INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Dental plaque accumulation around the gingival crevice and other oral surfaces is a predisposing factor for the initiation of periodontal diseases. Among bacterial species in plaque, Porphyromonas gingivalis, a putative major etiologic agent of periodontal diseases (28), has been shown to prevail in a variety of environments among surface components lining the oral cavity, such as mucosal membrane (7), healthy crevices (6), and supra- and subgingival plaques (30). Saliva coats all of these surfaces and is thought to be critical for the organism to adhere to and colonize the oral cavity (11, 13, 24).

Acidic proline-rich proteins (PRPs) have been reported to act as salivary receptors for several plaque-forming bacteria, such as Streptococcus gordonii, Streptococcus uberis, Streptococcus sanguis, Actinomyces viscosus, and Actinomyces naeslundii (11, 24). The mechanisms involved in these interactions are not fully understood. It was previously shown that fimbriae strongly bind to acidic PRPs and statherin by protein-protein interactions through definitive domains of the fimbriae (4) and salivary proteins (1-3, 14). The minimum active domain of PRP1 (a major variant of acidic PRPs) for binding to P. gingivalis fimbriae was found to be Pro-Gln-Gly-Pro-Pro-Gln (PQGPPQ). This peptide sequence is shared by a family of acidic and basic PRPs as a typical repeating sequence (1). The synthetic peptide PRP-C (pPRP-C), containing PQGPPQ, significantly inhibits the binding of fimbriae to salivary receptor proteins, i.e., acidic and basic PRPs and their size variants (1).

Recently, model systems using nonpathogenic oral streptococci were constructed for the secretion or surface expression of various biologically active proteins (16, 25-27). These trials were aimed toward therapy using recombinant organisms in place of commensal oral streptococci to induce protective host immune responses or to inhibit the adherence of pathogenic bacteria. Thus, replacement therapy could be a candidate for molecularly engineered vaccinations to prevent oral diseases.

The colonization of P. gingivalis is thought to be initiated by the direct anchoring of the organism to saliva-coated host surfaces or commensal plaque-forming bacteria (13, 15, 18, 24). In this study, S. gordonii was engineered to secrete the functional peptide pPRP-C by using a shuttle vector plasmid. We evaluated the inhibitory effects of the secreted peptide in the interactions and coaggregation of P. gingivalis with both the salivary component PRP1 and various oral streptococcal cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial culture conditions. P. gingivalis ATCC 33277 was grown and radiolabeled with [³H]thymidine as described previously (4). Streptococcus mitis ATCC 15909 and ATCC 15912, S. gordonii G9B, Streptococcus oralis ATCC 9811 and ATCC 10557, Streptococcus downei MFe 28, and S. sanguis ATCC 10556 were selected from our culture collections and were cultured as described previously (20). Bacterial cells were washed three times and suspended in an appropriate buffer for assay. Escherichia coli JM109 was grown in Luria-Bertani broth or medium containing 1.5% agar.

Preparation of P. gingivalis fimbriae and synthetic pPRP-C. Fimbriae were purified from P. gingivalis ATCC 33277 by the method of Yoshimura et al. (29) and iodinated as described previously (4). The synthetic pPRP-C, corresponding to the carboxy-terminal segment composed of 21 residues of PRP1, was synthesized and purified in a previous study (14). The amino acid sequence of pPRP-C is PQGPPPQGGRPQGPPQGQSPQ.

Preparation of polyclonal antibodies to synthetic pPRP-C. pPRP-C was polymerized by the addition of a cysteine residue at the amino terminus according to the m-maleimidobenzoyloxy succinimide method (12). Two clean rabbits were injected under the skin of the back with pPRP-C (0.67 mg in 0.3 ml of phosphate-buffered saline [pH 7.5]) emulsified with an equal volume of complete Freund's adjuvant for immunization. Beginning two weeks after the first immunization, these rabbits were given boosters four times with the same amount of the immunogen over 8 additional weeks. Two weeks after the last immunization, these rabbits were bled and their sera were separated by centrifugation. The reactivities of these sera to pPRP-C were confirmed and the immunoglobulin Gs (IgGs) were fractionated with a protein A affinity column (Amersham Pharmacia Biotech, Uppsala, Sweden).

Construction of recombinant S. gordonii. Shuttle vector plasmid pMNK-4 derived from pVA838 (19) was donated by T. Morita (Research Institute for Microbial Diseases, Osaka University). pMNK-4 was previously constructed to express and secrete Arthrobacter sp. dextranase by the insertion of dex linked to a DNA sequence encoding a promoter and a signal peptide of S. downei glucosyltransferase I, followed by the E. coli rrnBt1t2 terminator (16). DNA fragments containing the above promoter and signal peptide sequences were obtained by PCR with pMNK-4, a forward primer (5'-GCGCATGCGGATCGTC TATGGTAAAACAGAGAAGAA-3'), and a reverse primer (5'-TGCGCTAGCAACTGAAGCACCGAGA-3'). The forward and reverse primers incorporated the restriction enzyme sites of SphI and NheI, respectively (underlined). The PCR product was ligated into plasmid pGEM-T (Promega, Madison, Wis.) for DNA sequencing with the 373 DNA sequencing system (Perkin-Elmer Corp., Foster City, Calif.), and the resulting plasmid was digested with SphI and NheI.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Construction of shuttle vector pMNK-5. pMNK-5 was constructed by the tandem linkage of DNA fragments encoding a promoter and a signal peptide of S. downei glucosyltransferase I followed by the E. coli rrnBt1t2 terminator.

Transformation of S. gordonii with the construct. S. gordonii G9B was cultured in Todd-Hewitt broth supplemented with 0.2% yeast extract (THY broth) to the early logarithmic phase at 37°C. The harvested cells were heat shocked at 43°C and washed with 15% glycerol. pMNK-5 (1 µg/200 µl of cell suspension) was electroporated (1.75 kV, 25 µF, 400 , and 7 ms) into the competent cells. The cells were plated onto brain heart infusion agar plates containing 20 mM glucose and erythromycin (25 µg/ml) and were cultured at 37°C for 48 h. The positive transformants were screened by a direct PCR.

Preparation of r-pPRP-C secreted by the transformant. The recombinant S. gordonii was grown in THY broth containing erythromycin (25 µg/ml) at 37°C. The culture supernatant containing the secreted recombinant pPRP-C (r-pPRP-C) was dialyzed against deionized water overnight at 4°C through a membrane with a molecular weight cutoff of 500 (Spectrum Medical Industry Inc., Gardena, Calif.). The dialysate was filtrated through a membrane with a molecular weight cutoff of 20,000 (Toyo Roshi Co., Tokyo, Japan). The resulting solution was collected and freeze-dried for concentration.

Blot assay and determination of r-pPRP-C secreted by the transformant. The proteins and peptides were immobilized on a polyvinylidene difluoride membrane (Bio-Rad Laboratories) with a Bio-Dot apparatus (Bio-Rad Laboratories) for the dot blot assay with a horseradish peroxidase conjugate substrate kit (Bio-Rad Laboratories) as described previously (17). For the Western blotting, the samples were loaded for electrophoresis on a 15 to 25% polyacrylamide gel (Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan), and the proteins and peptides were transferred onto a Hybond enhanced chemiluminescence (ECL) nitrocellulose membrane (Amersham Pharmacia Biotech) as described previously (1). pPRP-C and r-pPRP-C were probed with anti-pPRP-C rabbit IgG (1:1,000) with anti-rabbit goat IgG (1:2,000) and an ECL Western blotting detection kit (Amersham Pharmacia Biotech). Bovine serum albumin was used as a negative control. The amount of r-pPRP-C secreted in the culture supernatant was measured with 96-well enzyme-linked immunosorbent assay plates (flat-bottom amino plate type A; Sumitomo Bakelite Co., Ltd., Tokyo, Japan) as previously described (2). Various known amounts of synthetic pPRP-C were dissolved in the culture supernatant of wild S. gordonii and then used as controls. The peptides were measured with anti-pPRP-C peptide sera (1:2,000). The antibodies that reacted with pPRP-C were detected by using a 1:1,000 dilution of horseradish peroxidase-conjugated anti-rabbit IgG (heavy plus light chain; Zymed Laboratories, Inc., San Francisco, Calif.) at room temperature for 2 h. The enzyme reaction proceeded with 3,3',5,5'-tetramethylbenzidine (Moss, Inc., Pasadena, Md.) as a substrate. The reaction was terminated with 0.5 M HCl, and the color intensity was measured at 450 nm. All assays were performed in triplicate on three separate occasions. Data were expressed as mean ± standard deviation. The peptide and protein content was measured by dry weight.

Binding assay using PRP1-coated HA beads. Assays of the binding of [³H]thymidine-labeled P. gingivalis cells and ¹²⁵I-labeled fimbriae to PRP1-coated HA beads (2 mg) were carried out with a 50 mM KCl buffer containing 1 mM KH₂PO₄, 1 mM CaCl₂, and 0.1 mM MgCl₂ (pH 6.8) as described previously (4). Synthetic pPRP-C (500 nmol) was dissolved in 1 ml of a 47-fold-concentrated supernatant of wild S. gordonii (40 mg/ml). ¹²⁵I-fimbriae (0.5 nmol) or [³H]thymidine-labeled cells (10⁸ cells) were added simultaneously with various inhibitors (200 µl) to PRP1-coated HA beads in glass tubes to a final volume of 400 µl. The specific binding was calculated by subtracting the nonspecific binding which was obtained by the preincubation of PRP1-coated HA beads with nonlabeled fimbriae (500 µl at 50 nmol/ml) at room temperature for 1 h. All assays were performed in triplicate on three separate occasions.

Coaggregation assay. The assays of coaggregation between P. gingivalis and some oral streptococci were performed according to the turbidimetric method of Nagata et al. (20). The progress of coaggregation was monitored at 37°C by measurement of the decrease in A₅₅₀ in 10 mM phosphate-buffered saline containing 0.15 M NaCl (pH 6.0).

Statistical analysis. The data were averaged (mean ± standard deviation), comparisons were performed with Student's t test, and P values of <0.01 were considered significant. The regression line was obtained by the least-squares method.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Secretion of r-pPRP-C by recombinant S. gordonii. It was confirmed by DNA sequencing that the combined structural gene containing the promoter-leader sequence, prpC, and the terminator was constructed in pMNK-5 in the correct frame and orientation. The transformant cells (recombinant S. gordonii) were grown in THY broth, and the culture supernatant was concentrated 47-fold (vol/vol) following dialysis. In the dot blot assay, the recombinant S. gordonii concentrate (20 µl at 40 mg/ml) clearly reacted with the anti-pPRP-C IgG while that of the wild cells showed no positive reaction (Fig. 2A). The molecular size of r-pPRP-C secreted by recombinant S. gordonii was examined by Western blotting. r-pPRP-C in the concentrate migrated as far as synthetic pPRP-C (Fig. 2B), indicating that the leader sequence was digested by the host cells. To determine the concentration of r-pPRP-C in the nonconcentrated recombinant S. gordonii supernatant, a regression line [y = (8.8 × 10²) x + (2.8 × 10³)] was obtained by dissolving various known amounts of the synthetic pPRP-C in the culture supernatant of wild S. gordonii (Fig. 2C). The supernatant gave an absorbance (A₄₅₀) of 0.040. Thus, the secreted amount of r-pPRP-C was estimated to be 8.6 µg/ml (4.3 nmol/ml) of the medium.

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Secretion of r-pPRP-C by recombinant S. gordonii. (A) Dot blot assay of r-pPRP-C secreted by recombinant S. gordonii. The culture supernatant of recombinant S. gordonii containing r-pPRP-C was concentrated 47-fold (vol/vol; 40 mg/ml) as stated in the text. Samples of 20 µl were adsorbed to a polyvinylidene difluoride membrane. The r-pPRP-C on the membrane was probed with anti-pPRP-C rabbit antibodies. Dots: 1, synthetic pPRP-C (1 mg/ml); 2, culture supernatant containing r-pPRP-C secreted by recombinant S. gordonii (40 mg/ml); 3, culture supernatant of wild S. gordonii (40 mg/ml); 4, 1% bovine serum albumin. (B) Western blot assay of r-pPRP-C in the culture supernatant of recombinant S. gordonii. The r-pPRP-C on the membrane was probed with anti-pPRP-C rabbit antibodies. Lanes: 1, r-pPRP-C in the culture supernatant secreted by recombinant S. gordonii (20 µl [40 mg/ml]); 2, synthetic pPRP-C (20 µl [1 mg/ml]); 3, culture supernatant of wild S. gordonii (20 µl [40 mg/ml]). Molecular sizes are on the left of the lanes. (C) Determination of the amount of r-pPRP-C in the nonconcentrated culture supernatant of recombinant S. gordonii by enzyme-linked immunosorbent assay. The regression line [y = (8.8 × 102) x + (2.8 × 103)] was obtained with various known amounts of synthetic pPRP-C dissolved in the nonconcentrated culture supernatant of wild S. gordonii (). The amount of the r-pPRP-C in the culture supernatant (100 µl) () was determined with the regression line.

Effect of the secreted r-pPRP-C. The inhibitory effect of r-pPRP-C in the 47-fold concentrate (40 mg/ml) was examined on the binding of ¹²⁵I-fimbriae and [³H]thymidine-labeled cells of P. gingivalis to PRP1-coated HA beads (Fig. 3). The amount of r-pPRP-C was estimated to be 5.05 nmol/mg of the concentrate. Synthetic pPRP-C (500 nmol) was dissolved in 1 ml of culture supernatant of wild S. gordonii which was also concentrated 47-fold (40 mg/ml) by the method used for recombinant S. gordonii. PRP1 in solution, which showed no inhibitory effect, and the concentrated supernatant of wild S. gordonii were used as negative controls. Synthetic pPRP-C markedly inhibited the binding of fimbriae and whole cells to PRP1 in a dose-dependent manner. Synthetic pPRP-C (final concentration, 250 nmol/ml) significantly inhibited the binding of fimbriae and whole cells by 93 and 90%, respectively. The r-pPRP-C concentrate also revealed significant inhibition of fimbrial and whole-cell binding (77 and 72%, respectively) when 8 mg was added. Since 8 mg of the concentrate was estimated to contain 40.4 nmol of r-pPRP-C, r-pPRP-C seems to be as effective as synthetic pPRP-C.

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Comparison of effects of r-pPRP-C and synthetic pPRP-C. The inhibitory effects of r-pPRP-C at various concentrations on the bindings of 125I-fimbriae and [3H]thymidine-labeled cells of P. gingivalis (3H-cells) to PRP1-coated HA beads were examined. Synthetic pPRP-C (500 nmol) was dissolved in 1 ml of 47-fold-concentrated supernatant (sup.) of wild S. gordonii (Sg) (40 mg/ml). 125I-fimbriae (0.5 nmol) or 3H-cells (108 cells) were simultaneously added with various inhibitors (200 µl) to PRP1-coated HA beads in glass tubes to a final volume of 400 µl. PRP1 (0.5 µmol/ml) and 47-fold-concentrated supernatant of wild S. gordonii (40 mg/ml) were negative controls. Inhibitory effects of synthetic pPRP-C and r-pPRP-C on the binding of 125I-fimbriae to PRP1-coated HA beads (A) and 3H-cells to PRP1-coated HA beads (B). The binding levels were calculated by subtracting the nonspecific binding level, which was obtained by the preincubation of HA beads with nonlabeled fimbriae (25 nmol).

Ability of synthetic pPRP-C to inhibit coaggregation. The coaggregation assay was performed with synthetic pPRP-C to examine the effect of pPRP-C on the coaggregation of P. gingivalis with oral streptococci. The synthetic pPRP-C clearly inhibited the coaggregation of P. gingivalis with various streptococcal strains in a dose-dependent manner as shown in Fig. 4. At a concentration of 133 nmol/ml, coaggregation with S. mitis ATCC 15909 was inhibited by 81%, and coaggregation with other strains was also significantly inhibited; S. mitis ATCC 15912, 73%; S. gordonii G9B, 70%; S. oralis ATCC 9811, 68%; S. oralis ATCC 10557, 60%; S. downei MFe 28, 54%; and S. sanguis ATCC 10556, 44%. These inhibitory effects were not further increased by the addition of pPRP-C above the final concentration.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Effect of synthetic pPRP-C on the coaggregation of P. gingivalis with oral streptococci. Aliquots (0.5 ml) of suspensions of P. gingivalis (A550 = 1.0; 5 × 108 cells) and each strain of oral streptococci (A550 = 0.35; 5 × 108 cells) and various concentrations of pPRP-C were simultaneously mixed to a total volume of 2 ml in a cuvette. The progress of coaggregation was monitored by the measurement of the decrease in A550 at 37°C in a temperature-controlled cuvette compartment with a magnetic stirrer. After the decrease in A550 had been recorded for 7.5 min, the change in A550/min was continuously calculated. The maximum value of the decrease in A550/min is the coaggregation activity. Lines: a (), S. mitis ATCC 15909; b (), S. mitis ATCC 15912; c (), S. gordonii G9B; d (), S. oralis ATCC 9811; e (), S. oralis ATCC 10557; f (), S. downei MFe 28; g (), S. sanguis ATCC 10556. All assays were performed in triplicate on three separate occasions.

Ability of r-pPRP-C to inhibit coaggregation. The turbidimetric method is suitable for evaluating the inhibitory effect on coaggregation (20). However, the r-pPRP-C concentrate was colored deep yellow due to the broth constituents, which prevented measurement of the decrease in A₅₅₀ by a spectrophotometer. Thus, the coaggregation was evaluated by the naked eye with a visual rating scale from  to 4+ (Table 1). All of the streptococcal cells examined aggregated with P. gingivalis cells with a score of 4+. Equal amounts of synthetic pPRP-C (67 nmol) and r-pPRP-C (67 nmol in 13.2 mg of the concentrate) were added as inhibitors. The coaggregation scores dropped to 2+ and 1+ with the addition of r-pPRP-C, which was as effective as synthetic pPRP-C. The effect of the concentrated culture supernatant of wild S. gordonii was negligible.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Salivary flow is able to prevent bacterial adherences to oral tissues (9, 10, 24); however, salivary components capable of interacting with P. gingivalis fimbriae appear to act as receptor components to promote the adherence of this organism to oral surfaces. The bindings of P. gingivalis fimbriae to salivary proteins (acidic and basic PRPs and statherin) are mediated by unique hidden receptors termed cryptitopes (4). These interactions occur by the exposure of the functional domain induced by conformational change only when the salivary proteins are fixed to surfaces, such as those of HA beads. Thus, these proteins in solution showed no ability to bind fimbriae. The synthetic pPRP-C, a carboxy-terminal fragment of PRP1, was not cryptic in nature (1, 14), and the peptide diminished P. gingivalis binding to salivary proteins on apatitic surfaces by a competitive inhibition of the binding domain of the fimbrillin molecule (1).

The r-pPRP-C was successfully secreted at a concentration of 8.6 µg/ml (4.3 nmol/ml) of the culture medium by recombinant S. gordonii. The nonconcentrated culture supernatant was used for the inhibition assays; however, the dose of r-pPRP-C was not sufficient to obtain significant effects (data not shown). It was quite hard to purify the r-pPRP-C from the supernatant with a good yield, so we assayed the concentrate. The components of the culture broth also showed weak inhibitory effects even after dialysis to remove the contaminants, as shown in Fig. 3. However, the results obtained here show that r-pPRP-C was as effective as synthetic pPRP-C. In this study, pPRP-C was found to have a prominent effect on the coaggregation of P. gingivalis with various streptococcal strains. For the adherence in the oral cavity, the major targets of P. gingivalis are saliva-coated surfaces and the various bacteria forming dental plaque. It is interesting that pPRP-C seems capable of inhibiting P. gingivalis interactions with both saliva and a major plaque-forming factor, streptococci. Meanwhile, recombinant S. gordonii can bind to other salivary components, such as amylase and mucin (24), even in the presence of pPRP-C. The mechanism involved in the inhibition of pPRP-C in the coaggregation is unknown. It has been suggested that the surface antigen (PAc) of S. mutans binds to salivary components through a proline-rich repeating region of the molecule (21). Streptococcal cells might possess specific proline-rich adhesive epitopes and/or proteins similar in amino acid sequence to pPRP-C on their surfaces.

Several secretion and expression systems have been constructed to produce bioactive proteins in nonpathogenic bacteria with plasmids transformed with foreign DNA, and efforts in oral biology have especially been concentrated on the inhibition of the glucosyltransferases (GTFs) promoting tooth decay by cariogenic S. mutans. Dextranase is an enzyme which hydrolyzes the cariogenic glucan produced by GTFs on the tooth surface; thus, the Arthrobacter gene encoding dextranase was introduced into S. gordonii by using the same shuttle vector as in a previous study (16). The transformants sometimes could be made to cease translating the foreign gene in the absence of selective antibiotics. Shiroza and Kuramitsu developed the resident plasmid integration method (26) and generated recombinant S. gordonii secreting cycloisomaltooligosaccharide glucanotransferase, a potent inhibitor of streptococcal GTF, by transformation with a foreign gene from Bacillus circulans (27). As an attempt to prevent periodontal disease, P. gingivalis fimbrillin polypeptide was engineered to be secreted by S. gordonii (25). The culture medium of the transformant was shown by immunoblotting to contain the fimbrillin peptide, and it was expected that the recombinant peptide would help to trigger an antibody response and to block fimbria-mediated adherence of P. gingivalis in vivo. These reports suggest that genetically engineered oral bacteria could be used to prevent dental caries and periodontal disease.

The efficiency of secretion and expression (e.g., the amount of secreted proteins) was not reported in the oral disease studies mentioned above; however, other reports have referred to secretion efficiencies. The -amylase gene from Bacillus amyloliquefaciens in multicopy plasmid pUB110 was inserted to transform B. subtilis, and the amount of the secreted enzyme was estimated to be 1.5 mg/ml of the culture medium (23). The same group inserted a hybrid gene encoding -amylase and human leukocyte alpha 2 interferon to transform B. subtilis and reported that the product was secreted at 0.5 to 1 µg/ml of the culture medium (22). The secretion efficiency of the present construct seems to be comparable to those in previous reports.

P. gingivalis often induces severe types of marginal periodontitis (8). Plaque control or oral prophylaxis is a crucial factor in preventing periodontal diseases; however, it is at present impossible to selectively control colonization by specific pathogens. pPRP-C might be an effective chemical agent in supragingival plaque control as well as in prevention of P. gingivalis-induced periodontitis. The usefulness of this model should be examined with animal models in order to form a better proposal for future studies.