Vol. 60, No. 4

INFECTION AND IMMUNITY, Apr. 1992, p. 1662-1670 0019-9567/92/041662-09$02.00/0 Copyright © 1992, American Society for Microbiology

Synthetic Peptides Analogous to the Fimbrillin Sequence Inhibit Adherence of Porphyromonas gingivalis JIN-YONG LEE, HAKIMUDDIN T. SOJAR, GURRINDER S. BEDI,* AND ROBERT J. GENCO Department of Oral Biology, State University of New York at Buffalo, Buffalo, New York 14214 Received 28 October 1991/Accepted 20 January 1992

Fimbriae are important in the adherence of many bacterial species to the surfaces they eventually colonize. Porphyromonas (Bacteroides) gingivalis fimbriae appear to mediate adherence to oral epithelial cells and the pellicle-coated tooth surface. The role and contribution of fimbriae in the binding of P. gingivalis to hydroxyapatite (HAP) coated with saliva as a model for the pellicle-coated tooth surface were investigated. 3H-labeled P. gingivalis or the radioiodinated purified fimbriae were incubated with 2 mg of HAP beads coated with whole human saliva (sHAP) and layered on 100% Percoll to separate unbound from sHAP-bound components. The radioactivity of the washed beads was a measure of the bound components. The binding of P. gingivalis 2561 (381) cells and that of purified fimbriae were concentration dependent and saturable at approximately 108 cells and 40 ,ug of fimbriae added, respectively. The addition of fimbriae inhibited binding of P. gingivalis to sHAP beads by 65%, while the 75-kDa protein, which is another major surface component of P. gingivalis 2561, did not show significant inhibition, suggesting that the fimbriae are important in adherence. Encapsulated and sparsely fimbriated P. gingivalis W50 did not bind to sHAP beads. On the basis of the predicted sequence of the fimbrillin, a structural subunit of fimbriae, a series of peptides were synthesized and used to localize the active fimbrillin domains involved in P. gingivalis adherence to sHAP beads. Peptides from the carboxyl-terminal one-third of the fimbrillin strongly inhibited P. gingivalis binding to sHAP beads. Active residues within the sequence of inhibitory peptide 226-245 (peptide containing residues 226 to 245) and peptide 293-306 were identified by using smaller fragments prepared either by trypsin cleavage of the peptide 226-245 or by synthesis of smaller segments of peptide 293-306. Hemagglutinin activity, lectinlike binding, and ionic interaction did not seem to be involved in this binding since lysine, arginine, carbohydrates, and calcium ions failed to affect the binding of P. gingivalis. The observation that poly-L-lysine, bovine serum albumin, and defatted bovine serum albumin, even at high concentrations, only partially blocked the binding of P. gingivalis indicates that hydrophobic interactions are not the major forces involved in P. gingivalis binding to sHAP beads. Protease inhibitors such as EDTA, leupeptin, pepstatin, 1,10-phenanthroline, and phenylmethylsulfonyl fluoride did not interfere with the binding of P. gingivalis. However, the binding of P. gingivalis to trypsin- or chymotrypsin-pretreated sHAP beads was reduced. Overall, these results suggest that fimbrillin has domains primarily confined in the carboxyl-terminal region of the protein which are responsible for binding P. gingivalis to surface-bound salivary components through specific protein-protein interactions. Specific fimbrillinmediated binding may be important in P. gingivalis attachment to oral surfaces coated with salivary components.

(25). Some adhesins of P. gingivalis have been purified and characterized previously (1, 13, 30, 36, 43). Although the exact role of P. gingivalis fimbriae is not known, available evidence suggests that fimbriae are important in adhesion. In a preliminary study, we found that fimbriated P. gingivalis 2561 bound strongly to cultured epithelial cells from human explants, epithelial cell lines in tissue culture, or human buccal epithelial cells. On the other hand, encapsulated and sparsely fimbriated strain W50 did not bind to these epithelial cells. Furthermore, Isogai et al. (17) have shown that Fab fragments of monoclonal antibodies to fimbriae inhibit binding of P. gingivalis to buccal epithelial cells. Fimbriae were also reported to mediate binding of P. gingivalis to Actinomyces viscosus, a pioneer colonizer of the oral cavity (12). The fimbriae of P. gingivalis are important in other interactions with host cells. It has been shown that P. gingivalis fimbriae can stimulate human gingival fibroblasts to produce thymocyte-activating factor (14) and activate mouse peritoneal macrophages, subsequently inducing gene expression and production of interleukin-1 (15). Two major protein components with molecular masses of

Porphyromonas (Bacteroides) gingivalis is a gram-negative, black-pigmented anaerobe associated with several periodontal diseases, including adult periodontitis, generalized juvenile periodontitis, periodontal abscesses, and refractory periodontitis (44). The pathogenic process of microbial infection includes gaining access to the host, adhering to and colonizing a unique niche, avoiding the host defense mechanisms, multiplying with subsequent infection, and destroying tissue. Adherence to mucosal surfaces is the first step whereby bacteria establish a close proximity to the mucosa. P. gingivalis adheres to a variety of surfaces, including epithelial cells (2, 17, 29), erythrocytes (1, 28-30), fibronectin-collagen complexes (26), salivary components like proline-rich proteins (11), and other bacteria (12, 34). Those elements which participate in the colonization process include bacterial surface adhesins, epithelial surface receptors, mucosal gel components such as mucins, and salivary coatings such as dental pellicle. Cell surface components which have been implicated in the adhesion properties of P. gingivalis include fimbriae, vesicles, and hemagglutinins *

Corresponding author. 1662

VOL. 60, 1992

SYNTHETIC PEPTIDE ANALOGS OF P. GINGIVALIS FIMBRILLIN

43 and 75 kDa, as determined by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE), are released when P. gingivalis 2561 cells are sheared or mildly sonicated (21, 43). Recently, we have developed procedures to purify native fimbriae and 75-kDa protein for use in the present studies of fimbrial function. Biochemical and immunogold localization studies indicate that the 43-kDa component is a subunit of fimbriae, while the 75-kDa component is a major outer membrane protein (36, 37). The purified fimbriae represent the major structural component(s) of the hairlike organelles extending from the surface of P. gingivalis as observed under the electron microscope (36). The term fimbrillin refers to the 43-kDa monomeric structural subunit of the fimbriae. In the present study, we have shown that fimbriae are required for adherence of P. gingivalis to hydroxyapatite (HAP) beads coated with saliva (sHAP beads). We have synthesized peptides with amino acid sequences identical to the deduced sequence from the cloned fimbrillin gene of P. gingivalis 381 (5), investigated the effect of the peptides on P. gingivalis binding, and localized the binding domains on the fimbrillin protein.

MATERIALS AND METHODS Bacterial cultural conditions and radiolabeling. Strains of P. gingivalis 2561 (381) and W50 were grown in half-strength (18 mg/ml) brain heart infusion broth (Difco) supplemented with 5 mg of yeast extract per ml, 5 ,ug of hemin per ml, and 0.2 ,ug of menadione per ml and buffered at pH 7.4. For radiolabeling, the cells were incubated with S ,uCi of [3H]thymidine (NEN Research Products) per ml at 37°C for 2 days in an anaerobic chamber (85% N2, 10% H2, 5% C02). The cells were centrifuged at 4,000 x g for 10 min at room temperature, washed three times in 50 mM KCl containing 1 mM KH2PO4, 1 mM CaCl2, and 0.1 mM MgCl2 buffered at pH 6 (buffered KCl) (4), and suspended in the same buffer. In the case of strain 2561, the washed cells were gently agitated with a Pasteur pipette and allowed to stand for 30 s to 1 min to separate aggregated cell clumps from well-dispersed cells. The well-dispersed floating cells were used for following P. gingivalis binding and binding inhibition assays. Fimbrial preparation and iodination. Pure fimbriae and a 75-kDa component of P. gingivalis 2561 were prepared, and their purity was confirmed by demonstration of a single band on an SDS-polyacrylamide gel stained with Coomassie blue or silver. Further evidence of purity comes from the finding of a single band on an SDS-polyacrylamide gel of fimbriae or the 75-kDa component labeled with 1251 and proved by autoradiography. Finally, immunoblots with rabbit antisera to fimbriae, fimbrillin, the 75-kDa component, or whole P. gingivalis cells showed a single band when tested with purified preparations of the fimbriae and the 75-kDa component. The purification procedure and criteria of purity were discussed in detail in previous publications (36, 37). Briefly, P. gingivalis 2561 from 1.5 liters of brain heart infusion culture was harvested by centrifugation at 8,000 x g for 25 min at room temperature. The pelleted cells were then washed with 20 mM Tris buffer (TB) (pH 7.4) containing 0.15 M NaCl and 10 mM MgCl2. The cells were resuspended in 30 ml of the same buffer and subjected to ultrasonication (Vibra Cell model VC250; Sonic and Materials Inc.) with a 3-mm microtip at a 20-W output with a pulse setting and a 50% duty cycle for 5 min in an ice bath. The supernatant obtained by centrifugation at 30,000 x g for 30 min at 4°C was brought to 40% saturation by the stepwise addition of solid ammonium sulfate and stirred at 4°C overnight. The precipitated pro-

1663

teins were collected by centrifugation at 10,000 x g for 25 min at 4°C, suspended in a small volume of 20 mM TB (pH 8.0), and dialyzed against the TB. This crude fimbrial preparation contained the fimbriae, the 75-kDa protein, and other minor components when assessed by SDS-PAGE after boiling at 100°C for 7 min in P-mercaptoethanol. A purified preparation of the fimbriae was then obtained by repetitive differential precipitation of the fimbrial preparation at pH 6.5 in the presence of 1% SDS and 0.2 M MgCl2 at 4°C for 16 to 18 h (36). A pure 75-kDa component was also prepared as an outer surface component of the organism (37). In brief, the dialyzed crude fimbrial preparation was further clarified by centrifugation at 100,000 x g for 1 h at 4°C in a Beckman LS8-80 ultracentrifuge, and the supernatant was adjusted to pH 5.0 with 1 M HCl and left to stand at 4°C for 16 to 18 h with continuous stirring. The precipitate containing mainly the 75-kDa protein was harvested by centrifugation at 10,000 x g for 30 min, dissolved in a small volume of 50 mM TB (pH 8.0), and dialyzed against the same buffer. The precipitation step was repeated once again to get a highly purified 75-kDa protein. The purified fimbriae were iodinated by using the chloramine T method (16) with modification. Ten micrograms of the fimbriae was labeled with 0.5 mCi of 1251-NaI (Amersham Corp.) in 0.5 M phosphate buffer (PB) (pH 7.2), in the presence of 10 ,ul of chloramine T (1 mg/ml) for 30 s, and iodination was terminated by adding 25 pul of sodium metabisulfite (2 mg/ml). After termination, 100 pA of PB containing 10% sucrose and 10% potassium iodide was added and the mixture was loaded onto a Sephadex G-75 column (1 by 25 cm) saturated with 1% bovine serum albumin (BSA) in PB. The column was washed extensively with PB before the iodinated mixture was loaded to avoid nonspecific loss of the fimbriae to the column. The iodinated peak was collected without BSA or other carrier protein. Preparation of synthetic fimbrillin peptides. The synthetic peptides prepared for this study correspond to the sequence of amino acids derived from the DNA sequence of the cloned fimbrillin gene for P. gingivalis 381 (5) (Table 1). Synthesis, purification, and characterization of peptides 49-68, 69-90, 81-98, and 99-110 were described in a previous study (21). The other synthetic peptides were prepared by using a solid-phase Fmoc peptide synthesis procedure with an Applied Biosystems model 431A peptide synthesizer. Briefly, the carboxyl-terminal amino acid was coupled to the 4-hydroxymethylphenoxymethyl copolystyrene-1% divinyl benzene resin (41) by using N,N'-dicyclohexylcarbodiimide as an activating reagent and 0.1 M dimethylaminopyridine in N,N'-dimethylformamide as a catalyst. The ax-amino group of the amino acid was protected by an Fmoc group which was subsequently deprotected at the beginning of every cycle by the addition of 20% piperidine. The carboxyl group of the subsequent amino acids was activated by the 1-hydroxybenzotrazole-N,N'-dicyclohexylcarbodiimide activation method (19) and mixed with the deprotected amino terminal of the growing peptide chain. Cleavage of the peptides from the resin was performed by using 95% trifluoroacetic acid (TFA)-5% thioanisol as a scavenger. Cleaved peptides were precipitated with diethylether and lyophilized. For most of the peptides prepared by Fmoc chemistry, >95% of the peptide appeared in one peak by analytical high-performance liquid chromatography (HPLC). However, peptides which showed less than 95% of the material in one peak were purified to homogeneity (>99% single peak) by preparative HPLC (Rainin, model Rabbit-HP) with a

INFECT. IMMUN.

LEE ET AL.

1664

TABLE 1. Synthetic peptides used in this study' Residue no.b

Sequence

1-16 22-41 42-61

AFGVGDDESKVAKLTV EQQEAIKSAENATKVEDIKC SAGQRTLVVMANTGAMELVG VVMANTGAMELVGKTLAEVK-C ALTTELTAENQEAAGLIMTAEP-C AAGLIMTAEPKTIVLKAG-C

49-68c

69-90 81-98 99-110 106-125 126-146 147-165 166-185 186-205 206-225 226-245 246-265 266-286 279-298 293-306 307-326 318-337

KNYIGYSGTGEG-C

GTGEGNHIENDPLKIKRVHA RMAFTEIKVQMSAAYDNIYTF VPEKIYGLTAKKQSNLFGA TLVNADANYLTGSLTTFNGA YTPANYANVPWLSRNYVAPA ADAPQGFYVLENDYSANGGT IHPTILCVYGKLQKNGADLA GADLAAAQAANWVDAEGKTY YPVLVNFNSNNYTYDSNYTPK YDSNYTPKNKIERNHKYDIK HKYDIKLTITGPGT NNPENPITESAHLNVQCTVA HLNVQCTVAEWVLVGQNATW

a The synthetic peptides were prepared to correspond to the fimbrillin sequence of the cloned gene (5). b In this study, numbering of the amino acid residues did not take into acount that the first 10 amino acids appeared as the leader sequence in the sequence of fimbrillin (5). c The peptides 49-68, 69-90, 81-98, and 99-110 were synthesized in the previous study (21) by using tertiary butyl oxycarbonyl-protected amino acids.

Hi-Pore RP-318 reverse-phase column (10 by 250 mm; BioRad). Ten to 15 mg of each peptide dissolved in 500 RI of 0.1% TFA-H20 was loaded onto the column, and separation was achieved by isocratic elution for 10 min with 10% 0.1% TFA-acetonitrile balanced with 0.1% TFA-H20, a subsequent linear gradient of 0.1% TFA-acetonitrile up to 60% for 70 min, and a gradient of 0.1% TFA-acetonitrile up to 100% at a flow rate of 4 ml/min. The composition and sequence of each peptide was confirmed by amino acid analysis and amino acid sequencing. For amino acid analysis, samples were hydrolyzed in 6 N HCl at 110°C for 22 h in evacuated, sealed tubes. The samples were dried in vacuo, and hydrolysates were analyzed with a Beckman 121 MB amino acid analyzer. a-Amino-,-guanidinopropionic acid was used as an internal standard. The amino acid sequence of the synthetic peptides was performed by automated stepwise sequencing on an Applied Biosystems model 477A gas-phase sequencer with an on-line model 120A phenylthiohydantoin (PTH) analyzer. The synthetic peptide samples (200 to 500 pM) were spotted on a glass-fiber disk previously loaded with Biobrene (Applied Biosystems) and precycled three times. PTH-amino acid analysis was performed on a PTH-Cl8 cartridge column (2.1 by 220 mm) packed with octylsilyl-type sorbent (5 p,m). Separation of PTH-amino acids was obtained at 55°C with a linear gradient of 11 to 40% solvent B (acetonitrile containing 500 nM dimethylphenylthiourea) at a flow rate of 0.2 ml/min. Solvent A, used as balance for the gradient, consisted of 5% aqueous tetrahydrofurane, 18 ml of 3 M acetate buffer (pH 3.8), and 2.5 ml of 3 M acetate buffer (pH 4.6), each per liter. Synthesis of smaller fragments and tryptic digestion of selective peptides. To further define and characterize the binding domain of the fimbrillin, peptide 226-245 (peptide containing residues 226 to 245) and peptide 293-306 were selected because the peptides showed a strong inhibitory effect on P. gingivalis binding to sHAP beads and they were

very soluble in buffered KCI. Segments of peptide 293-306 were synthesized, including peptide 293-300 (HKYDIKLT), peptide 300-306 (TITGPGT), and peptide 287-306 (NKIER

NHKYDIKLTITGPGT). Peptide 226-245 was trypsinized to obtain its smaller fragments, 226-236 and 240-245. Briefly, 20 mg of HPLCpurified peptide 226-245 was dissolved in 3 ml of 0.5% ammonium bicarbonate buffer and incubated with 400 ,ug of trypsin-tosyl-L-phenylalanyl chloromethyl ketone (TPCK) (Sigma). The solubility of the peptide in ammonium bicarbonate buffer at an optimal pH of 8 for trypsin was poor, so the pH of the buffer was brought to 5.6 and the reaction mixture was incubated for 10 h at room temperature and then further incubated with an additional 300 ,ug of trypsin-TPCK overnight at room temperature. After lyophilization, the peptide was dissolved in 0.1% TFA-H2O and centrifuged, and the supernatant was loaded and separated on a Hi-Pore RP 318 reverse-phase column. The two peptide fragments 226-236 (IHPTILCVYGK) and 240-245 (NGADLA) were obtained in separate peaks, and the purity of the fragments was confirmed by sequencing. Saliva collection. Parafilm-stimulated whole saliva from a single male donor was collected on ice and clarified by centrifugation at 10,000 x g for 10 min. The saliva was heated at 60°C for 30 min to inactivate endogenous enzymes and then centrifuged at 12,000 x g for 15 min to remove precipitated materials (4). This single batch of the saliva was aliquoted into Eppendorf tubes, stored at -20°C, and used for all of the experiments described here. sHAP. Spherical hydroxyapatite (HAP) beads were purchased from BDH Chemicals (Poole, England). sHAP beads were prepared by the method of Clark et al. (4). Briefly, 2 mg of HAP beads was added to a siliconized borosilicate culture tube (12 by 75 mm; PGC Scientifics), washed once with distilled water to remove fine particles, and equilibrated in 1 ml of buffered KCl containing 0.04% NaN3 with a gentle oscillating motion in a Roto-torque rotator (Cole-Parmer Instrument Co., model 7637) at room temperature overnight. The HAP beads were then washed twice with buffered KCl, treated with 100 ,ul of the clarified whole human saliva, containing 0.04% NaN3, rotated in the rotator at room temperature overnight, and then washed twice with buffered

KCI. Binding and binding inhibition assay. For the P. gingivalis binding assay, 2.5 x 107 to 2 x 108 cells of 3H-labeled P. gingivalis 2561 or W50 was added to a series of tubes containing sHAP beads and brought to a final volume of 400 ,ul with buffered KCl. This mixture was incubated at 20 rpm with a gentle oscillating motion in the rotator at room temperature for 1 h. The tubes were agitated gently for 5 s every 15 min to disperse aggregates and to maximize binding of P. gingivalis cells. After an incubation time of 1 h, the reaction mixture was layered on 1.5 ml of 100% Percoll (Sigma) in a new siliconized borosilicate tube to separate cells which were free from those bound to the HAP or sHAP beads. These experimental conditions of incubation time and Percoll concentration were optimized in preliminary tests to achieve the most stable and steady binding of P. gingivalis and to clearly separate unbound from bound cells. Unbound, free P. gingivalis cells floating on the Percoll layer were removed by aspiration with a 9-in. (ca. 23-cm)-long Pasteur pipette, and the beads with bound cells and the wall of the tube were then washed once with 0.5 ml of Percoll and then twice with 1 ml of buffered KCI. One milliliter of buffered KCI was added to the tube, and the beads were transferred to a vial so that the radioactivity of the beads could be

VOL. 60, 1992

SYNTHETIC PEPTIDE ANALOGS OF P. GINGIVALIS FIMBRILLIN

determined. The assay was performed in duplicate and was repeated once or twice on different days for each set of conditions; similar results were obtained for repeated experiments. Inhibition studies of P. gingivalis 2561 binding to sHAP beads were carried out with crude fimbriae, purified fimbriae, 75-kDa protein, synthetic peptides corresponding to regions of the fimbrillin, and various agents as inhibitors. A suspension of 300 .Ll of P. gingivalis 2561 (4 x 108 cells per ml) and 100 p,l of inhibitors at different concentrations was incubated with sHAP beads for 1 h. The sHAP beads in control tubes were equilibrated with 400 p.l of buffered KCI. In the case of the binding inhibition study with peptides that were insoluble in buffered KCI, 1 mg of the peptide was first dissolved in 10 to 30 ,u1 of dimethyl sulfoxide and then buffered KCI was added to a final volume of 1 ml. Controls of sHAP beads in buffered KCI with dimethyl sulfoxide were also made. Dimethyl sulfoxide in this range of concentrations did not measurably affect P. gingivalis 2561 binding under conditions of the assay. For the fimbria binding assay, a final concentration of iodinated fimbriae in a working dilution of buffered KCI (0.4 mg of the iodinated fimbriae per ml) was prepared by mixing the purified fimbriae (1 mg/ml) with the iodinated fimbriae, fivefold-concentrated buffered KCI, and distilled water in a ratio of 4:1.5:2:2.5 (by volume). The fimbria binding assay was performed in the same manner as the P. gingivalis whole-cell binding assay described above, except that 10 to 150 p.l of the iodinated fimbrial preparation was added to HAP or sHAP beads, bringing the final volume to 200 p.l with buffered KCI. RESULTS P. gingivalis binding to sHAP. P. gingivalis binding to sHAP beads was compared with binding to HAP beads equilibrated with buffered KCI. To obtain a concentrationdependent binding curve for P. gingivalis, 2.5 x 107 to 2 x 108 3H-labeled P. gingivalis 2561 or W50 cells was added to 2 mg of HAP or sHAP beads. P. gingivalis 2561 bound to sHAP beads in a concentration-dependent manner and reached a saturation level at 108 added P. gingivalis cells. At this saturation level, P. gingivalis 2561 binding to sHAP beads was 2.5 times greater than its binding to HAP beads. Approximately 10% of the total number of P. gingivalis 2561 cells bound to sHAP beads, while 4% bound to HAP beads (Fig. la). In contrast, the non- or poorly fimbriated strain W50 bound at very low levels to both sHAP and HAP beads

(Fig. lb).

Fimbria binding to sHAP. Purified fimbriae also bound to sHAP beads in a concentration-dependent manner when increasing amounts of the labeled fimbriae were added to the beads (Fig. 2). A saturation level of fimbria binding to 2 mg of sHAP beads was reached when approximately 40 p.g of the fimbriae was added. The binding of the fimbriae to sHAP beads was 1.5 to 2 times greater than it was to HAP beads in the range of the fimbriae used (Fig. 2). Inhibition of P. gingivalis binding to sHAP by fimbriae. To study the effect of fimbriae on binding of P. gingivalis, sHAP beads were incubated with radiolabeled P. gingivalis 2561 in the presence of increasing concentrations of the purified fimbriae. Two milligrams of sHAP beads was incubated with 8 to 200 p.g of the fimbriae in 100 p.l of the preparation and 300 p.l of the P. gingivalis suspension at 4 x 108 cells per ml. P. gingivalis binding to sHAP beads was inhibited by 65% when 200 p,g of fimbriae was added. A crude fimbrial

a

1665 sHAP

10 8cD

0 C)

6

-

HAP

4

.0

2

C)

0

b crx

10

-

8

Q: 4-

20

sHAP HAP ~~~~~~~

~

_ O .1

0

5

10

15

20

P. gingivalis cells added (107 ) FIG. 1. Binding curves of 3H-labeled P. gingivalis 2561 (a) and W50 (b) to buffered KCl-treated HAP and sHAP beads. Two milligrams of HAP beads equilibrated with buffered KCI or clarified whole human saliva was added to a siliconized borosilicate tube and incubated with different numbers of the labeled P. gingivalis (2.5 x 10' to 2 x 108 cells) in a total volume of 400 ,ul with a gentle, oscillating motion for 1 h at room temperature, and then the mixture was layered on 100% Percoll to separate floating unbound cells from the bead-bound cells. After washing, radioactivity of the beadbound cells was quantitated. The results shown are the mean values of duplicate samples representative of several experiments.

preparation containing fimbriae, 75-kDa protein, and other minor proteins inhibited binding of P. gingivalis to sHAP beads by 90% when 200 p.g of crude fimbriae was added (data not shown). In contrast, the P. gingivalis binding to sHAP beads was not significantly inhibited in the presence of high concentrations of the purified 75-kDa outer membrane protein (Fig. 3). Inhibition of P. gingivalis binding to sHAP by synthetic fimbrillin peptides. To localize the active binding domains of the fimbrial subunit (fimbrillin) to sHAP beads, a series of peptides were synthesized, corresponding to the full length of the cloned gene sequence of the fimbrillin. The inhibitory ability of each peptide against P. gingivalis 2561 binding was tested in reaction mixtures containing 2 mg of sHAP beads incubated with 300 p.l of the P. gingivalis suspension (4 x 108 cells per ml) and 100 pl (100 p.g) of each peptide. Several synthetic peptides, one from the amino-terminal sequence (peptide 126-145) and others from the carboxylterminal one-third (peptides 226-245, 266-288, 293-306, 307326, and 318-337), were strong inhibitors. Binding inhibition by these peptides ranged from 70 to 95% (Fig. 4). These peptides showed concentration-dependent inhibition of P. gingivalis binding to sHAP beads, which is illustrated for peptides 226-245 and 293-306 in Fig. 5a and 5b. Peptide 49-68 was moderately inhibitory at 100 p.g (Fig. 4).

1666

INFECT. IMMUN.

LEE ET AL. 6

0

sHAP

0-

Fimbriae

a

5

.0

-0 C -o 0 .0 .0

4

t-

2

0) C

._

75 kDa

op

CD E

HAP

o-w(p

-B

0L1

.c

0

I--.

0

10

I

I

I

20

30

40

I

I

50

-I

60

Fimbriae added (ig) FIG. 2. Binding curves of 251I-labeled purified fimbriae to HAP and sHAP beads. A final concentration of iodinated fimbriae in a working dilution of buffered KCI (0.4 mg/ml) was prepared by mixing purified fimbriae (1 mg/ml) with iodinated fimbriae, fivefoldconcentrated buffered KCI, and distilled water in a ratio of 4:1.5:2: 2.5 (by volume). Different amounts of iodinated fimbriae (4 to 60 pg) in a total volume of 200 ,ul were incubated with 2 mg of HAP beads equilibrated with buffered KCl or saliva. The fimbria binding assay was performed as described in the legend to Fig. 1. The results are the mean values of duplicate samples representative of several experiments.

0

1

2

3

4

5

Inhibitors added (nmol) FIG. 3. Inhibitory effect of purified fimbriae and 75-kDa outer membrane protein on the binding of P. gingivalis to sHAP beads. Two milligrams of sHAP beads was incubated with 1.2 x 10' cells of the labeled P. gingivalis 2561 in the presence of different amounts of fimbriae or 75-kDa protein in a total volume of 400 RI for 1 h at room temperature. After incubation, the reaction mixture was layered on 100% Percoll to separate the unbound cells from the sHAP beadbound cells and then washed with buffered KCI. The radioactivity of the washed sHAP bead-bound cells was measured. The results shown are representative of several experiments. The same molar concentrations of fimbriae and 75-kDa component were used for this experiment (8 to 200 ,g of fimbriae and 14 to 350 ,g of 75-kDa component in a total volume of 400 Rl). The molarity of fimbriae was calculated on the basis of an apparent molecular mass of 43 kDa for the fimbrial subunit, fimbrillin, as determined by SDS-PAGE (21, 36, 43).

Two strongly inhibitory peptides, 226-245 and 293-306, were selected for further study since they were soluble in buffered KCI and exhibited dose-dependent inhibition of binding of P. gingivalis cells. To further localize and define the binding sites of the fimbrillin, variants of these peptides were prepared and purified by HPLC. The smaller fragment, inhibitors EDTA, leupeptin, pepstatin, 1,10-phenanthroline, 226-236, derived by trypsin cleavage of peptide 226-245, and phenylmethylsulfonyl fluoride did not significantly inshowed a very strong inhibitory effect on binding of P. hibit the binding of P. gingivalis. On the other hand, pregingivalis to sHAP beads, and its inhibitory effect was treatment of the sHAP beads with trypsin or a-chymotrypsin comparable to that of the original intact peptide 226-245. In prior to being mixed with P. gingivalis cells reduced the contrast, the remaining trypsin-generated fragment 240-245 binding of P. gingivalis to sHAP beads by 53 and 52%, showed weak inhibition (Fig. Sa). respectively. The binding inhibitory abilities of two smaller peptides corresponding to fimbrillin residues 293 to 300 and 300 to 306 were compared with that of the parental peptide 293-306 (Fig. Sb). The inhibitory effect of peptide 293-300 on P. 318-337(D) gingivalis binding was comparable to that of the parental 307-326(D) 293-306, o peptide when tested at lower concentrations. The other, 0 279-298] 266-286(D) smaller peptide (300-306) showed a low binding inhibition 246-265 0) effect at an amount ranging from 2.5 to 100 ,ug. A larger 226-245 '0 206-225(D) peptide corresponding to residues 287 to 306 did not exhibit 186-205 0) 166-185(D) increased inhibition over that seen with the parental peptide CL 147-165 293-306; in fact, the inhibitory effect of peptide 287-306 was .0- 126-146(D) 106-1 25 E. comparable to that of the parental peptide on a molar basis. 99-110 0 81-98 Effects of various agents on P. gingivalis binding to sHAP. C, 69 90' 0 To better understand the nature of the noncovalent interac49-68' C 42-61 tions involved in P. gingivalis binding to sHAP beads, the C 22-41 1-16 effects of various agents on binding were evaluated (Table 2). Calcium ions and charged amino acids did not affect the 0 20 40 60 80 100 binding of P. gingivalis. POIY-L-lysine, BSA, and defatted P. gingivalis binding inhibition (%) BSA at high concentrations showed 36 to 43% binding FIG. 4. Inhibitory effect of synthetic fimbrillin peptides on the inhibition of P. gingivalis to sHAP beads. When P. gingivabinding of P. gingivalis to sHAP beads. Two milligrams of sHAP lis cells were incubated with a battery of neutral sugars and beads was incubated with 1.2 x 108 cells of P. gingivalis 2561 in the N-acetyl hexosamines, little or no inhibitory effect on P. presence of 100 ,ug of each peptide in a total volume of 400 ,ul for 1 gingivalis binding to sHAP beads was observed. However, h at room temperature. The binding inhibition assay was done as D-glucuronic acid, D-galacturonic acid, and sialic acid inhibdescribed in the legend to Fig. 3. (D), buffered KCl-insoluble ited the binding of P. gingivalis to sHAP beads by 21 to 33%, peptides were first dissolved in 10 to 30 ,ul of dimethyl sulfoxide and while neuraminidase did not affect binding. The protease diluted in 1 ml of buffered KCI. -1

-

VOL. 60, 1992

0

SYNTHETIC PEPTIDE ANALOGS OF P. GINGIVALIS FIMBRILLIN

a

100

226-236 226-245

Agents

0'I

40

a

20

240-245

0)

-20

(I)

.CZ

co

TABLE 2. Effect of various agents on the binding of P. gingivalis 2561 to sHAP

80

60

100 80

293-306 293-300

60

287-306

40

300-306

20 0

-20

0

20

40

60

80

100

Synthetic fimbrillin peptides (gg) FIG. 5. Inhibitory effect of derivatives of the peptides 226-245 and 293-306 on the binding of P. gingivalis to sHAP beads. (a) An inhibitory effect of the smaller fragments 226-236 and 240-245 obtained from trypsin cleavage of the peptide 226-245 was observed. (b) Smaller fragments (293-300 and 300-306) and the larger 20-mer peptide 287-306 were synthesized, and the inhibitory effect of the peptides on binding of P. gingivalis to sHAP beads was compared with that of the parental peptide 293-306.

DISCUSSION The attachment of oral bacteria to teeth and oral mucosal surfaces has received considerable attention since adherence is an essential step in colonization and infection (6, 9). HAP surfaces coated with saliva is a useful in vitro model to evaluate the attachment of bacteria to teeth, and possibly other surfaces coated with salivary components, and has been shown to correlate with in vivo adherence of some oral bacteria (4). In the present study, we employed this model to study the role of fimbriae in adhesion of P. gingivalis. One of the difficulties encountered during the assessment of P. gingivalis binding to sHAP beads is autoagglutination of P. gingivalis cells, which makes it difficult to separate aggregated clumps of bacteria from cells bound to sHAP beads. Percoll medium used for separation in this system improved reliability and reproducibility of the binding assay for P. gingivalis since the unbound P. gingivalis and visible large autoagglutinated cell clumps were readily separated from P. gingivalis cells bound to sHAP beads. In this study, we have shown that fimbriae mediate most of the binding of P. gingivalis to sHAP beads. Binding of the fimbriated strain P. gingivalis 2561 to sHAP beads was concentration dependent and saturable (Fig. la). In contrast, P. gingivalis W50, which was found to be sparsely fimbriated or to have an immunochemically unrelated fimbrialike structure (21), bound poorly to sHAP beads (Fig. lb). Further evidence for the role of fimbriae in the adherence of P. gingivalis comes from our finding that purified fimbriae also

1667

Final amt Inhibition (%) or concn

Buffered KCl (1 mM! CaCI2), control Buffered KCl (EGTA treated)b Buffered KCI with 5 mM CaCl2 L-Arginine 0.1 L-Lysine 0.1 Poly-L-lysine 0.25 BSA 1 Defatted BSA 1 0.1 D-Galacturonic acid D-Glucuronic acid 0.1 Sialic acid (N-acetyl neuraminic acid), sheep 0.1 0.1 N-Acetyl neuramin-lactose, bovine colostrum Colominic acid (poly-2,8-N-acetyl neuraminic 0.1 acid), Escherichia coli N-Acetylgalactosamine 0.1 0.1 N-Acetylglucosamine 0.1 Galactosamine Glucosamine 0.1 Fucose 0.1 Galactose 0.1 Glucose 0.1 Mannose 0.1 Lactose 0.1 Maltose 0.1 1 mM EDTA Leupeptin 2,uM Pepstatin 1,uM 1 mM 1,10-Phenanthroline 2 mM Phenylmethylsulfonyl fluoride 100 U Trypsin, bovine pancreas type XIII-TPCK 0.5 U a-Chymotrypsin, bovine pancreas type VII-TLCK Neuraminidase, Clostridium perfringens type V 5 mU

0 5 1 0 5 41 36 43 21 33 28 2 13

-5c -15 8 0 1 -7 -10 -15 -1 -8 24 19 -10 -8 13 53 52 21

a Final amount or concentration used in a total assay system of 400 Amounts are expressed in milligrams unless otherwise indicated.

pl.

b EGTA, ethylene glycol-bis(,B-aminoethyl ether)-N,N,N',N'-tetraacetic acid. c Minus indicates enhancement of P. gingivalis 2561 binding to sHAP beads.

bound to sHAP beads in a concentration-dependent manner, reaching a saturation level at 40 ,ug (approximately 1 nmol, determined on the basis of the molecular mass of 43 kDa for fimbrillin) of the fimbriae. To confirm the role of fimbriae in binding of P. gingivalis cells to sHAP beads, we showed that highly purified fimbriae, as well as synthetic peptides corresponding to several sequences found in the fimbrillin, inhibited binding of P. gingivalis cells to sHAP beads (Fig. 3 and 4). For example, binding inhibition of P. gingivalis 2561 to sHAP beads was observed when the purified fimbriae were added, and 65% inhibition was achieved with 200 ,ug of added fimbriae. However, the binding of P. gingivalis was not inhibited by the 75-kDa protein (Fig. 3), which is a major surface component of P. gingivalis 2561 (37). These observed results suggest that the fimbriae play an important role in specific binding of P. gingivalis to sHAP surfaces. The demonstration of direct binding of purified fimbriae to sHAP beads is further evidence of the role of fimbriae in binding of P. gingivalis cells to sHAP beads. The role of salivary components as receptors in this interaction is suggested by the finding that although fimbriae bind to bare HAP beads, their binding to sHAP beads is 1.5 to 2 times greater.

1668

LEE ET AL.

We found that P. gingivalis 2561 cells bound to sHAP beads to a greater extent than to HAP beads (Fig. la). This observation is consistent with the result reported by Cimasoni et al. (3) and with our finding that purified fimbriae bind to a greater extent to sHAP than to HAP beads (Fig. 2). Taken together, these findings suggest that fimbriae mediate binding of P. gingivalis to saliva-coated surfaces such as teeth. The adsorbed salivary components on HAP surfaces have been shown to impart a higher order of specificity for the adherence of oral bacteria (4). Saliva may also be important in P. gingivalis binding to other surfaces in the oral cavity such as buccal epithelial cells coated with a salivary pellicle in vivo (29). Synthetic peptides are useful for epitope or molecular mapping of proteins (18, 32, 40) and the development of antibodies (7) and are used as vaccines (23). In this study, to define the binding site of the fimbrillin, a series of peptides were synthesized corresponding to the entire length of the single polypeptide chain of fimbrillin which comprises the structural backbone of P. gingivalis fimbriae. In a binding inhibition assay, peptides corresponding to the carboxylterminal one-third of the fimbrillin sequence and one peptide from the amino-terminal half of the molecule strongly inhibited the binding of P. gingivalis cells to sHAP beads. These results suggest that different regions of the fimbrillin molecule, including several regions in the carboxyl terminus, are involved in active binding of P. gingivalis to sHAP beads. The inhibition by synthetic peptides appears to be specific since not all of the peptides tested inhibit binding of P. gingivalis to sHAP beads. Further evidence of specificity is suggested by the finding that these inhibitory peptides failed to inhibit the adherence of A. viscosus to sHAP beads (data not shown). It is interesting to observe that several separate, nonoverlapping synthetic peptides with apparently no sequence similarity each inhibited binding of P. gingivalis to sHAP beads by 70 to 95%, while 65% inhibition was observed with purified fimbriae. These findings may result from concentration effects or conformational and steric hindrance differences between fimbriae and synthetic peptides. On a molar basis, the concentration of synthetic peptides used in this study was 10-fold higher than the concentration of the fimbriae. Furthermore, the possibility of smaller peptides entering into the binding sites of salivary molecules more easily than whole fimbriae cannot be ruled out with the existing data. It is hypothesized that these different sites on the fimbrillin may be folded in a way to give a conformational domain(s) responsible for binding to salivary components. Only a few amino acids from each site may be involved in actual physical interaction with receptors on salivary components. This hypothesis is suggested by the use of either the tryptic fragments or the smaller synthetic peptides derived from the inhibitory peptides 226-245 and 293-306 (Fig. 5a and b) to further define the active inhibitory sites. It was found that the binding of P. gingivalis to sHAP beads is localized to residues 226 to 236 and 293 to 300, with corresponding amino acid sequences IHPTILCVYGK and HKYDIKLT. A structural study of these peptides and identification of the salivary components with which they react are under way to better understand the interaction of the binding domains of fimbrillin with salivary components, both bound and free in solution. Another possible explanation of the finding that several nonoverlapping synthetic fimbrillin peptides inhibit P. gingivalis cell binding to sHAP beads to an extent comparable to that of whole fimbriae is that stable P. gingivalis binding to

INFECT. IMMUN.

sHAP beads requires multivalent binding. It is proposed that several fimbrillin-active sites are required to interact with several receptor sites on one or more salivary molecules for stable P. gingivalis cell binding to occur and that inhibition of one or more of these interactions leads to dissociation of the bacterial cells from the surface. This hypothesis would explain why significant inhibition of P. gingivalis binding can be accomplished by any one of the active-site fimbrillin peptides as well as the whole fimbriae. The several possible explanations for the peptide inhibition data are under investigation. It appears that other adhesins besides the fimbriae are also involved in P. gingivalis binding to sHAP beads. This is suggested by the finding that the crude fimbrial preparation, which contains fimbriae as well as other surface components, showed a higher inhibitory effect (90%) on P. gingivalis binding to sHAP beads than that achieved by purified fimbriae. It is possible that there are fimbria-associated adhesins, other than the structural subunit, which mediate binding of P. gingivalis to surfaces. The 75-kDa protein does not appear to be involved in P. gingivalis binding to sHAP surfaces.

Adherence of P. gingivalis to oral surfaces is likely a complex set of interactions between bacterial cell surface adhesins and various host surface receptors. The negligible effect of L-lysine, L-arginine, and CaCl2 on P. gingivalis binding to sHAP beads suggests that binding of P. gingivalis to sHAP beads does not depend on strong ionic interactions. It has been shown that the hydrophobicity of bacteria is important in binding of the bacteria to certain surfaces (8, 24, 27, 42). Strains of P. gingivalis have been regarded as highly hydrophobic bacteria, and certain salivary components can be adsorbed to HAP beads and serve as potential hydrophobic receptors (8). The surface hydrophobicity of P. gingivalis is thought to be associated with the fimbriae and may enable P. gingivalis to bind to salivary pellicle or other bacterial surfaces (10, 31), as seen in the cases of other bacterial binding to host cells (24). Agents like poly-L-lysine, BSA, and defatted BSA, known to be very strong inhibitors of the hydrophobic phenomenon (33), failed to exert major inhibitory effects on P. gingivalis binding to sHAP beads. Also, not all hydrophobic peptides showed inhibitory effects on P. gingivalis binding. However, it is possible that, in our assay, the 100% Percoll may have eliminated nonspecific, lowaffinity hydrophobic binding. It appears that, under the conditions of the assay, binding of P. gingivalis to sHAP beads is mediated mainly by specific interactions of active regions of the fimbrillin. The importance of hydrophobic binding is yet to be determined. It was found that trypsinlike proteases contribute to the binding of P. gingivalis to A. viscosus (22) and to Streptococcus gordonii (39). Surface proteases may act as another family of adhesins for P. gingivalis. Recently, it has been reported that binding of P. gingivalis was greater to oral epithelial cells pretreated with trypsin or chymotrypsin (2). P. gingivalis exhibits strong trypsinlike and weak chymotrypsinlike activities (20, 35). It has been proposed that bacterial proteases may expose cryptic domains on A. viscosus surfaces and might function in concert with fimbriae as specific adhesins (12). Stinson et al. (38) have proposed that proteases may also act by clearing the surface adhesins of P. gingivalis which are fouled with salivary and serum molecules. In our study, protease inhibitors such as EDTA, leupeptin, pepstatin, 1,10-phenanthroline, and phenylmethylsulfonyl fluoride did not significantly affect the binding of P.

VOL. 60, 1992

SYNTHETIC PEPTIDE ANALOGS OF P. GINGIVALIS FIMBRILLIN

gingivalis to sHAP beads (Table 2), suggesting that the adhesins responsible for P. gingivalis binding to sHAP beads are not proteases. Rather, pretreatment of sHAP beads with trypsin or a-chymotrypsin greatly reduced the binding of P. gingivalis, suggesting that proteases may affect the salivary receptor molecules. Protease activity per se appears not to be responsible for fimbrial adhesion to sHAP beads. The hemagglutinins of P. gingivalis do not seem to be important adhesins in P. gingivalis binding to sHAP beads since L-lysine and L-arginine, which have been shown to be inhibitors of the hemagglutinin activity of P. gingivalis 381 (30), do not affect P. gingivalis binding. Nishikata et al. (28) reported that hemagglutinin possesses protease activity, showing that hemagglutinin activity, along with protease activity, was inhibited by arginine-related enzyme inhibitors such as leupeptin, antipain, and arginine itself, and by tosyl-L-lysine chloromethyl ketone (TLCK), an irreversible inhibitor of trypsinlike proteases. The inability of protease inhibitors to inhibit binding of P. gingivalis to sHAP further supports the interpretation that a proteaselike hemagglutinin is not involved in the binding of P. gingivalis to sHAP. A sugar-inhibitable lectin-receptor interaction also does not seem to be responsible for the binding of P. gingivalis to sHAP beads since a battery of monosaccharides and other sugars including sialic acid failed to significantly block binding. Pretreatment of oral epithelial cells with neuraminidase has been shown to change the binding ability of oral bacteria to the cells (2). However, in the present study, pretreatment of sHAP beads with neuraminidase did not affect the binding of P. gingivalis. In addition, the observation that sialic acid does not effectively inhibit P. gingivalis binding to sHAP beads suggests that most of the binding of whole-cell P. gingivalis to sHAP beads do not involve sialic acid residues either as part of the ligand or the receptor. Taken together, the data suggest that fimbriae are important in the binding of P. gingivalis 2561 to sHAP beads through protein-protein interactions. The interactions are likely to be between multiple binding domains of the fimbrillin and one or several salivary components. The nature of the salivary receptor(s) for the fimbrillin binding, the exact localization of the active sites on the fimbrillin, and the role of other adhesins in the binding of P. gingivalis to salivacoated surfaces will need to be elucidated for the complex but important events in its adherence and colonization to be fully understood. ACKNOWLEDGMENTS This study was supported in part by U.S. Public Health Service grants DE08240, DE07034, DE04898, and DE06514. REFERENCES 1. Boyd, J., and B. C. McBride. 1984. Fractionation of hemagglutinating and bacterial binding adhesins of Bacteroides gingivalis. Infect. Immun. 45:403-409. 2. Childs, W. C., III, and R. J. Gibbons. 1990. Selective modulation of bacterial attachment to oral epithelial cells by enzyme activities associated with poor oral hygiene. J. Periodontal Res. 25:172-178. 3. Cimasoni, G., M. Song, and B. C. McBride. 1987. Effect of crevicular fluid and lysosomal enzymes on the adherence of streptococci and bacteroides to hydroxyapatite. Infect. Immun. 55:1484-1489. 4. Clark, W. B., L. L. Bammann, and R. J. Gibbons. 1978. Comparative estimates of bacterial affinities and adsorption sites on hydroxyapatite surfaces. Infect. Immun. 19:846-853. 5. Dickinson, D. P., M. K. Kubiniec, F. Yoshimura, and R. J. Genco. 1988. Molecular cloning and sequencing of the gene

1669

encoding the fimbrial subunit protein of Bacteroides gingivalis. J. Bacteriol. 170:1658-1665. 6. Finlay, B. B., and S. Falkow. 1989. Common themes in microbial pathogenicity. Microbiol. Rev. 53:210-230. 7. Ghose, A. C., and F. Karush. 1988. Induction of polyclonal and monoclonal antibody responses to cholera toxin by the synthetic peptide approach. Mol. Immunol. 3:223-230. 8. Gibbons, R. J. 1984. Microbial ecology. Adherent interactions which may affect microbial ecology in the mouth. J. Dent. Res. 63:378-385. 9. Gibbons, R. J. 1989. Bacterial adhesion to oral tissue: a model for infectious diseases. J. Dent. Res. 68:750-760. 10. Gibbons, R. J., and I. Etherden. 1983. Comparative hydrophobicities of oral bacteria and their adherence to salivary pellicle. Infect. Immun. 41:1190-1196. 11. Gibbons, R. J., and D. I. Hay. 1988. Adsorbed salivary prolinerich proteins as bacterial receptors on apatitic surfaces, p. 143-169. In L. M. Switalski, M. Hook, and E. Beachey (ed.), Molecular mechanisms of microbial adhesion. Springer-Verlag, New York. 12. Goulbourne, P. A., and R. P. Ellen. 1991. Evidence that Porphyromonas (Bacteroides) gingivalis fimbriae function in adhesion to Actinomyces viscosus. J. Bacteriol. 173:5266-5274. 13. Grenier, D., and D. Mayrand. 1987. Functional characterization of extracellular vesicles produced by Bacteroides gingivalis. Infect. Immun. 55:111-117. 14. Hanazawa, S., K. Hirose, Y. Ohmori, S. Amano, and S. Kitano. 1988. Bacteroides gingivalis fimbriae stimulate production of thymocyte-activating factor by human gingival fibroblasts. Infect. Immun. 56:272-274. 15. Hanazawa, S., Y. Murakami, K. Hirose, S. Amano, Y. Ohmori, H. Higuchi, and S. Kitano. 1991. Bacteroides (Porphyromonas) gingivalis fimbriae activate mouse peritoneal macrophages and induce gene expression and production of interleukin-1. Infect. Immun. 59:1972-1977. 16. Hunter, W. M., and F. C. Greenwood. 1962. Preparation of iodine-131 labeled human growth hormone of high specific activity. Nature (London) 194:495-496. 17. Isogai, H., E. Isogai, F. Yosliimura, T. Suzuki, W. Kagota, and K. Takano. 1988. Specific inhibition of adherence of an oral strain of Bacteroides gingivalis 381 to epithelial cells by monoclonal antibodies against the bacterial fimbriae. Arch. Oral Biol. 33:479-485. 18. Kay, M. M. B. 1991. Molecular mapping of human band 3 anion transport regions using synthetic peptides. FASEB J. 5:109115. 19. Konig, W., and R. Geiger. 1970. A new method for synthesis of peptides: activation of the carboxyl group with dicyclohexylcarbodiimide using 1-hydroxybenzotrazoles as additives. Chem. Ber. 103:788-798. 20. Laughon, B. E., S. A. Syed, and W. J. Loesche. 1982. API ZYM system for identification of Bacteroides spp., Capnocytophaga spp., and spirochetes of oral origin. J. Clin. Microbiol. 15:97102. 21. Lee, J.-Y., H. T. Sojar, G. S. Bedi, and R. J. Genco. 1991. Porphyromonas (Bacteroides) gingivalis fimbrillin: size, aminoterminal sequence, and antigenic heterogeneity. Infect. Immun. 59:383-389. 22. Li, J., R. P. Ellen, C. I. Hoover, and J. R. Felton. 1991. Association of proteases of Porphyromonas (Bacteroides) gingivalis with its adhesion to Actinomyces viscosus. J. Dent. Res. 70:82-86. 23. Londofio, J. A., H. Gras-Masse, C. Dubeaux, A. Tartar, and P. Druilhe. 1990. Secondary structure and immunogenicity of hybrid synthetic peptides derived from two Plasmodium falciparum pre-erythrocytic antigens. J. Immunol. 145:1557-1563. 24. Magnusson, K.-E., J. Davies, T. Grundstrom, E. KihlstrOm, and S. Normarlk 1980. Surface charge and hydrophobicity of Salmonella, E. coli, gonococci in relation to their tendency to associate with animal cells. Scand. J. Infect. Dis. Suppl. 24:135-140. 25. Mayrand, D., and S. C. Holt. 1988. Biology of asaccharolytic black-pigmented Bacteroides species. Microbiol. Rev. 52:134152.

1670

LEE ET AL.

26. Naito, Y., and R. J. Gibbons. 1988. Attachment of Bacteroides gingivalis to collagenous substrata. J. Dent. Res. 67:1075-1080. 27. Nesbitt, W. E., R. J. Doyle, and K G. Taylor. 1982. Hydrophobic interactions and the adherence of Streptococcus sanguis to hydroxylapatite. Infect. Immun. 38:637-644. 28. Nishikata, M., F. Yoshimura, and Y. Nodasaka. 1989. Possibility of Bacteroides gingivalis hemagglutinin possessing protease activity revealed by inhibition studies. Microbiol. Immunol. 33:75-80. 29. Okuda, K., J. Slots, and R. J. Genco. 1981. Bacteroides gingivalis, Bacteroides asaccharolyticus, and Bacteroides melaninogenicus subspecies: cell surface morphology and adherence to erythrocytes and human buccal epithelial cells. Curr. Microbiol. 6:7-12. 30. Okuda, K., A. Yamamoto, Y. Naito, I. Takazoe, J. Slots, and R. J. Genco. 1986. Purification and properties of hemagglutinin from culture supernatant of Bacteroides gingivalis. Infect. Immun. 54:659-665. 31. Peros, W. J., I. Etherden, R. J. Gibbons, and Z. Skobe. 1985. Alteration of fimbriation and cell hydrophobicity by sublethal concentrations of tetracycline. J. Periodontal Res. 20:24-30. 32. Raja, R. H., G. Raucci, and M. Hook 1990. Peptide analogs to a fibronectin receptor inhibit attachment of Staphylococcus aureus to fibronectin-containing substrates. Infect. Immun. 58: 2593-2598. 33. Rosenberg, M., I. A. Buivids, and R. P. Ellen. 1991. Adhesion of Actinomyces viscosus to Porphyromonas (Bacteroides)gingivalis-coated hexadecane droplets. J. Bacteriol. 173:2581-2589. 34. Schwarz, S., R. P. Ellen, and D. A. Grove. 1987. Bacteroides gingivalis-Actinomyces viscosus cohesive interactions as measured by a quantitative binding assay. Infect. Immun. 55:23912397. 35. Slots, J. 1981. Enzymatic characterization of some oral and

INFECT. IMMUN.

36.

37.

38.

39. 40.

41. 42.

43.

44.

nonoral gram-negative bacteria with the API ZYM system. J. Clin. Microbiol. 14:288-294. Sojar, H. T., J.-Y. Lee, G. S. Bedi, M.-I. Cho, and R. J. Genco. 1991. Purification, characterization and immunolocalization of fimbrial protein from Porphyromonas (Bacteroides) gingivalis. Biochem. Biophys. Res. Commun. 175:713-719. Sojar, H. T., J.-Y. Lee, G. S. Bedi, M.-I. Cho, and R. J. Genco. 1991. Purification, characterization, and localization of a major membrane protein antigen from Porphyromonas (Bacteroides) gingivalis. Biochem. Int. 25:437-446. Stinson, M. W., G. G. Haraszthy, X. L. Zhang, and M. J. Levine. Submitted for publication. Stinson, M. W., K. Safulko, and M. J. Levine. 1991. Adherence of Porphyromonas (Bacteroides) gingivalis to Streptococcus sanguis in vitro. Infect. Immun. 59:102-108. Vordermeier, H.-M., P. Hoffmann, F. 0. Gombert, G. Jung, and W. G. Bessler. 1990. Synthetic peptide segments from the Escherichia coli porin OmpF constitute leukocyte activators. Infect. Immun. 58:2719-2724. Wang, S.-S. 1973. p-Alkoxybenzyl alcohol resin and p-alkoxybenzyloxycarbonyl-hydrazide resin for solid phase synthesis of protected peptide fragments. J. Am. Chem. Soc. 95:1328-1333. Weiss, E., M. Rosenberg, H. Judes, and E. Rosenberg. 1982. Cell-surface hydrophobicity of adherent oral bacteria. Curr. Microbiol. 7:125-128. Yoshimura, F., K. Takahashi, Y. Nodasaka, and T. Suzuki. 1984. Purification and characterization of a novel type of fimbriae from the oral anaerobe Bacteroides gingivalis. J. Bacteriol. 160:949-957. Zambon, J. J. 1990. Microbiology of periodontal disease, p. 147-160. In R. J. Genco, H. M. Goldman, and D. W. Cohen (ed.), Contemporary periodontics. The C.V. Mosby Co., St. Louis.