Cloning of Prevotella intermedia loci demonstrating multiple hemolytic domains.

A gene bank was created from Prevotella intermedia strain 27 chromosomal DNA, and a clone was isolated that conferred the expression of two separate modes of hemolytic activity in recombinant Escherichia coli. The original recombinant hemolytic strain (EB34) contained plasmid, pEB34, with a 5.6-kb insert from Sau 3 AI-digested P. intermedia strain 27 chromosomal DNA cloned into the Bam HI site of pUC18. EB34 and deletion subclones were tested for expression of hemolytic activity in a standard tube assay, measuring lysis of erythrocytes spectrophotometrically as a function of hemoglobin release. Cell suspensions of EB34 demonstrated a dose-dependent hemolytic activity, inhibitable by proteases, and heat treatment but not dependent on calcium ions, and not inhibitable by osmoprotectants. Cell-free lysates also demonstrated a heat inhibitable, dose dependent hemolytic activity. Sub-cloning experiments localized the hemolytic region of the insert to a 3.9-kb fragment under direction of the lac promoter. Sequence analysis of the entire insert revealed the presence of multiple open reading frames (1 to 3) in this region which correlated to different forms of hemolytic expression, such that subclones containing all open reading frames 1 to 3 demonstrated strong hemolytic phenotype on blood plates and in the tube assay. Subclones containing only ORF1 demonstrated hemolysis on plates, but not in the tube assay. Subclones containing only open reading frames 2 and 3, but not ORF1 demonstrated hemolysis in the tube assay but not on plates. Homology searches of DNA and protein databases have not revealed significant homologies with reported hemolysins or proteins in any of the open reading frames.

Identification of a fimbriae-associated haemagglutinin from Prevotella intermedia.

Prevotella intermedia, a putative periodontopathic microorganism, possesses various types of fimbriae (surface appendages). Some of these surface structures mediate the adherence of the organism to several mammalian erythrocytes, resulting in the agglutination of the erythrocytes. Prevotella intermedia fimbriae were solubilized and separated by preparative SDS gel electrophoresis followed by preparative isoelectric focusing to determine which fimbrial component(s) were responsible for the haemagglutinating activity exhibited by these bacteria. Heat treatment of isolated fractions which exhibited haemagglutinating activity revealed the presence of two types of haemagglutinating activity which were either heat sensitive or resistant. Analysis of isolated fractions, which exhibited haemagglutinating activity that were heat labile, by Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed the presence of a 3.8 kD peptide. The purified peptide had a pI of 4.8-5.0. Examination of fractions containing this peptide by electron microscopy showed the presence of fimbriae. Immunogold labelling showed that mouse antibodies raised against the 3.8 kD peptide bound specifically and extensively to P. intermedia fimbriae. It appears that this peptide is a fimbriae-associated haemagglutinin and may represent a major constituent of the fimbriae. Further, fractions exhibiting haemagglutinating activity that were heat resistant, which were recovered at a pH of 3.5 in preparative isoelectric focusing of fimbrial samples, did not possess any detectable major protein bands as shown by analytical gel electrophoresis. However, silver stained gels for the detection of lipopolysaccharide (LPS) revealed the presence of LPS-like components in these fractions. In addition, LPS isolated from whole cells showed a similar electrophoretic pattern and exhibited the haemagglutinating activity that was heat resistant. The results of this study strongly suggest that P. intermedia possesses the machinery to agglutinate erythrocytes, which may be a contributing factor in its colonization in vivo.

Identification of hemolytic activity in Prevotella intermedia.

Hemolysin production was measured in strains of Prevotella intermedia. Zones of beta-hemolysis were detected on agar plates supplemented with either sheep, rabbit or human erythrocytes. A standard tube assay was performed on cell suspensions of the organism to measure hemolytic activity, which was found to be dose dependent, eliminated by heat treatment, and saturable with increasing concentrations of blood. Growth-phase experiments suggested that hemolysin production was increased during logarithmic growth and was reduced during stationary phase. Cell fractionation, performed on several strains of P. intermedia, localized the activity in the outer membrane and in cell vesicles. The biological implication of this study is that P. intermedia, by virtue of its hemolytic activity, is capable of liberating the hemoglobin from erythrocytes, thereby acquiring an essential nutrient, iron, for its metabolism.

Coaggregation of Candida albicans with oral Fusobacterium species.

Nine strains of oral Fusobacterium were examined for their ability to coaggregate in vitro with four strains of the oral yeast. Candida albicans. All of the Fusobacterium nucleatum strains and Fusobacterium periodontium and Fusobacterium sulci coaggregated to various degrees with all of the Candida strains. Fusobacterium alocis, Fusobacterium mortiferum and Fusobactrium simiae strains did not coaggregate with any of the Candida strains. Exposure of the coaggregating Fusobacterium strains but not the Candida strains to heat, trypsin, and proteinase K eliminated coaggregation. Amphotericin B or trichodermin treatment of the yeast species had no effect. The reactions were inhibited by addition of 0.1 M mannose, glucosamine and alpha-methyl mannoside. All coaggregating pairs were disaggregated by the addition of sodium dodecyl sulfate, but nonionic detergents had no effect. The addition of 2.0 M urea completely reversed coaggregation. Candida strains were sensitive to periodate oxidation, whereas the Fusobacterium strains were stable to this treatment. All coaggregations occurred in the presence of saliva and appeared stronger than in buffer. These data suggest that the coaggregations involve either a protein or glycoprotein on the Fusobacterium surface, which may interact with carbohydrates or carbohydrate-containing molecules on the surface of the Candida. These observations expand the known range of intergeneric coaggregations occurring between human oral microbes and indicate that coaggregation of C. albicans and Fusobacterium species may be an important factor in oral colonization by this yeast. The authors believe this to be the first description of coaggregation concerning a carbohydrate component on the yeast cell and a protein component on the oral bacterial cell.

Fimbrial-mediated colonization of murine teeth by Actinomyces naeslundii.

Groups of mice fed diets high in sucrose or glucose were orally inoculated with 10(10), 10(9) or 10(8) colony-forming units of one of the following Actinomyces naeslundii strains possessing the type 1 (T1+) and/or the type 2 (T2+) fimbriae: T14VJ1 (T1+, T2+), 5519 (T1+), 5951 (T2+), and 147 (non-fimbriated). Ninety-six hours after inoculation their upper jaws were cultured to look at the implantation of each of these strains on the teeth. In mice fed a sucrose diet, regardless of the presence or absence of fimbriae, each bacterial strain colonized 100% of the mice at the highest inoculation doses of the infecting organism. But at a dose of 10(8), T14V-J1 was the only strain which colonized 100% (12/12) of the mice, 5519 colonized 10/11, 5951 colonized 9/11 and 147 colonized 7/11. These differences were not statistically significant. When mice were fed a high-glucose diet, 100% infection was achieved with strains T14V-J1, 5519 and 5951 only at the highest dose of 10(10) colony-forming units. Strain 147 colonized in 8/9 of the mice at that dosage. At lower dosages, no bacterial strain implanted in 100% of the mice. In the glucose experiment at a dose of 10(8), strains expressing the T1 fimbriae implanted significantly better than strains without the T1 fimbriae. At a dose of 10(9) colony-forming units, the parent strain T14V-J1 implanted significantly better than strains without the T1 fimbriae. Similarly, strain 5519 (T1+) implanted significantly better than 5951 and implanted better than 147, although the difference was not significant. These results suggest that while the presence of the T1 and T2 fimbriae may confer some advantage in the establishment of these organisms in vivo, even the strains without fimbriae were able to colonize. Strains T14VJ1 and 5519 were found to bind well to hydroxyapatite treated with mouse saliva, while strains 5951 and 147 did not. Only T2 fimbriated strains T14V-J1 and 5951 exhibited a lactose-reversible coaggreation with indigenous strains of enterococci that may contribute to the elevated levels of colonization of strain 5951 in vivo.

Prevotella intermedia fimbriae mediate hemagglutination.

Our earlier studies demonstrated that clinical strains of Prevotella intermedia, isolated from human periodontal pockets, possess various types of fimbria (surface appendages) as determined ultrastructurally. These bacteria have the ability to agglutinate selected mammalian erythrocytes. Hemagglutinating activity exhibited by these cells may be attributable to these surface structures. Strain 17, which possess fimbriae of 8 nm in diameter and readily agglutinates human, monkey, sheep, rabbit, and mouse erythrocytes was selected to determine whether these fimbriae possessed the hemagglutinating activity. Fimbriae were mechanically sheared, concentrated by ammonium sulfate precipitation, solubilized in 10 mM Tris-HCl buffer containing 0.5% deoxycholate and partly purified by ultracentrifugation in a 10-50% linear sucrose gradient. Isolated fimbriae banded at a density of 1.20-1.15 g/ml, appeared fairly uniform ultrastructurally, and possessed hemagglutinating activity. The hemagglutinating activity of P. intermedia whole cells and isolated fimbriae was reduced by treatment with proteases and eliminated by treatment with heat at 80 degrees C for 10 min. The optimal pH for the hemagglutination was 7.0. In the process of hemagglutination, P. intermedia whole cells and isolated fimbriae bound to rabbit erythrocytes as observed by: (a) decrease in the hemagglutinating activity of bacterial whole cells and isolated fimbriae after incubation with rabbit erythrocytes, (b) increase in hemagglutinating activity of the red cells used for absorption, and (c) the presence of P. intermedia whole cells and fimbriae on red cells after absorption as revealed by immunoelectron microscopy. Both the immune immunoglobulin G of the rabbit polyclonal anti-fimbriae antibody and the immune immunoglobulin G and its Fab fragments of the murine monoclonal anti-fimbriae antibody were effective inhibitors of hemagglutination mediated by isolated fimbriae. Immunogold labeling showed that the monoclonal antibody bound specifically to P. intermedia fimbriae. These results collectively suggest that the hemagglutination exhibited by P. intermedia may be attributable to the fimbriae or the fimbrial-associated components.

Coaggregation of Candida albicans with oral Actinomyces species.

Eight strains of Actinomyces were examined for their ability to coaggregate in vitro with four strains of Candida albicans. The Actinomyces coaggregated to various degrees with all of the Candida strains. Exposure of the Candida but not the Actinomyces to heat, trypsin, proteinase K, amphotericin B or trichodermin abolished coaggregation. All sugars tested did not inhibit any of the reactions. All coaggregating pairs were disaggregated by the addition of SDS, but nonionic detergents had no effect. The addition of urea or EDTA completely reversed coaggregation. Actinomyces strains were sensitive to periodate oxidation, whereas the Candida strains were unaffected. These data suggest that the coaggregations involve a protein on the Candida surface that may interact with carbohydrates or carbohydrate-containing molecules on the surface of the Actinomyces. These observations expand the known range of intergeneric coaggregations occurring between human oral microbes and indicate that coaggregation of C. albicans and Actinomyces may be an important factor in oral colonization by this yeast.

Inhibition of adherence of Actinomyces naeslundii (Actinomyces viscosus) T14V-J1 to saliva-treated hydroxyapatite by a monoclonal antibody to type 1 fimbriae.

A monoclonal antibody to Actinomyces naeslundii (A. viscosus) T14V-J1 type 1 fimbriae, capable of inhibiting the adherence of this bacterium to salivary proline-rich protein-treated hydroxyapatite, was generated by immunization of SWR mice with A. naeslundii 55-19, a strain derived from T14V-J1 that possess only type 1 fimbriae. Supernatants of hybridomas were screened for reactivity with purified type 1 fimbriae. An IgG monoclonal antibody, 86-49E, blocked the adsorption of the parent strain to proline-rich protein-treated hydroxyapatite by 77% with 1.0 microgram/ml of the monoclonal antibody; the Fab fragment derived from this monoclonal antibody inhibited adherence by 38% at the same concentration. Similarly, the adherence of strain 55-19 was inhibited by 100% and 64% to proline-rich protein-treated hydroxyapatite with 1.0 micrograms/ml of IgG and Fab fragments respectively. Control monoclonal antibody to the subunit of type 1 fimbriae, as well as to Actinobacillus actinomycetemcomitans caused only minimal adherence inhibition. Monoclonal antibody 86-49E also agglutinated both type 1 fimbriae-bearing strains of A. naeslundii T14V-J1 and 55-19 but not strains 59-51 and 147, which lack type 1 fimbriae. Further confirmation of the specificity of monoclonal antibody 86-49E was obtained using these fimbria-deficient mutant strains in an enzyme-linked immunosorbent assay, with the monoclonal antibody binding only to strains possessing type 1 fimbriae. Immunogold labeling in conjunction with electron microscopy suggested binding of monoclonal antibody 86-49E occurring near the distal end of the fimbriae. In contrast, when a monoclonal antibody specific for the type 1 fimbrial subunit but not capable of adherence inhibition was used together with 86-49E in double-labeling experiments, extensive labeling of the fimbriae by the subunit antibody was noted. These data suggest that a monoclonal antibody specific for the type 1 fimbriae of A. naeslundii that is capable of binding to a discrete site on the fimbriae has the capacity to inhibit the adsorption of this organism to saliva-treated hydroxyapatite.

Coaggregation of Prevotella intermedia with oral Actinomyces species.

Five strains of Prevotella intermedia were examined for their ability to coaggregate with various gram-positive and gram-negative species of oral bacteria. Two of the P. intermedia strains coaggregated with selected Actinomyces species, P. intermedia 27 with Actinomyces viscosus T14V and Actinomyces naeslundii ATCC 12104, PK606, PK984, and PK947, and P. intermedia 113 with Actinomyces odontolyticus WVU 1546 and Actinomyces israelii WVU 838. Exposure of both Prevotella strains but not the Actinomyces strains to heat, trypsin, or proteinase K abolished most coaggregations. All pairs were disaggregated by the addition of sodium dodecyl sulfate, but only those coaggregations involving P. intermedia 113 were reversed by the addition of 2.0 M urea. P. intermedia 27 was sensitive to periodate oxidation, whereas the partner strains were stable to this treatment. Most coaggregations occurred in the presence of saliva; however, reactions involving P. intermedia 27 were not as strong as those of buffer-suspended cells. Treatment of both P. intermedia 113 coaggregations pairs with proteinase K and the results obtained from suspensions of these pairs in saliva suggest that different surface molecules of this P. intermedia strain may mediate each of these coaggregations. These data suggest that all of these coaggregations involve either a protein or glycoprotein on the Prevotella strain, which may interact with carbohydrates or carbohydrate-containing molecules on the surface of the Actinomyces strain.

Isolation and characterization of Actinomyces viscosus mutants defective in binding salivary proline-rich proteins.

Recent studies have provided evidence for human salivary proline-rich proteins (PRPs) serving as potential receptors in the acquired pellicle for Actinomyces viscosus type 1 fimbriae. We report here the isolation of mutants derived from A. viscosus T14V-J1 which are defective in binding to PRPs partially purified from parotid gland saliva. Mutagenesis with ethyl methanesulfonate preceded enrichment for cells nonreactive with PRPs by successive adsorptions with PRP-treated latex beads. Screening was accomplished by random selection of 250 isolated colonies from each of four enrichment cycles and reaction with PRP-treated latex beads in microtiter plates. Two mutants of independent origin were examined for adherence to hydroxyapatite treated with either PRPs, proline-rich glycoproteins, deglycosylated proline-rich glycoproteins, or whole saliva. Additional surface properties that were examined included agglutination with polyclonal antisera to type 1 and type 2 fimbriae, agglutination by a monoclonal antibody to type 1 fimbriae that inhibits adherence of the parent strain to saliva-treated hydroxyapatite, the ability to bind monoclonal antibody to the type 1 fimbrial subunit, and lactose-reversible coaggregation with Streptococcus sanguis 34. Both mutants exhibited reduced binding to hydroxyapatite treated with whole saliva or salivary protein preparations but were still capable of reaction with antiserum to type 1 and type 2 fimbriae. In addition, these mutants possessed the ability to bind monoclonal antibody to the type 1 fimbrial subunit in amounts comparable to the amount bound by the parent strain but were not agglutinated by the adherence-inhibiting monoclonal antibody. When considered with previously published data, these results suggest that an adhesive molecule is probably associated with type 1 fimbriae and allows for the interaction of A. viscosus with constituents in the salivary pellicle.

Binding of colloidal gold-labeled salivary proline-rich proteins to Actinomyces viscosus type 1 fimbriae.

Salivary proline-rich proteins (PRPs), which were purified from parotid saliva, were adsorbed onto 15-nm-diameter gold particles to visualize specific binding of the salivary molecules to Actinomyces viscosus type 1 fimbriae. Negatively stained preparations incubated with PRP-gold conjugates but not bovine serum albumin-gold complexes bound specifically to bacteria possessing type 1 fimbriae, A. viscosus T14V-J1 and 5519. Binding of the PRP-gold probes to strains deficient in type 1 fimbriae, i.e., strains 5951 (type 2 fimbriae only) and 147 (no fimbriae), was negligible.

Pellicle receptors for Actinomyces viscosus type 1 fimbriae in vitro.

Actinomyces viscosus T14V-J1 and its fimbria-deficient mutant strain possessing type 1 fimbriae strongly aggregated with latex beads treated with acidic proline-rich protein 1, basic proline-rich proteins, and proline-rich glycoprotein and its deglycosylated derivative. These type 1+ strains did not aggregate with latex beads treated with other proteins, such as salivary amylase, salivary histidine-rich polypeptides, laminin, type 1 collagen, fibronectin, or C1q. The type 1+ strains also adsorbed well to experimental pellicles formed with acidic proline-rich protein 1, basic proline-rich proteins, and proline-rich glycoprotein and its deglycosylated derivative on hydroxyapatite (HA) surfaces. These interactions were inhibited with immunoglobulins and Fabs specific for type 1 fimbriae. Type 1- actinomyces exhibited feeble adsorption to latex beads or HA treated with any of the aforementioned proteins. Collectively, these data indicate that actinomyces type 1 fimbriae may specifically interact with several proline-rich salivary molecules, forming experimental pellicles on HA or polystyrene surfaces.

Hydrophobic interactions and the adherence of Streptococcus sanguis to hydroxylapatite.

Streptococcus sanguis demonstrated a high affinity for hydrocarbon solvents. When aqueous suspensions of the organism were mixed with either hexadecane or toluene, the cells tended to bind to the nonaqueous solvent. Increases in temperature resulted in a greater affinity of cells for hexadecane. Interaction between the cells and hexadecane was also enhanced by dilute aqueous sodium chloride and by low pH (pH less than 5). The results suggest that the cell surface of S. sanguis has hydrophobic properties. Isolated cell walls also tended to partition into the nonaqueous solvent. Amino acid analyses of the walls revealed the presence of several amino acids which possess hydrophobic side chains. It is likely that the hydrophobic amino acids associated with the cell wall contribute to the hydrophobicity of intact S. sanguis. When the adherence of S. sanguis to saliva-coated hydroxylapatite was measured, it was found that hydrophobic bond-disrupting agents, such as the Li+ cation, the SCN- anion, and sodium dodecyl sulfate, were capable of inhibiting the cell-hydroxylapatite union. In addition, it was observed that both urea and tetramethylurea were inhibitors of the adherence, although the latter reagent was the superior inhibitor. The results suggest that the adherence of S. sanguis to saliva-coated smooth surfaces is at least partially dependent on the formation of hydrophobic bonds between the cell and adsorbed salivary proteins. Hydrophobic bonding may contribute to cooperative interactions involving S. sanguis and saliva-coated hydroxylapatite (Nesbitt et al., Infect. Immun. 35:157-165, 1982).

Positive coooperativity in the binding of Streptococcus sanguis to hydroxylapatite.

The adherence of Streptococcus sanguis to hydroxylapatite beads has been analyzed by binding isotherms, Langmuir isotherms, and Scatchard plots. For saliva-coated beads, the Scatchard curves contained components with both positive and negative slopes. The results are interpreted as evidence for positive cooperativity in the binding process. Although all Scatchard curves were similar in shape, distinct differences were observed between saliva samples from different individuals. Salivary agglutinins against whole S. sanguis cells did not appear to influence the shapes of the curves or the extent of adherence. In addition, different strains of S. sanguis yielded similar Scatchard plots. When the binding of S. sanguis to buffer-coated hydroxylapatite beads was analyzed by Scatchard plots or binding isotherms, curves were generated which suggested that either direct ligand-ligand or nonspecific interactions were occurring. Hill plots of the adherence data yielded curves with slopes greater than unity for saliva-coated beads, providing additional support for the view that the interactions between S. sanguis and the pellicle involve cooperative phenomena. In contrast, a Hill plot for the binding data of S. sanguis to buffer-coated hydroxylapatite beads gave a curve with a slope of 0.91 +/- 0.07, suggesting negative cooperativity or limited specificity. When adherence data were plotted by the Langmuir method, curves were obtained which could not discriminate between the binding of the bacteria to the hydroxylapatite beads coated with either saliva or buffer. It was also observed that several different proteins and whole saliva tended to inhibit adherence. Scatchard plots, however, describing the binding of S. sanguis to the proteincoated beads were unique and revealed possible specific and nonspecific interactions. Scatchard analyses of binding data may be useful in understanding the mechanism(s) of adherence of streptococci to smooth surfaces.

Association of protein with the cell wall of Streptococcus mutans.

Cell walls from Streptococcus mutans were prepared by conventional technique and subjected to a series of extraction procedures involving classical protein solvents. The extracted walls contained several non-peptidoglycan amino acids and were also amenable to radiolabeling with [125I]sodium iodide and chloramine T. The cell walls could be chemically modified with tetranitromethane and diazo-1H-tetrazole, suggesting the presence of tyrosine or histidine or both. Flourescence spectra of the walls revealed the presence of either tyrosine or tryptophan. Several proteases, including pronase, trypsin, subtilisin, and proteinase K, removed some of the label from the walls. In contrast, treatment of the walls with salts or denaturants did not result in the solubilization of label. When the walls were solubilized with mutanolysin and subjected to chromatography, three peaks of radioactivity with apparent molecular weights of 73,000, 39,000, and 9,600 were observed. Wall digests subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a single band of radioactivity corresponding to an apparent molecular weight of 79,000. Isoelectric focusing of labeled wall digest gave rise to two major bands of radioactivity with isoelectric points of approximately 2.4 and 5.6. The results suggest that the cell wall of S. mutans contains tightly and possibley covalently bound polypeptide molecules. We propose that the cell wall polypeptides of S. mutans serve as factors in the attachment of the bacteria to smooth surfaces.