Portal de Periódicos da CAPES

Determination and Characterization of the Hemagglutinin-associated Short Motifs Found in Porphyromonas gingivalis Multiple Gene Products

1999; Elsevier BV; Volume: 274; Issue: 8 Linguagem: Inglês

10.1074/jbc.274.8.5012

1083-351X

Yasuko Shibata, Mitsuo Hayakawa, Hisashi Takiguchi, Teruaki Shiroza, Yoshimitsu Abiko,

Antimicrobial Peptides and Activities

Porphyromonas gingivalis is a Gram-negative anaerobic bacterial species implicated as an important pathogen in the development of adult periodontitis. In our studies ofP. gingivalis and ways to protect against periodontal disease, we have prepared the monoclonal antibody mAb-Pg-vc and its recombinant antibody, which are capable of inhibiting the hemagglutinating activity of P. gingivalis (Shibata, Y., Kurihara, K., Takiguchi, H., and Abiko, Y. (1998) Infect. Immun. 66, 2207–2212). To clarify the antigenically related hemagglutinating domains, we attempted to determine the minimum motifs responsible for P. gingivalis hemagglutinin. Initially, the 9-kilobase EcoRI fragment encoding the 130-kDa protein was cloned from the P. gingivalis chromosome using mAb-Pg-vc. Western blot analysis of nested deletion clones, the competition experiments using synthetic peptides, and the binding assay of the phage-displayed peptides using the mAb-Pg-vc allowed us to identify the minimum motifs, PVQNLT. Furthermore, the presence of multi-gene family coding for this epitope was confirmed via Southern blot analysis and PCR using the primers complementary to the domain corresponding to this epitope. It is suggested that the hemagglutinin-associated motif may be PVQNLT and that the gene families specifying this motif found in P. gingivalis chromosome encode many hemagglutinin and/or hemagglutinin-related proteases. Porphyromonas gingivalis is a Gram-negative anaerobic bacterial species implicated as an important pathogen in the development of adult periodontitis. In our studies ofP. gingivalis and ways to protect against periodontal disease, we have prepared the monoclonal antibody mAb-Pg-vc and its recombinant antibody, which are capable of inhibiting the hemagglutinating activity of P. gingivalis (Shibata, Y., Kurihara, K., Takiguchi, H., and Abiko, Y. (1998) Infect. Immun. 66, 2207–2212). To clarify the antigenically related hemagglutinating domains, we attempted to determine the minimum motifs responsible for P. gingivalis hemagglutinin. Initially, the 9-kilobase EcoRI fragment encoding the 130-kDa protein was cloned from the P. gingivalis chromosome using mAb-Pg-vc. Western blot analysis of nested deletion clones, the competition experiments using synthetic peptides, and the binding assay of the phage-displayed peptides using the mAb-Pg-vc allowed us to identify the minimum motifs, PVQNLT. Furthermore, the presence of multi-gene family coding for this epitope was confirmed via Southern blot analysis and PCR using the primers complementary to the domain corresponding to this epitope. It is suggested that the hemagglutinin-associated motif may be PVQNLT and that the gene families specifying this motif found in P. gingivalis chromosome encode many hemagglutinin and/or hemagglutinin-related proteases. It is recognized that the adherence of bacteria to host tissues is a prerequisite for colonization and one of the causative factors of bacterial pathogenesis. Porphyromonas gingivalis is a Gram-negative anaerobic bacteria that is isolated primarily from infectious periodontal pockets and considered to be the major pathogen for adult periodontitis (1Slots J. Genco R.J. J. Dent. Res. 1984; 63: 412-421Crossref PubMed Scopus (655) Google Scholar, 2Slots J. Bragd L. Wikström M. Dahlén G. J. Clin. Periodontol. 1986; 13: 570-577Crossref PubMed Scopus (434) Google Scholar). Colonization of this bacterium in gingival tissues is critical in the pathogenic process of periodontal disease resulting in tissue destruction. Therefore, a number of molecules including fimbriae (3Isogai H. Isogai E. Yoshimura F. Suzuki T. Kagota W. Takano K. Arch. Oral Biol. 1988; 33: 479-485Crossref PubMed Scopus (118) Google Scholar), potential molecular adhesins such as hemagglutinins (4Boyd J. McBride B.C. Infect. Immun. 1984; 45: 403-409Crossref PubMed Google Scholar), and lipopolysaccharides (5Holt S.C. Bramanti T.E. Crit. Rev. Oral Biol. Med. 1991; 2: 177-281Crossref PubMed Scopus (251) Google Scholar) responsible for colonization have been identified as the virulence factors. On the other hand, the cysteine proteases from this pathogen have been extensively investigated because these enzymes could play an important role in tissue destruction, activating host proenzymes, neutralizing host defense mechanisms, and providing essential amino acids for growth as well (6Kuramitsu H.K. Yoneda M. Madden T. Adv. Dent. Res. 1995; 9: 37-40Crossref PubMed Scopus (24) Google Scholar). Some of the enzymes are specific for cleavage of peptide bonds containing arginine (7Chen Z. Potempa J. Polanowski A. Wikström M. Travis J. J. Biol. Chem. 1992; 267: 18896-18901Abstract Full Text PDF PubMed Google Scholar, 8Nakayama K. Kadowaki T. Okamoto K. Yamamoto K. J. Biol. Chem. 1995; 270: 23619-23626Crossref PubMed Scopus (219) Google Scholar, 9Aduse-Opoku J. Muir J. Slaney J.M. Rangarajan M. Curtis M.A. Infect. Immun. 1995; 63: 4744-4754Crossref PubMed Google Scholar, 10Pavloff N. Potempa J. Pike R.N. Prochazka V. Kiefer M.C. Travis J. Barr P.J. J. Biol. Chem. 1995; 270: 1007-1010Abstract Full Text Full Text PDF PubMed Scopus (166) Google Scholar, 11Slakeski N. Cleal S.M. Reynolds E.C. Biochem. Biophys. Res. Commun. 1996; 224: 605-610Crossref PubMed Scopus (38) Google Scholar, 12Okamoto K. Misumi Y. Kadowaki T. Yoneda M. Yamamoto K. Ikehara Y. Arch. Biochem. Biophys. 1995; 316: 917-925Crossref PubMed Scopus (83) Google Scholar), others are specific for lysine residues (13Pike R. McGraw W. Potempa J. Travis J. J. Biol. Chem. 1994; 269: 406-411Abstract Full Text PDF PubMed Google Scholar, 14Okamoto K. Kadowaki T. Nakayama K. Yamamoto K. J. Biochem. (Tokyo). 1996; 120: 398-406Crossref PubMed Scopus (99) Google Scholar), whereas still others exhibit specificity for both amino acid residues (15Bedi G.S. Williams T. J. Biol. Chem. 1994; 269: 599-606Abstract Full Text PDF PubMed Google Scholar, 16Barkocy-Gallagher G.A. Han N. Patti J.M. Whitlock J. Progulske-Fox A. Lantz M.S. J. Bacteriol. 1996; 178: 2734-2741Crossref PubMed Google Scholar). Moreover, it has been proposed that the adherence of P. gingivalis to erythrocytes (13Pike R. McGraw W. Potempa J. Travis J. J. Biol. Chem. 1994; 269: 406-411Abstract Full Text PDF PubMed Google Scholar), fibrinogen (17Lantz M.S. Allen R.D. Vail T.A. Switalski L.M. Hook M. J. Bacteriol. 1991; 173: 495-504Crossref PubMed Google Scholar), fibronectin (18Lantz M.S. Allen R.D. Duck L.W. Blume J.L. Switalski L.M. Hook M. J. Bacteriol. 1991; 173: 4263-4270Crossref PubMed Google Scholar), collagenous substrata (19Naito Y. Gibbons R.J. J. Dent. Res. 1988; 67: 1075-1080Crossref PubMed Scopus (58) Google Scholar), and other bacteria (20Slots J. Gibbons R.J. Infect. Immun. 1978; 19: 254-264Crossref PubMed Google Scholar) is mediated, at least in part, by such proteases at the cell surface. Several investigators have indicated interesting results for the correlation of the protease and hemagglutinin activities inP. gingivalis (8Nakayama K. Kadowaki T. Okamoto K. Yamamoto K. J. Biol. Chem. 1995; 270: 23619-23626Crossref PubMed Scopus (219) Google Scholar, 21Nishikata M. Yoshimura F. Biochem. Biophys. Res. Commun. 1991; 178: 336-342Crossref PubMed Scopus (61) Google Scholar, 22Potempa J. Pavloff N. Travis J. Trend. Microbiol. 1995; 3: 430-434Abstract Full Text PDF PubMed Scopus (81) Google Scholar, 23Yoneda M. Kuramitsu H.K. Oral Microbiol. Immunol. 1996; 11: 129-134Crossref PubMed Scopus (18) Google Scholar): arginine-specific cysteine protease (arginine-gingipain)-deficient mutants exhibited decreased hemagglutinating activity, which suggests notable properties of this protease (8Nakayama K. Kadowaki T. Okamoto K. Yamamoto K. J. Biol. Chem. 1995; 270: 23619-23626Crossref PubMed Scopus (219) Google Scholar, 23Yoneda M. Kuramitsu H.K. Oral Microbiol. Immunol. 1996; 11: 129-134Crossref PubMed Scopus (18) Google Scholar). Recent enzymatic and molecular cloning analysis directly revealed that these proteases are composed of two distinct domains: one for proteolytic activity and the other for hemagglutinating activity (9Aduse-Opoku J. Muir J. Slaney J.M. Rangarajan M. Curtis M.A. Infect. Immun. 1995; 63: 4744-4754Crossref PubMed Google Scholar, 10Pavloff N. Potempa J. Pike R.N. Prochazka V. Kiefer M.C. Travis J. Barr P.J. J. Biol. Chem. 1995; 270: 1007-1010Abstract Full Text Full Text PDF PubMed Scopus (166) Google Scholar, 14Okamoto K. Kadowaki T. Nakayama K. Yamamoto K. J. Biochem. (Tokyo). 1996; 120: 398-406Crossref PubMed Scopus (99) Google Scholar, 16Barkocy-Gallagher G.A. Han N. Patti J.M. Whitlock J. Progulske-Fox A. Lantz M.S. J. Bacteriol. 1996; 178: 2734-2741Crossref PubMed Google Scholar, 22Potempa J. Pavloff N. Travis J. Trend. Microbiol. 1995; 3: 430-434Abstract Full Text PDF PubMed Scopus (81) Google Scholar, 24Madden T.E. Clark V.L. Kuramitsu H.K. Infect. Immun. 1995; 63: 238-247Crossref PubMed Google Scholar). The presence of hemagglutinating activity on the P. gingivalis cell surface was first reported by Okuda and Takazoe (25Okuda K. Takazoe I. Arch. Oral Biol. 1974; 19: 415-416Crossref PubMed Scopus (54) Google Scholar). Indeed, the specific genes for P. gingivalishemagglutinin, such as hagA, which have four repeat units of the hemagglutinin domain without the protease activity, have been isolated (26Han N. Whitlock J. Progulske-Fox A. Infect. Immun. 1996; 64: 4000-4007Crossref PubMed Google Scholar). It is reasonable to expect that some hemagglutinin molecules or other hemagglutinin domains encoded by many protease genes in P. gingivalis possess the ability to degrade a broad range of host proteins. Additionally, protoheme is an absolute requirement for the growth of P. gingivalis (27Chu L. Bramanti T.E. Ebersole J.L. Holt S.C. Infect. Immun. 1991; 59: 1932-1940Crossref PubMed Google Scholar), and it is probably derived from erythrocytes in the natural niche of the organism (28Shah H.N. Gharbia S.E. FEMS Microbiol. Lett. 1989; 61: 213-218Crossref Scopus (44) Google Scholar). Thus, the hemagglutinin molecule may be particularly important for the organism not only for the attachment to the gingival tissues but also to agglutinate and lyse erythrocytes to survive in vivo. Although it is now well known that genes coding for hemagglutinins and other proteases may share the domain specifying hemagglutinin and hence consisting of a multigene family, the minimum motif responsible for hemagglutinating activity in this family has only been speculated by some investigators (29Curtis M.A. Aduse-Opoku J. Slaney J.M. Rangarajan M. Booth V. Cridland J. Shepherd P. Infect. Immun. 1996; 64: 2532-2539Crossref PubMed Google Scholar, 30Aduse-Opoku J. Slaney J.M. Rangarajan M. Muir J. Young K.A. Curtis M.A. J. Bacteriol. 1997; 179: 4778-4788Crossref PubMed Google Scholar). Our approach to protect from periodontal diseases is to develop a passive immunization system whereby the colonization of P. gingivalis cells onto human host tissues could be blocked. Toward this goal, we first prepared a monoclonal antibody (mAb-Pg-vc) and a recombinant single chain variable fragment antibody that inhibited the hemagglutinating activity of P. gingivalis (31Shibata Y. Kurihara K. Takiguchi H. Abiko Y. Infect. Immun. 1998; 66: 2207-2212Crossref PubMed Google Scholar). Using these antibodies, P. gingivalis genomic library was screened, and the gene coding for the 130-kDa protein was isolated. The present investigation will describe the detailed properties of the 130-kDa protein, and its possible role in agglutination with erythrocytes will be discussed. P. gingivalis 381 was grown in Todd-Hewitt broth (Difco Laboratories, Detroit, MI) supplemented with hemin (0.2 μg/ml) and vitamin K1 (5 μg/ml) in an anaerobic atmosphere (80% N2, 10% H2, 10% CO2) for 24–48 h. Escherichia coli XL-1 Blue (Stratagene) was grown on LB (1% trypton, 0.5% yeast extract, 0.5% NaCl). For plasmid selection, 60 mg/liter of ampicillin was added to LB agar plates. Vesicles were isolated basically by the method of Grenier and Mayrand (32Grenier D. Mayrand D. Infect. Immun. 1987; 55: 111-117Crossref PubMed Google Scholar), with a slight modification as previously reported (33Hiratsuka K. Abiko Y. Hayakawa M. Ito T. Sasahara T. Takiguchi H. Arch. Oral Biol. 1992; 37: 717-724Crossref PubMed Scopus (50) Google Scholar). P. gingivalis 381 genomic DNA was prepared, digested with EcoRI, and inserted into the EcoRI site of λzapII. The constructed phage library was screened by mAb-Pg-vc, and the initial phage clone, HEM9, was isolated. This clone produced an immunoreactive 130-kDa protein (130-kDa hemagglutinating domain; 130k-HMGD). To obtain the minimum coding region for a 130k-HMGD, phage HEM9 DNA was digested with SacI. DNA fragments were subcloned into pBluescript II KS(+) vector (Stratagene), and the mAb-Pg-vc-reactive E. coli transformant harboring pHEM6 was isolated. For forward DNA sequencing, pHEM6 was digested with ApaI andEcoRI and incubated with exonucleases III and mung bean nuclease. Nested deletion clones were prepared following recircularization and transformation into E. coli cells. For reverse DNA sequencing, the SmaI fragment of pHEM6 was subcloned, and the nested deletion clones were constructed in the same way. DNA sequencing was performed with the Taq Dyedeoxy system (Applied Biosystems, Inc.) and analyzed on an ABI 373A DNA sequencer. Complete coverage of both strands and the contiguous linkage were achieved by Long Ranger gel solution (FMC Bioproducts) using the LI-COR 4000L infrared automated sequencer (ALOKA). The DNA sequence data were analyzed using the DNASIS programs (TaKaRa) and searched for in the DNA data bank. Sequence grade plasmid DNA was purified by the automatic DNA isolation system PI-50 (KURABO). Preparation of the rabbit polyclonal antibody raised against P. gingivalis vesicles and its monoclonal derivative, mAb-Pg-vc, has been described previously (31Shibata Y. Kurihara K. Takiguchi H. Abiko Y. Infect. Immun. 1998; 66: 2207-2212Crossref PubMed Google Scholar), both of which neutralize the hemagglutinating activity ofP. gingivalis vesicles. Recombinant 130k-HMGD was subjected to SDS-polyacrylamide gel electrophoresis, and the protein was purified by excising the gel corresponding to the recombinant 130-kDa protein band. A rabbit serum antibody against the recombinant 130k-HMGD was prepared after three intramuscular injections of the purified protein. Prior to usage, the rabbit serum was absorbed with sonicated E. coli cells to minimize nonspecific immunoreaction. The Ph.D.™ ligand screening system (7 peptides) was used to identify the epitopes recognized by mAb-Pg-vc. The sterile 65 × 15 mm polystyrene Petri dishes were coated with 1.5 ml of 0.01 m NaHCO3containing mAb-Pg-vc and stored overnight at 4 °C in humidified container. After the antibody solution was poured off, blocking buffer (0.1 m NaHCO3, pH 8.6, 5 mg/ml bovine serum albumin, 0.02% NaN3) was added and incubated for 1 h at 4 °C. Then these dishes were washed with TBST buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 0.5% Tween 20) six times, and 2 × 1011 phage in 1 ml of TBST buffer was added and rocked gently for 1 h at room temperature. These dishes were vigorously washed 10 times with TBST, and the bound phage was then eluted by lowering the pH with 1 ml of 0.2 mglycine-HCl (pH 2.2) containing 1 mg/ml bovine serum albumin. After neutralization with 150 μl of 1 m Tris-HCl (pH 9.1), the phages were amplified by infecting with E. coli NM522. Recovered phage particles were subjected to a repeated biopanning procedure in the same manner. After the third round of biopanning, a total of 10 plaques were selected, and the phages were amplified. To obtain the single-strand phage DNA, each culture supernatant was added to polyethylenglycol/NaCl, and the recovered pellets were incubated with iodide buffer (10 mm Tris-HCl, pH 8.0, 1 mm EDTA, 4 m NaI) to remove the phage protein. For sequencing, the primer (5′-CCC TCA TAG TTA GCG TAA CG) was used. To quantify the amounts of phages bound to the coated mAb-Pg-vc, enzyme-linked immunosorbent assay was carried out using horseradish peroxidase-conjugated anti-M13 antibody. The hemagglutinating activity of the vesicle fraction ofP. gingivalis 381 was assayed with washed rabbit erythrocytes in round-bottomed microtiter plates. 70 μl of vesicle solution (0.625 μg/ml) and 20 μl of antibody solution were transferred into microtiter wells and incubated for 30 min at room temperature. Then, 100 μl of 2% rabbit erythrocyte was added and incubated for 2 h at room temperature. The inhibition of hemagglutinating activity induced by adding synthetic peptides was carried out under the same conditions as mentioned above. To clone other genes coding for the epitope sites recognized by mAb-Pg-vc, the primers F (5′-TCC AAT GAA TTT GCT CCT) and R (5′-ATT TTC GAA TGA TTC GGA) were designed. These two primers were able to amplify the region corresponding to the base positions 1298–1447 in pHEM6. PCR 1The abbreviations PCRpolymerase chain reactionkbkilobase(s)ORFopen reading frameARSantibody recognition sitebpbase pair(s) was carried out by employing Pfu turbo (Stratagene) DNA polymerase and P. gingivalis genomic DNA as a template, and a reaction was performed for 30 cycles (94 °C for 1 min, 55 °C for 2 min, and 72 °C for 2 min). The PCR products were precipitated with 0.1 m of LiCl, 20 μg of glycogen, and 70% ethanol, ligated into the pCR-Script SK(+) vector (Stratagene), cloned, and then sequenced. polymerase chain reaction kilobase(s) open reading frame antibody recognition site base pair(s) P. gingivalisgenomic DNA was digested with endonucleases, and the fragments were separated in a 0.8% agarose gel. After alkaline denaturation, DNA fragments were transferred onto a Hybond N+ membrane (Amersham). Southern blot analysis was carried out using ECL labeling and indirect detection kits (Amersham) according to supplier's recommendation. For SDS-polyacrylamide gel electrophoresis, P. gingivalis whole cells, vesicles, and other recombinant samples were solubilized in 0.1% SDS solution, and subjected to 12% acrylamide-bisacrylamide gels as described by Laemmli (34Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (212360) Google Scholar). Western blots were prepared by electroblotting samples onto nitrocellulose for 1.5 h at 6 V in a transfer buffer (50 mm Tris-glycine, 20% methanol) using a semi-dry transfer cell system (Bio-Rad). Cloning of the gene coding for 130k-HMGD and isolation of chimeric plasmid, pHEM6, is outlined under "Experimental Procedures." To examine the antigenic properties of the recombinant protein produced by HEM6, Western blot analysis was carried out. As shown in Fig.1, both the rabbit polyclonal antibody and mAb-Pg-vc recognized the recombinant protein whose molecular mass was 130 kDa (A and B, lanes 3, respectively). This indicated that the subcloned 4.6-kb SacI fragment originating from the recombinant phage, HEM9, retained the same gene that encodes a recombinant 130-kDa protein responsible for hemagglutinin. The rabbit polyclonal antibody raised against the 130k-HMGD exhibited the superimposable Western blot profile (Fig.1 C) as that of mAb-Pg-vc (Fig. 1 B), suggesting that these two antibodies recognized the same epitopes of the hemagglutinin-related molecules of P. gingivalis. To fully characterize the 130k-HMGD, the 4.6-kb SacI fragment was completely sequenced. The GC composition of this fragment was 48%, a typical value for this organism (data not shown). Fig.2 (A and B) shows the gene organization and both nucleotide and deduced amino acid sequences, respectively. It was revealed that the 4.6-kbSacI fragment contained two open reading frames (ORFs).Figure 2The gene organization (A) and nucleotide and deduced amino acid sequences (B) of the 4.6-kb EcoRI-SacI fragment.In B, the two potential epitope sites, ARS-I and ARS-II, recognized by mAb-Pg-vc are boxed. The proposed hemoglobin-binding domain (39Nakayama K. Ratnayake D.B. Tsukuba T. Kadowaki T. Yamamoto K. Fujimura S. Mol. Microbiol. 1998; 27: 51-61Crossref PubMed Scopus (103) Google Scholar) is underlined.View Large Image Figure ViewerDownload (PPT) The upstream ORF contained no ATG initiation codon, which is preceded by a functional ribosome-binding site, ACATT, in this bacterium (35Kiyama M. Hayakawa M. Shiroza T. Nakamura S. Takeuchi A. Masamoto Y. Abiko Y. Biochim. Biophys. Acta. 1998; 1396: 39-46Crossref PubMed Scopus (37) Google Scholar). This ORF did possess the TAA termination codon, suggesting that the upstream ORF might be a 3′ portion of a putative gene, and expressed under the control of the plasmid-borne lacZ′ gene. The upstream ORF consisted of 1223 codons capable of coding for the 131.5-kDa protein, which is in good agreement with the experimentally determined molecular mass of recombinant 130k-HMGD. Detailed characterization of this ORF will be discussed later. The downstream ORF was composed of 212 codons coding for the 24.4-kDa protein. Following computer analysis, it was revealed that this gene product exhibited a high homology with the transposase of IS1126 from P. gingivalis W83 (36Maley J. Roberts I.S. FEMS Microbiol. Lett. 1994; 123: 219-224Crossref PubMed Scopus (27) Google Scholar). During the course of the sequencing experiments, a series of nested deletion clones were constructed. Using these clones, we effectively examined the region responsible for mAb-Pg-vc binding (Fig. 3). Western blot analysis of theE. coli lysate harboring each of the deletion plasmids permitted us to specify the antigenically active domains (data not shown). mAb-Pg-vc recognized Asn1-Gly734, Asn1-Arg602, Asn1-Asn482, and Asn1-Thr475. In addition to these clones,E. coli lysate harboring a 5′ portion ofEcoRI-BamHI fragment, which encompasses amino acid residues Asn1-Ser432, was also analyzed. No immunopositive signals were observed in this clone (data not shown). These results suggested that the epitope might reside within amino acid residues from Ser432 to Thr475. We designate this region as antibody recognition site-I (ARS-I). In 5′ deletion clones, both Ile603 and Pro783exhibited positive signals; however, no signals were observed in Ser989, suggesting that an additional epitope might exist in the C-terminal portion (783–988) of 130k-HMGD. This second epitope region is referred to as ARS-II. As mentioned above, two immunoreactive regions, ARS-I and ARS-II, were identified in recombinant 130k-HMGD. Comparison of the deduced amino acid sequences corresponding to these regions revealed the presence of three nearly resembling amino acid stretches, PVQNLT, LKWD(N)AP, and LS(N)ES(D)FEN. The first two are within ARS-I, whereas the remainder is outside of immunoreactive region. To confirm the exact epitope domain, a series of overlapping peptides, encompassing the amino acid residues from Lys423 to Thr473, were synthesized (Fig.4 A). The monoclonal antibody, mAb-Pg-vc, recognized the peptide-437AP. Using these peptides, inhibition of the vesicle-associated hemagglutinating activity was examined. As shown in Fig. 4 B, peptide-431EG (EGSNEFAPVQNL) and peptide-437AP (APVQNLTGSSVG) indicated an inhibitory effect on the vesicle-associated hemagglutinating activity. Although the other peptides KVTLKWDAPNGT, APNGTPNPNPNP, and NPNPGTT showed only limited inhibition, KVCKDVTVEGSN and SSVGQKVTLKWD had no effect. This strongly suggests that the region around the APVQNL structure possesses the ability to compete with vesicles for binding to erythrocytes. To confirm the epitope structure recognized by mAb-Pg-vc, phage clones expressing random hepta-residues were screened by biopanning as outlined under "Experimental Procedures." As a result, a total of seven independent phage clones were isolated. Predicted amino acid sequences are: FPVSQEL, HPVGNTS, KPLTIDT, THGPLSP, KHPTYRQ, YKLNPTR, and YTIGPPS. Among these, quantification of phages bound to coated mAb-Pg-vc by enzyme-linked immunosorbent assay indicated that following four hepta-peptides were immuno-positive; FPVSQEL, HPVGNTS, KPLTIDT, and YKLNPTR (data not shown). Comparison of these peptide sequences with that of 130k-HMGD suggested that PVQNLT (438–443) seemed to be the structure most responsible for the epitope site of mAb-Pg-vc (Fig. 4 C). This conclusion was supported by the results shown in Figs. 3 and 4 (A andB). Both the monoclonal antibody mAb-Pg-vc and the rabbit polyclonal antibody raised against 130k-HMGD recognized several proteins (Fig. 1). This suggested that P. gingivalis might possess gene families that encode for mAb-Pg-vc-reactive proteins. Alternatively, it was possible to explain that these multiple bands were degradation products of progenitor proteins. To confirm either possibility, Southern blot analysis was carried out. The P. gingivalis chromosome was digested with EcoRI orSacI, digested doubly with EcoRI andSacI, and transferred onto membrane. For a positive control, the 9-kb EcoRI fragment from HEM9 was also used. These blots were analyzed using the 4.6-kb SacI and the 150-bp (epi-150) fragments as probes. The latter fragment corresponds to Ser433-Asn482, which contained the proposed epitope structure, PVQNLT. As shown in Fig.5, both probes exhibited superimposable blotting profiles. As expected, the 9- and 4.6-kb bands were visualized in genomic EcoRI and EcoRI-SacI digests. In addition to these bands, several DNA fragments, ranging from 4.3 to 19.3 kb in size, could also be detected in genomicSacI digest. These data suggested that the P. gingivalis chromosome might contain multiple discrete loci coding for the mAb-Pg-vc epitope. Because Southern blot analysis suggested the presence of multiple discrete loci coding for the mAb-Pg-vc recognizable epitope, we attempted to clone these genes using a PCR technique (Fig.6). The two primers were designed so that the 150-bp fragment (epi-150), which was used as the probe in the Southern blot analysis, could be amplified when the 4.6-kbSacI fragment was used as a template (Fig. 6 C,lanes 2 and 4). When the P. gingivalischromosome was used as a template, several discrete DNA fragments were detected on the agarose gel (Fig. 6 C, lane 1). These PCR products were ligated with plasmid and transformed ontoE. coli cells. We randomly picked 48 white colonies, and all inserts were sequenced. As a result, following five independent clones were isolated with duplications (Fig. 6 D): epi-6 (165 bp), epi-11 (135 bp), epi-12 (150 bp), epi-41 (165 bp), and epi-22 (1.4 kb). The deduced amino acid sequence of epi-12 is exactly the same as that of 130k-HMGD and several other registered gene products with or without protease activity (GenBankTM accession numbers PGTLAGEN, PGU42210, and PGU75366; Swisprot accession number AF017059). In the other four clones, no identical sequences were found in the registered genes; however, each clone exhibited the following similarity; epi-06 (PGPRPR1), epi-11 (PGU41807), epi-41 (PGPRPR1), and epi-22 (PGU41807). Comparison of DNA sequences between each clone and that of the computer-aided counterpart protein indicated the presence of conservative (underlined) and nonconservative base-changes or deletion (boxed). In epi-06, for example, DNA sequence corresponding to underlined region (PNPNP) in counterpart protein PGPRPR1 is different from that of epi-06; however, translated amino acids were conserved. In addition, boxed NP sequence found in epi-06 was missing in PGPRP1. That the similar nucleotide disturbances were also observed in the remaining three clones rules out the possibility of contamination of strains used in the present study. P. gingivalis is a Gram-negative anaerobic bacterium that has become recognized as a major pathogen for adult periodontitis. One characteristic aspect of this bacterium is the formation of vesicle particles, budding from the parental P. gingivalis cell surface. One of the pathogenic properties resulted from released vesicles is the ability to agglutinate erythrocytes or hemagglutination. Using vesicle particles as antigen, we prepared two antibodies, anti-vesicle rabbit serum and mouse monoclonal antibody mAb-Pg-vc (31Shibata Y. Kurihara K. Takiguchi H. Abiko Y. Infect. Immun. 1998; 66: 2207-2212Crossref PubMed Google Scholar), capable of inhibiting hemagglutinating activity. In the present study, one of the antibodies, mAb-Pg-vc, was employed to screenP. gingivalis genomic library, and the 4.6-kbEcoRI-SacI fragment encoding for 130-kDa protein was successfully isolated. Complete nucleotide sequencing revealed that this fragment contained two ORFs; the upstream ORF is a 3′ portion of putative gene and is responsible for 130k-HMGD, whereas the downstream ORF specifies IS1126-like (30Aduse-Opoku J. Slaney J.M. Rangarajan M. Muir J. Young K.A. Curtis M.A. J. Bacteriol. 1997; 179: 4778-4788Crossref PubMed Google Scholar, 36Maley J. Roberts I.S. FEMS Microbiol. Lett. 1994; 123: 219-224Crossref PubMed Scopus (27) Google Scholar). Using the purified recombinant 130k-HMGD protein as the "landmark," we carried out three experiments: (i) Western blot analysis of nested deletion clones, (ii) competition experiments using synthetic peptides, and (iii) binding assay of phage-displayed peptide library. As a result, a possible amino acid stretch, PVQNLT, responsible for antigenicity was identified. Southern blot analysis of P. gingivalis genome with DNA fragment corresponded to this antigenic region as probe, the presence of gene family coding for PVQNLT traits was suggested, and further PCR experiment coupled with nucleotide sequencing confirmed this. Curtis et al. (29Curtis M.A. Aduse-Opoku J. Slaney J.M. Rangarajan M. Booth V. Cridland J. Shepherd P. Infect. Immun. 1996; 64: 2532-2539Crossref PubMed Google Scholar) and Kelly et al. (37Kelly C.G. Booth V. Kendal H. Slaney J.M. Curtis M.A. Lehner T. Clin. Exp. Immunol. 1997; 110: 285-291Crossref PubMed Scopus (51) Google Scholar) have reported that the synthetic polypeptide GVSPKVCKDVTVEGSNEFAPVQNLT (residues 907–931 of PrpR1) was recognized in the serum from a patient with periodontitis and that this domain must be related to bacterial colonization. These information gave authenticity to our results. Unexpectedly, however, 130k-HMGD exhibited no prominent hemagglutinating activity. This observation strongly suggests that vesicle-associated agglutination of erythrocytes consists of many steps. Attachment of vesicles onto the surface of erythrocyte may trigger the molecular cascade toward the hemagglutination, and native 130k-HMGD is involved in one of early events in these steps. To confirm this hypothesis, further experiments will be necessary. Using the DNA sequence data obtained in the present study, we performed a homology analysis of the registered genes. As a result, several genes could be identified with high homology. Among these, a total of 14 genes possessing the hemagglutinin domains were chosen, and their protein structures are shown in Fig. 7. Based on the protein structure, these proteins could be classified into five groups. The first group is prtK (W50), prtP(W12), and lysine-specific cysteine protease (W83), in which the hemagglutinin domains exhibit the highest homology (99.3, 99.0, and 98.8%, respectively) with that of 130k-HMGD. The second group (kgp and hagD from 381) is the homologue of the lysine gingipain K in which the C-terminal portion of residues (amino acids 890–1150) are different. The third group consists ofrgp-1 (H66), prtR (W50), and prpR1(W50), whose hemagglutinin domain is shorter than that of the first two groups. The fourth group is composed of prtH (W83),prtRII (381), rgp2 (H66), and rgpB and possesses only short hemagglutinin domain. The last protein is the HagA (381), which contains four repeating hemagglutinin domains in a single molecule. In a comparison of the protein structures of 130k-HMGD and the PrtK, it appeared that the "native" gene coding for 130k-HMGD is a homologue of the prtK and that the 3′ two-thirds of this gene was cloned in the present study. We predict P. gingivalis strain 381 might retain this gene. Additional computer analysis of the registered genes other than P. gingivalisprovided us the interesting hemagglutinin molecules (HA1 and HA2) originating from influenza virus. Of these, HA1 contains PVQNLT homologue residue, PLQNLT (38Donis R.O. Bean W.J. Kawaoka Y. Webster R.G. Virology. 1989; 169: 408-417Crossref PubMed Scopus (87) Google Scholar). This further supports the fact that PVQNLT might contribute to the immunogenicity in 130k-HMGD. Recently, Nakayama et al. (39Nakayama K. Ratnayake D.B. Tsukuba T. Kadowaki T. Yamamoto K. Fujimura S. Mol. Microbiol. 1998; 27: 51-61Crossref PubMed Scopus (103) Google Scholar) have reported that the hemoglobin receptor domain (HGP-15) was found in several P. gingivalis proteases including arginine-gingipain (rgp1), lysine-specific cysteine proteases (prtP, and kgp), and hemagglutinin (hagA), which was located next to the putative hemagglutinin region. Using a lysine-gingipain-deficient mutant in which the uptake of hemin was greatly retarded, Okamoto et al. (42Okamoto K. Nakayama K. Kadowaki T. Abe N. Ratnayake D.B. Yamamoto K. J. Biol. Chem. 1998; 273: 21225-21231Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar) indicated the defect of hemoglobin adsorption and heme accumulation. Interestingly enough, the duplicating immunoreactive amino acid stretch, PVQNLT, identified in the present study could be found in both sides of the HGP-15 in 130k-HMGD and other hemagglutinin-associated proteins as well. This fact is consistent with the idea that prior to penetration into erythrocyte, two PVQNLT stretches positioned in both side of HGP-15 might facilitate the attachment of P. gingivalis onto the erythrocyte cell surface. Once erythrocytes are coagulated, HGP-15 may function under the aid of the proteolytic activity equipped with the adjacent domain to obtain heme molecules. This gene structure seems to be ideal in this bacterium for the growth in the periodontal pocket to obtain the heme molecules as an iron source. Deduced amino acid sequence of 130k-HMGD, the identification of short motif present in this protein, and the multiplication process of this motif in P. gingivalis cells shed the light on the two arguments that can explain the pathogenicity of this bacterium. First, deduced amino acid sequence indicated that the PVQNLT stretch is followed by a PN repeat, which is composed of typical structure-braking residues, making immunoresponsible residues outside of the cell wall. In ARS-II, a proline-rich stretch (TTTPPPG), another structure-braking feature, did exist that may again facilitate localization of the PVQNLT toward the molecular surface position. Second, nucleotide sequencing of the 4.6-kb EcoRI-SacI fragment indicated the presence of IS1126-like next to the gene coding for 130k-HMGD. Because the insertion sequence has the ability to move on chromosome concomitant with a particular gene, it is likely that this IS sequence may help spread the gene coding for PVQNLT all around theP. gingivalis chromosome and that it contributes to make this bacterium a more virulent strain. In the present study, we identified the amino acid stretch, PVQNLT, responsible for hemagglutinating activity. Because the synthetic peptide, APVQNLTGSSVG, exhibited the inhibitory activity toward hemagglutinin and because these motifs are so specific and widely distributed at the P. gingivalis cell surface, this information might provide useful tools to establish a passive immunization system to prevent periodontal disease in human.