Structural Characterization of Peptide-Mediated Inhibition of Porphyro
http://www.100md.com
《感染与免疫杂志》
Department of Periodontics, Endodontics and Dental Hygiene, University of Louisville School of Dentistry, Louisville, Kentucky 40292
Department of Oral Biology, College of Dentistry, University of Florida, Gainesville, Florida 32610
ABSTRACT
Porphyromonas gingivalis is a periodontal pathogen whose primary niche is the anaerobic environment of subgingival dental plaque, but initial colonization of the oral cavity is likely to occur on supragingival surfaces that already support robust biofilm communities. Our studies have shown that P. gingivalis adheres to Streptococcus gordonii through interaction of the minor fimbrial antigen Mfa1 with a specific region of the streptococcal SspB polypeptide (residues 1167 to 1193) designated BAR. We show that a synthetic peptide comprising the BAR sequence potently inhibits P. gingivalis adherence to S. gordonii (50% inhibitory concentration = 1.3 μM) and prevents the development of P. gingivalis biofilms. However, a retroinverso peptide that possessed the same side chain topology as that of BAR was inactive, suggesting that interactions of Mfa1 with the peptide backbone of BAR are important for binding. A conformationally constrained analog of BAR inhibited P. gingivalis adherence and biofilm formation but at a lower specific activity than that of BAR. Therefore, to further define the structural features of the Mfa1-BAR interaction, we functionally screened combinatorial libraries of BAR in which active site residues (Asn1182, Thr1184, and Val1185) were replaced with each of the 19 common amino acids. Peptides containing positively charged amino acids at position 1182 or hydrophobic residues at position 1185 bound P. gingivalis more efficiently than did control peptides containing Asn and Val at these positions, suggesting that electrostatic and hydrophobic interactions may contribute to Mfa1-SspB binding. In contrast, replacement of Pro or Gly at these positions was detrimental to adherence, suggesting that perturbation of the BAR secondary structure influences activity. The net effect of substitutions for Thr1184 was less pronounced either positively or negatively than that at the other sites. These results define physicochemical characteristics of the interacting interface of Mfa1 and SspB and suggest that peptides or peptidomimetics with greater specific inhibitory activity than that of BAR can be developed. These compounds may represent potential therapeutics that target some of the first molecular interactions that allow P. gingivalis to colonize the oral cavity.
INTRODUCTION
Dental plaque is a complex and dynamic biofilm that accumulates through the sequential and ordered colonization of over 700 species of bacteria (13, 24, 25, 28). An oral biofilm comprised predominantly of gram-positive commensals such as the oral streptococci and Actinomyces spp. can exist in the oral cavity in the absence of overt disease (21, 25). However, populational shifts in the biofilm that lead to overrepresentation of acidophiles or of gram-negative obligate anaerobes may contribute to the onset and progression of the most common oral diseases, caries and periodontal disease (5). Indeed, adult periodontitis is associated with elevated levels of several gram-negative anaerobes in subgingival plaque, including the asaccharolytic, obligate anaerobe Porphyromonas gingivalis (26).
P. gingivalis is a secondary colonizer of the oral biofilm, and its primary niche is the anaerobic environment of subgingival plaque. In this environment, P. gingivalis interacts with a variety of other gram-negative obligate and facultative anaerobes, e.g., Fusobacterium nucleatum (11, 12), Treponema denticola (8, 9), and Tannerella forsythensis (29) via specific receptor-ligand interactions. However, the initial colonization of the oral cavity by P. gingivalis is likely to occur on supragingival surfaces that already support robust biofilm communities, and the successful colonization of this niche by P. gingivalis is contingent upon a variety of factors such as reduced oxygen tension and sufficient nutritional sources (1, 18). Consistent with this, P. gingivalis has been shown to adhere to primary colonizing organisms of supragingival surfaces such as Streptococcus gordonii (2, 6, 16).
The adhesion of P. gingivalis and S. gordonii is multimodal and involves at least two distinct receptor-ligand pairs. The long and short fimbriae of P. gingivalis have both been shown to be involved in this interaction (3). The structural subunit of the long fimbriae, FimA, has been shown to interact with cell surface glyceraldehyde-3-phosphate dehydrogenase of S. gordonii (19, 27), whereas the minor fimbrial protein Mfa1 interacts with streptococcal SspB (23), a cell surface protein in the antigen I/II family that is expressed by virtually all of the oral streptococci (10). Interestingly, neither intact P. gingivalis cells nor purified Mfa1 interacts with the antigen I/II protein of Streptococcus mutans, even though this protein is highly similar to SspB (2). This suggests that P. gingivalis may selectively colonize S. gordonii and the related oralis group streptococci over the mutans streptococci. In addition, Demuth et al. (6) and Cook et al. (4) showed that the Mfa1-SspB interaction is essential for the development of P. gingivalis biofilms on a streptococcal substrate and that biofilm growth exhibits the same selectivity for streptococcal species.
These initial colonization mechanisms utilized by P. gingivalis are potentially important targets for the development of therapeutic agents, since interfering with P. gingivalis adherence to S. gordonii may block the initial colonization of the supragingival biofilm by the organism and prevent it from reaching and multiplying in its primary niche in subgingival plaque. Our previous studies have focused on the mechanism of the Mfa1-SspB interaction, and we have identified a discrete region of the SspB protein that is essential for the interaction of purified Mfa1 and intact P. gingivalis cells with S. gordonii (6, 23). In this study, we show that a synthetic peptide encompassing the active site region of SspB (designated BAR) is a potent inhibitor (50% inhibitory concentration of approximately 1.3 μM) of P. gingivalis adherence to S. gordonii cells and blocks the formation of P. gingivalis biofilms. In addition, a combinatorial screening approach using BAR peptide libraries with substitutions at several active site amino acid residues identified specific characteristics of BAR that are required for P. gingivalis adherence.
MATERIALS AND METHODS
Growth of bacterial strains. P. gingivalis ATCC 33277 was grown in reduced Trypticase soy broth (Difco) supplemented with 0.5% yeast extract, 1 μg/ml (final concentration) menadione, and 5 μg/ml (final concentration) hemin. Twenty-five milliliters of medium was reduced for 24 h under anaerobic conditions by equilibration in an atmosphere consisting of 10% CO2, 10% H2, and 80% N2. Following equilibration, P. gingivalis was inoculated in the medium and grown for 48 h at 37°C under anaerobic conditions. S. gordonii DL-1 (22) was cultured aerobically without shaking in brain heart infusion broth supplemented with 1% yeast extract for 16 h at 37°C.
P. gingivalis biofilm formation with S. gordonii. The formation of P. gingivalis and S. gordonii biofilms was carried out essentially as previously described by Lamont et al. (15). Reactions were carried out using a Manostat Carter 4/8 cassette peristaltic pump (Fisher Scientific, Suwanee, GA) using 0.89-mm platinum-cured silicone tubing (Fisher Scientific, Suwanee, GA) and BST FC 71 flow cells (Biosurface Technologies Corp., Bozeman, MT). A single surface of a 15- x 40-mm cover glass (Fisher Scientific, Suwanee, GA) was coated with 0.22-μm filter-sterilized saliva and incubated at 37°C for 30 min. The saliva-coated cover glass was then washed with sterile 1x phosphate-buffered saline (PBS) at a flow rate of 6 ml per hour for 30 min.
S. gordonii DL-1 cells were harvested by centrifugation at 7,000 x g at 4°C and suspended in 10 ml of sterile 1x PBS. S. gordonii cells were labeled with 20 μl of hexidium iodide (1.6 mg/ml; Molecular Probes, Eugene, OR) at 25°C for 30 min in the dark and washed with PBS. To allow streptococci to attach to the saliva-coated cover glass, S. gordonii cells were delivered to the flow cell at the rate of 6 ml per hour for approximately 2 h. Following inoculation with S. gordonii, the flow cell was washed with sterile 1x PBS for 30 min at 6 ml per hour to remove nonadherent bacteria from the cover glass.
P. gingivalis cells were harvested by centrifugation, suspended in 10 ml of sterile 1x PBS, and introduced into the flow cell at a flow rate of 6 ml per hour for 2 h to allow P. gingivalis to adhere and accumulate on the streptococcal substrate. Flow cells were subsequently washed with sterile 1x PBS to remove nonadherent P. gingivalis cells. To visualize P. gingivalis, rabbit anti-P. gingivalis polyclonal antibody at 1:5,000 in 5 ml of sterile 1x PBS was made to flow into the cell at a rate of 6 ml/hour for approximately 1 h. The flow cell was washed with sterile 1x PBS for 1 h, reacted with anti-rabbit immunoglobulin G-fluorescein isothiocyanate (FITC) conjugate (Sigma, St. Louis, MO) in sterile 1x PBS (1:5,000) for 1 h at 6 ml/hour, and received a final wash with sterile 1x PBS as described above. Under these conditions, the streptococci did not form a confluent layer on the saliva-coated cover glass but were present in small clusters of cells. P. gingivalis subsequently formed distinct microcolonies on the immobilized streptococci which were quantified by confocal microscopy as described below.
P. gingivalis-S. gordonii biofilms were visualized using an Olympus Fluoview confocal laser scanning microscope (Olympus, Pittsburgh, PA) under 60x magnification using an argon laser for visualization of FITC labeling and the HeNe-G laser to visualize hexidium iodide-labeled streptococci. The number and height of FITC-labeled P. gingivalis microcolonies were determined from 30 to 60 randomly chosen frames using the FluoView software package provided by Olympus. Microcolony depth was determined by performing z-plane scans from 0 μm to 30 μm above the cover glass surface. P. gingivalis microcolonies that formed on S. gordonii in the absence of inhibitor ranged from 7 to 16 μm in depth under the experimental conditions used.
For biofilm inhibition experiments, BAR peptide was preincubated with P. gingivalis cells at concentrations of 0 to 10 μg per ml for 25°C for 30 min. The P. gingivalis-BAR suspension was introduced into the flow cell for 2 h at a flow rate of 6 ml per hour as described above. For some experiments using structural analogs of BAR, P. gingivalis cells were incubated with peptide concentrations of up to 50 μg per ml. P. gingivalis microcolonies that formed in the presence of inhibitor were analyzed by confocal microscopy as described above. Data were analyzed using GraphPad InStat3 (GraphPad Software Co.). A nonparametric analysis of variance using Dunn's multiple comparisons test was utilized to analyze the data acquired and determine pairwise statistical differences in colony number and depth between experimental samples and the control reaction mixture that did not contain inhibitor.
Synthesis of peptides and combinatorial peptide libraries. The peptides that were used in this study are shown in Fig. 1A. The BAR peptide, a retroinverso form of BAR, and a conformationally constrained analog of BAR were synthesized by BioSynthesis, Inc. (Lewisville, TX). All peptides were obtained at greater than 90% purity. The BAR peptide comprises amino acids 1167 to 1193 of the S. gordonii SspB polypeptide (2). The retroinverso analog of BAR (RI-BAR) is comprised of D-amino acids and was synthesized by reversing the chirality of the peptide bond backbone of BAR. The cyclic conformationally constrained analog of BAR (CR-BAR) was synthesized by substituting cysteine for leucines 1177 and 1191 of the SspB sequence, which flank the active site NITVK region of BAR that was previously identified (6).
Biased combinatorial libraries containing substitutions at each of the three potential active site residues of BAR are shown in Fig. 1B. Typically, a biased combinatorial is built around a lead compound (i.e., the BAR peptide) that has been characterized by previous structure-activity studies, and a limited number of positions within the lead compound sequence are randomized to generate a defined collection of structurally related peptides. Thus, a biased library facilitates the directed and rational analysis of a relatively large number of structurally related compounds and facilitates the rapid correlation of functional activity with structural properties of the lead compound. Within BAR, three residues were randomized, Asn1182, Thr1184, and Val1185. Asn1182 and Val1185 were shown by mutational analysis to be functionally relevant for P. gingivalis adherence and biofilm formation (6). In addition, secondary structure predictions indicated that these amino acids contribute to the local structure of BAR. Thr1184 was also randomized since it is present in the SspB sequence but not conserved in the functionally inactive homolog of SspB that is expressed by Streptococcus mutans (6).
P. gingivalis binding assay. The adherence of P. gingivalis to the BAR peptide analogs or to the combinatorial libraries was assayed as described by Chung et al. (3). Briefly, P. gingivalis cultures were centrifuged at 7,000 x g for 40 min, washed with reduced 1x PBS, and suspended in 5 ml of 0.1 M NaHCO3, pH 8.1. The bacterial cells were labeled with biotin (5 mg N-hydroxysuccinimidobiotin in 400 μl of dimethyl sulfoxide) and incubated at room temperature for 3 h. The biotinylated bacteria were harvested by centrifugation, washed twice with PBST (PBS supplemented with 0.05% Tween 20), and suspended in PBST to a density of 4 x 107 CFU/ml.
The BAR peptide analogs or samples of the combinatorial libraries were immobilized onto a nitrocellulose membrane using a dot blot apparatus, and the filter was blocked with 1.5% bovine serum albumin (in 1x PBS) for 2 h. For screening of the combinatorial libraries, 360 μg of each library mixture was immobilized onto the nitrocellulose filter. Thus, each spot contained approximately 1 μg of each peptide sequence that is present in the mixture. We previously showed that 1 μg of purified BAR yielded a readily detectable level of P. gingivalis adherence (6). A negative control consisted of PBST without peptide. The positive control was the peptide mixture from the combinatorial library that possessed the "z" amino acid that is present in BAR (e.g., N at residue 1182, T at residue 1184, or V at residue 1185), since only that mixture in the combinatorial library would contain the BAR peptide sequence. All adherence reactions were carried out in triplicate.
Membranes containing the immobilized combinatorial libraries were probed using biotinylated P. gingivalis for approximately 18 h at 37°C under anaerobic conditions and gently rinsed three times with PBST. Bound P. gingivalis was detected using streptavidin alkaline phosphatase conjugate (Bio-Rad) resuspended in PBST at a dilution of 1:5,000, followed by reaction with Sigma Fast BCIP/NBT (5-bromo-4-chloro-3-indolylphosphate/nitroblue tetrazolium)-buffered substrate (Sigma, St. Louis, MO). Digital images were generated from the resulting membranes using an Epson Stylus CX3810 flatbed scanner, and spot intensity was determined from the digital images using ImageJ software (NIH). The mean relative intensity of each experimental sample on the membrane was then normalized to the mean intensity of the positive-control spot (P. gingivalis adherence to the peptide mixture from the combinatorial library that possessed the "z" amino acid that is present in BAR) after subtraction of the mean background signal (density of the negative-control spot).
RESULTS
BAR inhibits formation of P. gingivalis biofilms. Our previous in vitro studies showed that P. gingivalis bound to a synthetic peptide (designated BAR; Fig. 1) comprising the adherence-promoting region of the SspB protein of S. gordonii (6, 15) and that this interaction is mediated by the short fimbrial protein Mfa1 of P. gingivalis (23). To determine if the BAR peptide could competitively inhibit P. gingivalis adherence to streptococcal cells and prevent biofilm development in an open flow system, P. gingivalis biofilms were formed on immobilized streptococci in the presence of increasing concentrations of BAR in flow cells. P. gingivalis microcolonies were then visualized by confocal scanning microscopy, and the number of microcolonies that formed per microscope field and the depth of each microcolony were recorded for each peptide concentration. As summarized in Table 1, an average of approximately six P. gingivalis microcolonies per frame were observed in the absence of BAR (control reaction). In contrast, preincubation of P. gingivalis with 10 μg per ml of BAR peptide almost completely eliminated formation of P. gingivalis microcolonies, suggesting that the peptide competitively inhibited the adherence of P. gingivalis cells to the immobilized streptococci. Furthermore, the depth of the P. gingivalis microcolonies that formed in the presence of the peptide inhibitor was significantly reduced (P < 0.001, Table 1) in the presence of 5 μg per ml and 10 μg per ml BAR. Thus, a soluble synthetic 27-mer peptide representing the adherence-promoting region of SspB effectively competes with the intact 170-kDa streptococcal cell surface protein for the P. gingivalis short fimbrial protein Mfa1.
Characterization of structural determinants required for P. gingivalis adherence to BAR. The specific inhibitory activity of BAR is likely influenced by a number of factors which, if controlled, might result in increased activity. For example, BAR contains five lysine residues and therefore may be highly susceptible to degradation by the Lys-gingipain expressed by P. gingivalis. We therefore synthesized a retroinverso analog of BAR, designated RI-BAR (Fig. 1), in which L-amino acids were replaced by D-amino acids and the direction of the peptide backbone of BAR was reversed. Thus, this peptide should possess the same side chain topology in solution as BAR does (7) but would likely be resistant to degradation by the P. gingivalis Lys-gingipain and potentially exhibit greater specific inhibitory activity than BAR does. However, our initial experiments using 1 to 10 μg of RI-BAR per ml as inhibitor had no effect on P. gingivalis biofilm formation (not shown). To determine if RI-BAR was active but at a lower specific activity than that of BAR, we repeated the experiments using higher concentrations of RI-BAR. As shown in Fig. 2, incubation of P. gingivalis cells with up to 50 μg per ml RI-BAR did not affect the development of microcolonies on the streptococcal substrate, indicating that RI-BAR is ineffective at inhibiting the adherence of P. gingivalis to immobilized streptococci, even at relatively high concentrations. These results suggest that native BAR, but not its retroinverso analog, can assume the functional conformation of SspB that is recognized by the P. gingivalis short fimbrial protein Mfa1.
Small peptides are structurally flexible in solution and are free to transiently assume many different conformations. To investigate more thoroughly the structural features that contribute to BAR inhibitory activity, a conformationally constrained analog of BAR and synthesized combinatorial BAR libraries was tested for biofilm-inhibitory activity. The conformationally constrained peptide (CR-BAR) maintained the chirality of BAR, but leucine residues at positions 1177 and 1191 were replaced with cysteine (Fig. 1). The resulting cyclic disulfide limits bond rotations within the peptide backbone of the intervening active site NITVK sequence that was identified by site-specific mutagenesis in our previous studies (6). As shown in Fig. 2, preincubation of P. gingivalis with CR-BAR resulted in a significant reduction in the number of P. gingivalis microcolonies that formed on the streptococcal substrate at a concentration of 50 μg per ml (P < 0.001). However, significant inhibition was not observed at 10 μg of CR-BAR per ml. Thus, the conformationally constrained peptide was active but exhibited lower specific inhibitory activity than the parent BAR peptide did.
Therefore, to better define the structural constraints of the NITVK region that are important for interaction with P. gingivalis, a combinatorial approach was undertaken to analyze BAR structure and function. To accomplish this, three combinatorial peptide libraries were synthesized in which one of the functionally active residues of NITVK (N1182, T1184, or V1185) was replaced by each of the 19 common amino acids (cysteine was excluded to prevent incorporation of disulfide bonds in the peptide mixtures; see Materials and Methods; Fig. 1). The remaining two amino acid positions were randomized by using equimolar mixtures of the 19 common amino acids (cysteine excluded) during synthesis. Thus, each of the combinatorial libraries comprised 19 distinct peptide mixtures that differed only in the amino acid residue that occupied the "z" position (Fig. 1). Each of the 19 peptide mixtures in a given library contained 361 different peptide sequences, arising from randomization of the two amino acids occupying the "x" positions. Thus, each library is comprised of 6,859 different peptide sequences.
The adherence of P. gingivalis cells to immobilized samples of each of the combinatorial peptide libraries was determined. As shown in Fig. 3A, several peptide mixtures in the combinatorial library designated N1182 bound to labeled P. gingivalis better than the positive-control mixture that possessed Asn at position 1182 (i.e., the residue present in BAR at this position). Interestingly, three of the amino acid substitutions at position 1182 that exhibited greater adherence activity were positively charged amino acid residues (Arg, Lys, and His), suggesting that a basic residue is optimal at this position in BAR. Substitution of Ser for Asn at 1182 also resulted in increased adherence of P. gingivalis. Screening of combinatorial library V1185 showed that three amino acid substitutions for Val resulted in increased adherence of P. gingivalis (Fig. 3C). Each of these residues was hydrophobic: isoleucine, phenylalanine, and tryptophan. This suggests that hydrophobic interactions may also in part drive the adherence of P. gingivalis to BAR. Indeed, the requirement for a hydrophobic amino acid in position 1185 appears to be quite pronounced, since replacement of most other amino acids at this position reduced P. gingivalis adherence (Fig. 3C). Finally, the overall effect of amino acid substitutions for Thr at position 1184 was less pronounced, either positively or negatively, than that observed at the other sites (compare y-axis scales of Fig. 3B with those of Fig. 3A and 3C). The three substitutions for Thr that resulted in greatest adherence were Arg, Lys, and Ser, similar to the results obtained with combinatorial library N1182.
For each of the combinatorial libraries, some amino acid substitutions resulted in decreased adherence activity and appear to be incompatible with the interaction of BAR with P. gingivalis. For example, substitution with Asp or Pro at any of the three positions (N1182, T1184, or V1185) was detrimental to P. gingivalis adherence. In addition, the presence of Gly at position 1182 or 1185 also reduced P. gingivalis adherence. These results are of interest since Gly and Pro are the amino acid residues that occupy positions 1182 and 1185 in the nonfunctional homolog of SspB that is expressed by S. mutans (6). Furthermore, substitution of Gly for Asn or Pro for Val by site-specific mutagenesis of the full-length SspB polypeptide resulted in a significant reduction of P. gingivalis adherence. Therefore, results from the combinatorial library screenings are consistent with previous structure-function data of the BAR-P. gingivalis interaction. Taken together, the outcome of the combinatorial screen of BAR (summarized in Fig. 4) suggests that both electrostatic and hydrophobic interactions may be required for efficient adherence of P. gingivalis to S. gordonii and consequently for formation of P. gingivalis biofilms. Our results also suggest that replacement of active residues with helix-breaking or -turn-inducing amino acids is detrimental for P. gingivalis adherence.
DISCUSSION
The human oral cavity presents a hostile environment to microorganisms as a result of the constant flow of saliva that contains numerous immune and nonimmune antimicrobial activities (20, 25). The early colonizing bacteria of the oral cavity that comprise supragingival plaque insulate themselves from these conditions by adhering tightly to oral tissues and/or to other oral organisms to form a multispecies biofilm (13, 28). The periodontal pathogen P. gingivalis faces an even greater challenge in colonizing the oral cavity since it is a fastidious obligate anaerobe that is asaccharolytic and acid sensitive and requires hemin for growth (17). Its primary niche is in the subgingival pocket, yet it must first colonize and survive the relatively aerobic environment of supragingival plaque. To accomplish this, our hypothesis is that P. gingivalis initially colonizes a specific niche in supragingival plaque by adhering tightly to organisms such as S. gordonii (16) and Fusobacterium nucleatum (1, 11) that may provide a reduced local environment that is not highly acidic. Since these specific interbacterial interactions may represent the first steps in P. gingivalis colonization, they are ideal targets for the development of therapeutic agents against adult periodontal disease.
Our previous studies have focused on the mechanism of P. gingivalis adherence to streptococci which leads to the subsequent development of biofilms on the streptococcal substrate. These studies showed that P. gingivalis adheres to S. gordonii cells by interaction of the short fimbrial subunit protein Mfa1 with a discrete region of the streptococcal SspB polypeptide (6, 23). Here we show that a synthetic peptide encompassing this region of SspB functions as a potent inhibitor of P. gingivalis adherence to S. gordonii and its subsequent formation of biofilms. From our results, we calculated that the number of P. gingivalis microcolonies that formed on the immobilized streptococcal substrate would be reduced by 50% in the presence of approximately 4 μg of peptide inhibitor per ml (50% inhibitory concentration, 1.3 μM). This value is consistent with the kinetics of adherence of purified Mfa1 protein to the BAR peptide as determined by enzyme-linked immunosorbent assay, where half-maximal binding occurred at approximately 0.8 μM BAR (23). This observation suggests that the interaction of Mfa1 with the BAR region of SspB is the primary driving force in P. gingivalis adherence to streptococci, which initiates the formation of biofilms.
Since P. gingivalis is a highly proteolytic organism, the specific inhibitory activity of BAR might be limited by proteolytic degradation. Indeed, based on its lysine content (5 of 27 residues), it is possible that the BAR peptide would be highly susceptible to degradation by the Lys-gingipain expressed by P. gingivalis. Since most proteases have not evolved to act on D-amino acid polypeptides (7), a retroinverso peptide that maintains the side chain topology of BAR might be expected to exhibit greater specific inhibitory activity. However, our results showed that a retroinverso derivative of BAR was inactive in preventing P. gingivalis adherence and formation of biofilms. A possible explanation for this result is that the association of Mfa1 with BAR requires crucial hydrogen bonding with the peptide backbone of BAR and reversing the chirality of the peptide bonds in BAR prevents these interactions from occurring. This has been previously suggested to represent a limitation of retroinverso peptidomimetics (7). We are currently exploring the possibility of selectively incorporating D-amino acids to limit proteolytic susceptibility of BAR while maintaining chirality in the remainder of the peptide backbone in order to facilitate potential interactions with Mfa1.
P. gingivalis adheres avidly to the SspB polypeptide of S. gordonii but does not interact with the highly related SpaP protein of S. mutans (2). Our previous studies showed that Asn1182, Thr1184, and Val1185 in BAR are not conserved in the SpaP sequence and that substitution of the corresponding S. mutans residues (Gly and Pro) for Asn1182 and Val1185, respectively, of SspB resulted in a nonfunctional S. gordonii protein (6). These results highlighted the functional role of the NITVK region of BAR in the interaction of P. gingivalis with SspB but were not informative about the physicochemical nature of the interaction of SspB and Mfa1. The biased combinatorial peptide libraries analyzed in the current study allowed us to examine the functional properties of BAR peptides containing all combinations of the common amino acids occupying the nonconserved residues (Asn1182, Thr1184, and Val1185) in the NITVK sequence. Interestingly, P. gingivalis efficiently bound to BAR-like peptides containing amino acid substitutions at both residues 1182 and 1185. The most active peptides were those containing a positively charged amino acid (Arg, Lys, or His) at position 1182 and a hydrophobic residue (Ile, Phe, or Trp) at 1185. This suggests that the SspB-Mfa1 interaction does not strictly depend on the specific amino acid occupying these positions in SspB but rather on the physical properties and characteristics of the amino acid residue.
Since the most active BAR peptide analogs contained a hydrophobic amino acid (Ile, Phe, Trp, or Val) at position 1185, we can speculate that the interacting interface of the P. gingivalis fimbrial protein Mfa1 likely contains a large hydrophobic pocket that can accommodate the bulky side chains of these residues. The fact that no other amino acids were functionally tolerated at this site in our combinatorial screen suggests that hydrophobic interactions may be essential for the association of Mfa1 with SspB. In addition, the presence of positively charged amino acids at position 1182 (Arg, Lys, or His) facilitated P. gingivalis adherence, whereas the replacement of N1182 with acidic residues (Asp and, to a lesser extent, Glu) reduced adherence. The preference for basic residues could be explained by the presence of negatively charged residues in the SspB binding pocket of Mfa1 that might contribute to binding through the formation of electrostatic interactions with SspB. However, the Mfa1-SspB interaction is clearly more complex than suggested above since the polar, but uncharged, residues Asn and Ser at position 1182 also promoted P. gingivalis adherence. While the exact structure of the interacting interface cannot be described in full without additional structural studies, our results suggest that both electrostatic and hydrophobic bonding may drive the protein-protein interaction of Mfa1 with SspB.
It is also of interest that several amino acids (e.g., Pro, Gly, and Asp) appeared to be generally detrimental to P. gingivalis adherence regardless of position in the NITVK sequence. The presence of Pro at any of the randomized positions in BAR, and Gly at residue 1182 and to a lesser extent at residue 1184, resulted in reduced adherence. In addition, substitution of Gly and Pro for N1182 and V1185 reduced adherence. This is in agreement with our previous site-specific mutagenesis data (6). Gly and Pro are both -turn-inducing or helix-breaking residues, and it is likely that their presence in BAR alters the structure of this region. Secondary structural predictions of the BAR peptide analogs are consistent with this conclusion. For example, residues 1167 to 1180 of BAR are predicted to be -helical, with the remainder of the peptide predicted to be -sheet. In contrast, structural predictions of a BAR analog containing Gly and Pro at positions 1182 and 1185, respectively, differ from BAR in that residues 1167 to 1180 are predicted to be -helical as described above, but this helix is followed by a -turn (comprising residues 1181 to 1187) and then a second -helix. The predicted turn in this peptide analog arises from the presence of three helix-interrupting amino acids Gly1182, Pro1185, and Gly1187 in the NITVK region. Thus, at least some of the amino acid substitutions that reduce P. gingivalis adherence may alter secondary structural elements of BAR, suggesting that the overall structure of the BAR region may also be important for recognition of SspB by Mfa1.
In summary, we have shown that the formation of P. gingivalis biofilms can be blocked by peptide-mediated inhibition of the interaction of Mfa1 with the streptococcal SspB protein. Analysis of peptide analogs and combinatorial peptide libraries based on the sequence of BAR suggests that both conformational determinants and the characteristics of specific amino acid residues in BAR are essential for this interaction to occur. These studies highlight molecular details of the interacting regions of the streptococcal SspB protein and the P. gingivalis minor fimbrial subunit Mfa1 and define physicochemical characteristics of the binding domains. These results suggest that peptides or peptidomimetics with greater specific inhibitory activity than that of BAR can be synthesized.
ACKNOWLEDGMENTS
We thank Masae Kuboniwa for her assistance with the P. gingivalis adherence assay and Douglas Darling for his help with the statistical analyses.
This study was supported by Public Health Service grant DE12505.
FOOTNOTES
Corresponding author. Mailing address: Department of Periodontics, Endodontics and Dental Hygiene, Research Group in Oral Health and Systemic Disease, University of Louisville School of Dentistry, 501 S. Preston St., Room 209, Louisville, KY 40292. Phone: (502) 852-3807. Fax: (502) 852-4052. E-mail: drdemu01@louisville.edu.
REFERENCES
1. Bradshaw, D. J., P. D. Marsh, G. K. Watson, and C. Allison. 1998. Role of Fusobacterium nucleatum and coaggregation in anaerobe survival in planktonic and biofilm oral microbial communities during aeration. Infect. Immun. 66:4729-4732.
2. Brooks, W., D. R. Demuth, S. Gil, and R. J. Lamont. 1997. Identification of a Streptococcus gordonii SspB domain that mediates adhesion to Porphyromonas gingivalis. Infect. Immun. 65:3753-3758.
3. Chung, W. O., D. R. Demuth, and R. J. Lamont. 2000. Identification of a Porphyromonas gingivalis receptor for the Streptococcus gordonii SspB protein. Infect. Immun. 68:6758-6762.
4. Cook, G. S., J. W. Costerton, and R. J. Lamont. 1998. Biofilm formation by Porphyromonas gingivalis and Streptococcus gordonii. J. Periodont. Res. 33:323-327.
5. Dalwai, F., B. J. Spratt, and J. Pratten. 2006. Modeling shifts in microbial populations in health and disease. Appl. Environ. Microbiol. 72:3678-3684.
6. Demuth, D. R., D. C. Irvine, J. W. Costerton, G. S. Cook, and R. J. Lamont. 2001. Discrete protein determinant directs the species-specific adherence of Porphyromonas gingivalis to oral streptococci. Infect. Immun. 69:5736-5741.
7. Fischer, P. M. 2003. The design, synthesis and application of stereochemical and directional peptide isomers: a critical review. Curr. Protein Pept. Sci. 4:339-356.
8. Grenier, D. 1992. Demonstration of bimodal coaggregation reaction between Porphyromonas gingivalis and Treponema denticola. Oral Microbiol. Immunol. 7:280-284.
9. Ikegami, A., K. Honma, A. Sharma, and H. K. Kuramitsu. 2004. Multiple functions of the leucine-rich repeat protein LrrA of Treponema denticola. Infect. Immun. 72:4619-4627.
10. Jenkinson, H. F., and D. R. Demuth. 1997. Structure, function and immunogenicity of streptococcal antigen I/II polypeptides. Mol. Microbiol. 23:183-190.
11. Kinder, S. A., and S. C. Holt. 1989. Characterization of coaggregation between Bacteroides gingivalis T22 and Fusobacterium nucleatum T18. Infect. Immun. 57:3425-3433.
12. Kolenbrander, P. E., and R. N. Andersen. 1989. Inhibition of coaggregation between Fusobacterium nucleatum and Porphyromonas (Bacteroides) gingivalis by lactose and related sugars. Infect. Immun. 57:3204-3209.
13. Kolenbrander, P. E., and J. London. 1993. Adhere today, here tomorrow: oral bacterial adherence. J. Bacteriol. 175:3247-3252.
14. Kuboniwa, M., G. D. Tribble, C. E. James, A. O. Kilic, L. Tao, M. C. Herzberg, S. Shizukuishi, and R. J. Lamont. 2006. Streptococcus gordonii utilizes several distinct gene functions to recruit Porphyromonas gingivalis into a mixed community. Mol. Microbiol. 60:121-139.
15. Lamont, R. J., A. El-Sabaeny, Y. Park, G. S. Cook, J. W. Costerton, and D. R. Demuth. 2002. Role of Streptococcus gordonii SspB protein in the development of Porphyromonas gingivalis biofilms on streptococcal substrates. Microbiology 148:1627-1636.
16. Lamont, R. J., S. G. Hersey, and B. Rosan. 1992. Characterization of the adherence of Porphyromonas (Bacteroides) gingivalis to oral streptococci. Oral Microbiol. Immunol. 7:193-197.
17. Lamont, R. J., and H. F. Jenkinson. 1998. Life below the gum line: pathogenic mechanisms of Porphyromonas gingivalis. Microbiol. Mol. Biol. Rev. 62:1244-1263.
18. Lamont, R. J., and H. F. Jenkinson. 2000. Subgingival colonization by Porphyromonas gingivalis. Oral Microbiol. Immunol. 15:341-349.
19. Maeda, K., H. Nagata, M. Kuboniwa, K. Kataoka, N. Nishida, M. Tanaka, and S. Shizukuishi. 2004. Characterization of binding of Streptococcus oralis glyceraldehyde 3-phosphate dehydrogenase to Porphyromonas gingivalis major fimbriae. Infect. Immun. 72:5475-5477.
20. Marsh, P. D. 1989. Host defenses and microbial homeostasis: role of microbial interactions. J. Dent. Res. 68:1567-1575.
21. Marsh, P. D. 1994. Microbial ecology of dental plaque and its significance in health and disease. Adv. Dent. Res. 8:263-271.
22. Pakula, R., and W. Walczak. 1963. On the nature of competence and transformable streptococci. J. Gen. Microbiol. 31:125-133.
23. Park, Y., M. R. Simionato, K. Sekiya, Y. Murakami, D. M. James, W. Chen, M. Hackett, F. Yoshimura, D. R. Demuth, and R. J. Lamont. 2005. Short fimbriae of Porphyromonas gingivalis and their role in coadhesion with Streptococcus gordonii. Infect. Immun. 73:3983-3989.
24. Paster, B. J., S. K. Boches, J. L. Galvin, R. E. Ericson, C. N. Lau, V. A. Levanos, A. Sahasrabudie, and F. E. Dewhirst. 2001. Bacterial diversity in human subgingival plaque. J. Bacteriol. 183:3770-3783.
25. Rosan, B., and R. J. Lamont. 2000. Dental plaque formation. Microbes Infect. 2:1599-1607.
26. Socransky, S. S., and A. D. Haffajee. 1992. The bacterial etiology of destructive periodontal disease: current concepts. J. Periodontol. 63:322-331.
27. Sohar, H. T., and R. J. Genco. 2005. Identification of glyceraldehyde 3-phosphate dehydrogenase of epithelial cells as a second molecule that binds to Porphyromonas gingivalis fimbriae. FEMS Immunol. Med. Microbiol. 45:25-30.
28. Whittaker, C. J., C. M. Klier, and P. E. Kolenbrander. 1996. Mechanisms of adhesion by oral bacteria. Annu. Rev. Microbiol. 50:513-552.
29. Yao, E. S., Lamont, R. J., S. P. Leu, and A. Weinberg. 1996. Interbacterial binding among pathogenic and commensal oral bacterial species. Oral Microbiol. Immunol. 11:35-41.(Carlo Amorin Daep, DeAnna M. James, Rich)
Department of Oral Biology, College of Dentistry, University of Florida, Gainesville, Florida 32610
ABSTRACT
Porphyromonas gingivalis is a periodontal pathogen whose primary niche is the anaerobic environment of subgingival dental plaque, but initial colonization of the oral cavity is likely to occur on supragingival surfaces that already support robust biofilm communities. Our studies have shown that P. gingivalis adheres to Streptococcus gordonii through interaction of the minor fimbrial antigen Mfa1 with a specific region of the streptococcal SspB polypeptide (residues 1167 to 1193) designated BAR. We show that a synthetic peptide comprising the BAR sequence potently inhibits P. gingivalis adherence to S. gordonii (50% inhibitory concentration = 1.3 μM) and prevents the development of P. gingivalis biofilms. However, a retroinverso peptide that possessed the same side chain topology as that of BAR was inactive, suggesting that interactions of Mfa1 with the peptide backbone of BAR are important for binding. A conformationally constrained analog of BAR inhibited P. gingivalis adherence and biofilm formation but at a lower specific activity than that of BAR. Therefore, to further define the structural features of the Mfa1-BAR interaction, we functionally screened combinatorial libraries of BAR in which active site residues (Asn1182, Thr1184, and Val1185) were replaced with each of the 19 common amino acids. Peptides containing positively charged amino acids at position 1182 or hydrophobic residues at position 1185 bound P. gingivalis more efficiently than did control peptides containing Asn and Val at these positions, suggesting that electrostatic and hydrophobic interactions may contribute to Mfa1-SspB binding. In contrast, replacement of Pro or Gly at these positions was detrimental to adherence, suggesting that perturbation of the BAR secondary structure influences activity. The net effect of substitutions for Thr1184 was less pronounced either positively or negatively than that at the other sites. These results define physicochemical characteristics of the interacting interface of Mfa1 and SspB and suggest that peptides or peptidomimetics with greater specific inhibitory activity than that of BAR can be developed. These compounds may represent potential therapeutics that target some of the first molecular interactions that allow P. gingivalis to colonize the oral cavity.
INTRODUCTION
Dental plaque is a complex and dynamic biofilm that accumulates through the sequential and ordered colonization of over 700 species of bacteria (13, 24, 25, 28). An oral biofilm comprised predominantly of gram-positive commensals such as the oral streptococci and Actinomyces spp. can exist in the oral cavity in the absence of overt disease (21, 25). However, populational shifts in the biofilm that lead to overrepresentation of acidophiles or of gram-negative obligate anaerobes may contribute to the onset and progression of the most common oral diseases, caries and periodontal disease (5). Indeed, adult periodontitis is associated with elevated levels of several gram-negative anaerobes in subgingival plaque, including the asaccharolytic, obligate anaerobe Porphyromonas gingivalis (26).
P. gingivalis is a secondary colonizer of the oral biofilm, and its primary niche is the anaerobic environment of subgingival plaque. In this environment, P. gingivalis interacts with a variety of other gram-negative obligate and facultative anaerobes, e.g., Fusobacterium nucleatum (11, 12), Treponema denticola (8, 9), and Tannerella forsythensis (29) via specific receptor-ligand interactions. However, the initial colonization of the oral cavity by P. gingivalis is likely to occur on supragingival surfaces that already support robust biofilm communities, and the successful colonization of this niche by P. gingivalis is contingent upon a variety of factors such as reduced oxygen tension and sufficient nutritional sources (1, 18). Consistent with this, P. gingivalis has been shown to adhere to primary colonizing organisms of supragingival surfaces such as Streptococcus gordonii (2, 6, 16).
The adhesion of P. gingivalis and S. gordonii is multimodal and involves at least two distinct receptor-ligand pairs. The long and short fimbriae of P. gingivalis have both been shown to be involved in this interaction (3). The structural subunit of the long fimbriae, FimA, has been shown to interact with cell surface glyceraldehyde-3-phosphate dehydrogenase of S. gordonii (19, 27), whereas the minor fimbrial protein Mfa1 interacts with streptococcal SspB (23), a cell surface protein in the antigen I/II family that is expressed by virtually all of the oral streptococci (10). Interestingly, neither intact P. gingivalis cells nor purified Mfa1 interacts with the antigen I/II protein of Streptococcus mutans, even though this protein is highly similar to SspB (2). This suggests that P. gingivalis may selectively colonize S. gordonii and the related oralis group streptococci over the mutans streptococci. In addition, Demuth et al. (6) and Cook et al. (4) showed that the Mfa1-SspB interaction is essential for the development of P. gingivalis biofilms on a streptococcal substrate and that biofilm growth exhibits the same selectivity for streptococcal species.
These initial colonization mechanisms utilized by P. gingivalis are potentially important targets for the development of therapeutic agents, since interfering with P. gingivalis adherence to S. gordonii may block the initial colonization of the supragingival biofilm by the organism and prevent it from reaching and multiplying in its primary niche in subgingival plaque. Our previous studies have focused on the mechanism of the Mfa1-SspB interaction, and we have identified a discrete region of the SspB protein that is essential for the interaction of purified Mfa1 and intact P. gingivalis cells with S. gordonii (6, 23). In this study, we show that a synthetic peptide encompassing the active site region of SspB (designated BAR) is a potent inhibitor (50% inhibitory concentration of approximately 1.3 μM) of P. gingivalis adherence to S. gordonii cells and blocks the formation of P. gingivalis biofilms. In addition, a combinatorial screening approach using BAR peptide libraries with substitutions at several active site amino acid residues identified specific characteristics of BAR that are required for P. gingivalis adherence.
MATERIALS AND METHODS
Growth of bacterial strains. P. gingivalis ATCC 33277 was grown in reduced Trypticase soy broth (Difco) supplemented with 0.5% yeast extract, 1 μg/ml (final concentration) menadione, and 5 μg/ml (final concentration) hemin. Twenty-five milliliters of medium was reduced for 24 h under anaerobic conditions by equilibration in an atmosphere consisting of 10% CO2, 10% H2, and 80% N2. Following equilibration, P. gingivalis was inoculated in the medium and grown for 48 h at 37°C under anaerobic conditions. S. gordonii DL-1 (22) was cultured aerobically without shaking in brain heart infusion broth supplemented with 1% yeast extract for 16 h at 37°C.
P. gingivalis biofilm formation with S. gordonii. The formation of P. gingivalis and S. gordonii biofilms was carried out essentially as previously described by Lamont et al. (15). Reactions were carried out using a Manostat Carter 4/8 cassette peristaltic pump (Fisher Scientific, Suwanee, GA) using 0.89-mm platinum-cured silicone tubing (Fisher Scientific, Suwanee, GA) and BST FC 71 flow cells (Biosurface Technologies Corp., Bozeman, MT). A single surface of a 15- x 40-mm cover glass (Fisher Scientific, Suwanee, GA) was coated with 0.22-μm filter-sterilized saliva and incubated at 37°C for 30 min. The saliva-coated cover glass was then washed with sterile 1x phosphate-buffered saline (PBS) at a flow rate of 6 ml per hour for 30 min.
S. gordonii DL-1 cells were harvested by centrifugation at 7,000 x g at 4°C and suspended in 10 ml of sterile 1x PBS. S. gordonii cells were labeled with 20 μl of hexidium iodide (1.6 mg/ml; Molecular Probes, Eugene, OR) at 25°C for 30 min in the dark and washed with PBS. To allow streptococci to attach to the saliva-coated cover glass, S. gordonii cells were delivered to the flow cell at the rate of 6 ml per hour for approximately 2 h. Following inoculation with S. gordonii, the flow cell was washed with sterile 1x PBS for 30 min at 6 ml per hour to remove nonadherent bacteria from the cover glass.
P. gingivalis cells were harvested by centrifugation, suspended in 10 ml of sterile 1x PBS, and introduced into the flow cell at a flow rate of 6 ml per hour for 2 h to allow P. gingivalis to adhere and accumulate on the streptococcal substrate. Flow cells were subsequently washed with sterile 1x PBS to remove nonadherent P. gingivalis cells. To visualize P. gingivalis, rabbit anti-P. gingivalis polyclonal antibody at 1:5,000 in 5 ml of sterile 1x PBS was made to flow into the cell at a rate of 6 ml/hour for approximately 1 h. The flow cell was washed with sterile 1x PBS for 1 h, reacted with anti-rabbit immunoglobulin G-fluorescein isothiocyanate (FITC) conjugate (Sigma, St. Louis, MO) in sterile 1x PBS (1:5,000) for 1 h at 6 ml/hour, and received a final wash with sterile 1x PBS as described above. Under these conditions, the streptococci did not form a confluent layer on the saliva-coated cover glass but were present in small clusters of cells. P. gingivalis subsequently formed distinct microcolonies on the immobilized streptococci which were quantified by confocal microscopy as described below.
P. gingivalis-S. gordonii biofilms were visualized using an Olympus Fluoview confocal laser scanning microscope (Olympus, Pittsburgh, PA) under 60x magnification using an argon laser for visualization of FITC labeling and the HeNe-G laser to visualize hexidium iodide-labeled streptococci. The number and height of FITC-labeled P. gingivalis microcolonies were determined from 30 to 60 randomly chosen frames using the FluoView software package provided by Olympus. Microcolony depth was determined by performing z-plane scans from 0 μm to 30 μm above the cover glass surface. P. gingivalis microcolonies that formed on S. gordonii in the absence of inhibitor ranged from 7 to 16 μm in depth under the experimental conditions used.
For biofilm inhibition experiments, BAR peptide was preincubated with P. gingivalis cells at concentrations of 0 to 10 μg per ml for 25°C for 30 min. The P. gingivalis-BAR suspension was introduced into the flow cell for 2 h at a flow rate of 6 ml per hour as described above. For some experiments using structural analogs of BAR, P. gingivalis cells were incubated with peptide concentrations of up to 50 μg per ml. P. gingivalis microcolonies that formed in the presence of inhibitor were analyzed by confocal microscopy as described above. Data were analyzed using GraphPad InStat3 (GraphPad Software Co.). A nonparametric analysis of variance using Dunn's multiple comparisons test was utilized to analyze the data acquired and determine pairwise statistical differences in colony number and depth between experimental samples and the control reaction mixture that did not contain inhibitor.
Synthesis of peptides and combinatorial peptide libraries. The peptides that were used in this study are shown in Fig. 1A. The BAR peptide, a retroinverso form of BAR, and a conformationally constrained analog of BAR were synthesized by BioSynthesis, Inc. (Lewisville, TX). All peptides were obtained at greater than 90% purity. The BAR peptide comprises amino acids 1167 to 1193 of the S. gordonii SspB polypeptide (2). The retroinverso analog of BAR (RI-BAR) is comprised of D-amino acids and was synthesized by reversing the chirality of the peptide bond backbone of BAR. The cyclic conformationally constrained analog of BAR (CR-BAR) was synthesized by substituting cysteine for leucines 1177 and 1191 of the SspB sequence, which flank the active site NITVK region of BAR that was previously identified (6).
Biased combinatorial libraries containing substitutions at each of the three potential active site residues of BAR are shown in Fig. 1B. Typically, a biased combinatorial is built around a lead compound (i.e., the BAR peptide) that has been characterized by previous structure-activity studies, and a limited number of positions within the lead compound sequence are randomized to generate a defined collection of structurally related peptides. Thus, a biased library facilitates the directed and rational analysis of a relatively large number of structurally related compounds and facilitates the rapid correlation of functional activity with structural properties of the lead compound. Within BAR, three residues were randomized, Asn1182, Thr1184, and Val1185. Asn1182 and Val1185 were shown by mutational analysis to be functionally relevant for P. gingivalis adherence and biofilm formation (6). In addition, secondary structure predictions indicated that these amino acids contribute to the local structure of BAR. Thr1184 was also randomized since it is present in the SspB sequence but not conserved in the functionally inactive homolog of SspB that is expressed by Streptococcus mutans (6).
P. gingivalis binding assay. The adherence of P. gingivalis to the BAR peptide analogs or to the combinatorial libraries was assayed as described by Chung et al. (3). Briefly, P. gingivalis cultures were centrifuged at 7,000 x g for 40 min, washed with reduced 1x PBS, and suspended in 5 ml of 0.1 M NaHCO3, pH 8.1. The bacterial cells were labeled with biotin (5 mg N-hydroxysuccinimidobiotin in 400 μl of dimethyl sulfoxide) and incubated at room temperature for 3 h. The biotinylated bacteria were harvested by centrifugation, washed twice with PBST (PBS supplemented with 0.05% Tween 20), and suspended in PBST to a density of 4 x 107 CFU/ml.
The BAR peptide analogs or samples of the combinatorial libraries were immobilized onto a nitrocellulose membrane using a dot blot apparatus, and the filter was blocked with 1.5% bovine serum albumin (in 1x PBS) for 2 h. For screening of the combinatorial libraries, 360 μg of each library mixture was immobilized onto the nitrocellulose filter. Thus, each spot contained approximately 1 μg of each peptide sequence that is present in the mixture. We previously showed that 1 μg of purified BAR yielded a readily detectable level of P. gingivalis adherence (6). A negative control consisted of PBST without peptide. The positive control was the peptide mixture from the combinatorial library that possessed the "z" amino acid that is present in BAR (e.g., N at residue 1182, T at residue 1184, or V at residue 1185), since only that mixture in the combinatorial library would contain the BAR peptide sequence. All adherence reactions were carried out in triplicate.
Membranes containing the immobilized combinatorial libraries were probed using biotinylated P. gingivalis for approximately 18 h at 37°C under anaerobic conditions and gently rinsed three times with PBST. Bound P. gingivalis was detected using streptavidin alkaline phosphatase conjugate (Bio-Rad) resuspended in PBST at a dilution of 1:5,000, followed by reaction with Sigma Fast BCIP/NBT (5-bromo-4-chloro-3-indolylphosphate/nitroblue tetrazolium)-buffered substrate (Sigma, St. Louis, MO). Digital images were generated from the resulting membranes using an Epson Stylus CX3810 flatbed scanner, and spot intensity was determined from the digital images using ImageJ software (NIH). The mean relative intensity of each experimental sample on the membrane was then normalized to the mean intensity of the positive-control spot (P. gingivalis adherence to the peptide mixture from the combinatorial library that possessed the "z" amino acid that is present in BAR) after subtraction of the mean background signal (density of the negative-control spot).
RESULTS
BAR inhibits formation of P. gingivalis biofilms. Our previous in vitro studies showed that P. gingivalis bound to a synthetic peptide (designated BAR; Fig. 1) comprising the adherence-promoting region of the SspB protein of S. gordonii (6, 15) and that this interaction is mediated by the short fimbrial protein Mfa1 of P. gingivalis (23). To determine if the BAR peptide could competitively inhibit P. gingivalis adherence to streptococcal cells and prevent biofilm development in an open flow system, P. gingivalis biofilms were formed on immobilized streptococci in the presence of increasing concentrations of BAR in flow cells. P. gingivalis microcolonies were then visualized by confocal scanning microscopy, and the number of microcolonies that formed per microscope field and the depth of each microcolony were recorded for each peptide concentration. As summarized in Table 1, an average of approximately six P. gingivalis microcolonies per frame were observed in the absence of BAR (control reaction). In contrast, preincubation of P. gingivalis with 10 μg per ml of BAR peptide almost completely eliminated formation of P. gingivalis microcolonies, suggesting that the peptide competitively inhibited the adherence of P. gingivalis cells to the immobilized streptococci. Furthermore, the depth of the P. gingivalis microcolonies that formed in the presence of the peptide inhibitor was significantly reduced (P < 0.001, Table 1) in the presence of 5 μg per ml and 10 μg per ml BAR. Thus, a soluble synthetic 27-mer peptide representing the adherence-promoting region of SspB effectively competes with the intact 170-kDa streptococcal cell surface protein for the P. gingivalis short fimbrial protein Mfa1.
Characterization of structural determinants required for P. gingivalis adherence to BAR. The specific inhibitory activity of BAR is likely influenced by a number of factors which, if controlled, might result in increased activity. For example, BAR contains five lysine residues and therefore may be highly susceptible to degradation by the Lys-gingipain expressed by P. gingivalis. We therefore synthesized a retroinverso analog of BAR, designated RI-BAR (Fig. 1), in which L-amino acids were replaced by D-amino acids and the direction of the peptide backbone of BAR was reversed. Thus, this peptide should possess the same side chain topology in solution as BAR does (7) but would likely be resistant to degradation by the P. gingivalis Lys-gingipain and potentially exhibit greater specific inhibitory activity than BAR does. However, our initial experiments using 1 to 10 μg of RI-BAR per ml as inhibitor had no effect on P. gingivalis biofilm formation (not shown). To determine if RI-BAR was active but at a lower specific activity than that of BAR, we repeated the experiments using higher concentrations of RI-BAR. As shown in Fig. 2, incubation of P. gingivalis cells with up to 50 μg per ml RI-BAR did not affect the development of microcolonies on the streptococcal substrate, indicating that RI-BAR is ineffective at inhibiting the adherence of P. gingivalis to immobilized streptococci, even at relatively high concentrations. These results suggest that native BAR, but not its retroinverso analog, can assume the functional conformation of SspB that is recognized by the P. gingivalis short fimbrial protein Mfa1.
Small peptides are structurally flexible in solution and are free to transiently assume many different conformations. To investigate more thoroughly the structural features that contribute to BAR inhibitory activity, a conformationally constrained analog of BAR and synthesized combinatorial BAR libraries was tested for biofilm-inhibitory activity. The conformationally constrained peptide (CR-BAR) maintained the chirality of BAR, but leucine residues at positions 1177 and 1191 were replaced with cysteine (Fig. 1). The resulting cyclic disulfide limits bond rotations within the peptide backbone of the intervening active site NITVK sequence that was identified by site-specific mutagenesis in our previous studies (6). As shown in Fig. 2, preincubation of P. gingivalis with CR-BAR resulted in a significant reduction in the number of P. gingivalis microcolonies that formed on the streptococcal substrate at a concentration of 50 μg per ml (P < 0.001). However, significant inhibition was not observed at 10 μg of CR-BAR per ml. Thus, the conformationally constrained peptide was active but exhibited lower specific inhibitory activity than the parent BAR peptide did.
Therefore, to better define the structural constraints of the NITVK region that are important for interaction with P. gingivalis, a combinatorial approach was undertaken to analyze BAR structure and function. To accomplish this, three combinatorial peptide libraries were synthesized in which one of the functionally active residues of NITVK (N1182, T1184, or V1185) was replaced by each of the 19 common amino acids (cysteine was excluded to prevent incorporation of disulfide bonds in the peptide mixtures; see Materials and Methods; Fig. 1). The remaining two amino acid positions were randomized by using equimolar mixtures of the 19 common amino acids (cysteine excluded) during synthesis. Thus, each of the combinatorial libraries comprised 19 distinct peptide mixtures that differed only in the amino acid residue that occupied the "z" position (Fig. 1). Each of the 19 peptide mixtures in a given library contained 361 different peptide sequences, arising from randomization of the two amino acids occupying the "x" positions. Thus, each library is comprised of 6,859 different peptide sequences.
The adherence of P. gingivalis cells to immobilized samples of each of the combinatorial peptide libraries was determined. As shown in Fig. 3A, several peptide mixtures in the combinatorial library designated N1182 bound to labeled P. gingivalis better than the positive-control mixture that possessed Asn at position 1182 (i.e., the residue present in BAR at this position). Interestingly, three of the amino acid substitutions at position 1182 that exhibited greater adherence activity were positively charged amino acid residues (Arg, Lys, and His), suggesting that a basic residue is optimal at this position in BAR. Substitution of Ser for Asn at 1182 also resulted in increased adherence of P. gingivalis. Screening of combinatorial library V1185 showed that three amino acid substitutions for Val resulted in increased adherence of P. gingivalis (Fig. 3C). Each of these residues was hydrophobic: isoleucine, phenylalanine, and tryptophan. This suggests that hydrophobic interactions may also in part drive the adherence of P. gingivalis to BAR. Indeed, the requirement for a hydrophobic amino acid in position 1185 appears to be quite pronounced, since replacement of most other amino acids at this position reduced P. gingivalis adherence (Fig. 3C). Finally, the overall effect of amino acid substitutions for Thr at position 1184 was less pronounced, either positively or negatively, than that observed at the other sites (compare y-axis scales of Fig. 3B with those of Fig. 3A and 3C). The three substitutions for Thr that resulted in greatest adherence were Arg, Lys, and Ser, similar to the results obtained with combinatorial library N1182.
For each of the combinatorial libraries, some amino acid substitutions resulted in decreased adherence activity and appear to be incompatible with the interaction of BAR with P. gingivalis. For example, substitution with Asp or Pro at any of the three positions (N1182, T1184, or V1185) was detrimental to P. gingivalis adherence. In addition, the presence of Gly at position 1182 or 1185 also reduced P. gingivalis adherence. These results are of interest since Gly and Pro are the amino acid residues that occupy positions 1182 and 1185 in the nonfunctional homolog of SspB that is expressed by S. mutans (6). Furthermore, substitution of Gly for Asn or Pro for Val by site-specific mutagenesis of the full-length SspB polypeptide resulted in a significant reduction of P. gingivalis adherence. Therefore, results from the combinatorial library screenings are consistent with previous structure-function data of the BAR-P. gingivalis interaction. Taken together, the outcome of the combinatorial screen of BAR (summarized in Fig. 4) suggests that both electrostatic and hydrophobic interactions may be required for efficient adherence of P. gingivalis to S. gordonii and consequently for formation of P. gingivalis biofilms. Our results also suggest that replacement of active residues with helix-breaking or -turn-inducing amino acids is detrimental for P. gingivalis adherence.
DISCUSSION
The human oral cavity presents a hostile environment to microorganisms as a result of the constant flow of saliva that contains numerous immune and nonimmune antimicrobial activities (20, 25). The early colonizing bacteria of the oral cavity that comprise supragingival plaque insulate themselves from these conditions by adhering tightly to oral tissues and/or to other oral organisms to form a multispecies biofilm (13, 28). The periodontal pathogen P. gingivalis faces an even greater challenge in colonizing the oral cavity since it is a fastidious obligate anaerobe that is asaccharolytic and acid sensitive and requires hemin for growth (17). Its primary niche is in the subgingival pocket, yet it must first colonize and survive the relatively aerobic environment of supragingival plaque. To accomplish this, our hypothesis is that P. gingivalis initially colonizes a specific niche in supragingival plaque by adhering tightly to organisms such as S. gordonii (16) and Fusobacterium nucleatum (1, 11) that may provide a reduced local environment that is not highly acidic. Since these specific interbacterial interactions may represent the first steps in P. gingivalis colonization, they are ideal targets for the development of therapeutic agents against adult periodontal disease.
Our previous studies have focused on the mechanism of P. gingivalis adherence to streptococci which leads to the subsequent development of biofilms on the streptococcal substrate. These studies showed that P. gingivalis adheres to S. gordonii cells by interaction of the short fimbrial subunit protein Mfa1 with a discrete region of the streptococcal SspB polypeptide (6, 23). Here we show that a synthetic peptide encompassing this region of SspB functions as a potent inhibitor of P. gingivalis adherence to S. gordonii and its subsequent formation of biofilms. From our results, we calculated that the number of P. gingivalis microcolonies that formed on the immobilized streptococcal substrate would be reduced by 50% in the presence of approximately 4 μg of peptide inhibitor per ml (50% inhibitory concentration, 1.3 μM). This value is consistent with the kinetics of adherence of purified Mfa1 protein to the BAR peptide as determined by enzyme-linked immunosorbent assay, where half-maximal binding occurred at approximately 0.8 μM BAR (23). This observation suggests that the interaction of Mfa1 with the BAR region of SspB is the primary driving force in P. gingivalis adherence to streptococci, which initiates the formation of biofilms.
Since P. gingivalis is a highly proteolytic organism, the specific inhibitory activity of BAR might be limited by proteolytic degradation. Indeed, based on its lysine content (5 of 27 residues), it is possible that the BAR peptide would be highly susceptible to degradation by the Lys-gingipain expressed by P. gingivalis. Since most proteases have not evolved to act on D-amino acid polypeptides (7), a retroinverso peptide that maintains the side chain topology of BAR might be expected to exhibit greater specific inhibitory activity. However, our results showed that a retroinverso derivative of BAR was inactive in preventing P. gingivalis adherence and formation of biofilms. A possible explanation for this result is that the association of Mfa1 with BAR requires crucial hydrogen bonding with the peptide backbone of BAR and reversing the chirality of the peptide bonds in BAR prevents these interactions from occurring. This has been previously suggested to represent a limitation of retroinverso peptidomimetics (7). We are currently exploring the possibility of selectively incorporating D-amino acids to limit proteolytic susceptibility of BAR while maintaining chirality in the remainder of the peptide backbone in order to facilitate potential interactions with Mfa1.
P. gingivalis adheres avidly to the SspB polypeptide of S. gordonii but does not interact with the highly related SpaP protein of S. mutans (2). Our previous studies showed that Asn1182, Thr1184, and Val1185 in BAR are not conserved in the SpaP sequence and that substitution of the corresponding S. mutans residues (Gly and Pro) for Asn1182 and Val1185, respectively, of SspB resulted in a nonfunctional S. gordonii protein (6). These results highlighted the functional role of the NITVK region of BAR in the interaction of P. gingivalis with SspB but were not informative about the physicochemical nature of the interaction of SspB and Mfa1. The biased combinatorial peptide libraries analyzed in the current study allowed us to examine the functional properties of BAR peptides containing all combinations of the common amino acids occupying the nonconserved residues (Asn1182, Thr1184, and Val1185) in the NITVK sequence. Interestingly, P. gingivalis efficiently bound to BAR-like peptides containing amino acid substitutions at both residues 1182 and 1185. The most active peptides were those containing a positively charged amino acid (Arg, Lys, or His) at position 1182 and a hydrophobic residue (Ile, Phe, or Trp) at 1185. This suggests that the SspB-Mfa1 interaction does not strictly depend on the specific amino acid occupying these positions in SspB but rather on the physical properties and characteristics of the amino acid residue.
Since the most active BAR peptide analogs contained a hydrophobic amino acid (Ile, Phe, Trp, or Val) at position 1185, we can speculate that the interacting interface of the P. gingivalis fimbrial protein Mfa1 likely contains a large hydrophobic pocket that can accommodate the bulky side chains of these residues. The fact that no other amino acids were functionally tolerated at this site in our combinatorial screen suggests that hydrophobic interactions may be essential for the association of Mfa1 with SspB. In addition, the presence of positively charged amino acids at position 1182 (Arg, Lys, or His) facilitated P. gingivalis adherence, whereas the replacement of N1182 with acidic residues (Asp and, to a lesser extent, Glu) reduced adherence. The preference for basic residues could be explained by the presence of negatively charged residues in the SspB binding pocket of Mfa1 that might contribute to binding through the formation of electrostatic interactions with SspB. However, the Mfa1-SspB interaction is clearly more complex than suggested above since the polar, but uncharged, residues Asn and Ser at position 1182 also promoted P. gingivalis adherence. While the exact structure of the interacting interface cannot be described in full without additional structural studies, our results suggest that both electrostatic and hydrophobic bonding may drive the protein-protein interaction of Mfa1 with SspB.
It is also of interest that several amino acids (e.g., Pro, Gly, and Asp) appeared to be generally detrimental to P. gingivalis adherence regardless of position in the NITVK sequence. The presence of Pro at any of the randomized positions in BAR, and Gly at residue 1182 and to a lesser extent at residue 1184, resulted in reduced adherence. In addition, substitution of Gly and Pro for N1182 and V1185 reduced adherence. This is in agreement with our previous site-specific mutagenesis data (6). Gly and Pro are both -turn-inducing or helix-breaking residues, and it is likely that their presence in BAR alters the structure of this region. Secondary structural predictions of the BAR peptide analogs are consistent with this conclusion. For example, residues 1167 to 1180 of BAR are predicted to be -helical, with the remainder of the peptide predicted to be -sheet. In contrast, structural predictions of a BAR analog containing Gly and Pro at positions 1182 and 1185, respectively, differ from BAR in that residues 1167 to 1180 are predicted to be -helical as described above, but this helix is followed by a -turn (comprising residues 1181 to 1187) and then a second -helix. The predicted turn in this peptide analog arises from the presence of three helix-interrupting amino acids Gly1182, Pro1185, and Gly1187 in the NITVK region. Thus, at least some of the amino acid substitutions that reduce P. gingivalis adherence may alter secondary structural elements of BAR, suggesting that the overall structure of the BAR region may also be important for recognition of SspB by Mfa1.
In summary, we have shown that the formation of P. gingivalis biofilms can be blocked by peptide-mediated inhibition of the interaction of Mfa1 with the streptococcal SspB protein. Analysis of peptide analogs and combinatorial peptide libraries based on the sequence of BAR suggests that both conformational determinants and the characteristics of specific amino acid residues in BAR are essential for this interaction to occur. These studies highlight molecular details of the interacting regions of the streptococcal SspB protein and the P. gingivalis minor fimbrial subunit Mfa1 and define physicochemical characteristics of the binding domains. These results suggest that peptides or peptidomimetics with greater specific inhibitory activity than that of BAR can be synthesized.
ACKNOWLEDGMENTS
We thank Masae Kuboniwa for her assistance with the P. gingivalis adherence assay and Douglas Darling for his help with the statistical analyses.
This study was supported by Public Health Service grant DE12505.
FOOTNOTES
Corresponding author. Mailing address: Department of Periodontics, Endodontics and Dental Hygiene, Research Group in Oral Health and Systemic Disease, University of Louisville School of Dentistry, 501 S. Preston St., Room 209, Louisville, KY 40292. Phone: (502) 852-3807. Fax: (502) 852-4052. E-mail: drdemu01@louisville.edu.
REFERENCES
1. Bradshaw, D. J., P. D. Marsh, G. K. Watson, and C. Allison. 1998. Role of Fusobacterium nucleatum and coaggregation in anaerobe survival in planktonic and biofilm oral microbial communities during aeration. Infect. Immun. 66:4729-4732.
2. Brooks, W., D. R. Demuth, S. Gil, and R. J. Lamont. 1997. Identification of a Streptococcus gordonii SspB domain that mediates adhesion to Porphyromonas gingivalis. Infect. Immun. 65:3753-3758.
3. Chung, W. O., D. R. Demuth, and R. J. Lamont. 2000. Identification of a Porphyromonas gingivalis receptor for the Streptococcus gordonii SspB protein. Infect. Immun. 68:6758-6762.
4. Cook, G. S., J. W. Costerton, and R. J. Lamont. 1998. Biofilm formation by Porphyromonas gingivalis and Streptococcus gordonii. J. Periodont. Res. 33:323-327.
5. Dalwai, F., B. J. Spratt, and J. Pratten. 2006. Modeling shifts in microbial populations in health and disease. Appl. Environ. Microbiol. 72:3678-3684.
6. Demuth, D. R., D. C. Irvine, J. W. Costerton, G. S. Cook, and R. J. Lamont. 2001. Discrete protein determinant directs the species-specific adherence of Porphyromonas gingivalis to oral streptococci. Infect. Immun. 69:5736-5741.
7. Fischer, P. M. 2003. The design, synthesis and application of stereochemical and directional peptide isomers: a critical review. Curr. Protein Pept. Sci. 4:339-356.
8. Grenier, D. 1992. Demonstration of bimodal coaggregation reaction between Porphyromonas gingivalis and Treponema denticola. Oral Microbiol. Immunol. 7:280-284.
9. Ikegami, A., K. Honma, A. Sharma, and H. K. Kuramitsu. 2004. Multiple functions of the leucine-rich repeat protein LrrA of Treponema denticola. Infect. Immun. 72:4619-4627.
10. Jenkinson, H. F., and D. R. Demuth. 1997. Structure, function and immunogenicity of streptococcal antigen I/II polypeptides. Mol. Microbiol. 23:183-190.
11. Kinder, S. A., and S. C. Holt. 1989. Characterization of coaggregation between Bacteroides gingivalis T22 and Fusobacterium nucleatum T18. Infect. Immun. 57:3425-3433.
12. Kolenbrander, P. E., and R. N. Andersen. 1989. Inhibition of coaggregation between Fusobacterium nucleatum and Porphyromonas (Bacteroides) gingivalis by lactose and related sugars. Infect. Immun. 57:3204-3209.
13. Kolenbrander, P. E., and J. London. 1993. Adhere today, here tomorrow: oral bacterial adherence. J. Bacteriol. 175:3247-3252.
14. Kuboniwa, M., G. D. Tribble, C. E. James, A. O. Kilic, L. Tao, M. C. Herzberg, S. Shizukuishi, and R. J. Lamont. 2006. Streptococcus gordonii utilizes several distinct gene functions to recruit Porphyromonas gingivalis into a mixed community. Mol. Microbiol. 60:121-139.
15. Lamont, R. J., A. El-Sabaeny, Y. Park, G. S. Cook, J. W. Costerton, and D. R. Demuth. 2002. Role of Streptococcus gordonii SspB protein in the development of Porphyromonas gingivalis biofilms on streptococcal substrates. Microbiology 148:1627-1636.
16. Lamont, R. J., S. G. Hersey, and B. Rosan. 1992. Characterization of the adherence of Porphyromonas (Bacteroides) gingivalis to oral streptococci. Oral Microbiol. Immunol. 7:193-197.
17. Lamont, R. J., and H. F. Jenkinson. 1998. Life below the gum line: pathogenic mechanisms of Porphyromonas gingivalis. Microbiol. Mol. Biol. Rev. 62:1244-1263.
18. Lamont, R. J., and H. F. Jenkinson. 2000. Subgingival colonization by Porphyromonas gingivalis. Oral Microbiol. Immunol. 15:341-349.
19. Maeda, K., H. Nagata, M. Kuboniwa, K. Kataoka, N. Nishida, M. Tanaka, and S. Shizukuishi. 2004. Characterization of binding of Streptococcus oralis glyceraldehyde 3-phosphate dehydrogenase to Porphyromonas gingivalis major fimbriae. Infect. Immun. 72:5475-5477.
20. Marsh, P. D. 1989. Host defenses and microbial homeostasis: role of microbial interactions. J. Dent. Res. 68:1567-1575.
21. Marsh, P. D. 1994. Microbial ecology of dental plaque and its significance in health and disease. Adv. Dent. Res. 8:263-271.
22. Pakula, R., and W. Walczak. 1963. On the nature of competence and transformable streptococci. J. Gen. Microbiol. 31:125-133.
23. Park, Y., M. R. Simionato, K. Sekiya, Y. Murakami, D. M. James, W. Chen, M. Hackett, F. Yoshimura, D. R. Demuth, and R. J. Lamont. 2005. Short fimbriae of Porphyromonas gingivalis and their role in coadhesion with Streptococcus gordonii. Infect. Immun. 73:3983-3989.
24. Paster, B. J., S. K. Boches, J. L. Galvin, R. E. Ericson, C. N. Lau, V. A. Levanos, A. Sahasrabudie, and F. E. Dewhirst. 2001. Bacterial diversity in human subgingival plaque. J. Bacteriol. 183:3770-3783.
25. Rosan, B., and R. J. Lamont. 2000. Dental plaque formation. Microbes Infect. 2:1599-1607.
26. Socransky, S. S., and A. D. Haffajee. 1992. The bacterial etiology of destructive periodontal disease: current concepts. J. Periodontol. 63:322-331.
27. Sohar, H. T., and R. J. Genco. 2005. Identification of glyceraldehyde 3-phosphate dehydrogenase of epithelial cells as a second molecule that binds to Porphyromonas gingivalis fimbriae. FEMS Immunol. Med. Microbiol. 45:25-30.
28. Whittaker, C. J., C. M. Klier, and P. E. Kolenbrander. 1996. Mechanisms of adhesion by oral bacteria. Annu. Rev. Microbiol. 50:513-552.
29. Yao, E. S., Lamont, R. J., S. P. Leu, and A. Weinberg. 1996. Interbacterial binding among pathogenic and commensal oral bacterial species. Oral Microbiol. Immunol. 11:35-41.(Carlo Amorin Daep, DeAnna M. James, Rich)