEscherichia coli (type 1 pilus) and P. aeruginosa (PAK and PAO1 strains). Both the pili of P. aeruginosa PAK and PAO1 and antibodies to these pili failed to inhibit P. cepacia binding to mucin. Thus, P. cepacia adhesion to mucin is mediated by a pilin-associated 22-kDa protein which differs from epithelial-cell-binding pilin proteins ofP. aeruginosa. We postulate that the 22-kDa adhesin may play a role in the virulence of P. cepacia lung infections of patients with CF.
pilin-associated protein with an Mr of 22,000 mediates binding and that this protein does not cross-react with antipilin antibodies to Pseudomonas aeruginosa, a common colonizer of lungs of patients with CF.
Pseudomonas cepacia is an opportunistic pathogen which colonizes the lungs of some patients with cystic fibrosis (CF) (5) and some immunocompromised patients (3, 15). The prevalence of P. cepacia infection in some CF clinics has gradually increased over the past 20 years (8, 26), as has premature mortality from the "cepacia syndrome" (27). Very little information concerning the early pathogenic events of P. cepacia infection is available, although in vitro adherence to respiratory cells has been described. Saiman et al. (20) reported that outer membrane proteins of P. cepacia mediate binding, while Kuehn et al. (11, 12) suggested that polar pili are responsible. In other articles (23, 24), we report that P. cepacia isolates from 19 patients treated at the Hospital for Sick Children, Toronto, Ontario, Canada, exhibited specific binding to carbohydrates of mucins purified from the respiratory and intestinal tracts of patients with and without CF, respectively. No differences in the binding capacities of the two mucins were observed. Isolates from those patients with the most severe lung damage tended to bind most strongly to mucins. Since a mucus blanket normally covers the epithelial surfaces of the respiratory tract in vivo and is poorly cleared from lungs of patients with CF, we have speculated that P. cepacia binding to this reservoir of stagnant mucus may be a potential virulence factor. The goal of the present study was to identify the bacterial adhesin responsible for the specific attachment of P. cepacia to mucin receptors. Our results indicate that a unique *
MATERIALS AND METHODS
Bacteria, growth conditions, and radiolabelling. Fifty-three isolates of documented P. cepacia from sputum cultures of 22 patients with CF were provided to us by M. Karmali, Department of Microbiology, Hospital for Sick Children. The lyophilized primary cultures were suspended in sterile water, grown on brain heart infusion (BHI) agar overnight at 37°C, and stored in 5% trisodium citrate containing 40% glycerol at -70°C before use. In preparation for mucin binding assays, stock cultures of six isolates were subcultured on BHI agar and a single colony was inoculated into 10 ml of Trypticase soy broth containing 0.1 mCi of [3H]sodium acetate (ICN Biomedicals Inc., Costa Mesa, Calif.) (specific activity, 2.5 Ci/mmol). After overnight growth at 37°C in an orbit shaker (Canlab, Mississauga, Ontario, Canada) at 150 rpm, bacteria were harvested by centrifugation (10,000 x g for 10 min at 4°C), washed three times, and suspended in phosphate-buffered saline (Na2 HPO4-NaH2PO4 [0.01 M] in 0.15 M NaCI) (PBS; pH 7.2) containing 0.1% bovine serum albumin (BSA; globulin free; Sigma Chemical Co., St. Louis, Mo.). The specific activity of the radiolabelled P. cepacia was 2.06 x 10' dpm/CFU. Pili from P. aeruginosa (strains PAK and PAO1) were purified by the method of Paranchych et al. (17). A polyclonal antibody to PAK pili and monoclonal antibodies
Corresponding author. 1434
VOL. 60, 1992
PK3B (specific for PAK pili) and PK34C (specific for the C-terminal PAK and PA01 pilin epithelial-cell-binding domains) (4), were kindly provided by R. T. Irvin, University of Alberta, Edmonton, Alberta, Canada. Isolation and purification of mucin. Mucins from intestinal secretions of persons without CF and sputum of persons with CF were purified as described earlier (19). Since P. cepacia binds equally in vitro to both mucins (23, 24), and since intestinal mucin was available in much greater quantity, intestinal mucin was used for the binding experiments described. A previously described polyclonal antibody for human intestinal mucin (14, 18) was used in many experiments to detect mucin binding either to bacterial cells or to individual bacterial components (see below). Isolation of the mucin-binding adhesin of P. cepacia. The method of Paranchych et al. (17) for purification of pili was used with some modifications. From three mucin-binding isolates of P. cepacia (designated PC 5, PC 7, and PC 24) which had been subcultured and grown overnight on BHI agar, bacterial colonies were inoculated into 4 liters of Trypticase soy broth and grown for 36 h at 37°C on a gyratory shaker (New Brunswick Scientific Co., Inc.) at 150 rpm. Bacteria were harvested by centrifugation (5,000 x g for 20 min), washed once with PBS, and suspended in 200 ml of 5 mM Tris HCI (pH 7.2) containing 200 nM phenylmethylene sulfonic acid, 2 mM EDTA, and 0.02% sodium azide. The suspension was homogenized in a polytron at 2,000 rpm for 10 min at 4°C and centrifuged at 10,000 x g for 30 min, and ammonium sulfate (final concentration, 50%) was added to the supernatant. After 17 h at 40C, the pellet was collected by centrifugation, suspended in the Tris HCI buffer described above, and recentrifuged at 5,000 x g for 10 min. The supernatant was adjusted to 20% saturation with ammonium sulfate, left for 2 h at 4°C, and recentrifuged. The final pellet was washed, suspended in H20, and stored at -200C. This preparation was designated the adhesin-enriched fraction. It contained several proteins, as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), although one band at 22 kDa was of special interest. The adhesin was further purified by subjecting the enriched fraction to SDS-PAGE (15% polyacrylamide gels) (13) and electroeluting the mucin-binding band (Mr, 22,000) in a Bio-Rad Laboratories (Richmond, Calif.) electroeluter (model 422) with 0.05 M ammonium bicarbonate containing 0.1% SDS. From 4 liters of Trypticase soy broth, the yield of purified adhesin was approximately 2 to 4 mg. Purification of the same 22-kDa protein from one P. cepacia isolate which was known not to bind to mucin was attempted, but no product was obtained. SDS-PAGE, electroblotting, Western blot (immunoblot) analyses, and mucin overlays. Crude homogenates of P. cepacia (or fractions generated during the purification of the mucin-binding adhesin), were prepared from overnight broth cultures by homogenization in a polytron for 30 s and centrifugation. The supernatants were applied to 12.5 or 15% polyacrylamide gels and subjected to SDS-PAGE (13). Proteins were transferred to nitrocellulose membranes (28); the membranes were blocked with 3% BSA and incubated for 1 h at 370C with specific antibodies or purified mucin as indicated in Results. Detection of bound mucin by an antimucin antibody or bound antiadhesin antibody (see below) relied on the use of goat anti-rabbit immunoglobulin G (IgG) (1:3,000 dilution) conjugated to alkaline phosphatase (BioRad) and a specific substrate color reagent (2). P. cepacia binding to mucin. Adherence of 'H-labelled P. cepacia to purified mucin preparations was determined by a
MUCIN-BINDING ADHESIN OF P. CEPACIA
1435
microtiter binding assay described previously (21, 22). Background binding to BSA-coated wells was subtracted from all mucin-binding values before presentation of data. When the 22-kDa adhesin protein or antibodies were used as inhibitors, they were preincubated for 1 h at 37°C with either mucin or 3H-labelled bacteria, respectively, before the binding assay. Electron microscopy of P. cepacia. Isolates were grown overnight at 37°C on BHI agar, lifted onto a copper-coated grid, stained for 30 s with 0.5% uranyl acetate, air dried, and examined with a Philips 400T transmission electron microscope at 80 kV. For immunogold localization of antibodies, the copper-coated grids were inverted, floated for 30 min on a drop of 3% BSA and then on antiserum (1:20 dilution), and washed four times with PBS containing 1% BSA. Bound antibody was detected by colloidal gold (particle size, 15 nm) conjugated to goat anti-rabbit IgG. Specimens were then counterstained with 0.05% uranyl acetate. For localization of mucin binding, the bacteria were placed on grids as before and floated sequentially on BSA, mucin (120 ,g of protein per ml), antibody to mucin (1:2,000 dilution), and finally the second gold-conjugated anti-rabbit IgG. Antiadhesin antibody production. A polyclonal antibody to the adhesin-rich fraction was developed in two New Zealand White rabbits by subcutaneous injections of 100 ,g of the antigen in complete adjuvant followed by injections in incomplete adjuvant after 2, 6, and 10 weeks. Antiserum was obtained 2 weeks after the last injection by way of a venous catheter with the rabbit under general anesthesia. The antiserum was tested by immunoblotting of SDS-PAGE bands of homogenates of isolate PC 5, one of the known mucinbinding isolates. Two bands (18 and 22 kDa) reacted, although the 22-kDa protein was the dominant antigenic species. Antibodies reactive with the 18-kDa protein were removed by sequential adsorption of the antisera with two poorly piliated P. cepacia isolates which had no specific mucin-binding capacity. The final antiserum preparation was designated antibody to the 22 kDa adhesin. Amino acid analyses. Adhesin preparations were hydrolyzed for 24 h in 6 M HCI at 110°C and analyzed for amino acid composition by the Picotag amino acid high-performance liquid chromatography analyzer system (7). RESULTS In previous studies (23, 24), we found that 38 of 53 clinical isolates of P. cepacia obtained from the sputa of 19 of 22 patients with CF exhibited specific saturable binding to purified mucins (respiratory mucins from patients with CF or intestinal mucins from patients without CF). For the present study, three isolates (designated PC 5, PC 7, and PC 24) were chosen for identification of the bacterial adhesin which mediates mucin binding. Three additional isolates (PC 11, PC 12, and PC 61) which did not exhibit mucin binding were used as controls. Electron microscopy. Both mucin-binding and non-mucinbinding isolates were negatively stained and examined by electron microscopy (Fig. 1). Only mucin-binding isolates exhibited substantial numbers of surface pili (Fig. 1A). The length of pili on 10 bacteria from 10 different fields was measured and was found to vary from 0.7 to 1 ,um. As judged visually, the densities of pili in the three isolates correlated positively with their mucin-binding capacity (i.e., PC 7 > PC 24 > PC 5). In contrast to the polar location of pili of P. aeruginosa (6) and of P. cepacia reported by Kuehn et al. (11, 12) and Saiman et al. (20), the pili of our P. cepacia
1436
INFECT. IMMUN.
SAJJAN AND FORSTNER
B
FIG. 1. Electron microscopy of negatively stained P. cepacia. P. cepacia grown on BHI agar was adsorbed on a copper-coated grid, stained with 0.5% uranyl acetate for 30 s, and air dried. The grids were observed under a Philips 400T electron microscope at 80 kV. (A) PC 7, a mucin-binding, piliated isolate. Magnification, x48,600. (B) PC 12, a non-mucin-binding, nonpiliated isolate. Magnification, x45,600.
isolates were distributed over the entire surface of the cells. Non-mucin-binding isolates expressed very few or no surface pili (Fig. 1B). When purified mucin was added to P. cepacia isolates and then immunolocalized by colloidal gold staining, mucin was found to be attached to surface pili as shown in Fig. 2A. The colloidal gold particles were distributed sparsely and at irregular intervals over the pili, suggesting that the adhesin for mucin may be only a minor pilin component. Nonpiliated
isolates gave no evidence of mucin binding, and a negative control (BSA instead of mucin) also showed no gold labelling
(Fig. 2B).
SDS-PAGE of P. cepacia and mucin blotting. Supernatant solutions (10 ,ug of protein) of P. cepacia crude homogenates were subjected to SDS-PAGE (PC 7 shown in Fig. 3A). Protein bands were blotted onto nitrocellulose and incubated sequentially with mucin, antibody to mucin, and a second antibody-enzyme conjugate as described in Materials and
FIG. 2. Immunolocalization of purified mucin to P. cepacia. Bacteria grown on BHI agar were adsorbed on a copper-coated grid, fixed in 1% glutaraldehyde, blocked with 3% BSA, and then incubated with mucin (120 ,ug of protein per ml) at 37°C for 30 min. After washing, the bound mucin was detected by use of a specific antibody to mucin followed by colloidal gold-conjugated goat anti-rabbit IgG. The grids were counterstained with 0.5% uranyl acetate. (A) Mucin binding to PC 7. Magnification, x 125,280. (B) BSA control with PC 7. Magnification, x 111,240.
MUCIN-BINDING ADHESIN OF P. CEPACI
VOL. 60, 1992
A
1437
kDa
B
3 l
.
4 7 0, -"'P
47I gIo
o33
o P _
24"-
1 6"
"N
no
-
422
-
b 4 22 16"-
FIG. 3. SDS-PAGE of P. cepacia homogenate and purified adhesin. P. cepacia (PC 7) homogenate supernatant (10 pg of protein) (A) and purified adhesin (2 ,ug) (B) were electrophoresed on SDS12.5% polyacrylamide gels and stained with Coomassie blue. The left lane in each panel represents molecular mass standards (in kilodaltons).
Methods. The major mucin-binding component (Fig. 4A) was a doublet at an Mr of 22,000, although minor positive bands (and one more prominent one at -69 kDa) were also observed. Since the 69-kDa band was present in nonpiliated isolates (which do not bind specifically to mucin) as well as in piliated isolates (which do), we judged that the 69-kDa component was not a specific mucin adhesin. In contrast, only the piliated isolates (PC 5, PC 7, and PC 24) showed a mucin-binding component at 22 kDa, from which we concluded that this might be a specific pilin adhesin. Adhesin antibody localization. (i) Western blotting. After
A
B
110" _ 47" " we,
.: tt
':
.:
v,
INFECTION AND IMMUNITY, Apr. 1992, p. 1434-1440 0019-9567/92/041434-07$02.00/0 Copyright X 1992, American Society for Microbiology
Identification of the Mucin-Binding Adhesin of Pseudomonas cepacia Isolated from Patients with Cystic Fibrosis S. U. SAJJAN AND J. F. FORSTNER*
Research Institute, Department of Biochemistry, The Hospital for Sick Children, University of Toronto, Toronto, Ontario M5G 1X8, Canada Received 30 October 1991/Accepted 30 January 1992
In previous experiments, we have shown that isolates of Pseudomonas cepacia from sputa of patients with cystic fibrosis (CF), particularly those with severe lung infection, exhibited specific binding to purified respiratory or intestinal mucins (U. Sajjan, M. Corey, M. Karmali, and J. Forstner, J. Clin. Invest. 89:648-656, 1992). The present report describes the identification of the adhesin as a protein located on fimbriae of mucin-binding P. cepacia. From a total of 53 isolates available (from 22 patients with CF), we used three mucin-binding and three non-mucin-binding isolates for our experiments. Sodium dodecyl sulfatepolyacrylamide gel electrophoresis of crude P. cepacia homogenates was performed, the separated proteins were blotted onto nitrocellulose and overlaid with purified mucin, and mucin-binding components were detected with an antimucin antibody and then a second-antibody-alkaline phosphatase conjugate system. Only mucin-binding isolates exhibited a positively stained band at an Mr of 22,000. The 22-kDa protein was purified, and a polyclonal antibody specific for it was developed in rabbits. By electron microscopy and immunogold labelling, both the antibody and mucin (separately) were localized to pili present over the entire surface of the bacterial cells. Non-mucin-binding isolates did not have (or had very few) pili and did not stain with either mucin or the antibody to the 22-kDa protein. The purified 22-kDa protein and its antibody were each able to inhibit piliated P. cepacia binding to mucin. The amino acid composition of the 22-kDa protein was dissimilar to those of the major pilin proteins ofEscherichia coli (type 1 pilus) and P. aeruginosa (PAK and PAO1 strains). Both the pili of P. aeruginosa PAK and PAO1 and antibodies to these pili failed to inhibit P. cepacia binding to mucin. Thus, P. cepacia adhesion to mucin is mediated by a pilin-associated 22-kDa protein which differs from epithelial-cell-binding pilin proteins ofP. aeruginosa. We postulate that the 22-kDa adhesin may play a role in the virulence of P. cepacia lung infections of patients with CF.
pilin-associated protein with an Mr of 22,000 mediates binding and that this protein does not cross-react with antipilin antibodies to Pseudomonas aeruginosa, a common colonizer of lungs of patients with CF.
Pseudomonas cepacia is an opportunistic pathogen which colonizes the lungs of some patients with cystic fibrosis (CF) (5) and some immunocompromised patients (3, 15). The prevalence of P. cepacia infection in some CF clinics has gradually increased over the past 20 years (8, 26), as has premature mortality from the "cepacia syndrome" (27). Very little information concerning the early pathogenic events of P. cepacia infection is available, although in vitro adherence to respiratory cells has been described. Saiman et al. (20) reported that outer membrane proteins of P. cepacia mediate binding, while Kuehn et al. (11, 12) suggested that polar pili are responsible. In other articles (23, 24), we report that P. cepacia isolates from 19 patients treated at the Hospital for Sick Children, Toronto, Ontario, Canada, exhibited specific binding to carbohydrates of mucins purified from the respiratory and intestinal tracts of patients with and without CF, respectively. No differences in the binding capacities of the two mucins were observed. Isolates from those patients with the most severe lung damage tended to bind most strongly to mucins. Since a mucus blanket normally covers the epithelial surfaces of the respiratory tract in vivo and is poorly cleared from lungs of patients with CF, we have speculated that P. cepacia binding to this reservoir of stagnant mucus may be a potential virulence factor. The goal of the present study was to identify the bacterial adhesin responsible for the specific attachment of P. cepacia to mucin receptors. Our results indicate that a unique *
MATERIALS AND METHODS
Bacteria, growth conditions, and radiolabelling. Fifty-three isolates of documented P. cepacia from sputum cultures of 22 patients with CF were provided to us by M. Karmali, Department of Microbiology, Hospital for Sick Children. The lyophilized primary cultures were suspended in sterile water, grown on brain heart infusion (BHI) agar overnight at 37°C, and stored in 5% trisodium citrate containing 40% glycerol at -70°C before use. In preparation for mucin binding assays, stock cultures of six isolates were subcultured on BHI agar and a single colony was inoculated into 10 ml of Trypticase soy broth containing 0.1 mCi of [3H]sodium acetate (ICN Biomedicals Inc., Costa Mesa, Calif.) (specific activity, 2.5 Ci/mmol). After overnight growth at 37°C in an orbit shaker (Canlab, Mississauga, Ontario, Canada) at 150 rpm, bacteria were harvested by centrifugation (10,000 x g for 10 min at 4°C), washed three times, and suspended in phosphate-buffered saline (Na2 HPO4-NaH2PO4 [0.01 M] in 0.15 M NaCI) (PBS; pH 7.2) containing 0.1% bovine serum albumin (BSA; globulin free; Sigma Chemical Co., St. Louis, Mo.). The specific activity of the radiolabelled P. cepacia was 2.06 x 10' dpm/CFU. Pili from P. aeruginosa (strains PAK and PAO1) were purified by the method of Paranchych et al. (17). A polyclonal antibody to PAK pili and monoclonal antibodies
Corresponding author. 1434
VOL. 60, 1992
PK3B (specific for PAK pili) and PK34C (specific for the C-terminal PAK and PA01 pilin epithelial-cell-binding domains) (4), were kindly provided by R. T. Irvin, University of Alberta, Edmonton, Alberta, Canada. Isolation and purification of mucin. Mucins from intestinal secretions of persons without CF and sputum of persons with CF were purified as described earlier (19). Since P. cepacia binds equally in vitro to both mucins (23, 24), and since intestinal mucin was available in much greater quantity, intestinal mucin was used for the binding experiments described. A previously described polyclonal antibody for human intestinal mucin (14, 18) was used in many experiments to detect mucin binding either to bacterial cells or to individual bacterial components (see below). Isolation of the mucin-binding adhesin of P. cepacia. The method of Paranchych et al. (17) for purification of pili was used with some modifications. From three mucin-binding isolates of P. cepacia (designated PC 5, PC 7, and PC 24) which had been subcultured and grown overnight on BHI agar, bacterial colonies were inoculated into 4 liters of Trypticase soy broth and grown for 36 h at 37°C on a gyratory shaker (New Brunswick Scientific Co., Inc.) at 150 rpm. Bacteria were harvested by centrifugation (5,000 x g for 20 min), washed once with PBS, and suspended in 200 ml of 5 mM Tris HCI (pH 7.2) containing 200 nM phenylmethylene sulfonic acid, 2 mM EDTA, and 0.02% sodium azide. The suspension was homogenized in a polytron at 2,000 rpm for 10 min at 4°C and centrifuged at 10,000 x g for 30 min, and ammonium sulfate (final concentration, 50%) was added to the supernatant. After 17 h at 40C, the pellet was collected by centrifugation, suspended in the Tris HCI buffer described above, and recentrifuged at 5,000 x g for 10 min. The supernatant was adjusted to 20% saturation with ammonium sulfate, left for 2 h at 4°C, and recentrifuged. The final pellet was washed, suspended in H20, and stored at -200C. This preparation was designated the adhesin-enriched fraction. It contained several proteins, as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), although one band at 22 kDa was of special interest. The adhesin was further purified by subjecting the enriched fraction to SDS-PAGE (15% polyacrylamide gels) (13) and electroeluting the mucin-binding band (Mr, 22,000) in a Bio-Rad Laboratories (Richmond, Calif.) electroeluter (model 422) with 0.05 M ammonium bicarbonate containing 0.1% SDS. From 4 liters of Trypticase soy broth, the yield of purified adhesin was approximately 2 to 4 mg. Purification of the same 22-kDa protein from one P. cepacia isolate which was known not to bind to mucin was attempted, but no product was obtained. SDS-PAGE, electroblotting, Western blot (immunoblot) analyses, and mucin overlays. Crude homogenates of P. cepacia (or fractions generated during the purification of the mucin-binding adhesin), were prepared from overnight broth cultures by homogenization in a polytron for 30 s and centrifugation. The supernatants were applied to 12.5 or 15% polyacrylamide gels and subjected to SDS-PAGE (13). Proteins were transferred to nitrocellulose membranes (28); the membranes were blocked with 3% BSA and incubated for 1 h at 370C with specific antibodies or purified mucin as indicated in Results. Detection of bound mucin by an antimucin antibody or bound antiadhesin antibody (see below) relied on the use of goat anti-rabbit immunoglobulin G (IgG) (1:3,000 dilution) conjugated to alkaline phosphatase (BioRad) and a specific substrate color reagent (2). P. cepacia binding to mucin. Adherence of 'H-labelled P. cepacia to purified mucin preparations was determined by a
MUCIN-BINDING ADHESIN OF P. CEPACIA
1435
microtiter binding assay described previously (21, 22). Background binding to BSA-coated wells was subtracted from all mucin-binding values before presentation of data. When the 22-kDa adhesin protein or antibodies were used as inhibitors, they were preincubated for 1 h at 37°C with either mucin or 3H-labelled bacteria, respectively, before the binding assay. Electron microscopy of P. cepacia. Isolates were grown overnight at 37°C on BHI agar, lifted onto a copper-coated grid, stained for 30 s with 0.5% uranyl acetate, air dried, and examined with a Philips 400T transmission electron microscope at 80 kV. For immunogold localization of antibodies, the copper-coated grids were inverted, floated for 30 min on a drop of 3% BSA and then on antiserum (1:20 dilution), and washed four times with PBS containing 1% BSA. Bound antibody was detected by colloidal gold (particle size, 15 nm) conjugated to goat anti-rabbit IgG. Specimens were then counterstained with 0.05% uranyl acetate. For localization of mucin binding, the bacteria were placed on grids as before and floated sequentially on BSA, mucin (120 ,g of protein per ml), antibody to mucin (1:2,000 dilution), and finally the second gold-conjugated anti-rabbit IgG. Antiadhesin antibody production. A polyclonal antibody to the adhesin-rich fraction was developed in two New Zealand White rabbits by subcutaneous injections of 100 ,g of the antigen in complete adjuvant followed by injections in incomplete adjuvant after 2, 6, and 10 weeks. Antiserum was obtained 2 weeks after the last injection by way of a venous catheter with the rabbit under general anesthesia. The antiserum was tested by immunoblotting of SDS-PAGE bands of homogenates of isolate PC 5, one of the known mucinbinding isolates. Two bands (18 and 22 kDa) reacted, although the 22-kDa protein was the dominant antigenic species. Antibodies reactive with the 18-kDa protein were removed by sequential adsorption of the antisera with two poorly piliated P. cepacia isolates which had no specific mucin-binding capacity. The final antiserum preparation was designated antibody to the 22 kDa adhesin. Amino acid analyses. Adhesin preparations were hydrolyzed for 24 h in 6 M HCI at 110°C and analyzed for amino acid composition by the Picotag amino acid high-performance liquid chromatography analyzer system (7). RESULTS In previous studies (23, 24), we found that 38 of 53 clinical isolates of P. cepacia obtained from the sputa of 19 of 22 patients with CF exhibited specific saturable binding to purified mucins (respiratory mucins from patients with CF or intestinal mucins from patients without CF). For the present study, three isolates (designated PC 5, PC 7, and PC 24) were chosen for identification of the bacterial adhesin which mediates mucin binding. Three additional isolates (PC 11, PC 12, and PC 61) which did not exhibit mucin binding were used as controls. Electron microscopy. Both mucin-binding and non-mucinbinding isolates were negatively stained and examined by electron microscopy (Fig. 1). Only mucin-binding isolates exhibited substantial numbers of surface pili (Fig. 1A). The length of pili on 10 bacteria from 10 different fields was measured and was found to vary from 0.7 to 1 ,um. As judged visually, the densities of pili in the three isolates correlated positively with their mucin-binding capacity (i.e., PC 7 > PC 24 > PC 5). In contrast to the polar location of pili of P. aeruginosa (6) and of P. cepacia reported by Kuehn et al. (11, 12) and Saiman et al. (20), the pili of our P. cepacia
1436
INFECT. IMMUN.
SAJJAN AND FORSTNER
B
FIG. 1. Electron microscopy of negatively stained P. cepacia. P. cepacia grown on BHI agar was adsorbed on a copper-coated grid, stained with 0.5% uranyl acetate for 30 s, and air dried. The grids were observed under a Philips 400T electron microscope at 80 kV. (A) PC 7, a mucin-binding, piliated isolate. Magnification, x48,600. (B) PC 12, a non-mucin-binding, nonpiliated isolate. Magnification, x45,600.
isolates were distributed over the entire surface of the cells. Non-mucin-binding isolates expressed very few or no surface pili (Fig. 1B). When purified mucin was added to P. cepacia isolates and then immunolocalized by colloidal gold staining, mucin was found to be attached to surface pili as shown in Fig. 2A. The colloidal gold particles were distributed sparsely and at irregular intervals over the pili, suggesting that the adhesin for mucin may be only a minor pilin component. Nonpiliated
isolates gave no evidence of mucin binding, and a negative control (BSA instead of mucin) also showed no gold labelling
(Fig. 2B).
SDS-PAGE of P. cepacia and mucin blotting. Supernatant solutions (10 ,ug of protein) of P. cepacia crude homogenates were subjected to SDS-PAGE (PC 7 shown in Fig. 3A). Protein bands were blotted onto nitrocellulose and incubated sequentially with mucin, antibody to mucin, and a second antibody-enzyme conjugate as described in Materials and
FIG. 2. Immunolocalization of purified mucin to P. cepacia. Bacteria grown on BHI agar were adsorbed on a copper-coated grid, fixed in 1% glutaraldehyde, blocked with 3% BSA, and then incubated with mucin (120 ,ug of protein per ml) at 37°C for 30 min. After washing, the bound mucin was detected by use of a specific antibody to mucin followed by colloidal gold-conjugated goat anti-rabbit IgG. The grids were counterstained with 0.5% uranyl acetate. (A) Mucin binding to PC 7. Magnification, x 125,280. (B) BSA control with PC 7. Magnification, x 111,240.
MUCIN-BINDING ADHESIN OF P. CEPACI
VOL. 60, 1992
A
1437
kDa
B
3 l
.
4 7 0, -"'P
47I gIo
o33
o P _
24"-
1 6"
"N
no
-
422
-
b 4 22 16"-
FIG. 3. SDS-PAGE of P. cepacia homogenate and purified adhesin. P. cepacia (PC 7) homogenate supernatant (10 pg of protein) (A) and purified adhesin (2 ,ug) (B) were electrophoresed on SDS12.5% polyacrylamide gels and stained with Coomassie blue. The left lane in each panel represents molecular mass standards (in kilodaltons).
Methods. The major mucin-binding component (Fig. 4A) was a doublet at an Mr of 22,000, although minor positive bands (and one more prominent one at -69 kDa) were also observed. Since the 69-kDa band was present in nonpiliated isolates (which do not bind specifically to mucin) as well as in piliated isolates (which do), we judged that the 69-kDa component was not a specific mucin adhesin. In contrast, only the piliated isolates (PC 5, PC 7, and PC 24) showed a mucin-binding component at 22 kDa, from which we concluded that this might be a specific pilin adhesin. Adhesin antibody localization. (i) Western blotting. After
A
B
110" _ 47" " we,
.: tt
':
.:
v,
