We reported previously that Muc1 mucin on the epithelial cell surface is an adhesion site for Pseudomonas aeruginosa (Lillehoj EP, Hyun SW, Kim BT, Zhang XG, Lee DI, Rowland S, and Kim KC. Am J Physiol Lung Cell Mol Physiol 280: L181–L187, 2001). The present study was designed to identify the adhesin(s) responsible for bacterial binding to Muc1 mucin using genetic and biochemical approaches. Chinese hamster ovary (CHO) cells stably transfected with a Muc1 cDNA (CHO-Muc1) or empty plasmid (CHO-X) were compared for adhesion of P. aeruginosa strain PAK. Our results showed that 1) wild-type PAK and isogenic mutant strains lacking pili (PAK/NP) or flagella cap protein (PAK/fliD) demonstrated significantly increased binding to CHO-Muc1 cells, whereas flagellin-deficient (PAK/fliC) bacteria were no more adherent to CHO-Muc1 than CHO-X cells, and 2) P. aeruginosa adhesion was blocked by pretreatment of bacteria with antibody to flagellin or pretreatment of CHO-Muc1 cells with purified flagellin. We conclude that flagellin is an adhesin of P. aeruginosa responsible for its binding to Muc1 mucin on the epithelial cell surface.
- cystic fibrosis
pseudomonas aeruginosa is a gram-negative opportunistic pathogen associated with pneumonia in immunocompromised patients and chronic lung disease, such as cystic fibrosis. The molecular interactions ofP. aeruginosa with respiratory epithelial cells is complex and likely involves multiple types of ligand-receptor contacts (reviewed in Refs. 1, 17, 36,38, 39). Several P. aeruginosaadhesins have been described, including pili (9, 16, 22,42), flagella (12, 39), and lipopolysaccharide (LPS) (16). Some of these studies have emphasized the importance of epithelial cell glycosphingolipids, such as asialoGM1, for binding to bacterial pili (41). In addition, specific binding of flagellar proteins to secreted respiratory mucins appears to be physiologically relevant to P. aeruginosa colonization of the lung (3, 8).
P. aeruginosa adherence to secreted mucins prompted us to investigate whether cell-associated mucins also serve as receptors for bacterial adhesion. Five genes encoding membrane-tethered mucins have been identified (15, 20, 30, 46-48), among which MUC1 mucin is abundantly expressed by respiratory epithelial cells, such that its strategic location in the upper airways permits interaction with inspired microorganisms. Indeed, we demonstrated previously that P. aeruginosa specifically binds to hamster Muc1 mucin, judging from in vitro saturable binding kinetics and competitive inhibition by the homologous ligand (24). In the study reported here, we examined the molecular nature of the P. aeruginosa adhesin and identified flagellin, the major flagellar protein, as responsible for binding to Muc1 mucin.
All materials were purchased from Sigma (St. Louis, MO) unless stated otherwise. Rabbit antiserum to PAK flagellin was kindly provided by Dr. Dan Wozniak (Wake Forest University, Winston-Salem, NC). Rabbit antiserum to PAK pilin was kindly provided by Dr. Randy Irvin (University of Alberta, Edmonton, AB, Canada). Mouse monoclonal antibody (M4E8) to PAK LPS (serotype 6) was kindly supplied by Dr. Joseph Lam (University of Guelph, Guelph, ON, Canada).
PAK is a wild-type, nonmucoid, piliated, and motile strain. Its isogenic mutant PAK/NP is a derivative of PAK in which thepilA gene encoding the pilin protein is replaced by homologous recombination with a mutant gene interrupted by a tetracycline resistance cassette (41). Strains PAK/fliC (12) and PAK/fliD(3) are nonmotile derivatives in which the fliCgene encoding flagellin or the fliD gene encoding the flagellar cap protein is replaced with homologous genes interrupted by gentamicin resistance cassettes. PAK, PAK/NP, and PAK/fliCwere kindly provided by Dr. Alice Prince (Columbia University, New York, NY). PAK/fliD was a generous gift from Dr. Reuben Ramphal (University of Florida, Gainesville, FL).
CHO-Muc1 and CHO-X cells.
The hamster Muc1 expression plasmid pcDNA/Muc1 was constructed by insertion of a full-length hamster Muc1 cDNA (34) into the pcDNA3 vector (Invitrogen, San Diego, CA) (24). The pcDNA/Muc1 plasmid and pcDNA3 vector alone were transfected separately into Chinese hamster ovary (CHO) cells (CCL61, American Type Culture Collection, Manassas, VA) using the cationic liposome Lipofectin (Life Technologies, Rockville, MD), and stable clones were identified and characterized. A clone expressing hamster Muc1 mucin was referred to as CHO-Muc1 and a clone derived from the vector alone as CHO-X. The latter was used as a negative control in the P. aeruginosa binding assay. Both cell lines were maintained in Ham's F-12-Dulbecco's modified Eagle's medium containing 200 μg/ml G418, 5% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Life Technologies).
P. aeruginosa binding assay.
Adhesion of P. aeruginosa to transfected CHO cells was performed according to the method described by Rostand and Esko (40) and modified by Lillehoj et al. (24). Briefly, cells were cultured for 20–24 h in 24-well microplates in an inoculum of 2 × 105 cells per well. The monolayer was washed twice with phosphate-buffered saline (PBS), pH 7.2, fixed for 10 min with 2.5% (vol/vol) glutaraldehyde in PBS at room temperature, and washed three times with PBS. Bacteria were metabolically radiolabeled in sulfate-free M9 medium containing 10 μCi/ml Na2 35SO4 (1,175 Ci/mmol, 100 mCi/ml, carrier-free; American Radiolabeled Chemicals, St. Louis, MO) for 16 h at 37°C, washed twice, resuspended in PBS containing 2 mg/ml glucose, and quantified photometrically (0.50 optical density at 600 nm = 5 × 108 cells). Fixed CHO cells were incubated with 2 × 107 P. aeruginosa/well in 0.5 ml for 40 min at 37°C and washed three times with PBS, adhered bacteria were lysed with 1.0 ml of 2% sodium dodecyl sulfate (SDS), and radioactivity was measured by liquid scintillation counting. In some experiments, bacteria were pretreated for 30 min at 37°C with PAK flagellin antiserum or cells were pretreated with purified flagellin as potential binding inhibitors.
Purification of flagellin.
P. aeruginosa PAK flagellin was purified according to the procedure of Allison et al. (2). Briefly, overnight liquid cultures in mineral salts medium [7.0 g K2HPO4, 3.0 g KH2PO4, 1.0 g (NH4)2SO4, 50 mg MgSO4 · 7H2O, 2.5 mg FeCl3 · H2O, 7.4 mgl-methionine, and 4.0 g sodium succinate per liter, pH 7.0] were centrifuged at 5,000 g for 15 min at 4°C, and the bacterial pellet was resuspended in 10 mM potassium phosphate, pH 7.0, and blended at a low speed in a commercial blender (Black and Decker, Towson, MD) for 30 s at room temperature to shear the flagella. The suspension was centrifuged at 16,000 g for 15 min at 4°C to remove bacteria, and the resulting supernatant was centrifuged at 100,000 g in a centrifuge (model L8-70, Beckman, Duarte, CA) with a 42.1 rotor for 3 h at 4°C. In some experiments, flagellin was further purified by sequential acid pH dissociation of flagella, centrifugation at 100,000 g for 1 h at 4°C, and neutral pH reassociation of the supernatant as described elsewhere (7, 43). Protein concentrations were determined by the Coomassie blue technique of Bradford (6) with crystalline bovine serum albumin as standard (Bio-Rad, Richmond, CA). A mock flagellin preparation was prepared in an identical manner using an extract of PAK/fliC bacteria.
SDS-PAGE and immunoblot analysis.
SDS-PAGE was performed according to the method of Laemmli (19). Separated proteins were stained with 0.25% Coomassie blue or transblotted to polyvinylidene difluoride membrane as described previously (23), and the membrane was blocked with 5% nonfat dry milk in 10 mM Tris · HCl, pH 7.4, containing 1.5% Tween 20 and reacted with rabbit antisera to PAK flagellin or pilin, each diluted 1:3,000 in blocking buffer, or monoclonal antibody to PAK LPS diluted 1:20 in blocking buffer. After the membrane was washed, bound antibody was detected with horseradish peroxidase-conjugated goat anti-rabbit IgG or goat anti-mouse IgG (KPL, Gaithersburg, MD) diluted 1:10,000 in blocking buffer and enhanced chemiluminescence reagent (Amersham Pharmacia Biotech, Arlington Heights, IL).
The number of P. aeruginosa adhered to CHO cells was calculated in the following order: 1) determination of the number of bacteria bound to cells in each well by converting disintegrations per minute to actual number of bacteria, 2) determination of the number of epithelial cells at the end of the assay using a separate parallel group not exposed to bacteria, and3) assessment of the significance of the difference between groups using Student's t-test for unpaired samples or analysis of variance. Differences were considered significant atP < 0.05.
P. aeruginosa lacking flagellin exhibit diminished adherence to CHO-Muc1 cells.
Prior studies have documented P. aeruginosa pili, flagella, and cap protein in adherence to epithelial cell glycolipids and secreted mucins (3, 12, 22, 39, 41, 42). To examine the possible involvement of these bacterial components in binding to Muc1 mucin, we compared adherence of wild-type PAK with isogenic mutants devoid of pilin (PAK/NP), flagellin (PAK/fliC), or cap protein (PAK/fliD) to CHO-X and CHO-Muc1 cells. As shown in Fig. 1, binding of wild-type PAK, PAK/NP, and PAK/fliD to CHO-Muc1 cells was nearly twice that to CHO-X cells. In contrast, binding of flagellin-defective PAK/fliC to the two cells was equivalent. Because PAK/fliC and PAK/fliD are nonmotile, the lack of adherence of PAK/fliC to CHO-Muc1 cells due to its lack of motility was ruled out. Thus P. aeruginosa flagellin appeared to constitute the molecular adhesin binding to Muc1 mucin.
Flagellin antiserum inhibits P. aeruginosa binding to CHO-Muc1 cells.
To confirm the role of flagellin as an adhesin for Muc1 mucin, wild-type PAK were preincubated with flagellin antiserum for 30 min before the binding assay. As shown in Fig.2, flagellin antiserum blocked P. aeruginosa binding to CHO-Muc1 cells in a dose-dependent manner. Compared with PBS treatment, preincubation with flagellin antiserum decreased PAK binding by 61.2% at 6.0 μl/ml (P < 0.001) and completely abolished binding at 20 μl/ml (P < 0.0001). Bacterial adherence after pretreatment with flagellin antiserum also was significantly reduced compared with adhesion after preincubation with equivalent amounts of normal rabbit serum (P < 0.004 at 6.0 μl/ml and P< 0.0002 at 20 μl/ml; data not shown).
Purified flagellin inhibits P. aeruginosa binding to CHO-Muc1 cells.
Flagellin was purified from wild-type PAK by differential centrifugation as described previously (2) and analyzed by immunoblotting and Coomassie blue staining after separation by SDS-PAGE. As shown in Fig. 3 A, immunoblot analysis with rabbit antiserum against PAK flagellin revealed a prominent 53-kDa immunoreactive band (lane 2) that comigrated with a Coomassie blue-stained band detected in the same flagellin preparation (lane 6). Importantly, no protein bands were detected by immunoblotting with antiserum against PAK pilin (lane 4), indicating the absence of pilin (17.8 kDa), a protein often copurifying with flagellin by the centrifugation technique (14). However, immunoblot analysis using monoclonal antibody against PAK LPS revealed the presence of LPS (Fig.3 B, lane 2) with a banding pattern essentially identical to that previously reported (10). The majority of LPS was removed by sequential acid pH dissociation, ultracentrifugation, and neutral pH reassociation as described by Brett et al. (7) (Fig. 3 B, lane 3). Although the recovery of flagellin from this additional purification was low (∼25%), essentially pure flagellin was obtained. As shown in Fig.4, adhesion of P. aeruginosato CHO-Muc1 cells was significantly decreased after pretreatment of cells for 30 min with purified flagellin compared with PBS-treated cells (P < 0.05 at 0.02 μg/ml and P< 0.01 at 0.2 μg/ml). To control for nonspecific inhibition, a mock flagellin preparation was generated from flagellin-deficient PAK/fliC bacteria in the same manner used to prepare flagellin from wild-type PAK. Pretreatment of CHO-Muc1 cells with 0.2 μg/ml mock flagellin did not affect P. aeruginosa binding compared with PBS-treated cells (Fig. 4 E, P> 0.05). On the basis of the foregoing results, we conclude thatP. aeruginosa flagellin is an adhesin for Muc1 mucin.
We reported previously that expression of hamster Muc1 mucin on the surface of transfected CHO cells significantly augmented the binding of P. aeruginosa strains PAK (nonmucoid) and CF-3 (mucoid) (24). Because the increase in bacterial adhesion was completely abolished by removal of the extracellular domain of Muc1 mucin, either by treatment of CHO-Muc1 cells with human neutrophil elastase or deletion mutagenesis, these results indicated that Muc1 mucin is an adhesion site for P. aeruginosa and prompted us to conduct the experiments described here to identify the cognate bacterial adhesin. Our results with reduced bacterial adherence by flagellin-deficient bacteria and binding inhibition by flagellin antiserum and purified flagellin collectively lead us to conclude thatP. aeruginosa flagellin is the adhesin interacting with Muc1 mucin at the molecular level. We chose to employ the PAK strain for the present studies, since it is nonmucoid and, thus, more typical ofP. aeruginosa strains initially isolated from cystic fibrosis lungs (see below). Although the amount of P. aeruginosa that we observed binding to CHO-Muc1 cells (0.2–5.1 bacteria/cell or 0.4–10 × 105bacteria/well) may appear quantitatively low, it compares favorably with that previously documented in the literature on several occasions. For example, using 24-well plates with confluent monolayers, Feldman et al. (12) reported (2.0–30) × 105colony-forming units per well of PAK wild-type, PAK/NP, or PAK/fliC bound to 1HAEo− cells. Although the structural features of Muc1 mucin accounting for its interaction with flagellin remain to be elucidated, the GalNAcβ(1–4)Gal carbohydrate sequence of asialoGM1, previously suggested to mediate binding to P. aeruginosapilin (44), is probably not involved, since Muc1 mucin glycosyl moieties lack this disaccharide structure (5). Moreover, because the PAK and PAO1 strains exhibited adherence to CHO-Muc1 cells (data not shown), both a-type (PAK) and b-type (PAO1) flagellin are likely capable of binding to Muc1 mucin.
Numerous prior studies demonstrating the role of P. aeruginosa flagella in the pathogenesis of various lung diseases, both experimental (18, 21, 31) and clinical (25), support our observations. In a neonatal mouse model of pneumonia, Tang et al. (45) showed that flagella ofP. aeruginosa strain PAO1 are among several bacterial virulence factors contributing to bacteremia, disease, and death. Although the P. aeruginosa flagellum is a virulence factor associated with disease pathogenesis, it also serves as a target for nonopsonic bacterial clearance by host phagocytic cells (27). In a later study, Feldman et al. (12) used PAK wild-type and isogenic mutants to document reduced incidence of pneumonia and absence of mortality associated with murine lung infection by PAK/fliC compared with PAK and PAK/NP infections. In a clinical setting, in vivo selection against flagella production may be one mechanism by which P. aeruginosapersists in the cystic fibrosis lung. Sequential bacterial isolations from infected patients revealed that initial isolates obtained shortly after colonization were nonmucoid, flagellated, and motile, whereas mucoid bacterial variants lacking motility and resistant to phagocytosis emerged later during the course of the disease (26). The mechanism by which flagella synthesis is downregulated is not known. However, it seems reasonable to suggest that the initial interaction between flagella and the epithelial cell surface may be an early signal that activates host innate defense responses, ultimately leading to selection of nonmotile bacterial strains. In this regard, Muc1 mucin on the epithelial cell surface may play an important role in this process, considering the fact that its cytoplasmic domain contains multiple phosphorylation sites implicated in signal transduction pathways of inflammation (4, 28, 29, 34,50).
One of the hallmark features of ligand-receptor interactions is the ability of unlabeled ligand to competitively inhibit radiolabeled ligand binding to its cognate receptor in a dose-dependant manner (49). We therefore investigated the ability of P. aeruginosa flagellin to block adhesion of 35S-labeled bacteria to CHO-Muc1 cells. As shown in Fig. 4, purified flagellin blocked P. aeruginosa binding to CHO-Muc1 cells. Whereas binding decreased in a dose-dependent fashion up to 0.2 μg/ml, to our surprise, flagellin concentrations >2.0 μg/ml enhanced P. aeruginosa binding to control CHO-X cells, likely through as yet unidentified interactions with nonmucin components present on the cell surface. Interestingly, a similar enhancement of attachment ofSalmonella enteritidis to ovarian granulosa cells by purified flagellin was reported by Popielzrczyk et al. (37). Although flagellin was free of detectable bacterial protein contamination, minor levels of LPS were apparent (Fig.3 B). However, this degree of LPS contamination did not appear to be responsible for inhibition of bacterial adherence, since an extract of PAK/fliC bacteria (devoid of flagella) that was prepared identically to wild-type flagellin and contained an equal amount of LPS as estimated by immunoblotting with LPS antibody was unable to block PAK binding to CHO-Muc1 cells (Fig. 4).
Flagella are the most complex organelles known in bacteria, assembled from three separate substructures (filament, hook, and basal body), each containing a complex array of different proteins (13). Although our initial results (Fig. 1) suggested that flagellin was the adhesin responsible for binding to Muc1 mucin, we tested several alternative explanations to account for the lack of adherence displayed by flagella-deficient P. aeruginosa. To rule out the possibility that a nonflagellin component of the flagellum was responsible for binding to Muc1 mucin, we used rabbit antiserum against P. aeruginosa PAK flagellin and inhibited the interaction between bacteria and CHO-Muc1 cells. Second, we considered that loss of bacterial motility by flagellin-defective or antiserum-treated bacteria was responsible for reduced binding interaction. Because PAK/fliC is nonmotile and treatment of motile bacteria with flagellin antiserum rendered them nonmotile (21, 32, 43), we used gramicidin D to rule out the possibility that lack of adhesion to CHO-Muc1 cells was simply due to loss of the bacteria's ability to swim. Gramicidin D is an effective inhibitor of Pseudomonas motility, causing 100% inhibition within 2 min at 1.5 × 10−4 M, a concentration not affecting O2 uptake (11). At this concentration, gramicidin D had no effect on P. aeruginosabinding to CHO-Muc1 cells (data not shown). Collectively, these results extend our previous observations that Muc1 mucin on the cell surface is an adhesion site for P. aeruginosa (24) by showing that the major bacterial component responsible for adhesion to Muc1 mucin is flagellin. Flagellin and Muc1 mucin can thus be added to the list of >100 adhesin-receptor pairs known to exist between bacteria and eukaryotic cells (33).
This work was supported by National Heart, Lung, and Blood Institute Grant RO1 HL-47125 and a grant from the Cystic Fibrosis Foundation.
Address for reprint requests and other correspondence: K. C. Kim, Dept. of Pharmaceutical Sciences, School of Pharmacy, University of Maryland, 20 N. Pine St., Baltimore, MD 21201.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society