Binding of nonmucoid Pseudomonas aeruginosa to normal human intestinal mucin and respiratory mucin from patients with cystic fibrosis.

Lung infections due to Pseudomonas aeruginosa and Pseudomonas cepacia are common in patients with cystic fibrosis. Initial colonization is due to nonmucoid P. aeruginosa, while later mucoid variants emerge and are associated with chronic infection. P. cepacia colonization tends to be more prevalent in older patients. The present study was conducted to discover whether highly purified mucins (from cystic fibrosis sputum and control intestinal secretions) exhibited specific binding of nonmucoid P. aeruginosa. In vitro solid phase microtiter binding assays (with or without a blocking agent) as well as solution phase assays were conducted. Bacteria bound to both mucins via bacterial pili, but no differences in binding capacity were noted between the mucins. Unlike P. cepacia (described in the accompanying manuscript) there was also no preferential binding of P. aeruginosa to mucins versus bovine serum albumin, casein, gelatin, or a host of structurally unrelated proteins and glycoproteins. Carbohydrate hapten inhibition studies did not suggest the existence of specific mucin carbohydrate receptors for P. aeruginosa. In solid phase assays a low concentration (0.05 M) of tetramethylurea abolished P. aeruginosa bacterial binding to both mucins as well as to BSA, whereas in solution phase assays mucin binding to bacteria was not completely disrupted by tetramethylurea. Specific monoclonal antipilus antibodies did not inhibit binding to a greater extent than did Fab fragments of normal mouse IgG. Binding of strains PAO1 and PAK (and isolated PAK pili) to buccal epithelial cells was not influenced by the presence of mucin in binding assay mixtures. Our findings do not support the widely held notion that specific mucin receptors are responsible for the attachment of P. aeruginosa pili, nor do they support the idea that there is a competitive interference by mucins of bacterial binding to respiratory cells. In patients with cystic fibrosis, it would seem unlikely therefore that initial colonization of the lungs by P. aeruginosa is due to a 'selective tropism' of these bacteria for respiratory mucin.

Crystal structure of Pseudomonas aeruginosa PAK pilin suggests a main-chain-dominated mode of receptor binding.

Fibers of pilin monomers (pili) form the dominant adhesin of Pseudomonas aeruginosa, and they play an important role in infections by this opportunistic bacterial pathogen. Blocking adhesion is therefore a target for vaccine development. The receptor-binding site is located in a C-terminal disulphide-bonded loop of each pilin monomer, but functional binding sites are displayed only at the tip of the pilus. A factor complicating vaccination is that different bacterial strains produce distinct, and sometimes highly divergent, pilin variants. It is surprising that all strains still appear to bind a common receptor, asialo-GM1. Here, we present the 1.63 A crystal structure of pilin from P. aeruginosa strain PAK. The structure shows that the proposed receptor-binding site is formed by two beta-turns that create a surface dominated by main-chain atoms. Receptor specificity could therefore be maintained, whilst allowing side-chain variation, if the main-chain conformation is conserved. The location of the binding site relative to the proposed packing of the pilus fiber raises new issues and suggests that the current fiber model may have to be reconsidered. Finally, the structure of the C-terminal disulphide-bonded loop will provide the template for the structure-based design of a consensus sequence vaccine.

Interaction of the receptor binding domains of Pseudomonas aeruginosa pili strains PAK, PAO, KB7 and P1 to a cross-reactive antibody and receptor analog: implications for synthetic vaccine design.

The four synthetic peptide antigens, PAK 128-144, PAO 128-144, KB7 128-144 and P1 126-148, correspond in amino acid sequence to the C-terminal receptor binding regions of four strains (PAK, PAO, KB7, P1) of Pseudomonas aeruginosa pilin. The NMR solution structures of the trans forms of the peptides show conserved beta-turns which have been implicated in antibody and receptor recognition. The interactions between these peptides and a cross-reactive monoclonal antibody, PAK-13, have been studied using two-dimensional (1)H NMR spectroscopy in order to map the antigenic determinants recognized by the antibody. Residues for which spectral changes were observed upon antibody binding differed from peptide to peptide but were mostly confined to one or both of the turn regions and to the hydrophobic pockets. Conformational changes in the beta-turns and hydrophobic pockets of these peptides upon antibody binding were also monitored by examination of the pattern of nuclear Overhauser effects (NOEs) versus transferred nuclear Overhauser effects (TRNOEs) for the free versus the bound peptides. Although TRNOEs developed strongly between side chain resonances in the hydrophobic pockets of the peptides, no additional backbone TRNOEs were observed in the presence of antibody, suggesting no major conformational changes in the secondary structures of the peptides upon binding. This implies a flexible antibody combining site, a feature which is discussed with respect to cross-reactivity, strain specificity, and the design of a synthetic peptide vaccine effective against a broad spectrum of P. aeruginosa strains. The binding of the PAK peptide to a disaccharide receptor analog, (beta GalNAc(1-4)beta Gal), was also studied using (1)H NMR in order to map the "adhesintope" recognized by the receptor. Spectral changes observed in the peptide spectrum with the binding of receptor were similar to those seen for the binding of antibody, suggesting that the epitope recognized by the antibody is structurally coincident with the adhesintope recognized by the receptor. The relevancy of this result is discussed with respect to immunogenicity versus pathogenicity, and the proper design of a vaccine which could prevent the mutational escape of the pathogen away from the host's defence systems.

Antigen-antibody interactions: elucidation of the epitope and strain-specificity of a monoclonal antibody directed against the pilin protein adherence binding domain of Pseudomonas aeruginosa strain K.

The C-terminal region of Pseudomonas aeruginosa strain K (PAK) pilin comprises both an epitope for the strain-specific monoclonal antibody PK99H, which blocks pilus-mediated adherence, and the adherence binding domain for buccal and tracheal epithelial cells. The PK99H epitope was located in sequence 134-140 (Asp-Glu-Gln-Phe-Ile-Pro-Lys) by using a single alanine replacement analysis on the 17-residue synthetic peptide corresponding to the PAK C-terminal sequence 128-144. Indeed, a 7-residue peptide corresponding to this sequence was shown to have a similar binding affinity to that of the native conformationally constrained (disulfide bridged) 17-residue peptide. This epitope was found to contain two critical residues (Phe137 and Lys140) and one nonessential residue (Gln136). Interestingly, the peptide, Phe-Ile-Pro-Lys, which constitutes the four most important side chains for antibody binding did not bind to PK99H. It was of interest to investigate the structural basis of the strain-specificity of PK99H utilizing naturally occurring pilin sequences. Therefore, all different residues found in the sequence corresponding to the PK99H epitope of the four other strains (PAO, CD4, K122-4, and KB7) were substituted one at a time in the PAK sequence and the changes in binding affinity of these analogs to the antibody PK99H were determined by competitive ELISA. The strain-specificity of PK99H for strains PAO, K122-4, and KB7 can be explained by the accumulated sequence changes in these strains, and at least two amino acid changes were required to explain the strain-specificity of PK99H. Similarly, cross-reactivity of PK99H with CD4 can be explained by the fact that there was only one side chain responsible for decreasing binding affinity compared to the PAK sequence.

Anti-peptide monoclonal antibody imaging of a common binding domain involved in muscle regulation.

Multiple-component regulatory protein systems function through a generalized mechanism where a single regulatory protein or ligand binds to a variety of receptors to modulate specific functions in a physiologically sensitive context. Muscle contraction is regulated by the interaction of actin with troponin I (TnI) or myosin in a Ca(2+)-sensitive manner. Actin utilizes a single binding domain (residues 1-28) to bind to residues 104-115 of TnI (Van Eyk JE, Sönnichsen FD, Sykes BD, Hodges RS, 1991, In: Rüegg JC, ed, Peptides as probes in muscle research, Heidelberg, Germany: Springer-Verlag, pp 15-31) and to myosin subfragment 1 (S1, an enzymatic fragment of myosin containing both the actin and ATP binding sites) (Van Eyk JE, Hodges RS, 1991, Biochemistry 30:11676-11682) in a Ca(2+)-sensitive manner. We have utilized an anti-TnI peptide (104-115) monoclonal antibody, Mab B4, that binds specifically to TnI, to image the common binding domain of actin and thus mimic the activity of actin including activation of the S1 ATPase activity and TnI-mediated regulation of the S1 ATPase. Mab B4 has also been utilized to identify a receptor binding domain on myosin (residues 633-644) that is recognized by actin. Interestingly, Mab B4 binds to the native protein receptors TnI and S1 with relative affinities of 100- and 25,000-fold higher than the binding affinity to the 12-residue peptide immunogen. Thus, anti-peptide monoclonal antibodies prepared against a receptor binding domain can mimic the ligand binding domain and be utilized as a powerful tool for the detailed analysis of complex multiple-component regulatory systems.

Use of synthetic peptides in characterization of microbial adhesins.

The characterization of microbial adhesins is greatly facilitated by the use of synthetic peptides. Synthetic peptides can be used to identify specific antigenic epitopes, to delineate receptor-binding domain of adhesins, and to facilitate the characterization of the adhesin, and they allow for a direct examination of structure-binding relationships.

Corneal epithelial glycoproteins exhibit Pseudomonas aeruginosa pilus binding activity.

Adherence of Pseudomonas aeruginosa to the cornea is a requisite step in the pathogenesis of bacteria-induced corneal disease. P. aeruginosa is capable of attaching to host epithelial cells by its pili, but there is little information regarding the epithelial receptors of this adhesin in the cornea. Using nitro-cellulose blotting of polyacrylamide gels of solubilized adult mouse corneal epithelium, four major proteins (molecular weights: 38, 42, 57, and 66 kD) and several minor proteins were identified that bound purified pili from strain PAK and its hyperpiliated mutant PAK/PR1. These proteins were identified by immunoblotting either with pilus-specific monoclonal antibodies, XLR-3 and PK 3B, or using peptide PAK 128-144 (OX). The glycosylated nature of the proteins was determined using similar gel electrophoresis of corneal epithelial proteins, blotting onto nitrocellulose, and staining the blots with lectins conjugated to either horseradish peroxidase or alkaline phosphatase. All four major pilus-binding proteins were stained with concanavalin A lectin (mannose and glucose) and either wheat germ agglutinin lectin (WGA, specific for sialic acid and N-acetylglucosamine) or succinylated WGA lectin (only N-acetylglucosamine). Staining for peanut agglutinin lectin (galactose beta(1-3) N-acetylgalactosamine) was seen for the 42-, 57-, and 66-kD proteins. The importance of the carbohydrate portions of these corneal proteins in pili binding was confirmed by preincubation of corneal epithelial blots with periodate or pili with sialic acid, both of which abolished the pili binding. These studies indicate that corneal epithelial pilus-binding proteins are glycoproteins in nature and that sialic acid may be a constituent of these pilus-specific receptors in the adult mouse corneal epithelium.

Immunochemical examination of the Pseudomonas aeruginosa glycocalyx: a monoclonal antibody which recognizes L-guluronic acid residues of alginic acid.

Mice immunized with Formalin-fixed mucoid Pseudomonas aeruginosa cells developed an immune response directed, in part, towards the P. aeruginosa glycocalyx. The polyclonal mouse sera produced good immunofluorescent staining of the P. aeruginosa glycocalyx and cell surface. A library of 250 hybridoma cell lines which produced monoclonal antibodies directed against P. aeruginosa was established. Twelve clones (4.8%) produced antibody which reacted with alginate in an enzyme-linked immunosorbent assay (ELISA). Clone Ps 53 was chosen for further study, cloned, and an ascites tumor established. Clone Ps 53 was chosen for further study because the antibody produced demonstrated a specificity similar to that of a recently isolated heparin--rat-lung lectin which recognizes alginates of the Homma nontypable P. aeruginosa strains. The Ps 53 clone produced an immunoglobulin M which reacted with P. aeruginosa alginate and produced good immunofluorescent staining of the P. aeruginosa glycocalyx. The Ps 53 monoclonal antibody has an apparent specificity for L-guluronic residues in ELISA. Competitive binding studies with various alginates and monosaccharides suggest that the C6 carboxyl group of uronic acids are recognized by the antibody and that the antigen-binding site is fairly large and may recognize a particular sequence or epitope of alginic acid which is rich in L-guluronic acid. The Ps 53 monoclonal antibody did not react uniformily with all P. aeruginosa alginates but did react with all of the alginates of the Homma nontypable strains tested, suggesting that acetylation or various modifications found in P. aeruginosa alginates may interfere with antibody binding and define specific epitopes. The Ps 53 antibody also reacted with purified outer membrane, indicating that some alginate or L-guluronic acid is intimately associated with outer membrane.

Adhesion of Pseudomonas aeruginosa to human buccal epithelial cells: evidence for two classes of receptors.

The adhesion of Pseudomonas aeruginosa strain 492c to trypsinized and untrypsinized buccal epithelial cells (BECs) was studied. Kinetic analysis of the adhesion data, employing a Langmuir absorption isotherm, indicated the presence of two classes of binding sites on untrypsinized BECs: a high affinity - low copy number site (apparent association constant (Ka) approximately equal to 1.57 X 10(-8) mL/cell with ca. 29 binding sites/cell) and a low affinity - high copy number class of binding sites (Ka approximately equal to 4.78 X 10(-10) mL/cell with ca. 264 binding sites/cell). The low affinity - high copy number class of sites was found to be trypsin sensitive. A single class of binding sites was found on trypsinized BECs exhibiting a high affinity - low copy number (Ka approximately equal to 3.70 X 10(-7) mL/cell with ca. 31 binding sites/cell). Positive cooperativity in binding of P. aeruginosa strain 492c to the low affinity - high copy number class site on untrypsinized BECs was demonstrated by analysis of Hill plots of the adhesion data. Sugar inhibition data using a preincubation methodology showed an inhibition of adhesion to trypsinized BECs in the presence of N-acetylneuraminic acid and D-arabinose, while these same two sugars enhanced adhesion to untrypsinized BECs. D-Galactose and N-acetylglucosamine enhanced adhesion to both types of BECs though the latter did to different extents. D-Fucose only inhibited adhesion to untrypsinized BECs.(ABSTRACT TRUNCATED AT 250 WORDS)

The role of bacterial surface structures in pathogenesis.

Modern research has revealed that the true surfaces of animal cells consist of polysaccharide chains that are linked to proteins hydrophobically anchored in the membrane and protrude to form a dense glycocalyx. It has become increasingly clear that most pathogenic bacteria must position themselves at the surface of their "target" cell in order to exert their toxic or otherwise deleterious effects. The true surface of most pathogenic bacteria has also been recently shown to consist of a protruding mass of polysaccharide chains--the bacterial glycocalyx--that is composed of teichoic acids in many gram-positive species and of acid polysaccharides in many gram-negative organisms. Through this bacterial glycocalyx certain cell surface proteins and organized protein structures (e.g., pili) are known to project, so that the bacterial surface is a mosaic of polysaccharides and proteins; both of these types of molecules have been implicated in instances of specific pathogenic adhesion. Besides their role in specific adhesion to target cells, these surface components interpose a highly charged, and often very extensive, barrier that can prevent the penetration of antibodies and antibiotics to their target sites in the bacterial cell. They may also frustrate mucociliary clearance, phagocytosis, and other clearance mechanisms of the host. We will discuss the chemical and physical nature of these bacterial surface components that mediate pathogenic adhesion and counteract host defense mechanisms sufficiently to allow infections to become established.

The binding of Pseudomonas aeruginosa outer membrane ghosts to human buccal epithelial cells.

The binding of outer membrane (OM) ghosts derived from Pseudomonas aeruginosa strain 492c to human buccal epithelial cells (BECs) was examined. Electron microscopic examination of the binding of OM ghosts to BECs revealed direct OM ghost-BEC interaction. Equilibrium analysis of the binding of OM ghosts to trypsinized BECs employing the Langmuir adsorption isotherm indicated the number of binding sites (N) to be 1.3 X 10(-4) micrograms protein per BEC with an apparent association constant (Ka) of 3.4 X 10(-2) mL/microgram protein. The Langmuir analysis of binding of OM ghosts to untrypsinized BECs was complex, suggesting two possible classes of receptors, a high affinity-low copy number class (Ka, 7.8 X 10(-2)mL/microgram protein; N, 8.6 X 10(-5) microgram protein per BEC) and a low affinity-high copy number class (Ka, 3.7 X 10(-3)mL/microgram protein; N, 9.2 X 10(-4)microgram protein per BEC). Sugar inhibition studies incorporating D-galactose enhanced binding to each BEC type. N-Acetylneuraminic acid and N-acetylglucosamine both enhanced binding of OM ghosts to untrypsinized BECs, while inhibiting binding to trypsinized BECs. D-Arabinose inhibited binding to both BEC types. Binding of OM ghosts to both BEC types was greatly inhibited by D-fucose, while L-fucose only greatly inhibited binding to untrypsinized BECs. These sugar inhibition data demonstrated a difference in the binding of OM ghosts to trypsinized and untrypsinized BECs and possibly reveal the nature of the receptor(s), free of possible bacterial metabolic effects.(ABSTRACT TRUNCATED AT 250 WORDS)

Structural and biochemical examination of ghosts derived from a deep rough (heptose-deficient lipopolysaccharide) strain and a smooth strain of Escherichia coli.

Outer membrane derived 'ghosts' can be readily generated from both smooth and deep rough (heptose-deficient LPS) strains of Escherichia coli 08. MORPHOlogical and biochemical studies confirmed that 'ghosts' of both strains are composed of protein (four major proteins), LPS, and phospholipid (cardiolipin and phosphatidylethanolamine) in the form of a single membrane of roughly the same shape as intact normal cells. The ghost membrane cleaves only slightly in freeze-etch preparations of ghosts derived from the smooth strain as compared to the extensive cleavage plane of ghosts derived from the rough strain. The asymmetrical distribution of ghost proteins was visualized, by critical point drying and shadowing with platinum, as a relatively smooth outer surface with some discernible particles (10-15 nm) and an extremely particulate inner surface (10-15-mm particles. Ghosts derived from the smooth strain retained their structure following chloroform-methanol extraction, while ghosts derived from the rough strain fragmented with chloroform-methanol extraction. Evidence is presented that LPS-protein interactions as well as protein-protein interactions are significant in maintaining the ghost structure.

Cell surface-localized alkaline phosphatase of Escherichia coli as visualized by reaction product deposition and ferritin-labeled antibodies.

When cells of a wild-type Eschericia coli O8 strain bearing a complete lipopolysaccharide were incubated for alkaline phosphatase reaction product and examined by electron microscopy, the depostion of lead salts was to be observed primarily within the periplasmic space. A similar treatment of cells derived from this strain, which bears a highly abbreviated lipopolysaccharide, showed a mixed cell surface and periplasmic localization of reaction product, suggesting a surface association of a portion of the enzyme. To further explore this possibility, ferritin-antibody conjugates against the active enzyme and its irreversibly dissociated subunits were prepared and allowed to react with cells of both strains. The results obtained from these experiments revealed the presence of both the active enzyme and inactive subunits of the enzyme at the cell surface of the mutant strain. The evidence obtained offers further proof of the validity of the reaction product deposition technique and indicates that alkaline phosphatase may be associated with some component of the outer membrane in this organism. The observation of enzyme subunits at the cell surface further suggests that an association of these subunits with structural components of the cell envelope may provide a locus at which they may dimerize to form active enzyme.

Immunocytological investigation of protein synthesis in Escherichia coli.

Ferritin-conjugated specific antibodies have been used to localize beta-galactosidase and both the monomer and active dimer of alkaline phosphatase in frozen thin sections of cells of Escherichia coli O8 strain F515. The even distribution of the ferritin marker throughout cells that had been induced for beta-galactosidase synthesis, frozen, sectioned, and exposed to ferritin-anti-beta-galactosidase conjugate showed that this enzyme was present throughout the cytoplasm of these cells. Frozen thin sections of cells that had been derepressed for the synthesis of alkaline phosphatase were exposed to both ferritin-anti-alkaline phosphatase monomer and ferritin-anti-alkaline phosphatase dimer conjugates, and the ferritin markers showed a peripheral distribution of both the monomer and the dimer of this enzyme. This indicates that alkaline phosphatase is present only in the peripheral regions of the cell and argues against the existence of a cytoplasmic pool of inactive monomers of this enzyme. This peripheral location of both the monomers and dimers of alkaline phosphatase supports the developing concensus that this enzyme is, like other wall-associated enzymes, synthesized in association with the cytoplasmic membrane and vectorially transported to the periplasmic area, where it assumes its tertiary and quaternary structure and acquires its enzymatic activity.

Tris(hydroxymethyl)aminomethane buffer modification of Escherichia coli outer membrane permeability.

The effect of tris(hydroxymethyl)aminomethane (Tris) buffer on outer membrane permeability was examined in a smooth strain (D280) and in a heptose-deficient lipopolysaccharide strain (F515) of Escherichia coli O8. Tris buffer (pH 8.00) was found to increase outer membrane permeability on the basis of an increased Vo of whole-cell alkaline phosphatase activity and on the basis of sensitivity to lysozyme and altered localization pattern of alkaline phosphatase. The Tris buffer-mediated increase in outer membrane permeability was found to be dependent upon the extent of exposure to and concentration of the Tris buffer. The Tris buffer effects were demonstrated not to be due to allosteric activation of cell-associated alkaline phosphatase and were specific for Tris buffer. Exposure of cells to Tris resulted in the release of a limited amount of cell envelope component. Investigators utilizing Tris buffer are cautioned that Tris is not physiologically inert and that it may interact with the system under investigation.

Cell envelope protection of alkaline phosphatase against acid denaturation in Escherichia coli.

The effects of a highly acidic environment on the cell-associated alkaline phosphatase activities of a smooth and a rough strain of Escherichia coli O8 have been examined. The observation that cell-associated enzyme is denatured to a lesser degree than purified enzyme suggests that the association of the enzyme with the cell envelope affords it some degree of protection from potentially disruptive agents in the environment. The degree of protection afforded the enzyme from pH denaturation appears to be dependent upon the presence of a complete lipopolysaccharide in the outer membrane of these strains. An abbreviation of the chemical structure of this cell envelope component produces a change in the outer membrane, resulting in increased susceptibility of the cells to a battery of antibiotics and to lysozyme and in a small, but significant, change in the sensitivity of the cell envelope-associated alkaline phosphatase to the denaturing effect of an acidic environment.

Equilibrium analysis of binding of Candida albicans to human buccal epithelial cells.

The effect of growth temperature on the binding of Candida albicans to human buccal epithelial cells (BECs) was examined using an equilibrium of binding analysis. Candida albicans was cultured in M9 medium either for 12 h at 25 degrees C or for 9 h at 25 degrees C and then shifted to 37 degrees C for 3 h. The temperature shift did not result in germ tube formation; however, the adherence of C. albicans to BECs was altered. Shifting temperature increased the yeast's ability to bind to BECs. A Langmuir adsorption isotherm was used to calculate the maximum number of available binding sites (N) and the apparent association constants of binding (Ka) for all resolvable adhesin-receptor interactions. Three classes of adhesin-receptor interactions were resolved when the yeast was cultured at 25 degrees C and included a low copy number site (N = 3.0 cfu/BEC; Ka = 2.11 X 10(-6) mL/cfu), a medium copy number site (N = 23.6 cfu/BEC, Ka = 8.21 X 10(-7) mL/cfu), and a high copy number site (N = 91.7 cfu/BEC, Ka = 3.35 X 10(-8) mL/cfu). Two classes of adhesin-receptor interactions were resolved when the incubation temperature was shifted to 37 degrees C: a low copy number site (N = 4.5 cfu/BEC, Ka = 3.98 X 10(-6) mL/cfu) and a high copy number site (N = 150.5 cfu/BEC, Ka = 8.47 X 10(-8) mL/cfu). Augmented C. albicans adherence to BECs due to the elevated growth temperatures appears to result from a temperature-regulated alteration in the C. albicans adhesin that recognizes a high copy number receptor site with relatively low affinity.

Aberrant nuclear morphology in the mouse myeloma SP2/0 cell line.

The mouse myeloma SP2/0 cell line when grown in supplemented Dulbecco's modified Eagle's media spontaneously produced aberrant nucleated cells which increased in frequency with cell culture age. These cells underwent cytological changes associated with apoptosis, that is, the condensation of chromatin followed by karyorrhexis and the production of small apoptotic bodies. Aberrant cells were induced by media changes, centrifugation, and temperature shocking. The rapid induction of aberrant cells by a media change suggests that the mechanism of fragmentation was not associated with cell division.

Mechanism of adhesion of Alysiella bovis to glass surfaces.

Alysiella bovis adheres to surfaces by means of short, ruthenium red-staining, rod-like fimbriae. The fimbriae remain associated with the cell envelope of A. bovis, even when sonicated or exposed sequentially to toluene, Triton X-100, lysozyme, ribonuclease, and deoxyribonuclease. Adhesion of outer membrane-derived cell wall ghosts of A. bovis to glass was inhibited by IO4-, sodium dodecyl sulfate, urea, pronase, and trypsin. Protease treatment digested the fimbriae from the distal end, and exposure to sodium dodecyl sulfate depolymerized the fimbriae. Exposure of ghosts to 1% sodium dodecyl sulfate preferentially solubilized a 16,500-dalton protein which was subsequently purified by gel filtration and demonstrated to be a glycoprotein (ca. 17% carbohydrate). Antibodies raised against the 16,500-dalton glycoprotein agglutinated whole cells and inhibited adhesion of ghosts to glass.