INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Porphyromonas gingivalis has been implicated as a major etiological agent in periodontitis (6), which becomes a chronic state typified by extensive tissue degradation, with eventual bone erosion and tooth loss (53). A number of virulence factors have been implicated in the pathogenicity of P. gingivalis, with the cysteine proteinases having received considerable attention due to their ability to activate and/or degrade a broad range of host proteins (42). The cysteine proteinases referred to as gingipains cleave synthetic and natural substrates after arginine or lysine residues. Recent in vitro data support the role of gingipains R in the inflammatory process by the generation of C5a and the activation of prokallikrein, with subsequent bradykinin release from kininogen (7, 43). This cascade of events induced by gingipains R has been postulated to contribute to the pathogenic potential of P. gingivalis in vivo (20, 21, 54). Gingipain K may also be involved in disease progression, primarily through binding and degradation of fibrinogen and fibronectin (7, 40, 48).

Arginine-specific gingipains are encoded by two related genes, while lysine-specific gingipains are encoded by only a single genetic locus (reviewed in reference 12). For gingipain R1 and gingipain K, the translated part of each gene, rgp1 and kgp, respectively, encodes a prepropeptide, a catalytic domain, and a hemagglutinin domain, and the initial polyprotein is apparently subject to posttranslational processing. The major forms of high-molecular-mass gingipains, gingipain R1 and gingipain K, released by P. gingivalis HG66 are purified as noncovalent complexes of the catalytic domain with several polypeptide chains (GP44, GP19, GP17, and GP27) derived from the nascent hemagglutinin domain via putative proteolytic processing after one lysine and three arginine residues (39). The hemagglutinin domains of gingipain R1 and gingipain K are nearly identical in primary sequence in each enzyme. The catalytic domains of both gingipain R1 and gingipain K share only limited identity (27%) scattered throughout the polypeptide chain, except for an identical 30-amino-acid residue fragment located in the C-terminal part of the molecule. In contrast to gingipain R1, gingipain R2 is expressed as a preproenzyme missing the majority of hemagglutinin domain but otherwise closely related to the rgp1 gene product (29). The N-terminal two-thirds of the primary structure of mature gingipain R2, containing the active-site cysteine residue and putative histidine residue of the catalytic dyad, is nearly identical with the catalytic domain of gingipain R1, but the two structures differ considerably at the C terminus (31).

In P. gingivalis W50, an alternative proteolytic processing of the initial polypeptide encoded by the prpR1 gene, a gene basically identical to rgp1, apparently leads to assembly of a slightly different noncovalent complex composed of the catalytic (a-chain) and the hemagglutinin (b-chain) domains (2). Posttranslational modification of the catalytic domain can give rise to three biochemically distinct enzymes, a dimer and two catalytic monomers, one unmodified and one lipopolysaccharide modified (45). In this regard, the 95-kDa gingipain R1 is an equivalent of the dimer, while the 50-kDa gingipain R2 is analogous to unmodified catalytic monomer, a product of homologous gene designated as prpR2 in strain W50 (45).

Antibodies specific for gingipains are produced in patients with adult periodontitis patients, with the majority being reactive with antigenic determinants within the hemagglutinin/adhesion domain of gingipain R1 (22). However, the function of these antibodies has not been elucidated. Patients with a history of destructive periodontal disease frequently demonstrate an elevated immunoglobulin G (IgG) response to P. gingivalis; however, these antibodies are apparently ineffective at limiting continued disease progression (16, 28, 51, 55). Indeed, in several animal studies, induction of an immune response to P. gingivalis antigens has been demonstrated to actually exacerbate disease (reviewed in reference 26). In this report, we describe animal experiments which were designed to determine the protective effect of P. gingivalis-specific antibodies produced against functionally defined peptide fragments derived from the catalytic and hemagglutinin/adhesion domains of the 95-kDa gingipain R1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria. P. gingivalis A7436 is the virulent strain used in our previous studies in the mouse chamber model (13, 14). Strain A7436 expresses gingipains R1 and R2 and contains the rgp1 and rgp2 genes as detected by enzymatic activity and Northern blot analysis, respectively (unpublished data). P. gingivalis A7436 was typically maintained on anaerobic blood agar plates (Difco) and grown for 2 to 3 days under anaerobic conditions as previously described (14). For mouse challenge studies, P. gingivalis cultures were grown for 18 h in anaerobic broth MIC (Difco), washed, and adjusted to a final cell density of 10¹⁰ CFU/ml. Mice were infected with approximately 10⁹ CFU of P. gingivalis A7436 directly into subcutaneous chambers implanted in mice as described below. Exact CFU used for each experiment was determined by serially diluting the inoculum and plating on anaerobic blood agar followed by incubation for 7 days under anaerobic conditions.

Antigen preparation. Whole-cell antigens used for immunization were prepared by centrifugation of P. gingivalis cultures for 10 min at 10,000 × g at room temperature and resuspended in 1/10 the original volume in anaerobic broth. Bacterial cells (approximately 10⁸ CFU, as determined by serially diluting cultures prior to heat inactivation) were heated to 95°C for 10 min; the heat-treated preparations were plated on anaerobic blood agar and incubated for 7 days under anaerobic conditions to confirm effective killing. Gingipains were purified from strain HG66 (39), and multiple antigenic peptide (MAP)-conjugated peptides were synthesized (49) at the University of Georgia Core Facility (Athens, Ga.). The amount of antigen used for immunization was quantified by dry weight for the MAP-conjugated peptides and by bicinchoninic acid protein assay kit (Sigma Chemical Co., St. Louis, Mo.) for the purified gingipains.

Experimental animals. Female BALB/c mice 4 to 6 weeks of age (Charles River Laboratories) were used for these experiments. The total number of mice used for each group is indicated in Table 1. Approval to conduct these studies was obtained from the Animal Use Committee at Morehouse School of Medicine (Atlanta, Ga.). Mice were immunized by injection of each immunogen (50 µg/mouse in Freund's complete adjuvant) in subcutaneous chambers implanted in mice (13). Animals immunized with heat-killed P. gingivalis received an initial immunization corresponding to approximately 10⁸ CFU (as determined prior to heat inactivation). Control mice were immunized with Freund's adjuvant only. Seven days after primary immunization, mice were boosted (10 times) at 3-day intervals with the 95-kDa gingipain R1, the 50-kDa gingipain R2, or MAP-conjugated peptides (50 µg/mouse in Freund's incomplete adjuvant). Animals immunized with P. gingivalis were boosted (10 times) at 3-day intervals with heat-killed P. gingivalis corresponding to approximately 10² CFU (as determined prior to heat inactivation). At 14, 21, and 28 days postimmunization, chamber fluid was removed with a hypodermic needle and syringe to assess the IgG response to P. gingivalis and gingipains.

Chamber fluid analysis. Aliquots of fluid from each chamber (50 µl) were streaked for isolation onto anaerobic blood agar plates and cultured at 37°C for 7 days under anaerobic conditions (14). Results were expressed as the number of positive cultures obtained from chamber samples/total number of chambers sampled. Samples from individual mice were also pooled for each group and serially diluted. Samples were then plated on anaerobic blood agar plates and cultured at 37°C under anaerobic conditions for 7 days. Results are expressed as CFU/milliliter for each group.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Virulence of P. gingivalis in immunized mice. Female BALB/c mice were immunized with either the 95-kDa gingipain R1, the 50-kDa gingipain R2, or MAP-conjugated gingipain R-derived peptides (Fig. 1), by direct injection into stainless steel chambers implanted subcutaneously, and subsequently challenged by injection of live P. gingivalis into chambers (13). Nonimmunized animals, or animals immunized with a scrambled peptide A control and challenged with P. gingivalis, developed ulcerated necrotic lesions on their abdomens and exhibited severe cachexia with ruffled hair, hunched bodies, and weight loss; 14 of 22 and 5 of 8 mice, respectively, died (Table 1). In contrast, animals immunized with MAP-conjugated peptide A, corresponding to the N terminus of the catalytic domain of gingipains R (Fig. 1), followed by challenge with P. gingivalis were protected from abscess formation and death (Table 1). Similar results were obtained for animals that had been immunized with either whole P. gingivalis cells, the 95-kDa gingipain R1, or the 50-kDa gingipain R2. However, immunization with a peptide corresponding to sequences within the adhesion/hemagglutinin domain of gingipain R1 (peptide D) (Fig. 1), which has been shown to be directly involved in the hemagglutinin activity of this gingipain (5), followed by challenge with P. gingivalis, did not protect animals (Table 1). Similar results were obtained following challenge with P. gingivalis in mice immunized with peptides corresponding to either a sequence encompassing the catalytic cysteine residue of gingipains R (peptide B) or a homologous sequence within the catalytic domains of gingipains R and gingipain K (peptide C) (data not shown). Immunization with either peptide A, the 95-kDa gingipain R1, the 50-kDa gingipain R2, or P. gingivalis whole cells, followed by challenge with live bacteria, resulted in fewer animals from which P. gingivalis was cultured compared to the nonimmunized group. Following challenge with P. gingivalis, viable organisms were cultured from chamber fluid obtained from 20 of 22 nonimmunized mice up to the time of death (Table 2). Viable organisms were also cultured from animals immunized with peptides B, C, and D (Table 2 and data not shown). Following challenge with live bacteria, in the nonimmunized group, P. gingivalis levels increased relative to the initial inoculum (10⁸ to 10¹² CFU) throughout the course of the experiments (Table 2). In animals that were immunized with peptide A, peptide D, gingipain R1, gingipain R2, or whole cells, followed by challenge with P. gingivalis, we observed an initial increase in P. gingivalis CFU in those chambers that were positive for P. gingivalis; however, by day 7, the CFUs had decreased to <10⁶ (Table 2). These results indicate that immunization with a peptide corresponding to the N-terminal catalytic domain of gingipains R or with active proteinases or whole bacteria can limit the ability of P. gingivalis to invade in the mouse chamber model.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Structure of pro-gingipain R1, with indicated location of peptides used for immunization. The initial transcript of the rgp1 gene consists of a propeptide, a catalytic domain, and an adhesion/hemagglutinin domain (38). During translocation onto the P. gingivalis surface, the polyprotein undergoes proteolytic processing, resulting in the formation of the mature 95-kDa gingipain R1, either in membrane-bound or soluble forms consisting of a noncovalent complex of catalytic polypeptide and fragments of the adhesion/hemagglutinin domain (41). The adhesion/hemagglutinin domain is conveniently divided into subdomains (HGPs) of 44, 15, 17, and 27 kDa, according to putative proteolytic processing after one lysine and three arginine residues (arrowheads). The hemagglutination active site (peptide D) (5) is a part of a triplicate amino acid sequence repeat present in the HGP44, HGP17, and HGP27 subdomains. The triplicate repeats of the 50-amino-acid sequence within the adhesion/hemagglutinin domain are represented by hatched boxes numbered beneath the structure. Gingipain R2 is also translated as a proenzyme, nearly identical in sequence to the catalytic domain of gingipain R1 but missing the entire adhesion/hemagglutinin domain. The structure of the gingipain K initial polyprotein is similar to that of gingipain R1, with the adhesion/hemagglutinin domain being virtually identical. The gingipain K initial transcript is apparently subject to posttranslational processing by gingipains R (34). The catalytic domains of both gingipains share only limited identity (27%) scattered throughout the polypeptide chain, except for an identical 30-amino-acid residue fragment (peptide C). The cleavage of the propeptide releasing the N-terminal sequence of active gingipains R is shown by an arrow. Arrowheads indicate putative proteolytic processing sites leading to assembly of the soluble or membrane-bound enzyme (molecular mass 95 of kDa) in the form of a noncovalent complex of the catalytic domain with indicated, active fragments of the adhesion/hemagglutinin domain.

IgG response to P. gingivalis. Immunization with the N-terminal peptide induced a moderate IgG response to the 95-kDa gingipain R1 and the 50-kDa gingipain R2 as detected in chamber fluid samples (Table 3). The absence of a response to whole cells may be due to the lack of exposure of this epitope on cell surfaces so that the N-terminal sequence of the membrane-associated gingipain R1 catalytic domain is not available for antibody binding (41). The IgG response obtained following immunization with peptide fragment D, representing sequences within the adhesion/hemagglutinin domain of the 95-kDa gingipain R1, was comparable to that induced by the N-terminal peptide; however, protection against P. gingivalis challenge was not observed when this peptide was used as an immunogen (Tables 1 and 2). Immunization with the 95-kDa gingipain R1 induced a high IgG titer to all antigens examined except for the 50-kDa gingipain R2 (Table 3). The low titer to the 50-kDa gingipain R2 may be due to the absence of the highly immunogenic adhesion/hemagglutinin domain in this enzyme (24). Indeed, immunization with whole cells induced a good response to the 95-kDa gingipain R1 and gingipain K, with essentially no binding to the 50-kDa gingipain R2. For all immunization groups, postchallenge serum IgG titers were higher than the chamber fluid IgG titers obtained 3 weeks postimmunization but prior to challenge and reflected the effect of challenge with P. gingivalis (data not shown).

View larger version (50K): [in this window] [in a new window]   FIG. 2.   Competitive ELISA. Chamber fluid from mice immunized with heat-killed P. gingivalis was preincubated with increasing concentration of the 95-kDa gingipain R1 (light bars) and gingipain K (dark bars) as competing antigens before the mixture was added to a microtitration plate coated with whole P. gingivalis cells. The amount of antibody specifically bound to bacterial surface antigens was determined by subsequent binding of peroxidase-labeled goat anti-mouse IgG antibodies. When the 50-kDa gingipain R2 was used as the competing antigen, levels of residual IgG binding to P. gingivalis were as follows: 0.05 ng, 102.9%; 0.10 ng, 98.3%; 1.0 ng, 96.8%; 10 ng, 93.8%, 100 ng, 100.2%; 2,000 ng, 91.9%; and 10,000 ng, 96.8%.

Western blot analysis. Chamber fluid from mice immunized with the N-terminal peptide of the catalytic domain of gingipains R reacted with the catalytic domain (50 kDa) of gingipain R1, with gingipains R present in vesicles and bacterial membrane fractions, and with the 50-kDa gingipain R2 (Fig. 3A). A similar pattern was observed when chamber fluid from animals immunized with whole gingipain R2 was used (Fig. 3E). The lack of reactivity with gingipain K is in agreement with antibody specificity results (Table 3). Although the adhesion domain-derived peptide induced a poor IgG response as detected by ELISA, we found reactivity to several proteins by Western blot analysis (Fig. 3C). The 50-kDa gingipain R2 was not recognized by this antibody due to the lack of an adhesion domain. However, reactivity could be detected with the 27-kDa domain of the 95-kDa gingipain R1 and gingipain K, as well as proteins migrating in the range of 60 to 70 kDa in vesicle and membrane preparations. Significantly, the putative sequences of the adhesion domains present in the 44- and 17-kDa subunits (Fig. 1) did not bind antibody (Fig. 3C). Immunization with the 95-kDa gingipain R1 resulted in antibodies with specificity predominantly directed against the 44-kDa adhesion/hemagglutinin domain of gingipain R1 and the 42-kDa domain of gingipain K (Fig. 3B). These domains were also recognized in vesicle and membrane preparations. Additional protein bands recognized by this antiserum included the 32- and 17-kDa proteins in gingipain K, as well as their apparent equivalents in vesicles and membranes. However, the 95-kDa gingipain R1 catalytic domain was only weakly recognized, and the 50-kDa gingipain R2 was not recognized at all. These results are in agreement with previous studies in which the catalytic domains of gingipains R were poorly recognized in antisera obtained from rabbits or chickens immunized with the entire gingipain R1 molecule (12).

View larger version (32K): [in this window] [in a new window]   FIG. 3.   Western blot analysis of chamber fluid samples. Purified gingipains (95-kDa gingipain R1 [HRGP]), (50-kDa gingipain R2 [RGP-2]), and (gingipain K [KGP]) and samples of P. gingivalis vesicles and membranes (1.0 µg of purified gingipains or 5 µg of protein in vesicles and membrane fractions) were boiled, resolved by SDS-PAGE, and transferred to nitrocellulose. The nitrocellulose was transiently stained with Ponceau S; the positions of RGP-2 and polypeptide chains constituting the HRGP complex were then marked (dots to the right of an appropriate lane) and incubated in chamber fluid obtained from mice immunized with either the N-terminal peptide of the catalytic domain of RGPs (peptide A; 1,000-fold dilution) (A), HRGP (1,000-fold dilution) (B), the peptide derived from sequences within the adhesion/hemagglutinin domain of HRGP (peptide D; 100-fold dilution) (C), heat-killed P. gingivalis (1,000-fold dilution) (D), or RGP-2 (1,000-fold dilution) (E). Alkaline phosphatase-labeled goat anti-mouse IgG was then added, and blots were developed. Low-molecular weight (LMW) standards were loaded in the gels depicted in panels A and C. The molecular mass values for the markers from top to bottom are denoted by lines and correspond to 94, 67, 43, 30, 21, and 14 kDa.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Previous studies have established a role for gingipains R in the ability of P. gingivalis to colonize and invade in several animal models (11, 23, 31). In a previous study we found that inhibition of the enzymatic activity of the P. gingivalis cysteine proteinases including gingipains R severely limits the ability of P. gingivalis to colonize and invade in the mouse chamber model (12). A P. gingivalis rgpA rgpB double mutant has recently been shown to exhibit a marked reduction in virulence, indicating a major role for gingipains in the virulence potential of P. gingivalis (31). In the present study, we found that immunization of mice with the 95-kDa gingipain R1, the 50-kDa gingipain R2, or a peptide derived from the N-terminal sequence of the catalytic domain of gingipains R (peptide A) resulted in protection from P. gingivalis invasion. The protection observed following neutralization of these enzymes is most likely due to the ability of these enzymes to contribute to the virulence potential of P. gingivalis in a multifactorial manner. It is known that gingipains R may act as processing proteinases responsible for self-maturation as well as for maturation of gingipain K, fimbrillin, and a 75-kDa outer membrane protein (30, 35). Since fimbrillin, gingipains R, and gingipain K have been documented in the virulence potential of P. gingivalis (24, 25, 32, 50), gingipains R may participate in a central role in the pathogenesis of periodontal disease via production of pathophysiologically significant proteins.

The ability to modulate the colonization or pathogenicity of P. gingivalis has been demonstrated in several animal models using whole-cell and component vaccines. Whole-cell and polysaccharide protein conjugate vaccines have been demonstrated to protect against the invasive capabilities of P. gingivalis in mice (13, 47). Immunization of rats with purified P. gingivalis fimbriae and fimbrial peptides also has been shown to protect against periodontal destruction induced by infection with P. gingivalis in a rat periodontitis model (25). There is also evidence that immunization can influence the colonization of P. gingivalis in several primate models (3, 8, 19). However, in several animal models, induction of an immune response to P. gingivalis antigens has been demonstrated to actually exacerbate disease (reviewed in reference 26). Thus, it is critical that an immune response be generated to antigens which can function in a protective capacity. Indeed, it has been documented that although patients with a history of destructive periodontal disease frequently demonstrate an elevated IgG antibody response to P. gingivalis, these antibodies are apparently ineffective at limiting continued disease progression (9, 16, 26, 46, 51, 55).

The results reported in this study indicate that in mice the major IgG response is targeted to sequences within the adhesion/hemagglutinin domain of the 95-kDa gingipain R1 and gingipain K. It has recently been demonstrated that antibodies specific for P. gingivalis gingipains R were produced in patients with adult periodontitis, with the majority of the antibodies reactive with antigenic determinants within the hemagglutinin/adhesion domain of the 95-kDa gingipain R1 (22). We have also observed that patients with severe, untreated periodontitis mount a strong IgG response to this same domain (12). The strong reactivity against this part of gingipains may result from the fact that sequences within the hemagglutinin/adhesion domain are present in at least two other P. gingivalis proteins. These sequences are repeated four times in the hagA gene product (18) and once in a TonB-dependent protein encoded by the tla gene (1). It is possible that such a specific response to sequences within the adhesion/hemagglutinin domain of gingipains mounted in patients with periodontitis actually serve to divert the immune response from other protective antigens. In the mouse model, antibodies with this specificity may function to limit the invasion of P. gingivalis. However, in human subjects, where the local inflammatory response can lead to bone and tissue destruction within the periodontal ligament, the production of antibodies may function to aggravate local tissue damage.

Recent studies have demonstrated that the inhibition of hemagglutination in vivo by administration of a monoclonal antibody specific for the hemagglutinin domain of the prpRI gene (P. gingivalis W50 equivalent of the rgp1 gene) product results in the inhibition of P. gingivalis colonization (5). In the present study, we found that immunization with a peptide derived from sequences within the hemagglutinin domain did not induce a protective immunological response against P. gingivalis; however, the peptide used for our immunization studies did not contain a FEED motif. This sequence is present in several microbial adhesins and has been demonstrated to be directly responsible for fibronectin binding (36, 44, 52); the absence of the FEED motif in the peptide used in our studies could explain the lack of a protective response. Thus, in future studies it will be critical to examine peptides derived from other regions of the hemagglutinin domain.

Although immunization with the entire gingipains R induced an IgG response to the cell surface enzyme in intact bacteria, those antibodies generated following immunization with the N-terminal peptide were unable to recognize the mature protein in whole cells, suggesting that this epitope was not exposed in whole cells. Rabbit antisera generated to the N-terminal portion of the catalytic domain of the 95-kDa gingipain R1 and the 50-kDa gingipain R2 also did not recognize gingipain R1 in membrane or vesicle preparations unless samples were dissociated by boiling (41), again suggesting that this epitope is not exposed in whole cells or vesicles. The protective effect induced following immunization with the N-terminal peptide may result from IgG binding within the sequence of progingipain transiently expressed on the cell surface. We postulate that this binding interferes with the processing and or folding of progingipain and in this way hinders the pathogenic potency of P. gingivalis.

In summary, immunization with a MAP-conjugated peptide corresponding to the N-terminal domain of gingipains R results in the production of antibodies which afford protection of mice from P. gingivalis infection in the mouse chamber model. These findings support the hypothesis that inhibition of the maturation and/or catalytic activity of gingipains Rs can inhibit invasion of P. gingivalis in mice. Future studies are aimed at further defining specific domains of gingipains R which function in inducing a protective response in a periodontitis model of infection.