Immunological characterization and localization of a Porphyromonas gingivalis BApNA-hydrolyzing protease possessing hemagglutinating activity

ELSEVIER

FEMS Microbiology Letters 131 (1995) 211-217

Immunological characterization and localization of a Porphyromonas gingivalis BApNA-hydrolyzing protease possessing hemagglutinating activity Daisuke Hinode ayb,Kaname Masuda a, Masami Yoshioka a, Kiyoko Watanabe ‘, Toshio Umemoto ‘, Daniel Grenier b, Denis Mayrand b, Ryo Nakamura ap* ’ Department of Preventive Dentistry, School of Dentistry, The Universi@ of Tokushima, Tokushima 770, Japan b Groupe de Recherche en .&ologie Buccale, Facult.5 de Mkdecine Dentaire, Universiti Laval, Qu&ec, Canada GlK 7P4 ’ Department of Oral Microbiology, Kanagawa Dental College, Yokosuka, Kanagawa 238, Japan Received 3 May 1995; revised 3 July 1995; accepted 5 July 1995

Abstract A monoclonal antibody (mAb-PC) was produced against a BApNA-hydrolyzing protease possessing hemagglutinating activity (Pase-C) from Porphyromonas gingivalis. Other P. gingivalis BApNA-hydrolyzing enzymes (Pase-B and Pase-S) did not react with this antibody. By ELISA or SDS-PAGE and Western immunobfotting analysis, rnAb-PC recognized all P. gingivalis and P. endodontulis strains tested but did not recognize other members of the Porphyromonas genus nor other putative periodontopathogenic organisms. Pase-C, extracellular vesicles (ECV) and human strains of P. gingiualis showed two major immunoreactive bands (44 kDa and 40 kDa), whereas a different pattern was obtained with animal strains of P. gingioalis. Biotinylarginyl chloromethane, an irreversible inhibitor of trypsin-like proteases, did not affect the reactivity of Pase-C with mAb-PC on immunoblot. By reversed-phase electromnicroscopy following inununogold labeling, the antibody was shown to bind to the cell surface of P. gingivalis. mAb-PC inhibited the hemagglutinating activity of both P. gingivalis cells and ECV whereas a monoclonal antibody against LPS of P. gingiualis did not. These results suggest that Pase-C is located on the cell surface of P. gingivalis and may participate in erythrocyte binding. Keywords:

BApNA-hydrolyzing

protease; Porphyromonas

gingiualis;

Monoclonal antibody; Hemagglutinating activity; Periodontal

disease

1. Introduction Porphyromonas gingivalis is known as an important periodontopathogenic anaerobe because it is fre-

BA pNA = N-a-benzoyl-DL-arginine pAbbreviations: nitroanilide * Corresoondine author. Tel: +81 (886) 31 7337: Fax: +81 (886) 314i25.

”

0378-1097/95/$09.50

quently isolated from periodontal pockets of patients with adult periodontitis [l] and produces several hydrolytic enzymes that may be associated with periodontal destruction [1,2]. Among these enzymes, BApNA-hydrolyzing proteases are considered as a potential virulence factor. Proteases with a similar specificity were also detected in other suspected periodontopathogens such as Bacteroides forsythus [3] and Treponema dentico2a [4]. Taking advantage of this characteristic, a diagnostic test for periodontal

0 1995 Federation of European Microbiological Societies. All rights reserved

SSDI 0378-1097(95)00261-8

212

D. Hinode el al. / FEMS Microbiology Letters 131 (1995) 21 l-21 7

disease has been developed [5]. We have previously demonstrated that P. gingiualis produces three types of BApNA-hydrolyzing proteases, each of which has different biological activities [2,6,7]. One of these proteases, Pase-C which is a membrane-associated, high molecular mass cysteine protease, was very active towards collagen and immunoglobulins [2]. This protease also possessed hemagglutinating activity [7], which is thought to play a role in colonization of oral tissues by this bacteria. Several investigators have reported the purification and characterization of cysteine proteases having BApNAhydrolyzing activity such as the gingipain [8] and the porphypain [9]. Recently, Madden et al. [lo] and Pavloff et al. 1111reported the sequence of different BA pNA-hydrolyzing protease genes. The purpose of this study was to use a monoclonal antibody for characterizing Pase-C.

2. Materials and methods 2.1. Organisms and growth conditions The following anaerobic bacteria were used for this study: P. gingiualis 381, ATCC 33277, ATCC 49417, HW 24D-1, W83, W50, WSO/BEl, Chien 5B (isolated from dog), Chat 1 (isolated from cat), Porphyromonas salivosa NCK 11632, Porphyromonas endodontalis ATCC 35406, Porphyromonas asaccharolytica ATCC 25260, Actinobacillus actinomycetemcomitans ATCC 29522, Y4, Bacteroides forsythus ATCC 43037, Prevotella intermedia 163, and Treponema denticola ATCC 33520. The cells of B. forsythus were directly harvested from culture plates of blood agar medium. T. denticola was grown on a TYGVS medium by the method of Ohta et al. 141.All the other strains were grown in a brain heart infusion broth or Todd Hewitt broth supplemented with yeast extract (O.l%), vitamin K, (1 @g ml-‘), and hemin (10 pg ml-‘) to early stationary growth phase at 37°C in an anaerobic chamber.

2.2. Preparation BApNA-hydrolyzing

of extracellular vesicles and proteases from P. gingivalis

Extracellular vesicles (ECV) were prepared from P. gingiualis 381 [12]. Three different BApNA-hy-

drolyzing proteases, Pase-B (membrane-associated), Pase-C (membrane-associated) and Pase-S (soluble) were prepared from the culture supematant of P. gingiualis 381 as described previously 16,131. 2.3. Antibody preparation Monoclonal antibody against Pase-C was prepared as follows: BALB/c mice were immunized with the purified Pase-C, then the spleen cells were collected and fused with SP2/0 myeloma cells. Culture supematants of hybridoma cells were tested against Pase-C by a modified ELISA [14] and the positive cell line was selected. After the repeated subcloning by limiting dilution in 96-well tissue culture plates, a stable hybridoma producing high titer antibody was chosen and grown in large quantities in vitro. The monoclonal antibody against Pase-C (mAb-PC) was isolated from the supernatant by affinity chromatography using the Affi-gel protein A MAPS II kit (Bio-Rad Laboratories, Richmond, CA) and then divided into small aliquots and stored at -80°C until used. The immunoglobulin isotype of the prepared monoclonal antibody (mAb-PC) was identified as subclass IgGl using the Mouse Typer Sub-Isotyping Kit (Bio-Rad). 2.4. Detection of Pase-C in bacterial cells by ELZSA and immunogold labeling

The specificity of the mAb-PC was first analyzed by ELISA [14]. Cell suspensions of various oral anaerobes were adjusted to 0.13 at 630 nm in phosphate-buffered saline (PBS) and 200 ~1 of each sample was placed into the wells of a plate as antigens. For the immunolocalization study, P. gingiualis 381 cells fixed with 2.5% glutaraldehyde were placed in 1% bovine serum albumin in PBS for 5 min, and then the cells were removed and immediately incubated (1 h at room temperature) in a solution containing mAb-PC or mouse IgG (as a control) in PBS. Cells were then washed three times with PBS, and incubated in a 20 times-diluted solution of Immunogold conjugate EM goat anti-mouse IgG (British Bio Cell International, Cardiff, UK) at room temperature for 1 h prior to wash extensively. The cells were applied on Formvar-coated 200-mesh copper grids and blotted dry with a filter paper, stained

D. Hinode et al. / FEh4S Microbiology Letters 131 (1995) 21 l-21 7

with 2% uranyl acetate and examined by transmission electronmicroscopy (Hitachi H-800, Tokyo, Japan).

B 116Km

SDS-PAGE was performed using 12.5% resolving gels [15]. Bacterial cells were treated with 10% trichloroacetic acid (4”C/ovemight) in order to prevent proteolysis during the preparation of the cell lysate. Washed bacterial pellets were then mixed with reduced SDS-sample buffer (0.0625 M Tris-HCl buffer, pH 6.8, containing 1% SDS, 5% 2mercaptoethanol and 10% glycerol), sonicated and boiled for 5 min. Bacterial cells and ECV which had been electrophoresed were transferred onto a 0.2 pm nitrocellulose membrane. After being probed with mAb-PC for 1 h and washed, the membrane was incubated with horseradish peroxidase-conjugated goat anti-mouse IgG cl:3000 dilution, Bio-Rad) followed by development according to the procedures described in the Bio-Rad technical bulletin supplied with the assay kit. Purified proteases were treated either at 37°C for 30 min in non-reduced SDS-sample buffer or at 100°C for 5 min in reduced SDSsample buffer. Inactivated Pase-C with biotinylarginyl chloromethane (Bio-Arg-CH,Cl, Biosyn Ltd., Belfast, UK) was prepared according to the method described previously [7], followed by im-

66Kb

2

3

4

5

C

200Kh

2.5. SDS-PAGE and immunoblotting

1

213

42 K,

3OKb

1 2 3 4

1 2 3 4

Fig. 1. SDS-PAGE and Western immunoblotting analysis of P. gingivalis proteases with mAb-PC (panels B and Cl. Protein bands were visualized with Colloidal Gold Total Protein Stain (Bio-Rad) (panel A). Each sample was pretreated at 37°C for 30 min with non-reduced SDS-sample buffer (A and B) or sample of Pase-C were preincubated at 100°C for 5 min in reduced SDSsample buffer (Cl. Lane 1: Pase-B (0.2 &; Lane 2: Pase-C (0.2 pg); Lane 3: Pase-S (0.2 pg); Lane 4: acetone precipitate of the culture supematant (3 pg.l.

munoblotting with mAb-PC as described above. The immunoreactivity of inactivated Pase-C was compared to the non-treated Pase-C.

6

7

8

9

10111213

Fig. 2. SDS-PAGE and Western immunoblotting analysis of cells of different strains of Porphyromonas (10 pg) and ECV of P. gingivalis 381 strain (5 CL&) with mAb-PC. Samples were preincubated at 100°C for 5 min in reduced SDS-sample buffer. Lane 1: ECV; Lane 2: P. gingivalk 381; Lane 3: P. gingivalis ATCC 33277; Lane 4: P. gingivalis ATCC 49417; Lane 5: P. gingivalis HW 24D-1; Lane 6: P. gingivalis W83; Lane 7: P. gingivalis WSO; Lane 8: P. gingivalis Chat 1; Lane 9: P. gingivalis Chien .5B; Lane 10: P. gingivalis WSO/BEI; Lane 11: P. endodontalis ATCC 35406; Lane 12: P. salivosa NCTC 11632; Lane 13: P. asaccharolytica ATCC 25260.

214

D. Hinode et al. / FEMS Microbiology Letters 131 (1995) 21 I-21 7

2.6. Inhibition study of hemagglutination

Hemagglutinating activity was assayed using P. gingivalis 381 cells or ECV with washed sheep red blood cells. Bacterial suspensions were washed twice in PBS and adjusted to A,,, = 1.0. Four times diluted cell suspension or 10 pg ml-’ ECV were mixed with an equal volume of saline in round-bottom.microtiter plate, then 50 ~1 of 0.5% red blood cells was added and the plate was kept overnight at 4°C. Inhibition of hemagglutination by antibodies was measured under the above conditions. mAb-PC, mouse IgG (Miles Scientific, Naperville, IL) or purified monoclonal antibody against LPS of P. gingivalis (mAb-LPS, subclass IgGl, kind gift from C. Mouton, UniversitC Laval, Canada) was added instead of saline, and incubated at room temperature for 15 min before mixing with sheep red blood cells.

a non-pigmented mutant from W50. These reactive bands and the bands of ECV were similar to bands detected in the purified Pase-C. Another strong reactive band (27 kDa) was also detected in cells of P. gingiualis W83 and W50 (lane 6 and 7). Both P. gingivalis Chat 1 (lane 8) and Chien SB (lane 9) showed three strong immunoreactive bands (32, 30 and 23 kDa). P. salivosa and P. asaccharolytica did not react with mAb-PC whereas P. endodontalis reacted as a diffise band with an apparent molecular weight of 110 kDa (lane 111. In order to confirm the reactivity of mAb-PC with P. gingivalis cells and the cell surface location of Pase-C, reverse-phase immunoelectron microscopy was carried out after colloidal gold labeling of the antibody. The gold

3. Results Fig. 1 shows the immunoreactivity of purified P. proteases (Pase-B, -C and -S) and the acetone precipitate of a culture supernatant (starting material for protease purification) with the mAb-PC following SDS-PAGE and Western immunoblotting. The mAb-PC reacted with Pase-C (Fig. 1, panel B, lane 2) and the reactive band of the acetone precipitate was consistent with the protein band of Pase-C (panel B, lane 4). There is a drastic difference in the migration of Pase-C depending on the pretreatment of this enzyme. Two sharp bands were seen at 44 kDa and 40 kDa when the sample was pretreated under heating and reducing conditions (panel C>. However, Pase-B and PaseS did not react with the mAb-PC. All P. gingivalis strains showed strong reactivity with mAb-PC by ELISA, whereas other anaerobes possessing BApNA-hydrolyzing activity (B. forsythus and T. denticola) did not react (data not shown). Fig. 2 shows the immunoblotting pattern of cells of different members of the Porphyromonas genus, solubilized by heating in reduced SDS-sample buffer. Although the molecular size of the minor bands were variable depending on the strains, two common strong immunoreactive bands were seen in all human strains of P. gingivalis except for P. gingivalis WSO/BEl, gingivalis BApNA-hydrolyzing

Fig. 3. Reaction of P. gingiualis 381 with mAb-PC (A) or mouse IgG (B). Immunogold conjugate EM goat anti-mouse IgG was used to visualize mAb-PC on the bacterial surface by transmission electromnicroscopy. Bar = 200 nm.

D. Hinode et al. / FEMS Microbiology Letters I31 (1995) 21 l-21 7

a

215

be due to the level of hemagglutinating activity present in the preparations. No inhibition of hemagglutination was observed with mouse IgG and mAbLPS.

b 4. Discussion d e

Fig. 4. Inhibition of hemagglutination by the mAb-PC. Preparations for hemagglutination assay in panels A and B were P. gingivalis 381 cells and P. gingivalis 381 ECV, respectively. The number on the column indicates the amount of immunoglobulin protein as follows: 1: 5.0 pg/well; 2: 1.0 pg/well; 3: 0.2 pg/well. The letter on the row indicates the immunoglobulin used as putative inhibitor. a: negative control (no P. gingivalis preparation); b: positive control (no immunoglobulin); c: mAb-PC; d: non-immune IgG, e: mAb-LPS.

beads were observed on the bacterial cell surface, indicating that the Pase-C exists at the cell surface membrane of the bacterium (Fig. 3A). No beads attached to the cells in the absence of mAb-PC (Fig. 3B). Pase-C possessed strong degradative activity of IgG as shown previously [2]. For this reason, we chose the inhibition assay by immunoblotting using Bio-Arg-CH,Cl, an irreversible inhibitor of trypsinlike proteases. No differences in immunoreactive bands (mAb-PC) were found for Pase-C inactivated with Bio-Arg-CH,Cl compared to a non-treated Pase-C (control) (data not shown). To determine the inhibition of hemagglutination by mAb-PC, two bacterial fractions (cells and ECV) were used because bacterial cells may possess other factors different from Pase-C, responsible for hemagglutination. mAb-PC exhibited a typical inhibition of hemagglutination on both preparations (Fig. 4). Under the assay performed, hemagglutination of cells was inhibited by 1 kg of mAb-PC whereas hemagglutination of ECV was inhibited by 5 pg of mAb-PC. This difference in amount of mAb-PC required for the inhibition of the two preparations was considered to

Although numerous reports on monoclonal antibodies generated against components of P. gingivalis have been published, only one concerned proteases [16]. We have recently shown, using a polyclonal antibody against Pase-S, that Pase-C was immunologically distinct from Pase-B and Pase-S, both of which were clostripain-like proteases from P. gingivalis [7]. The monoclonal antibody (mAb-PC) which we have produced in this study reacted with all strains of P. gingivalis and was found to be highly specific to Pase-C. These results support previous data indicating an immunological distinction of the three BApNA-hydrolyzing proteases and also suggest that all strains of P. gingivalis produce a similar Pase-C protease. Lantz et al. [17] have reported the surface location of a cysteine protease with an apparent molecular weight of 150 kDa and possessing arginyl/lysyl peptidase activity, which is similar to the activity of Pase-C. Pase-C also possesses a high molecular mass (> 120 kDa) when the sample was pretreated with non-reduced SDS sample buffer, then loaded and electrophoresed. The location of Pase-C was demonstrated in ECV as well as on the cell surface of P. gingivalis. These data suggest that Pase-C may be similar to the protease studied by Lantz et al. [17]. Surprisingly, SDS-PAGE immunoblotting pattern of P. gingiualis strains from animals was different from that of human strains of P. gingivalis. We demonstrated previously that beagle dogs had Porphyromonas species analogous to P. gingivalis in their oral cavities, although there are some biological and serological differences between them [18]. The molecule of Pase-C may also be different in the two biotypes. P. endodontalis, originally isolated from endodontic infection sites, cross-reacted with mAbPC. Wallace et al. [16] showed that a monoclonal antibody against a thiol-activated BApNA-hydrolyzing protease with an apparent molecular weight of 43

216

D. Hinode et al. / FEMS Microbiology Letters 131 (1995) 21 l-21 7

kDa failed to react with the same strain of P. endodontalis that we used. Therefore, their monoclonal antibody may recognize a different epitope from our mAb-PC or a protease different from ours, The 44-kDa immunoreactive band seen in Fig. 1, panel C, corresponded to an immunoreactive band detected in both Pase-C and P. gingiualis 381 cells using a polyclonal antibody raised against a recombinant antigen which was produced on the basis of the first 13 N-terminal amino acid sequence of the putative hemagglutinin of P. gingivalis [8] (data are not shown, this polyclonal antibody was a kind gift from F. Yoshimura, Aichigakuin University, Japan). Hoover and Yoshimura [19] recently showed that transposon-induced pigment-deficient mutants of P. gingiualis which exhibited simultaneous deficiencies in BApNA-hydrolyzing activity and hemagglutination activity, lacked a reaction with the polyclonal antibody against putative hemagglutinin as described above. Interestingly, P. gingiualis WSO/BEl, which is also an avirulent non-pigmented variant and lacks BApNA-hydrolyzing activity and hemagglutinating activity, did not react with mAb-PC whereas the parent strain (P. gingiualis W50) exhibited the immunoreaction. Previously, our group [20] and another group [21] have suggested that BA pNA-hydrolyzing activity and hemagglutinating activity are part of the same molecule and that the same active site in this molecule participates in the substrate binding and the erythrocyte binding. Our mAb-PC did not recognize the active site in the protease because no apparent inhibition of immunoreactivity was observed when the protease was treated with Bio-Arg-CH,Cl, an active site-directed inhibitor. On the other hand, the mAb-PC inhibited the hemagglutinating activity of P. gingiualis cells as well as of ECV, whereas mAb-LPS and non-immune IgG did not (Fig. 4). This suggests that the major hemagglutinating activity of P. gingivalis is related to Pase-C. From these results, it can be speculated that hemagglutination inhibition exhibited by mAb-PC was due to an immunological reaction in an area close to the active site. The active site of the protease may not directly react with the monoclonal antibody but the antibody may act as a physical barrier to prevent the contact between erythrocytes and the active binding site. However, it is also possible to consider that the

epitopes involved in protease and hemagglutinating activity are present at a different site on the bifunctional molecule, as suggested by other groups [lO,ll]. Further studies will be needed to confirm these hypotheses. Mouton and co-workers [22,23] have reported that the hemagglutinating adhesin (HAAg2) of P. gingiualis is present in outer membrane preparations as a complex of polypeptides with apparent molecular weight of 49 and 43 kDa. Their monoclonal antibodies against the HA-Ag2 also allowed inhibition of hemagglutination. mAb-PC may also recognize HA-Ag2 although the estimated molecular weights of HA-Ag2 and Pase-C are slightly different. The properties of several BApNA-hydrolyzing enzymes are not necessarily the same. For instance, the enzyme of B. forsythus only possessed activity against small synthetic substrates [3] whereas enzymes of P. gingiualis possessed a large spectrum of activities [1,2]. Clinical trials to utilize the P. gingivalis BApNA-hydrolyzing activity for the diagnosis of periodontal disease is being developed by measuring this activity in subgingival plaque and/or periodontal fluids [5]. The specificity of mAb-PC could distinguish Pase-C from other BApNA-hydrolyzing proteases of P. gingivalis and from trypsin-like proteases produced by other oral bacteria or host cells in clinical samples. Therefore, rnAb-PC could be very useful in determining the disease activity through monitoring of the Pase-C. A combination assay involving the determination of both Pase-C and BApNA-hydrolyzing activities might be effective for a more detailed diagnosis of periodontitis. Acknowledgements This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan, by the Medical Research Council of Canada and by the Budget Special de la Recherche, Fact& des Sciences et de Genie, Universite Lava]. References 111 Mayrand, D. and Holt, SC. (1988) Biology of asaccharolytic black-pigmented Bacteroides species. Microbial. Rev. 52, 134-152.

D. Hinode et al. / FEMS Microbiology [2] Nakamura, R., Hinode, D., Terai, H. and Morioka, M. (1991) Extracellular enzymes of Porphyromonas (Bacteroides) gingiualis in relation to periodontal destruction. In: Periodontal Disease: Pathogens and Host Immune Responses (Hamada, S., Holt, S.C. and McGhee, J.R., Eds.), pp. 129-141, Quintessence Publishing Co., Tokyo. [3] Grenier, D. (1995) Characterization of the trypsin-like activity of Bacteroides forsythus. Microbiology 141, 921-926. [4] Ohta, K., Makinen, K.K. and Loesche, W.J. (1986) Purification and characterization of an enzyme produced by Treponema denticola capable of hydrolyzing synthetic trypsin substrates. Infect. Immun. 53, 213-220. [.5] Loesche, W.J., Bretz, W.A., Kerschensteiner, D., Stoll, J., Socransky, S.S., Hujoel, P. and Lopatin, D.E. (1990) Development of a diagnostic test for anaerobic periodontal infections based on plaque hydrolysis of benzoyl-DL-argininenaphthylamide. J. Clin. Microbial. 28, 1551-1559. [6] Hinode, D., Hayashi, H. and Nakamura, R. (1991) Purification and characterization of three types of proteases from culture supematants of Porphyromonas gingiualis. Infect. Immun. 59, 3060-3068. [7] Hinode, D., Masuda, K., Yoshioka, M., Hayashi, H., Nakamura, R., Grenier, D. and Mayrand, D. (1995) Biological and antigenic characterization of three BApNA-hydrolyzing proteases from the culture supematant of Porphyromonas gingiualis. Oral Microbial. Immunol. (in press). [S] Pike, R., McGraw, W., Potempa, J., and Travis, J. (1994) Lysine-and arginine-specific proteinases from Porphyromonas gingiualis. J. Biol. Chem. 269,406-411. [9] Ciborowski, P., Nishikata, M., Allen, R.D. and Lantz, M.S. (1994) Purification and characterization of two forms of a high-molecular-weight cysteine proteinase (Porphypain) from Porphyromonas gingiualis. J. Bacterial. 176, 4549-4557. [lo] Madden, T.E., Clark, V.L. and Kuramitsu, H.K. (1995) Revised sequence of Porphyromonas gingiualis PrtT cysteine protease/hemagglutinin gene: homology with streptococcal pyrogenic exotoxin B/streptococcal proteinase. Infect. Immun. 63, 238-247. [ll] Pavloff, N., Potempa, J., Pike, R.N., Prochazka, V., Kiefer, M.C. Travis, J. and Barr, P.J. (1995) Molecular cloning and structural characterization of the Arg-gingipain proteinase of Porphyromonas gingivalis. J. Biol. Chem. 270, 1007-1010. [12] Grenier, D. and Mayrand, D. (1987) Functional characterization of extracellular vesicles produced by Bacteroides gingiualis. Infect. Immun. 55, 111-117.

Letters 131 (1995) 21 l-21 7

217

[13] Hinode, D., Masuda, K., Yoshioka, M., Hayashi, H. and Nakamura, R. (1994) Purification of an extracellular soluble protease possessing BApNA hydrolytic activity from culture supernatants of Porphyromonas gingiualis. J. Dent. Health 44, 308-314. [14] Ebersole, J.L., Frey, D.E., Taubman, M.A., Smith, D.J., Socransky, S.S. and Tanner, A.C.R. (1984) Serological identification of oral Bacteroides spp. by enzyme-linked immunosorbent assay. J. Clin. Microbial. 19, 639-644. [15] Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. [16] Wallace, A., Rigg, G., Hyman, S.C., James, R. and Roberts, IS. (1992) Generation and characterization of a monoclonal antibody specific for the major thiol-activated cysteine proteinase of Porphyromonas gingiualis W83. L&t. Appl. Microhiol. 15, 202-206. [17] Lantz, MS., Allen, R.D. Ciborowski, P. and Holt, S.C. (1993) Purification and immunolocalization of a cysteine protease from Porphyromonas gingiualis. J. Periodont. Res. 28, 467-469. [18] Yamasaki, T., Nagata, A., Kiyoshige, T., Sato, M. and Nakamura, R. (1990) Black-pigmented, asaccharolytic Bacteroides species resembling Porphyromonas gingiualis (Bacteroides gingiualis) from beagle dogs. Oral Microbial. Immunol. 5, 332-335. [19] Hoover, C.I. and Yoshimura, F. (1994) Transposon-induced pigment-deficient mutants of Porphyromonas gingiualis. FEMS Microbial. Lett. 124, 43-48. [20] Hayashi, H., Nagata, A., Hinode, D., Sato, M. and Nakamura, R. (1992) Survey of a receptor protein in human erythrocytes for hemagglutinin of Porphyromonas gingiualis. Oral Microbial. Immunol. 7, 204-211. [21] Nishikata, M. and Yoshimura, F. (1991) Characterization of Porphyromonas (Bacteroides) gingiualis hemagglutinin as a protease. Biochem. Biophys. Res. Commun. 178, 336-342. [22] Mouton, C., Bouchard, D., Deslauriers, M. and Lamonde, L. (1989) lmmunochemical identification and preliminary characterization of a nonfimbrial hemagglutinating adhesin of Bacteroides gingiualis. Infect. Immun. 57, 566-573. [23] Deslauriers, M. and Mouton, C. (1992) Epitope mapping of hemagglutinating adhesin HA-Ag2 of Bacteroides (Porphyromonas) gingivalis. Infect. Immun. 60, 2791-2799.