Adherence of Streptococcus mutans and Streptococcus sanguis to salivary components bound to glass.

Adherence of radiolabeled Streptococcus mutans and Streptococcus sanguis to saliva-treated glass surfaces was studied under conditions which minimized bacteria-glass interactions. Treatment of glass with an alkylsilane solution decreased nonspecific bacterial adherence and enhanced adsorption of radiolabeled salivary components to these surfaces. Addition of Triton X-100 to the bacterial suspensions also reduced nonspecific adherence to siliconized glass, but did not affect adherence to salivary components attached to siliconized glass. Calcium stimulated S. mutans adherence to saliva-free glass, but inhibited adherence to saliva-treated glass. S. sanguis adherence to either saliva-free or saliva-treated glass was inhibited slightly at high calcium ion concentrations. Adherence of streptococci to saliva-treated glass exhibited saturation kinetics, and the numbers of binding sites on the experimental salivary pellicle and the affinity constants for bacteria-saliva attachment were determined. Preincubation of the streptococci with whole saliva decreased their capacity to adhere to saliva-treated glass, but not to saliva-free glass. Bacteria adherent to saliva-treated glass surfaces were readily desorbed by washing with saliva. The addition of homologous antisera, ammonium sulfate-precipitated immunoglobulins, or Fab fragments to the bacterial suspensions inhibited cell adherence to saliva-treated glass.     

Comparison of Streptococcus mutans and Streptococcus sanguis receptors for human salivary agglutinin.

Oral streptococci vary in their susceptibility to salivary agglutinin-mediated aggregation. To understand the molecular basis of this specificity, the structure and function of receptors for agglutinin from Streptococcus mutans KPSK2 (MSL-1) and Streptococcus sanguis M5 (SSP-5) were compared. Immunological screening of an S. mutans KPSK2 genomic DNA library yielded two identical clones expressing a streptococcal protein that co-migrated with a 220 kDa peptide in SDS extracts from this organism. This protein inhibited agglutinin-mediated aggregation of S. mutans KPSK2 in a dose-dependent manner. The MSL-1 gene is homologous to the S. mutans SpaP and pac genes although single base substitutions alter several amino acids. MSL-1 is also similar to the agglutinin receptor (SSP-5) cloned from S. sanguis M5. All three proteins, MSL-1, P1, and SSP-5 share at least one epitope since monoclonal and polyclonal anti-SSP-5 antibodies react with both MSL-1 and P1. However, other monoclonal antibodies are specific for SSP-5 and appear to react with a peptide domain exhibiting little homology to MSL-1 or P1. Sugar inhibition studies showed that agglutinin-mediated aggregation of S. mutans KPSK2 was most potently inhibited by fucose and lactose. Sialic acid, a potent inhibitor of S. sanguis aggregation, had no effect on the interaction of agglutinin with S. mutans KPSK2. These results suggest that while the MSL-1 and SSP-5 proteins are genetically and immunologically related, their specificity for binding sites on agglutinin differs.     

Cloning of a Streptococcus sanguis adhesin which mediates binding to saliva-coated hydroxyapatite.

Chromosomal DNA from a salivary aggregating strain of Streptococcus sanguis 12 was partially digested with PstI and ligated into the plasmid vector pUC18 and transformed into Escherichia coli JM83. A total of 1,700 recombinant clones of E. coli were examined by a colony immunoassay with antisera raised against either S. sanguis 12 whole cells or S. sanguis 12 surface fibrils. Five clones which reacted with one or the other antiserum were shown to be unique by Western blotting (immunoblotting) and restriction endonuclease digestion. One recombinant plasmid pSA2 expressed two proteins with Mrs of 20,000 and 36,000. The 36,000-Mr protein has been designated SsaB. Both proteins were purified to homogeneity by Sephadex G-75 and ion-exchange chromatography. The proteins were present in mutanolysin digests of whole-cell lysates of S. sanguis 12 and in the non-saliva-aggregating variant 12na and the hydrophilic variant 12L. Polyclonal antiserum raised against the SsaB protein reacted strongly with the cell surfaces of S. sanguis 12 and 12na but not with that of 12L. SsaB inhibited the adhesion of S. sanguis 12na to saliva-coated hydroxyapatite, indicating that the adhesin mediates the binding to the pH-sensitive receptor.     

Recombination-deficient Streptococcus sanguis.

A UV-sensitive derivative was obtained from Streptococcus sanguis Challis. The organism could be transformed with a number of small streptococcal plasmids at frequencies equal to, or 1 logarithm below, the transformation frequencies for the parent organism. However, transformation with chromosomal DNA was greatly impaired in the UV-sensitive derivative.     

Fibrinogen binding and resistance to phagocytosis of Streptococcus sanguis expressing cloned M protein of Streptococcus pyogenes.

The biological properties of Streptococcus pyogenes M protein cloned and expressed in S. sanguis were investigated. The spm-5 gene previously cloned into Escherichia coli was subcloned into the E. coli-S. sanguis shuttle plasmid pVA838 to produce a newly constructed plasmid, pBK100. Cells of S. sanguis transformed with pBK100 expressed 53-, 55-, and 58-kilodalton polypeptides reacting with type 5 M protein antiserum in immunoblots. The M protein was expressed on the surface of S. sanguis cells as shown by the capacity of the intact cells to (i) inhibit the reactivity of anti-type 5 antibodies with purified M protein as demonstrated by enzyme-linked immunosorbent assay; (ii) inhibit the opsonization by M5 antisera of type 5 S. pyogenes; (iii) express M-protein-like fibrils on the surface of the organisms that react with M5 antisera as revealed by immunoelectron microscopy; (iv) bind plasma fibrinogen and, as a consequence, resist phagocytosis by human blood neutrophils; and (v) be rendered susceptible to phagocytosis by opsonic M5 antisera. These results provide additional evidence that streptococcal M proteins bind host proteins as a ploy to evade host defense mechanisms.     

Adherence of Streptococcus sanguis to saliva-coated hydroxyapatite: evidence for two binding sites.

The characteristics of bacterial adherence to saliva-coated hydroxyapatite were examined for a salivary aggregating strain of Streptococcus sanguis, strain 12, and for its nonaggregating variant, strain 12na. Both strains were found to adhere in similar numbers to saliva-coated hydroxyapatite that had been preincubated at 4 degrees C overnight. Preincubation of saliva-coated hydroxyapatite overnight at 37 degrees C reduced subsequent adherence of S. sanguis 12 by approximately 10%, whereas adherence of S. sanguis 12na was reduced by over 80%. Preincubation at 37 degrees C in the presence of neuraminidase reduced adherence of S. sanguis 12 by over 90% and caused some additional reduction in adherence of S. sanguis 12na. The data were analyzed with Langmuir isotherms, Scatchard plots, and Hill plots. Some evidence of cooperativity was seen. A peak in the Scatchard plot for S. sanguis 12 binding to saliva-coated hydroxyapatite preincubated at 4 degrees C disappeared after preincubation at 37 degrees C, suggesting the loss of a salivary receptor. Many more organisms were found to bind when adherence was measured by assays counting the number of organisms remaining in suspension after the beads had settled. These weakly binding organisms, which were removed by washing, demonstrated adherence characteristics similar to those of the firmly bound organisms. Both strains were strongly hydrophobic. It is proposed that the binding of S. sanguis 12 and 12na involves two types of receptor on the salivary pellicle. One type of receptor is stable at 37 degrees C, but sensitive to neuraminidase; the second type is inactivated by prolonged incubation at 37 degrees C. S. sanguis 12 may bind to both types of receptor, whereas S. sanguis 12na binds only to the second type. The neuraminidase-sensitive receptor might be involved in saliva-mediated aggregation.     

Characterization of an adhesion antigen of Streptococcus sanguis G9B.

An antigen possessing the attributes of an adhesion has been identified in Streptococcus sanguis G9B. Cell surface components were extracted from G9B and a spontaneously occurring nonadherent mutant of G9B, strain Adh-, with a 2 mM barbital buffer, pH 8.6. The extract of G9B but not of Adh- absorbed more than 80% of the adhesion-inhibitory activity of anti-G9B immunoglobulin G (IgG). Immunoblots revealed 80- and 52-kilodalton (kDa) antigens present in the G9B extract but not in the Adh- extract. Absorption of anti-G9B IgG with Adh- and G9B barbital extracts showed a correlation between the loss of the 80- and 52-kDa antibodies and the loss of adhesion-inhibitory activity. An antibody prepared against the 80-kDa antigen excised from sodium dodecyl sulfate-polyacrylamide gels recognized the 80- and 52-kDa antigens and another antigen of 62 kDa but did not inhibit adhesion. However, an antibody from an electroblot containing the native protein from which the 80-kDa and related antigens were derived (the 80-kDa antigen complex) inhibited adhesion to the same extent as anti-G9B IgG. Periodate oxidation of the G9B barbital extract modified the 80-kDa antigen complex and resulted in the loss of 40% of its absorbing activity. The barbital extract also contained an endogenous enzyme responsible for producing the 62- and 52-kDa antigens from the 80-kDa protein and which, under optimal conditions, degraded the antigen completely, resulting in the loss of antibody-absorbing activity. The 80-kDa antigen complex has a molecular mass of more than 200 kDa in native polyacrylamide gels and a pI of 4.1 to 4.8. These observations suggest that the adhesion antigen in S. sanguis G9B is a large glycoprotein from which an 80-kDa antigen complex is derived.     

Antigens of Streptococcus sanguis

An antigenic analysis of the alpha-hemolytic streptococci isolated from dental plaque was performed by use of antisera against a strain of Streptococcus sanguis (M-5) which was isolated from dental plaque. Immunoelectrophoretic and Ouchterlony tests of Rantz and Randall extracts of 45 strains gave positive reactions with the M-5 antisera. These strains represented 60% of the strains tested. The number of antigens which could be identified in these extracts varied from one to five and were designated a to e. The a antigen was found in 36 of the strains tested, including reference strains of S. sanguis and the group H streptococci. The strains reacting with the M-5 antisera were divided into two majors types: type I consisted of 23 strains in which the a antigen was found alone or with one or more of the c, d, and e antigens; type II consisted of 13 strains in which both the a and b antigens were found with or without one or more of the c, d, and e antigens. The remaining strains contained, either singly or in combination, the b, c, d, and e antigens but not the a antigen. Biochemical tests of representatives of each serotype and reference strains indicated that strains reacting with M-5 antisera were S. sanguis. These findings suggest that S. sanguis strains share common physiological and serological properties.     

Response of a Streptococcus sanguis strain to arginine-containing peptides.

For dental plaque organisms such as Streptococcus sanguis, the ecological importance of the ability to utilize arginine as an energy source has been established in previous studies. The present investigation was undertaken to determine the ability of a strain of S. sanguis to process unsubstituted arginine-containing peptides. The organism was grown under glucose-limited conditions in a chemically defined medium, and peptide was added to washed, resting cells in a pH-stat at pH 7.0. Filtrates taken at appropriate time intervals were assayed for peptide, free amino acids, and metabolites. Irrespective of the position of the arginine residue, all peptides tested were attacked, although those that possessed a C-terminal arginine (including a tetrapeptide) were processed at a faster rate than were those in which arginine was N terminal. However, C-terminal arginine was cleaved only slowly from a peptide containing 24 residues. In each case, most of the released arginine was converted to ornithine via the arginine deiminase pathway. Such peptidase activities appeared to occur at or near the cell surface and were probably constitutive. It was found that the organism grew in chemically defined medium containing arginine that was present solely in the form of a tripeptide, and also that a strain of S. mutans possessed only a limited ability to attack arginine-containing peptides and was unable to utilize the released arginine.     

Effect of immunization on susceptibility to experimental Streptococcus mutans and Streptococcus sanguis endocarditis.

It has been asserted that humoral immunity is an important potentiating factor in pathogenesis of infective endocarditis, in that prior immunization to certain bacteria may predispose the host to endocarditis caused by those organisms. If so, possible future vaccination of humans with streptococcal antigens for the prevention of dental caries might increase the susceptibility of the population to streptococcal endocarditis. To examine this hypothesis further, we immunized rabbits with killed Streptococcus sanguis or Streptococcus mutans. After complement-fixing antibody had developed, the rabbits were tested for susceptibility to experimental infective endocarditis. Rabbits with high titers of complement-fixing antibody to the infecting organism developed streptococcal endocarditis less often (13%) than animals with lower titers (69%; P less than 0.0002). These findings do not support the hypothesis that pre-immunization predisposes to infective endocarditis and lend no credence to the concept that vaccination of human subjects against dental caries might increase their susceptibility to streptococcal endocarditis. On the contrary, the results of these experiments indicate that specific antibody can confer relative immunity to infective endocarditis.     

Cloning and expression of a Streptococcus sanguis surface antigen that interacts with a human salivary agglutinin.

Human saliva contains a high-molecular-weight glycoprotein (agglutinin) which binds to specific streptococci in a calcium-dependent reaction leading to the formation of bacterial aggregates. We report the cloning of a gene encoding a surface antigen from Streptococcus sanguis M5 and show that the expressed protein inhibits agglutinin-mediated aggregation and specifically binds the salivary agglutinin in a calcium-dependent fashion. Clones isolated from the immunological screening of S. sanguis M5 genomic libraries with polyclonal antibodies against whole cells were assayed for the ability to compete with S. sanguis for agglutinin. One clone, pSSP-5, expressed antigens of 165 and 130 kilodaltons (kDa) possessing this activity. A 3-kilobase-pair (kbp) insert fragment from this clone was used to screen a genomic library in lambda EMBL3 which resulted in the isolation of clone SSP-5A. This clone contained an insert of 17 kb and expressed proteins of 170 to 205 kDa that reacted with the anti-S. sanguis antibodies. Subcloning of a 5.3-kbp EcoRI-BamHI fragment from SSP-5A produced pEB-5, which expressed streptococcal components that were indistinguishable from SSP-5A. The streptococcal antigen was purified by gel permeation and ion exchange chromatography and shown to potently compete with S. sanguis M5 cells for agglutinin. The antigen also bound purified salivary agglutinin in the presence of 1 mM CaCl2. This binding was inhibited by EDTA. Both the SSP-5 antigen and a 205-kDa protein in surface protein extracts from S. sanguis M5 cross-reacted with antibodies directed against antigen B from S. mutans and SpaA from S. sobrinus 6715. These results indicate that a 205-kDa surface protein that is antigenically related to SpaA and antigen B is involved in the binding of salivary agglutinin to S. sanguis M5.     

Adherence of Streptococcus sanguis to conformationally specific determinants in fibronectin.

The adherence of Streptococcus sanguis to specific receptors exposed or deposited at the site of endothelial damage may play an important role in the development of infective endocarditis. Adherence of the Challis strain of S. sanguis to gelatin (or collagen) and gelatin-binding components of plasma was examined with an enzyme-linked immunosorbent assay. S. sanguis adhered poorly to immobilized gelatin and to molecular or fibrillar collagen. However, in the presence of fresh human plasma, the adherence of S. sanguis to all three substrates increased as much as eightfold. Removal of gelatin-binding proteins eliminates the ability of plasma to enhance adherence of S. sanguis to the substrates. Addition of purified human plasma fibronectin (Fn) to the absorbed plasma restored the adherence-promoting ability in a dose-dependent manner. A similar dose-dependent increase in S. sanguis adherence was observed when increasing concentrations of Fn alone were added to the gelatin-coated assay wells. S. sanguis adherence to immobilized fibronectin could not be inhibited by preincubating either the bacteria or the gelatin-coated assay wells with Fn or by including excess soluble Fn in the assay mixture. Studies with peptides purified from trypsin digests of Fn indicated that the 160- to 180-kilodalton (kDa) fragments which retain both the gelatin-binding and the cell-binding regions of the intact molecule support adherence of S. sanguis to gelatin. The 160- to 180-kDa fragments inhibited the interaction of S. sanguis with immobilized Fn. In contrast, intact Fn and the 31-kDa amino-terminal fragment were unable to inhibit the adherence when used in equivalent or greater molar amounts. These in vitro results suggest that in the presence of whole plasma, S. sanguis binds to immobilized gelatin or collagen via Fn bound to the immobilized substrates. Our finding that adherence of S. sanguis to immobilized Fn can occur in the presence of large concentrations of Fn, whether in plasma or purified, indicates that a S. sanguis-binding domain is cryptic in the Fn molecule while in solution and is exposed by a conformational change when the Fn becomes bound to gelatin-coated plastic. The ability of peptide fragments of Fn to inhibit S. sanguis adherence is consistent with this hypothesis.     

Characterization of hemin binding activity of Streptococcus pneumoniae.

Streptococcus pneumoniae is a causative agent of bacterial pneumonia, otitis media, meningitis, and bacteremia. It causes considerable morbidity and mortality throughout the world, especially among children, the elderly, and immunocompromised individuals. We have demonstrated previously that the growth of S. pneumoniae is limited under iron-depleted conditions and can be restored by the addition of either hemin or hemoglobin. In the present study, we showed that S. pneumoniae had the ability to bind hemin and that the level of hemin binding activity was not affected by supplementation of the growth medium with iron. Approximately 70 to 80% of the hemin binding activity was mediated by proteinase-resistant components, and the remainder was mediated by proteins. Hemin binding proteins were located in both soluble extract and envelope fractions of pneumococcal cells. By batch affinity chromatography, a major hemin binding polypeptide with an apparent molecular mass of 43 kDa was identified in the cell lysate of S. pneumoniae. Polyclonal antibodies against this polypeptide were raised. By immunoblot analysis, this hemin binding polypeptide was localized in the envelope and did not exhibit any variation in molecular weight among all serotypes tested. The subcellular distribution of hemin binding activity may have functional implications.     

Role of granulocytes in experimental Streptococcus sanguis endocarditis.

We investigated the role of granulocytes during the induction and course of experimental Streptococcus sanguis endocarditis in rabbits by depleting blood granulocytes with nitrogen mustard. The induction of the endocarditis was not influenced by granulocytopenia: the 50% infectious dose was 5.4 X 10(4) colony-forming units in normal and granulocytopenic rabbits. However, granulocytopenia influenced the curse of the endocarditis, as shown by a significant increase in the number of colony-forming units per gram of vegetation (P less than 0.02) from 24 to 72 h after the injection of 10(5) colony-forming units of S. sanguis. This rise did not occur in the control rabbits. Furthermore, bacteremia was significantly higher in the granulocytopenic rabbits (P less than 0.05) during the first 48 h compared with the control rabbits. This was not because of altered clearance of the streptococcus inoculum or seeding of streptococci from extracardiac bacterial foci. We concluded that granulocytes have no measurable effect on the induction of S. sanguis endocarditis, but during the course of the endocarditis, granulocytes keep the endocardial infection in check.     

Specificity of salivary-bacterial interactions: role of terminal sialic acid residues in the interaction of salivary glycoproteins with Streptococcus sanguis and Streptococcus mutans.

Four highly purified salivary glycoproteins were used to study salivary-bacterial interactions. One pair of glycoproteins was mucin-like in composition, whereas the second pair was not. By an agglutination assay, it was found that only the mucin-glycoproteins agglutinated Streptococcus sanguis and S. mutans. Removal of sialic acid from these molecules resulted in a loss of agglutination of S. sanguis but not of S. mutans. The agglutination phenomenon was shown to require a salivary macromolecule of at least 150,000 daltons.     

Characterization of a rat salivary sialoglycoprotein complex which agglutinates Streptococcus mutans.

Rat saliva agglutinated Streptococcus mutans Ingbritt and NCTC 10449 and Streptococcus sanguis NCTC 7864 but not S. mutans NCTC 10921, GS 5, or LM 7, Streptococcus sobrinus 6715-13 or OMZ 65, or Streptococcus cricetus HS 6, as measured turbidometrically. The specificity of agglutination by rat saliva was the same as that by human saliva. Agglutination was associated with a mucin complex (rat salivary agglutinin complex [rS-A]) of sulfated sialoglycoproteins, with a trace of associated lipid and an apparent Mr of 1.6 X 10(6), isolated by gel-filtration Fast Protein Liquid Chromatography. The complex was dissociated in a high-ionic-strength buffer containing 6 M urea and then fractionated by gel filtration and anion-exchange Fast Protein Liquid Chromatography into four sulfated sialoglycoprotein components, designated rS-A-1Q1, rS-A-1Q2, rS-A-1Q3, and rS-A-2, with rS-A-1Q2 being polydisperse through differential glycosylation of the polypeptide backbone. The dissociation destroyed agglutination activity. The polypeptide backbones contained up to 42% serine plus threonine and up to 40% glycine plus alanine plus proline plus valine. The carbohydrate moiety of the rS-A sialoglycoproteins consisted of N-acetylgalactosamine, sialate, galactose, fucose, N-acetylglucosamine, and small amounts of mannose, with the predominant sugar being N-acetylgalactosamine. Agglutination was inhibited by 1 mM EDTA but was restored by 1.5 mM CaCl2. Agglutination was also inhibited by 5 mM CaCl2; nonimmune sera; cationic polymers; and wheat germ, lentil, soybean, and peanut lectins. However, agglutination was not affected by lipoteichoic acid, various anionic proteins, or various sugars. Neuraminidase treatment of rS-A did not affect activity, but tryptic digestion of S. mutans did prevent agglutination. The results are consistent with calcium bridging the negative groups within the rS-A complex and allowing the approach of rS-A to the bacterial cell surface to effect a specific conformational attachment.     

Immunoelectron microscopic identification and localization of Streptococcus sanguis with peroxidase-labeled antibody: localization of Streptococcus sanguis in intact dental plaque.

Streptococcus sanguis has been localized ultrastructurally within intact dental plaque by means of an indirect technique which utilizes horseradish peroxidase-labeled antibody. The technique allows for complete diffusion of the reagents to all portions of the plaque specimens. Control procedures can be carried out on serial sections of plaque with a bacterial composition similar to that of the experimental specimen. The 30-mum-thick sections can be examined in the light microscope to localize areas specifically labeled with peroxidase prior to cutting ultra-thin sections for electron microscopy. This study demonstrated that specific bacteria can be localized within intact dental plaque. The results also indicated that S. sanguis grows in dental plaque as columnar shaped microcolonies perpendicular to the tooth surfaces. Growth appears to be by cell division rather than deposition of new cells at the surfaces. Despite their relatively good structural preservation, the cells in the deeper (older) layers of plaque appear to have lost some of their antigenic activity in comparison to the cells near the surface.     

Lactose-reversible coaggregation between oral actinomycetes and Streptococcus sanguis.

Freshly isolated strains of oral actinomycetes were obtained from human dental plaque and were tested for the ability to coaggregate with common laboratory stock strains of Streptococcus sanguis. Strains belonging to the genera Actinomyces, Arachnia, Bifidobacterium, and Bacterionema were isolated. Only members of the genus Actinomyces coaggregated with the streptococci, and only Actinomyces viscosus and Actinomyces naeslundii exhibited lactose-reversible interactions. A total of 61 strains, consisting of all of the A. viscosus isolates and 86% of the A. naeslundii isolates, coaggregated; 87% inhibited lactose-reversible coaggregation. On the basis of this property and the altered ability of strains to coaggregate after heat treatment of the cells, we delineated four coaggregation groups. The other 13% of the strains constituted a fifth group, which was characterized by a pattern of closely related interactions that were not reversed by lactose. Compared with previously characterized coaggregation properties determined with stock culture strains of actinomycetes, more than 80% of these fresh isolates exhibited identical coaggregation properties. Thus, most of the coaggregation between freshly isolated oral actinomycetes and streptococci involves lactose-reversible cell-cell interactions, which suggests that such coaggregation is mediated by a network of lectin-carbohydrate interactions between complementary cell surface structures on the two cell types.     

Platelet-interactive products of Streptococcus sanguis protoplasts

To isolate a more native, platelet-interactive macromolecule (class II antigen) of Streptococcus sanguis, cultured protoplasts were used as a source. Protoplasts were optimally prepared from fresh washed cells by digestion with 80 U of mutanolysin per ml for 75 min at 37 degrees C while osmotically stabilized in 26% (wt/vol) raffinose. Osmotically stabilized forms were surrounded by a 9-nm bilaminar membrane, as shown by transmission electron microscopy. Protoplasts were cultured in chemically defined synthetic medium and osmotically stabilized by ammonium chloride. Spent culture media were harvested daily for 7 days. Each day, soluble proteins were isolated from media, preincubated with platelet-rich plasma, and tested for inhibition of platelet aggregation induced by S. sanguis cells. Products released from S. sanguis protoplasts and reactive with an anti-class II antigen immunoaffinity matrix were able to inhibit S. sanguis-induced platelet aggregation. As resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, anti-class II-reactive protoplast products included silver-stained bands of 67, 79, 115, 216, and 248 kDa. The 115-kDa protein fraction was isolated by gel filtration and ion-exchange chromatography. This form of the class II antigen contained N-formylmethionine at its amino terminus. Rhamnose constituted 18.2% of the total residual dry weight and nearly half of its carbohydrate content. Diester phosphorus constituted 1% of this fraction. After trypsinization of the protoplast products from either preparation, a 65-kDa protein fragment was recovered. This protoplast protein fragment and the S. sanguis cell-derived 65-kDa class II antigen, previously implicated in the induction of platelet aggregation, were shown to be functionally and immunologically identical.