.-irchsoral Bioi.Vol. 37.So. 9. PP.717724. 1992 Printedtn Great Bntain. All rightsreserved
0003-9969’92 $5.00+ 0.00 CopyrightC 1992PergamonPressLtd
ROLE OF PORPHYROMONAS GINGIVALIS 40-kDa OUTER MEMBRANE PROTEIN IN THE AGGREGATION OF P. GINGIVALIS VESICLES AND ACTINOMYCES VISCOSUS K. HIRATSC’KA,’ Y. ABIKO,’ M. HAYAKAWA,’T. ITO,* H. SASAHARA’ and H. TAKIGUCHI’ Departments of ‘Biochemistry and IOral Diagnostics, Nihon University School of Dentistry at Matsudo. 2-870-I Sakaecho-nishi, Matsudo, Chiba 271, Japan (Accepted
30 March
1992)
Summary-Porphyrontonns~ gingicolis, an important pathogen in periodontitis, produces extracellular vesicles that aggregate with Acfinomyces uiscosur cells. A 40-kDa outer membrane protein (OMP)-coding gene from P. ginginalis was cloned and the protein was found to be localized in these vesicles. The recombinant 40-kDa OMP did not show aggregation activity. However. affinity-purified antibody against the recombinant protein significantly inhibited aggregation of P. gingicalis vesicles with A. ciscosus cells. The antibody also inhibited cellular coaggregation of several strains of P. gingicalis with A. ciscosus cells, but not with other periodontal pathogens. Moreover, aggregation of A. ciscosus cells with P. gingiwlis vesicles was inhibited in a dose-dependent manner by pre-treatment of the A. rircosw cells with the recombinant protein. These findings suggest that the 40-kDa OMP may be an important aggregation factor of P. gingicalis. Key
words: Porphyromonas
gingicalis,
outer membrane protein. aggregation, coaggregation. vesicle, gene
cloning.
et al., 1990). This gene codes for a 40-kDa protein with a high hydrophobic amino acid content (43%). Moreover, antiserum against the purified recombinant 40-kDa outer-membrane protein reacted with a polypeptide of similar size in the outer-membrane fraction and in vesicles of P. gingivalis. Our purpose now was to investigate the relation of the 40-kDa outer-membrane protein to the aggregation factor of
ISTRODUCTION Coaggregation or interbacterial aggregation between specific pairs of bacteria plays an important role in the establishment of bacterial micro-communities. Thor formation of coaggregates by mixing cell suspensions of oral bacteria was first demonstrated by Gibbons and Nygaard (1970), who observed highly specific partnerships. Porph~romonus gingkafis is the predominant isolate in lesions of advanced adult periodontitis in man (Slots, 1982). This microorganism has several potentially virulent factors, including those that can agglutinate erythrocytes, those that can attach to buccal epithelial cells or saliva-coated hydroxyapatite, and those that can cause coaggregation of a number of Gram positive organisms (Okuda, Slots and Genco, 1981). They also showed that the presence of dental plaque containing actinomyces and other Gram-positive bacteria may be essential for the attachment and colonization of P. gingicafis cells after their introduction into the mouth. Extracellular vesicles released from the outer membrane of P. gingicalis are involved in coaggregation (Grenier and Mayrand, 1987; Singh, Grenier and McBride, 1989). Ellen and Grove (1989) reported that P. gingicalis vesicles can aggregate Actinomyces ciscosus cells by binding to cell walls and surface fibrils. Recently, we succeeded in cloning a gene for outer-membrane protein from P. gingiuafis (Abiko Abbreviations:
BSA, bovine serum albumin, ELISA, enzyme-linked immunoabsorbent assay, PBS-, Dulbecco’s phosphate-buffered saline without Ca2+ and Mg2+.
P. gingivalis. hUTERIALS Bacteria
AND METHODS
and growth conditions
P. gingiualis 38 I was obtained from stock strains at
the Research Laboratory of Oral Biology, Sunstar Inc., Osaka, Japan. P. gingiualis ATCC 33277, A. ciscosus ATCC 19246, Capnocytophaga ochracea ATCC 33596, Eikenella corrodens ATCC 23834 and Fusobacterium nucleatum ATCC 23726 were obtained from the American Type Culture Collection. P. gingicafis Sub3 was kindly provided by Dr K. Okuda, Department of Microbiology, Tokyo Dental College, Chiba, Japan. All bacterial species were grown in brain-heart infusion (BBL Microbiology Systems, Cockeysville, MD, U.S.A.)--O.25% yeast extract supplemented with haemin (lOpg/ml) and vitamin K (1 pg/ml). All cultures were incubated at 37°C in an anaerobic chamber containing 80% N,. 10% H, and 10% C02. Cells were harvested by centrifugation (10,000 g for 30 min), washed three times with PBS- (pH 7.4), and could be stored at -2O’C in PBS--glycerol (1: 1) until used without any change in their ability to coaggregate. The cell suspensions were washed, suspended in PBS-, and adjusted to an optical density of 1.0 at
717
718
K.
HIRATSUKA er al
550 nm in a spectrophotometer (UV-l60A, Shimadzu Co., Kyoto, Japan). The suspensions were shaken vigorously before the assays were made. Preparation of vesicles
Vesicles were isolated by the method of Grenier and Mayrand (1987), with some modification. In brief, P. gingitalis 381 cells from a 10 litre diffusate culture (3-day culture) were removed from the growth medium by centrifugation (10,OOOg for 30 min). The supematant containing the vesicles was concentrated to 250ml by passage through an ultrafiltration system (Millipore Co., Bedford, MA, U.S.A.) with a membrane having a molecular weight cut-off of 10,000. This sample was dialysed against 50 mM tris-HCI (pH 9.5), containing 0.5 mM dithiothreitol, at 4C overnight to solubilize the pilli. The vesicles were collected by centrifugation (90,OOOg) for 2 h and suspended in PBS-. This suspension was dialysed against PBS- at 4C overnight and kept at -20°C until used. Afinitv purification of antibody against recombinant 40-kDa protein
Recombinant 40-kDa outer-membrane protein was purified to homogeneity from the cell sonicate of the recombinant clone by the method of Kawamoto et al. (1991). Antiserum against the recombinant protein was obtained by immunizing rabbits (body weight around 2.5 kg) by injection of 0.25 mg of the purified recombinant protein in 0.5 ml of PBS- with 0.5 ml of Freund’s incomplete adjuvant (ICN ImmunoBiologicals, Tokyo, Japan), at several sites on the back once a week for 4 weeks. Three days after the last injection, blood was drawn from the ear vein. Immune serum was passed through a diethyl-aminoethyl cellulose column (DE52, Whatman Ltd, Maidstone. Kent, U.K.), and the eluant was applied to an immunoaffinity column (Glycosyl-hardgel; ICN ImmunoBiologicals) bearing the purified recombinant protein. After extensive vvashing with PBS-, affinity-purified antibody was eluted with 0.2 M glycine-HCI buffer (pH 2.3) and dialysed against PBS-. Protein determination
The amounts of protein in test samples were routinely determined using a protein assay kit (Bio-Rad Laboratories, Richmond, CA, U.S.A.) based on Bradford’s method (1976). Bovine serum albumin (Albumin, fractionv; Boehringer Mannheim, Germany) was used as the reference standard. Aggregation (coaggregation) assays
The flocculation slide assay was a modification of the method of Ellen and Grove (1989). On a flocculation slide, cell suspension (50~1) was mixed with equal volumes of PBS- and vesicle suspension that had been diluted 1: 1000 from the stock solution (approx. 0.7 mg, ml). This mixture (total volume, 150~1) was rotated in an incubator at 37’C for 1Omin (60 rev/min), and the slide was examined by eye. The radioactivity assay was a modification of the method of Kinder and Holt (1989). Bacterial cells labelled with [‘HI-thymidine (ICN, Irvine, CA, U.S.A.) were used. After labelling, the cells were
harvested, washed, and adjusted to an optical density of 1.0 at 55Onm as described above. A mixture of labelled cell suspension (50 p I), vesicles or unlabelled partner cell suspension (50 PI), and PBS- (200~1; final volume, 300~1) was made and put into microcentrifuge tubes (treated with 0.05% Tween 20), then incubated at 37°C for 30 min with slow rotation, and centrifuged at 42g for 1 min. A loo-p1 portion of the supernatant was then removed for liquid scintillation counting. As a control, the level of autoaggregation (self-aggregation) of the radiolabelled strain was determined by combining PBS- (250~1) and the radiolabelled cell suspension (50 ~1) described above. The percentage of autoaggregation and radioactive input (determined by sampling directly from the labelled cell suspension into a scintillation vial) and percentage of coaggregation were then calculated. All determinations were done in triplicate. The ability of the affinity-purified antibody against the recombinant 40-kDa outer-membrane protein to inhibit aggregation or coaggregation was determined by pre-incubation of the antibody with P. gingivalis 38 1 vesicles or whole cells at 37’C for 30 min. As a control, normal rabbit IgG (chromatographically purified rabbit IgG, Zymed Laboratories Inc., CA, U.S.A.) and heat-treated, affinity-purified antibody (IOO’C for 10 min) were used. Similarly, the ability of the recombinant protein to inhibit aggregation was determined by pre-incubation of this substance with A. ciscosus cells at 37’C for 30min. Indirect cellular ELBA
To examine whether the recombinant 40-kDa protein was bound to A. riscosus cells directly, a modification of an indirect cellular ELISA was used. In brief, suspensions of the A. ciscosus cells were incubated with the serially diluted recombinant 40-kDa outer-membrane protein in PBS- containing 0.5% BSA at 37’C for 2 h in microcentrifuge tubes. After blocking with BSA-PBS-, these tubes were incubated with 5 ilg of affinity-purified antibody against the recombinant protein in BSA-PBS- at 37’C for 2 h, and then with added peroxidase-conjugated goat anti-rabbit IgG (1: 5000 dilution; Organon Teknika Corp., West Chester, PA, U.S.A.) in BSA-PBS- at 37’C for 1 h. One-tenth mg/ml of o-phenylenediamine dihydrochloride (Sigma Chemical Co., MO, U.S.A.) and HzOz in 0.1 M citrate buffer (pH 4.5) were added to develop the colour reaction, and the reactions were determined at an absorbance of 492 nm. After each incubation throughout this assay, washes of PBS- containing 0.02% Triton X-100 were used to remove unbound additions. Heat-treated (100°C for 10 min) and protease-treated @roteinase K; Boehringer) (0.1 pg:ml, 37C for 2 h) recombinant proteins were used as a control. All measurements were done in duplicate. RESULTS
To examine adequately the effects of antibody to the outer-membrane protein of P. ginghalis on the aggregation activity of P. gingivalis vesicles with A. riscosus cells, the antibody should be highly purified and must be specific for the antigen. As rabbit antiserum usually contains non-specific antibodies,
719
Aggregation and 40-kDa outer membrane protein
even when highly purified antigen is used, the antiserum reacts weakly with E. coli and P. gingivalis components. Therefore, we made an affinity-purification of the monospecific antibody by affinitythe purified using column chromatography, recombinant protein as a ligand. The purified antibody reacted only with the recombinant 40-kDa protein and with the identical protein in the vesicles of P. gingicalis (data not shown). The flocculation slide assay showed that A. viscosus ATCC 19246 cells did not aggregate by themselves [Plate Fig. l(A)], whereas their aggregation was significant when P. gingivalis 381 vesicles were added to the cell suspension [Plate Fig. l(B)]. As the 40-kDa outer-membrane protein is found in the vesicles, we examined the recombinant of this protein to find out if it aggregates with A. viscosus cells. No aggregation activity was observed [Plate Fig. l(C)]. The aggregation activity of P. gingivalis vesicles toward A. ciscosus cells was inhibited in a dose-dependent manner by the affinity-purified antibody against the recombinant protein [Plate Fig. l(D)-(F)]. In contrast, the same concentration of control normal rabbit IgG or heat-treated (100°C for 30 min) affinitypurified antibody did not inhibit aggregation activity [Plate Fig. l(G) and (H)]. Inhibition of aggregation activity by the antibody was confirmed by phasecontrast microscopy. Plate Fig. 2 shows that affinity purified antibody against the recombinant 40-kDa outer-membrane protein prevented the appearance of large clumps of A. viscosus ATCC 19246 cells formed by P. gingicalis 381 vesicle aggregation. We chose the indirect cellular ELISA to clarify whether receptors for the 40-kDa outer-membrane protein of P. gingivalis vesicles were present on A. ciscosus cell surfaces, using affinity-purified antibody and then peroxidase-conjugated goat antirabbit IgG (Text Fig. 3). The amount of binding to A. ciscosus cells, expressed as absorbance at 492 nm, increased in a dose-dependent manner when the recombinant protein was added. In contrast, binding to A. ciscosus cells with 5Opg of the heat-treated recombinant protein and 50 pg of the pronase-treated protein were 79.6 and 63.2%, respectively, less than binding with the same concentration of the non-heat-treated recombinant protein. We then examined the effects of cell exposure to the recombinant protein on vesicle-elicited aggregation activity (Plate Fig. 4). Pre-treatment of A. viscosus cells with recombinant protein reduced cell aggregation in a dose-dependent manner, As it is difficult to measure aggregation activity with the flocculation slide assay, a radioactivity assay was also made. As shown in Text Fig. 5, there was a significant, dose-dependent inhibition of P. gingicalis vesicle-produced A. viscosus cell aggregation by the affinity-purified antibody against the recombinant 40-kDa protein. In contrast, normal rabbit IgG had no effect on cell aggregation. Table 1 shows the coaggregation of A. viscosus cells with various species of oral bacteria and the inhibition of this coaggregation by the affinitypurified antibody against the recombinant 40-kDa protein. All strains, F. nucleatum ATCC 23726, E. corrodens ATCC 23834, C. ochracea ATCC 33596, and P. gingicalis ATCC 33277, 381, and Su63,
Table 1. Effects of the affinity-purified antibody against recombinant 40-kDa protein on coaggregation between A. viscosusATCC 19246cells and other oral bacteria cells
Bacteria strain F. nuclearurn ATCC 23726 E. corrodens ATCC 23834+ C. ochracea ATCC 33596t P. gingicalis Su63 P. gingicalis ATCC 33277t P. gingicalis 381
A-pAb (IJg)*
Coaggregation with labelled A. ciscosuscells (%)
0
10.0 20.0 0 10.0 20.0 0 10.0 20.0 0 10.0 20.0 0 10.0 20.0 0 10.0 20.0
96.8 + 0.6 96.9 f. 0.3 95.4 + 0.5 86.5 t_ 6.2 90.5 * 0.7 87.7 & 0.2 80.9 c 7.0 89.0 * 1.3 84.9 + 0.9 89.1 f. 0.1 5.9 _+1.2 2.1 + 2.2 85.3 &-5.1 50.3 f 6.4 15.8 : 1.7 68.4 rt:0.3 51.2 * 2.8 26.0 k 2.8
‘Affinity-purified antibody (A-pAb) against recombinant 40-kDa orotein. Antibody treatment of cells was at 37°C for 30 &in with rotation: fType strain. Calculated values; see Materials and Methods for details. The results are reported as mean + SD.
strongly coaggregated with A. viscosus ATCC 19246 cells. However, the affinity-purified antibody against 40-kDa outer-membrane protein inhibited the coaggregation with P. gingicalis strains only, and different percentages of coaggregation inhibition were observed among the three P. gingivalis strains. DISCUSSION
Extracellular vesicles from bacterial cells can act as supplementary forms of receptors which have the potential to mediate attachment of a heterogeneous group of organisms to dental plaque; these vesicles would therefore contribute to the microbial complexity of the plaque. When P. gingivalis cells were grown under haemin limitation, large numbers of vesicles could be seen, both surrounding the cell surface and free in the environment (McKee et al., 1986). The vesicles probably contribute significantly to improving conditions for the growth of P. gingivalis by constructing an insoluble substratum on the pellicles for coaggregation (Singh et al., 1989). This would then protect the cells from immune factors and from antibiotics (Ellen and Grove, 1989). A. ciscosus is present in the earliest stages of dental plaque formation (Theilade et al., 1974; Tinanoff, Gross and Brady, 1976; Socransky et al., 1977) and is regarded as a potential pathogen in root surface caries and periodontal diseases (Jordan and Keyes, 1964; Jordan and Hammond, 1972). Slots and Gibbons (1978) suggested that the adherence of the Gram-negative pathogen P. gingivalis to teeth was enhanced by the presence of Gram-positive plaque bacteria like A. ciscosus. Schwarz, Ellen and Grove (1987) provided evidence supporting this in an in vitro assay using saliva-coated hydroxyapatite and ‘actinobeads’, (saliva-coated hydroxyapatite further coated
720
K. HIRATSWA et al.
Aggregation and 40-kDa outer membrane protein 100 r
Q
E
z b
In O.l-
2
* A. Concentration
..
OLl,
I
’
0.6 1.3 2.5
5.0
Concentration
(pg)
Fig. 5. Effects of pretreatment of P. gingiualis 381 vesicles with various concentrations of the affinity-purified antibody against recombinant 40-kDa protein (-) on the subsequent aggregation of A. uiscosus AYCC 19246. Normal rabbit antibody (. .) was used as a control. The percentage of aggregation is calculated as that in the absence of antibodies, which was taken as 100%. Standard errors of triplicate determinations are indicated by the vertical lines.
(119)
Fig. 3. Attachment of various concentrations of recombinant 40-kDa protein (0) to A. ciscosus ATCC 19246cell surface. As a control, 5Opg of protease-treated (*) and heat-treated (m) recombinant proteins were used. Attachment is indicated as the difference of absorbance, at 495 nm, between incubation with and without the recombinant protein. All assays were done in duplicate.
with A. okcosus).
721
Furthermore, the actinomyces may be key intermediates in the succession from a Grampositive flora to one rich in various Gram-negative species. There have been other studies of the coaggregation of P. gingivalis and A. viscosus. P. gingivalis cells bound to both saliva-coated hydroxyapatite and actinobeads (Schwarz et al., 1987), and the presence of A. uiscosus significantly increased the number of bound P. gingivalis cells (Ellen, Schwarz-Faulkner and Grove, 1988). Heat treatment and proteolysis of P. gingivafis cells significantly reduced their adherence to the actinobeads, and P. gingivalis lipopolysaccharide partially inhibited binding to both saliva-coated hydroxyapatite and actinobeads but did not interact directly with A. uiscosus cells alone (Ellen
ef al., 1988). Moreover, vesicles also had a concentration-dependent inhibitory effect on the subsequent adherence of P. gingivalis whole cells to the actinobeads (Ellen and Grove, 1989). Ellen and Grove also found that P. gingiualis vesicles aggregated A. ciscosus and Acrinomyces naeslundii. These aggregations occurred over a pH range of 5-9 and did not seem to be affected by high salt concentrations or by the presence of carbohydrates; treatment of the vesicles with proteases, sodium dodecyl sulphate, and high temperatures reduced or eliminated aggregation (Bourgeau and Mayrand, 1990). These data suggest that the same substance between P. gingiualis cells and the bacterial vesicles participates in the interaction with A. viscosus cells, and that the active aggregation component of P. gingiualis whole cells and vesicles is probably proteinaceous. However, the structural elements of this component have not yet been elucidated. A number of studies have examined the coaggregation factor in A. viscosws cells. Investigations of the adherence mechanism of A. uiscosus T14V have revealed two antigenically distinct types of fimbriae that differ in their functional properties. Type-2
Plate I Fig. 1. Effects of the affinity-purified (A-p) antibody against the recombinant 40-kDa protein on aggregation of A. uiscosus ATCC 19246 cells with P. gingiralis 381 vesicles. A uiscosur cells alone (A), a mixture of A. uiscoms cells and P. gingiualis 38 I vesicles (B), and a mixture of A. uiscosus cells and 50 pg of recombinant 40-kDa protein (C) are shown at the end of the flocculation slide assay. After preincubation of P. gingicalis 381 vesicles with 1.3 I.cg(D). 2.5 pg (E), and 5.0 pg (F) of A-p antibody at 37’C for 30 min, the mixtures were incubated with A. uiscosus ATCC 19246 cells. As controls, 5.0 pg of heat-treated (100°C. 10 min) A-p antibody (G) and 5.0 pg of normal rabbit IgG (H) were used. Fig. 2. Phase-contrast photomicrographs of A. ciscosus ATCC 19246 cells (A), a mixture of A. uiscosus ATCC 19246 cells and P. gingidis 381 vesicles (B), and a mixture of A. uiscosus ATCC 19246 cells and P. gingiuolis 381 vesicles pre-incubated with the affinity-purified (A-p) (5.0 pg) against the recombinant 40-kDa protein (C). Inhibition of aggregation by the A-p antibody is very evident. x 400 Fig. 4. Effects of recombinant 40.kDa protein on aggregation of A. ciscosus ATCC 19246 cells with P. gingiualis 381 vesicles. A. ciscosur cells were preincubated at 37°C for 30 min with PBS- buffer (A), or with recombinant 40-kDa protein 6.Opg (B), 12.0 pg (C) or 24.Opg (D); a P. gingiualis 381 vesicle suspension was then added to each mixture, after which the mixtures were incubated for IO min.
72
K. H~UTSUKAet al.
fimbriae are the sites of a lactose-sensitive lectin (Cisar ef al., 1981; Revis er al., 1982). They are responsible for the coaggregation of actinomycete cells with certain plaque streptococci (Cisar, Kolenbrander and McIntire. 1979; Cisar, Sandberg and Mergenhagen, 1984; Kolenbrander and Williams, 1981; McIntire et al., 1978) and for the haemagglutination of neuraminidase-treated erythrocytes (Costello et al., 1979; Ellen er al., 1980). In contrast, type-l fimbriae mediate direct attachment of A. ciscosus TllV to saliva-coated hydroxyapatite; this adherence is not influenced by lactose (Wheeler, Clark and Birdsell, 1979) but by antibody against type-l fimbriae (Wheeler and Clark, 1980). Actinomyces receptors for the vesicles are not uniquely associated with fibril antigens or epitopes that are not shared by the two species, e.g., type-l fibril antigen, as aggregation with P. gingiuafis vesicles is not limited to typical A. riscosus and A. naeslnndii strains. This finding may indicate the presence of a common receptor for A. viscosus and A. naesfundii (Ellen and Grove, 1989; Bourgeau and Mayrand, 1990). Thus, the structural components of the bacterial coaggregation factor that is reactive with P. gingicafis cells and vesicles also remain to be clarified. We show that the aggregation activity of P. gingivafis vesicles was inhibited by the affinity-purified antibody against the recombinant 40-kDa outermembrane protein. However, because of the problem of steric effect, antibody-blocking experiments alone are not sufficient to explain the relationship between the JO-kDa outer-membrane protein and the aggregation factor. Depending on the abundance of the 40-kDa protein in the vesicle membrane, it is conceivable that a high proportion of the vesicle surface would be coated by the monospecific antibody and that this in itself would interfere with the process of aggregation, whether or not the antibody recognized the principal attachment protein. More definitive evidence was the demonstration that pre-treatment of A. viscosus whole cells with the recombinant protein reduced subsequent aggregation; it thus appears that this protein is specifically recognized and bound by A. riscosus. Moreover, the indirect cellular ELISA demonstrated that the recombinant 40-kDa protein was bound to the cell surface of A. viscosus. Slots and Gibbons (1978) reported that A. ciscosus T14 coaggregated with periodontal disease-related micro-organisms, such as F. nucfeatum, E. corrodens, Cupnocytophuga, and P. gingivafis. Ebisu, Nakae and Okada (1988) demonstrated that the coaggregation of E. corrodens with A. viscosus was mediated by some lectin-like substance on the cell wall or capsule of E. corrodens cells. We were interested to find out whether the affinity-purified antibody against the 40-kDa recombinant protein would affect coaggregation between these micro-organisms and A. viscosus. The purified antibody inhibited coaggregation with all P. gingivufis strains tested, but not with F. nucleatum, E. corrodens, or C. ochrucea, suggesting that this antibody is specific to P. gingivulis strains. Why the percentage inhibition of A. ciscosusP. gingivulis coaggregation by the purified antibody differed among P. gingivufis strains is not clear. Using a DNA hybridization method, we have detected wild and reference strains of P. gingivufis with a gene
encoding the IO-kDa protein of P. gingivulis 381 as a strain-specific probe (unpublished data). Thus, the result of affinity-purified antibody inhibition may depend on differences in the concentration of the 40-kDa outer-membrane protein in the source of genomic expression of this protein in different P. gingivufis strains. There is now growing evidence that a P. gingivulis trypsin-like enzyme plays a part in bacterial adherence to a variety of surfaces in the subgingival environment (Childs and Gibbons. 1988; Naito and Gibbons, 1988; Li et al., 1991). Li et al. (1991) reported that P. gingivafis proteases contributed to the cohesion of P. gingivafis and A. ciscosus. In contrast, Bourgeau and Mayrand (1990) reported that an anti-P. gingivalis protease antibody (obtained from B. C. McBride) had no detectable effect on the aggregation of A. viscosus with P. gingivufis vesicles. To clarify these discrepancies, we examined the relationship between the 40-kDa outer-membrane protein and protease. Our recombinant 40-kDa protein had no trypsin-like enzyme activity, and affinitypurified antibody against it did not inhibit the trypsin-like enzyme activity of P. gingivafis cells (unpublished data). In addition, protease inhibitors (tosyl+lysine chloromethyl ketone, P-chloromercuriphenyl sulphonate, and tosyl-L-phenylalanine chloromethyl ketone; Sigma Chemical Co.), which significantly inhibited the adherence of P. gingivulis to actinobeads (Li et al., 1991) had no inhibitory action on the binding of the recombinant protein to A. oiscosus cell surfaces (data not shown). Thus, we found no clear correlation between the 40-kDa outermembrane protein and protease on P. gingivufis cells. Similarly, arginine has an important role in the adherence of P. gingivafis (Inoshita et al., 1986; Okuda et al., 1986; Bourgeau and Mayrand, 1990; Nagata et al., 1990). L-arginine effectively inhibits aggregation of A. viscosus with P. gingivafis vesicles (Bourgeau and Mayrand, 1990). Li et ul. (1991) suggested that arginine-containing peptides may be present in or near functional domains mediating P. gingivafis-A. viscosus interactions. We suggest that, as the recombinant protein contains only 1.6 mol% of arginine, there is also no relationship between the recombinant 40-kDa outer-membrane protein and arginine (Abiko et al., 1990). The recombinant 40-kDa protein of P. gingivalis cannot aggregate A. viscosus cells. Why this is so is unknown, but some explanations can be offered. (1) Aggregation molecules have a polymeric structure and require at least two binding sites to bridge the target cells; the recombinant 40-kDa protein did not express a polymeric structure in E. coli cells. (2) Membrane-associated proteins sometimes lose their activity during solubilization and purification, as the tertiary or quaternary structures needed for their activity can easily be lost during these processes; the 40-kDa protein may require membrane structures such as phospholipids and carbohydrates to express its aggregation activity. (3) The P. gingivalis aggregation factor may consist of multiple components, with the 40-kDa protein being only one of these. We thus found that P. gingiculis 40-kDa outermembrane protein was one of the structural components in the interaction between P. gingivafis and
Aggregation and 40-kDa outer membrane protein A. ciscosus.
The availability of large quantitites of this recombinant protein would therefore be useful. Furthermore, isolating the gene encoding this protein will also permit detailed studies on the structure and function of the coaggregation factor, using DNA sequencing and site-directed mutagenesis. Such studies are now in progress in our laboratory. Acknowledgemenrs-This work was supported grants from The Science Research Promotion Japan Private School Foundation and Foundation for Aging and Health from the Public Welfare.
in part by Fund of the the Japan Ministry of
723
Jordan H. V. and Hammond B. F. (1972) Filamentous bacteria isolated from human root surface caries. Archs orul Biol. 17, 1333-1342.
Jordan H. V. and Keyes P. H. (1964) Aerobic, grampositive, filamentous bacteria as etiologic agents of experimental periodontal disease in hamsters. Archs oral Biol. 9, 401-41 I. Kawamoto Y., Hayakawa M. and Abiko Y. (1991) Purification and immunochemical characterization of a recombinant outer membrane protein from Bacteroides gingicalis. Inr. J. Biochem. 23, 1053-1061. Kinder S. A. and Holt S. C. (1989) Characterization of coaggregation between Bacteroides gingicalis T22 and Fusobacterium nucleatum TIS. Infect. Immun. 57, 3425-3433.
REFERENCES Abiko Y., Hayakawa M.. Aoki H.. Kikuchi T., Shimatake H. and Takiguchi H. (1990) Cloning of a Bucferoides gingicalis outer membrane protein gene in Escherichia Coli. Archs oral Biol. 35, 689-695.
Bourgeau G. and Mayrand D. (1990) Aggregation of Actinomyces strains by extracellular vesicles produced by Bacteroides gingivalis. Can. J. Microbial. 36, 362-W.
Bradford M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analyt. Biochem. 72, 248-254.
Childs W. C. and Gibbons R. J. (1988) Use of percoll density gradients for studying the attachment of bacteria to oral eoithelial cells. J. dent. Res. 67. 826-830. Cisar J. 0.: Kolenbrander P. E. and McIntire F. C. (1979) Specificity of coaggregation reactions between human oral streptococci and strains of Acfinomyces riscosus or Acfinomyces naeslundii. Infect. Immun. 24, 742-752.
Cisar J. 0.. Barsumian E. L., Curl S. H., Vatter A. E., Sandberg A. L. and Siraganian R. P. (1981) Detection and localization of a lectin on Acfinomyces viscosus Tl4V by monoclonal antibodies. J. Immunol. 127, 1318-1322. Cisar J. 0.. Sandberg A. L. and Mergenhagen S. E. (1984) The function and distribution of different fimbriae on strains of Acfinomyces ciscosus and Actinomyces naeslundii. J. dent. Res. 63, 393-396.
Costello A. H., Cisar J. 0.. Kolenbrander P. E. and Gabriel 0. (1979) Neuraminidase-dependent hemagglutination of human erythrocytes by human strains of Acrinomyces ciscosus and Acfinomyces naeslundii. Infect. Immun. 26, 563-572.
Ebisu S.. Nakae H. and Okada H. (1988) Coaggregation of Eikenella corrodens with oral bacteria mediated by bacterial lectin-like substance. Adv. dent. Res. 2, 323-327. Ellen R. P. and Grove D. A. (1989) Bacteroides gingicalis vesicles bind to and aggregate Acfinomyces viscosus. Infect. Immun. 57, 1618-1620. Ellen R. P., Fillery E. D., Chan K. H. and Grove D. A. (1980) Sialidase-enhanced lectin-like mechanism for Acrinomyces riscosus and Actinomyces naeslundii hemagglutination. Znfict. Immun. 27, 335-343. Ellen R. P., Schwarz-Faulkner S. and Grove D. A. (1988) Coaggregation among peridontal pathogens, emphasizing Bacteroides gingivalis-Actinomyces viscosus cohesion on a saliva-coated mineral surface. Can. J. Microbial. 34, 299-306.
Gibbons R. J. and Nygaard M. (1970) Interbacterial aggregation of plaque bacteria. Archs oral Biol. 15, 1397-1400. Grenier D. and Mayrand D. (1987) Functional characterization of extracellular vesicles produced by Bacferoides giflgicalis. fnfecf. Immun. 55, I I l-l 17. Inoshita E., Amano A., Hanioka T., Tamagawa H., Shizukuishi S. and Tsunemitsu A. (1986) Isolation and some properties of exohemagglutinin from the culture medium of Bacferoides gingivalis 381. Infect. Immun. 52, 421-427.
Kolenbrander P. E. and Williams B. L. (1981) Lactosereversible coaggregation between oral actinomycetes and Streptococcus sanguis. Infect. Immun. 33, 95-102.
Li J., Ellen R. P., Hoover C. I. and Felton J. R. (1991) Association of proteases of Porphyromonas (Bacteroides) gingiralis with its adhesion to Acfinomyces ciscosus. J. dent. Res. 70, 82-86.
McIntire F. C., Vatter A. E., Baros J. and Arnold J. (1978) Mechanism of coaggregation between Acfinomyces ciscosus Tl4V and Streptococcus sanguis 34. In_fect. Immun. 21, 978-988.
McKee A. S., McDermid A. S., Baskerville A.. Dowsett A. B., Ellwood D. C. and Marsh P. D. (1986) Effect of hemin on physiology and virulence of Bacferoides gingivalis WSO. Infect. Immun. 52, 349-355. Nagata H., Murakami Y., Inoshita E., Shizukuishi S. and Tsunemitsu A. (1990) Inhibitory effect of human plasma and saliva on coaggregation between Bacreroides gingicalis and Streptococcus mitis. J. dent. Res. 69, 1476-1479.
Naito Y. and Gibbons R. J. (1988) Attachment of Bacteroides gingicalis to collagenous substrata. J. dent. Res. 67, 1075-1080.
Okuda K., Slots J. and Genco R. J. (1981) Bacteroides gingiralis, Bacferoides asaccharolyticus and Bacteroides melaninogenicus subspecies: cell surface morphology and adherence to erythrocytes and human buccal epithelial cells. C’urr. Microbial. 6, 7-12. Okuda K., Yamamoto A., Naito Y.. Takazoe I., Slots J. and Genco R. J. (1986) Purification and properties of hemagglutinin from culture supernatant of Bacteroides gingivalis. Infect. Immun. 54, 659-665.
Revis G. J., Vatter A. E., Crowle A. J. and Cisar J. 0. (1982) Antibodies against the Ag2 fimbriae of Actinomyces ciscosus Tl4V inhibit lactose-sensitive bacterial adherence. Infect. Immun. 36, 1217-1222. Schwarz S.. Ellen R. P. and Grove D. A. (1987) Bacteroides gingiralis-Acfinomyces uiscosus cohesive interactions as measured by a quantitative binding assay. Infect. h~nnm. 55, 2391-2397. Singh U.. Grenier D. and McBride 8. C. (1989) Bacferoides gingicalis vesicles mediate attachment of streptococci to serum-coated hydroxyapatite. Oral Microbial. Immun. 4, 199-203. Slots J. (1982) Importance of black-pigmented Bacteroides in human periodontal disease. In Host-Parasite Interactions in Peridontal Disease (Eds Genco R. J. and Margenhagen S. E.), pp. l-12. Am. Sot. Microbial., Washington, DC. Slots J. and Gibbons R. J. (1978) Attachment of Bacteroides melaninogenicus subsp. asaccharolyticus to oral surfaces and its possible role in colonization of the mouth and of peridontal pockets. hfect. fmmun. 19, 254-264.
Socransky S. S., Manganiello A. D., Propas D., Oram V. and van Houte J. (1977) Bacteriological studies of developing supragingival dental plaque. J. periodont. Res. 12, 90-106.
724
K. HIRATXJKAet al.
Theilade E., Fejerskov 0.. Prachyabrued W. and Kilian M. (1974) Microbiologic study on developing plaque in human fissures. Stand. J. dent. Rex 82, 420-427.
Tinanoff N., Gross A. and Brady J. M. (1976) Development of plaque on enamel. Parallel investigations. J. peridonf. Res. 11, 197-209.
Wheeler T. T. and Clark W. B. (19gO) Fibril-mediated adherence of Acfinom,vces uiscosus to saliva-treated hydroxyapatite. Infecr. Immun. 28, 577-584. Wheeler T. T., Clark W. B. and Birdsell D. C. (1979) Adherence of Acrinomws ciscosus Tl4V and T14AV to hydroxyapatite surfaces in vitro and human teeth in oivo. Infect. Immun. 25, 1066-1074.
0003-9969’92 $5.00+ 0.00 CopyrightC 1992PergamonPressLtd
ROLE OF PORPHYROMONAS GINGIVALIS 40-kDa OUTER MEMBRANE PROTEIN IN THE AGGREGATION OF P. GINGIVALIS VESICLES AND ACTINOMYCES VISCOSUS K. HIRATSC’KA,’ Y. ABIKO,’ M. HAYAKAWA,’T. ITO,* H. SASAHARA’ and H. TAKIGUCHI’ Departments of ‘Biochemistry and IOral Diagnostics, Nihon University School of Dentistry at Matsudo. 2-870-I Sakaecho-nishi, Matsudo, Chiba 271, Japan (Accepted
30 March
1992)
Summary-Porphyrontonns~ gingicolis, an important pathogen in periodontitis, produces extracellular vesicles that aggregate with Acfinomyces uiscosur cells. A 40-kDa outer membrane protein (OMP)-coding gene from P. ginginalis was cloned and the protein was found to be localized in these vesicles. The recombinant 40-kDa OMP did not show aggregation activity. However. affinity-purified antibody against the recombinant protein significantly inhibited aggregation of P. gingicalis vesicles with A. ciscosus cells. The antibody also inhibited cellular coaggregation of several strains of P. gingicalis with A. ciscosus cells, but not with other periodontal pathogens. Moreover, aggregation of A. ciscosus cells with P. gingiwlis vesicles was inhibited in a dose-dependent manner by pre-treatment of the A. rircosw cells with the recombinant protein. These findings suggest that the 40-kDa OMP may be an important aggregation factor of P. gingicalis. Key
words: Porphyromonas
gingicalis,
outer membrane protein. aggregation, coaggregation. vesicle, gene
cloning.
et al., 1990). This gene codes for a 40-kDa protein with a high hydrophobic amino acid content (43%). Moreover, antiserum against the purified recombinant 40-kDa outer-membrane protein reacted with a polypeptide of similar size in the outer-membrane fraction and in vesicles of P. gingivalis. Our purpose now was to investigate the relation of the 40-kDa outer-membrane protein to the aggregation factor of
ISTRODUCTION Coaggregation or interbacterial aggregation between specific pairs of bacteria plays an important role in the establishment of bacterial micro-communities. Thor formation of coaggregates by mixing cell suspensions of oral bacteria was first demonstrated by Gibbons and Nygaard (1970), who observed highly specific partnerships. Porph~romonus gingkafis is the predominant isolate in lesions of advanced adult periodontitis in man (Slots, 1982). This microorganism has several potentially virulent factors, including those that can agglutinate erythrocytes, those that can attach to buccal epithelial cells or saliva-coated hydroxyapatite, and those that can cause coaggregation of a number of Gram positive organisms (Okuda, Slots and Genco, 1981). They also showed that the presence of dental plaque containing actinomyces and other Gram-positive bacteria may be essential for the attachment and colonization of P. gingicafis cells after their introduction into the mouth. Extracellular vesicles released from the outer membrane of P. gingicalis are involved in coaggregation (Grenier and Mayrand, 1987; Singh, Grenier and McBride, 1989). Ellen and Grove (1989) reported that P. gingicalis vesicles can aggregate Actinomyces ciscosus cells by binding to cell walls and surface fibrils. Recently, we succeeded in cloning a gene for outer-membrane protein from P. gingiuafis (Abiko Abbreviations:
BSA, bovine serum albumin, ELISA, enzyme-linked immunoabsorbent assay, PBS-, Dulbecco’s phosphate-buffered saline without Ca2+ and Mg2+.
P. gingivalis. hUTERIALS Bacteria
AND METHODS
and growth conditions
P. gingiualis 38 I was obtained from stock strains at
the Research Laboratory of Oral Biology, Sunstar Inc., Osaka, Japan. P. gingiualis ATCC 33277, A. ciscosus ATCC 19246, Capnocytophaga ochracea ATCC 33596, Eikenella corrodens ATCC 23834 and Fusobacterium nucleatum ATCC 23726 were obtained from the American Type Culture Collection. P. gingicafis Sub3 was kindly provided by Dr K. Okuda, Department of Microbiology, Tokyo Dental College, Chiba, Japan. All bacterial species were grown in brain-heart infusion (BBL Microbiology Systems, Cockeysville, MD, U.S.A.)--O.25% yeast extract supplemented with haemin (lOpg/ml) and vitamin K (1 pg/ml). All cultures were incubated at 37°C in an anaerobic chamber containing 80% N,. 10% H, and 10% C02. Cells were harvested by centrifugation (10,000 g for 30 min), washed three times with PBS- (pH 7.4), and could be stored at -2O’C in PBS--glycerol (1: 1) until used without any change in their ability to coaggregate. The cell suspensions were washed, suspended in PBS-, and adjusted to an optical density of 1.0 at
717
718
K.
HIRATSUKA er al
550 nm in a spectrophotometer (UV-l60A, Shimadzu Co., Kyoto, Japan). The suspensions were shaken vigorously before the assays were made. Preparation of vesicles
Vesicles were isolated by the method of Grenier and Mayrand (1987), with some modification. In brief, P. gingitalis 381 cells from a 10 litre diffusate culture (3-day culture) were removed from the growth medium by centrifugation (10,OOOg for 30 min). The supematant containing the vesicles was concentrated to 250ml by passage through an ultrafiltration system (Millipore Co., Bedford, MA, U.S.A.) with a membrane having a molecular weight cut-off of 10,000. This sample was dialysed against 50 mM tris-HCI (pH 9.5), containing 0.5 mM dithiothreitol, at 4C overnight to solubilize the pilli. The vesicles were collected by centrifugation (90,OOOg) for 2 h and suspended in PBS-. This suspension was dialysed against PBS- at 4C overnight and kept at -20°C until used. Afinitv purification of antibody against recombinant 40-kDa protein
Recombinant 40-kDa outer-membrane protein was purified to homogeneity from the cell sonicate of the recombinant clone by the method of Kawamoto et al. (1991). Antiserum against the recombinant protein was obtained by immunizing rabbits (body weight around 2.5 kg) by injection of 0.25 mg of the purified recombinant protein in 0.5 ml of PBS- with 0.5 ml of Freund’s incomplete adjuvant (ICN ImmunoBiologicals, Tokyo, Japan), at several sites on the back once a week for 4 weeks. Three days after the last injection, blood was drawn from the ear vein. Immune serum was passed through a diethyl-aminoethyl cellulose column (DE52, Whatman Ltd, Maidstone. Kent, U.K.), and the eluant was applied to an immunoaffinity column (Glycosyl-hardgel; ICN ImmunoBiologicals) bearing the purified recombinant protein. After extensive vvashing with PBS-, affinity-purified antibody was eluted with 0.2 M glycine-HCI buffer (pH 2.3) and dialysed against PBS-. Protein determination
The amounts of protein in test samples were routinely determined using a protein assay kit (Bio-Rad Laboratories, Richmond, CA, U.S.A.) based on Bradford’s method (1976). Bovine serum albumin (Albumin, fractionv; Boehringer Mannheim, Germany) was used as the reference standard. Aggregation (coaggregation) assays
The flocculation slide assay was a modification of the method of Ellen and Grove (1989). On a flocculation slide, cell suspension (50~1) was mixed with equal volumes of PBS- and vesicle suspension that had been diluted 1: 1000 from the stock solution (approx. 0.7 mg, ml). This mixture (total volume, 150~1) was rotated in an incubator at 37’C for 1Omin (60 rev/min), and the slide was examined by eye. The radioactivity assay was a modification of the method of Kinder and Holt (1989). Bacterial cells labelled with [‘HI-thymidine (ICN, Irvine, CA, U.S.A.) were used. After labelling, the cells were
harvested, washed, and adjusted to an optical density of 1.0 at 55Onm as described above. A mixture of labelled cell suspension (50 p I), vesicles or unlabelled partner cell suspension (50 PI), and PBS- (200~1; final volume, 300~1) was made and put into microcentrifuge tubes (treated with 0.05% Tween 20), then incubated at 37°C for 30 min with slow rotation, and centrifuged at 42g for 1 min. A loo-p1 portion of the supernatant was then removed for liquid scintillation counting. As a control, the level of autoaggregation (self-aggregation) of the radiolabelled strain was determined by combining PBS- (250~1) and the radiolabelled cell suspension (50 ~1) described above. The percentage of autoaggregation and radioactive input (determined by sampling directly from the labelled cell suspension into a scintillation vial) and percentage of coaggregation were then calculated. All determinations were done in triplicate. The ability of the affinity-purified antibody against the recombinant 40-kDa outer-membrane protein to inhibit aggregation or coaggregation was determined by pre-incubation of the antibody with P. gingivalis 38 1 vesicles or whole cells at 37’C for 30 min. As a control, normal rabbit IgG (chromatographically purified rabbit IgG, Zymed Laboratories Inc., CA, U.S.A.) and heat-treated, affinity-purified antibody (IOO’C for 10 min) were used. Similarly, the ability of the recombinant protein to inhibit aggregation was determined by pre-incubation of this substance with A. ciscosus cells at 37’C for 30min. Indirect cellular ELBA
To examine whether the recombinant 40-kDa protein was bound to A. riscosus cells directly, a modification of an indirect cellular ELISA was used. In brief, suspensions of the A. ciscosus cells were incubated with the serially diluted recombinant 40-kDa outer-membrane protein in PBS- containing 0.5% BSA at 37’C for 2 h in microcentrifuge tubes. After blocking with BSA-PBS-, these tubes were incubated with 5 ilg of affinity-purified antibody against the recombinant protein in BSA-PBS- at 37’C for 2 h, and then with added peroxidase-conjugated goat anti-rabbit IgG (1: 5000 dilution; Organon Teknika Corp., West Chester, PA, U.S.A.) in BSA-PBS- at 37’C for 1 h. One-tenth mg/ml of o-phenylenediamine dihydrochloride (Sigma Chemical Co., MO, U.S.A.) and HzOz in 0.1 M citrate buffer (pH 4.5) were added to develop the colour reaction, and the reactions were determined at an absorbance of 492 nm. After each incubation throughout this assay, washes of PBS- containing 0.02% Triton X-100 were used to remove unbound additions. Heat-treated (100°C for 10 min) and protease-treated @roteinase K; Boehringer) (0.1 pg:ml, 37C for 2 h) recombinant proteins were used as a control. All measurements were done in duplicate. RESULTS
To examine adequately the effects of antibody to the outer-membrane protein of P. ginghalis on the aggregation activity of P. gingivalis vesicles with A. riscosus cells, the antibody should be highly purified and must be specific for the antigen. As rabbit antiserum usually contains non-specific antibodies,
719
Aggregation and 40-kDa outer membrane protein
even when highly purified antigen is used, the antiserum reacts weakly with E. coli and P. gingivalis components. Therefore, we made an affinity-purification of the monospecific antibody by affinitythe purified using column chromatography, recombinant protein as a ligand. The purified antibody reacted only with the recombinant 40-kDa protein and with the identical protein in the vesicles of P. gingicalis (data not shown). The flocculation slide assay showed that A. viscosus ATCC 19246 cells did not aggregate by themselves [Plate Fig. l(A)], whereas their aggregation was significant when P. gingivalis 381 vesicles were added to the cell suspension [Plate Fig. l(B)]. As the 40-kDa outer-membrane protein is found in the vesicles, we examined the recombinant of this protein to find out if it aggregates with A. viscosus cells. No aggregation activity was observed [Plate Fig. l(C)]. The aggregation activity of P. gingivalis vesicles toward A. ciscosus cells was inhibited in a dose-dependent manner by the affinity-purified antibody against the recombinant protein [Plate Fig. l(D)-(F)]. In contrast, the same concentration of control normal rabbit IgG or heat-treated (100°C for 30 min) affinitypurified antibody did not inhibit aggregation activity [Plate Fig. l(G) and (H)]. Inhibition of aggregation activity by the antibody was confirmed by phasecontrast microscopy. Plate Fig. 2 shows that affinity purified antibody against the recombinant 40-kDa outer-membrane protein prevented the appearance of large clumps of A. viscosus ATCC 19246 cells formed by P. gingicalis 381 vesicle aggregation. We chose the indirect cellular ELISA to clarify whether receptors for the 40-kDa outer-membrane protein of P. gingivalis vesicles were present on A. ciscosus cell surfaces, using affinity-purified antibody and then peroxidase-conjugated goat antirabbit IgG (Text Fig. 3). The amount of binding to A. ciscosus cells, expressed as absorbance at 492 nm, increased in a dose-dependent manner when the recombinant protein was added. In contrast, binding to A. ciscosus cells with 5Opg of the heat-treated recombinant protein and 50 pg of the pronase-treated protein were 79.6 and 63.2%, respectively, less than binding with the same concentration of the non-heat-treated recombinant protein. We then examined the effects of cell exposure to the recombinant protein on vesicle-elicited aggregation activity (Plate Fig. 4). Pre-treatment of A. viscosus cells with recombinant protein reduced cell aggregation in a dose-dependent manner, As it is difficult to measure aggregation activity with the flocculation slide assay, a radioactivity assay was also made. As shown in Text Fig. 5, there was a significant, dose-dependent inhibition of P. gingicalis vesicle-produced A. viscosus cell aggregation by the affinity-purified antibody against the recombinant 40-kDa protein. In contrast, normal rabbit IgG had no effect on cell aggregation. Table 1 shows the coaggregation of A. viscosus cells with various species of oral bacteria and the inhibition of this coaggregation by the affinitypurified antibody against the recombinant 40-kDa protein. All strains, F. nucleatum ATCC 23726, E. corrodens ATCC 23834, C. ochracea ATCC 33596, and P. gingicalis ATCC 33277, 381, and Su63,
Table 1. Effects of the affinity-purified antibody against recombinant 40-kDa protein on coaggregation between A. viscosusATCC 19246cells and other oral bacteria cells
Bacteria strain F. nuclearurn ATCC 23726 E. corrodens ATCC 23834+ C. ochracea ATCC 33596t P. gingicalis Su63 P. gingicalis ATCC 33277t P. gingicalis 381
A-pAb (IJg)*
Coaggregation with labelled A. ciscosuscells (%)
0
10.0 20.0 0 10.0 20.0 0 10.0 20.0 0 10.0 20.0 0 10.0 20.0 0 10.0 20.0
96.8 + 0.6 96.9 f. 0.3 95.4 + 0.5 86.5 t_ 6.2 90.5 * 0.7 87.7 & 0.2 80.9 c 7.0 89.0 * 1.3 84.9 + 0.9 89.1 f. 0.1 5.9 _+1.2 2.1 + 2.2 85.3 &-5.1 50.3 f 6.4 15.8 : 1.7 68.4 rt:0.3 51.2 * 2.8 26.0 k 2.8
‘Affinity-purified antibody (A-pAb) against recombinant 40-kDa orotein. Antibody treatment of cells was at 37°C for 30 &in with rotation: fType strain. Calculated values; see Materials and Methods for details. The results are reported as mean + SD.
strongly coaggregated with A. viscosus ATCC 19246 cells. However, the affinity-purified antibody against 40-kDa outer-membrane protein inhibited the coaggregation with P. gingicalis strains only, and different percentages of coaggregation inhibition were observed among the three P. gingivalis strains. DISCUSSION
Extracellular vesicles from bacterial cells can act as supplementary forms of receptors which have the potential to mediate attachment of a heterogeneous group of organisms to dental plaque; these vesicles would therefore contribute to the microbial complexity of the plaque. When P. gingivalis cells were grown under haemin limitation, large numbers of vesicles could be seen, both surrounding the cell surface and free in the environment (McKee et al., 1986). The vesicles probably contribute significantly to improving conditions for the growth of P. gingivalis by constructing an insoluble substratum on the pellicles for coaggregation (Singh et al., 1989). This would then protect the cells from immune factors and from antibiotics (Ellen and Grove, 1989). A. ciscosus is present in the earliest stages of dental plaque formation (Theilade et al., 1974; Tinanoff, Gross and Brady, 1976; Socransky et al., 1977) and is regarded as a potential pathogen in root surface caries and periodontal diseases (Jordan and Keyes, 1964; Jordan and Hammond, 1972). Slots and Gibbons (1978) suggested that the adherence of the Gram-negative pathogen P. gingivalis to teeth was enhanced by the presence of Gram-positive plaque bacteria like A. ciscosus. Schwarz, Ellen and Grove (1987) provided evidence supporting this in an in vitro assay using saliva-coated hydroxyapatite and ‘actinobeads’, (saliva-coated hydroxyapatite further coated
720
K. HIRATSWA et al.
Aggregation and 40-kDa outer membrane protein 100 r
Q
E
z b
In O.l-
2
* A. Concentration
..
OLl,
I
’
0.6 1.3 2.5
5.0
Concentration
(pg)
Fig. 5. Effects of pretreatment of P. gingiualis 381 vesicles with various concentrations of the affinity-purified antibody against recombinant 40-kDa protein (-) on the subsequent aggregation of A. uiscosus AYCC 19246. Normal rabbit antibody (. .) was used as a control. The percentage of aggregation is calculated as that in the absence of antibodies, which was taken as 100%. Standard errors of triplicate determinations are indicated by the vertical lines.
(119)
Fig. 3. Attachment of various concentrations of recombinant 40-kDa protein (0) to A. ciscosus ATCC 19246cell surface. As a control, 5Opg of protease-treated (*) and heat-treated (m) recombinant proteins were used. Attachment is indicated as the difference of absorbance, at 495 nm, between incubation with and without the recombinant protein. All assays were done in duplicate.
with A. okcosus).
721
Furthermore, the actinomyces may be key intermediates in the succession from a Grampositive flora to one rich in various Gram-negative species. There have been other studies of the coaggregation of P. gingivalis and A. viscosus. P. gingivalis cells bound to both saliva-coated hydroxyapatite and actinobeads (Schwarz et al., 1987), and the presence of A. uiscosus significantly increased the number of bound P. gingivalis cells (Ellen, Schwarz-Faulkner and Grove, 1988). Heat treatment and proteolysis of P. gingivafis cells significantly reduced their adherence to the actinobeads, and P. gingivalis lipopolysaccharide partially inhibited binding to both saliva-coated hydroxyapatite and actinobeads but did not interact directly with A. uiscosus cells alone (Ellen
ef al., 1988). Moreover, vesicles also had a concentration-dependent inhibitory effect on the subsequent adherence of P. gingivalis whole cells to the actinobeads (Ellen and Grove, 1989). Ellen and Grove also found that P. gingiualis vesicles aggregated A. ciscosus and Acrinomyces naeslundii. These aggregations occurred over a pH range of 5-9 and did not seem to be affected by high salt concentrations or by the presence of carbohydrates; treatment of the vesicles with proteases, sodium dodecyl sulphate, and high temperatures reduced or eliminated aggregation (Bourgeau and Mayrand, 1990). These data suggest that the same substance between P. gingiualis cells and the bacterial vesicles participates in the interaction with A. viscosus cells, and that the active aggregation component of P. gingiualis whole cells and vesicles is probably proteinaceous. However, the structural elements of this component have not yet been elucidated. A number of studies have examined the coaggregation factor in A. viscosws cells. Investigations of the adherence mechanism of A. uiscosus T14V have revealed two antigenically distinct types of fimbriae that differ in their functional properties. Type-2
Plate I Fig. 1. Effects of the affinity-purified (A-p) antibody against the recombinant 40-kDa protein on aggregation of A. uiscosus ATCC 19246 cells with P. gingiralis 381 vesicles. A uiscosur cells alone (A), a mixture of A. uiscoms cells and P. gingiualis 38 I vesicles (B), and a mixture of A. uiscosus cells and 50 pg of recombinant 40-kDa protein (C) are shown at the end of the flocculation slide assay. After preincubation of P. gingicalis 381 vesicles with 1.3 I.cg(D). 2.5 pg (E), and 5.0 pg (F) of A-p antibody at 37’C for 30 min, the mixtures were incubated with A. uiscosus ATCC 19246 cells. As controls, 5.0 pg of heat-treated (100°C. 10 min) A-p antibody (G) and 5.0 pg of normal rabbit IgG (H) were used. Fig. 2. Phase-contrast photomicrographs of A. ciscosus ATCC 19246 cells (A), a mixture of A. uiscosus ATCC 19246 cells and P. gingidis 381 vesicles (B), and a mixture of A. uiscosus ATCC 19246 cells and P. gingiuolis 381 vesicles pre-incubated with the affinity-purified (A-p) (5.0 pg) against the recombinant 40-kDa protein (C). Inhibition of aggregation by the A-p antibody is very evident. x 400 Fig. 4. Effects of recombinant 40.kDa protein on aggregation of A. ciscosus ATCC 19246 cells with P. gingiualis 381 vesicles. A. ciscosur cells were preincubated at 37°C for 30 min with PBS- buffer (A), or with recombinant 40-kDa protein 6.Opg (B), 12.0 pg (C) or 24.Opg (D); a P. gingiualis 381 vesicle suspension was then added to each mixture, after which the mixtures were incubated for IO min.
72
K. H~UTSUKAet al.
fimbriae are the sites of a lactose-sensitive lectin (Cisar ef al., 1981; Revis er al., 1982). They are responsible for the coaggregation of actinomycete cells with certain plaque streptococci (Cisar, Kolenbrander and McIntire. 1979; Cisar, Sandberg and Mergenhagen, 1984; Kolenbrander and Williams, 1981; McIntire et al., 1978) and for the haemagglutination of neuraminidase-treated erythrocytes (Costello et al., 1979; Ellen er al., 1980). In contrast, type-l fimbriae mediate direct attachment of A. ciscosus TllV to saliva-coated hydroxyapatite; this adherence is not influenced by lactose (Wheeler, Clark and Birdsell, 1979) but by antibody against type-l fimbriae (Wheeler and Clark, 1980). Actinomyces receptors for the vesicles are not uniquely associated with fibril antigens or epitopes that are not shared by the two species, e.g., type-l fibril antigen, as aggregation with P. gingiuafis vesicles is not limited to typical A. riscosus and A. naeslnndii strains. This finding may indicate the presence of a common receptor for A. viscosus and A. naesfundii (Ellen and Grove, 1989; Bourgeau and Mayrand, 1990). Thus, the structural components of the bacterial coaggregation factor that is reactive with P. gingicafis cells and vesicles also remain to be clarified. We show that the aggregation activity of P. gingivafis vesicles was inhibited by the affinity-purified antibody against the recombinant 40-kDa outermembrane protein. However, because of the problem of steric effect, antibody-blocking experiments alone are not sufficient to explain the relationship between the JO-kDa outer-membrane protein and the aggregation factor. Depending on the abundance of the 40-kDa protein in the vesicle membrane, it is conceivable that a high proportion of the vesicle surface would be coated by the monospecific antibody and that this in itself would interfere with the process of aggregation, whether or not the antibody recognized the principal attachment protein. More definitive evidence was the demonstration that pre-treatment of A. viscosus whole cells with the recombinant protein reduced subsequent aggregation; it thus appears that this protein is specifically recognized and bound by A. riscosus. Moreover, the indirect cellular ELISA demonstrated that the recombinant 40-kDa protein was bound to the cell surface of A. viscosus. Slots and Gibbons (1978) reported that A. ciscosus T14 coaggregated with periodontal disease-related micro-organisms, such as F. nucfeatum, E. corrodens, Cupnocytophuga, and P. gingivafis. Ebisu, Nakae and Okada (1988) demonstrated that the coaggregation of E. corrodens with A. viscosus was mediated by some lectin-like substance on the cell wall or capsule of E. corrodens cells. We were interested to find out whether the affinity-purified antibody against the 40-kDa recombinant protein would affect coaggregation between these micro-organisms and A. viscosus. The purified antibody inhibited coaggregation with all P. gingivufis strains tested, but not with F. nucleatum, E. corrodens, or C. ochrucea, suggesting that this antibody is specific to P. gingivulis strains. Why the percentage inhibition of A. ciscosusP. gingivulis coaggregation by the purified antibody differed among P. gingivufis strains is not clear. Using a DNA hybridization method, we have detected wild and reference strains of P. gingivufis with a gene
encoding the IO-kDa protein of P. gingivulis 381 as a strain-specific probe (unpublished data). Thus, the result of affinity-purified antibody inhibition may depend on differences in the concentration of the 40-kDa outer-membrane protein in the source of genomic expression of this protein in different P. gingivufis strains. There is now growing evidence that a P. gingivulis trypsin-like enzyme plays a part in bacterial adherence to a variety of surfaces in the subgingival environment (Childs and Gibbons. 1988; Naito and Gibbons, 1988; Li et al., 1991). Li et al. (1991) reported that P. gingivafis proteases contributed to the cohesion of P. gingivafis and A. ciscosus. In contrast, Bourgeau and Mayrand (1990) reported that an anti-P. gingivalis protease antibody (obtained from B. C. McBride) had no detectable effect on the aggregation of A. viscosus with P. gingivufis vesicles. To clarify these discrepancies, we examined the relationship between the 40-kDa outer-membrane protein and protease. Our recombinant 40-kDa protein had no trypsin-like enzyme activity, and affinitypurified antibody against it did not inhibit the trypsin-like enzyme activity of P. gingivafis cells (unpublished data). In addition, protease inhibitors (tosyl+lysine chloromethyl ketone, P-chloromercuriphenyl sulphonate, and tosyl-L-phenylalanine chloromethyl ketone; Sigma Chemical Co.), which significantly inhibited the adherence of P. gingivulis to actinobeads (Li et al., 1991) had no inhibitory action on the binding of the recombinant protein to A. oiscosus cell surfaces (data not shown). Thus, we found no clear correlation between the 40-kDa outermembrane protein and protease on P. gingivufis cells. Similarly, arginine has an important role in the adherence of P. gingivafis (Inoshita et al., 1986; Okuda et al., 1986; Bourgeau and Mayrand, 1990; Nagata et al., 1990). L-arginine effectively inhibits aggregation of A. viscosus with P. gingivafis vesicles (Bourgeau and Mayrand, 1990). Li et ul. (1991) suggested that arginine-containing peptides may be present in or near functional domains mediating P. gingivafis-A. viscosus interactions. We suggest that, as the recombinant protein contains only 1.6 mol% of arginine, there is also no relationship between the recombinant 40-kDa outer-membrane protein and arginine (Abiko et al., 1990). The recombinant 40-kDa protein of P. gingivalis cannot aggregate A. viscosus cells. Why this is so is unknown, but some explanations can be offered. (1) Aggregation molecules have a polymeric structure and require at least two binding sites to bridge the target cells; the recombinant 40-kDa protein did not express a polymeric structure in E. coli cells. (2) Membrane-associated proteins sometimes lose their activity during solubilization and purification, as the tertiary or quaternary structures needed for their activity can easily be lost during these processes; the 40-kDa protein may require membrane structures such as phospholipids and carbohydrates to express its aggregation activity. (3) The P. gingivalis aggregation factor may consist of multiple components, with the 40-kDa protein being only one of these. We thus found that P. gingiculis 40-kDa outermembrane protein was one of the structural components in the interaction between P. gingivafis and
Aggregation and 40-kDa outer membrane protein A. ciscosus.
The availability of large quantitites of this recombinant protein would therefore be useful. Furthermore, isolating the gene encoding this protein will also permit detailed studies on the structure and function of the coaggregation factor, using DNA sequencing and site-directed mutagenesis. Such studies are now in progress in our laboratory. Acknowledgemenrs-This work was supported grants from The Science Research Promotion Japan Private School Foundation and Foundation for Aging and Health from the Public Welfare.
in part by Fund of the the Japan Ministry of
723
Jordan H. V. and Hammond B. F. (1972) Filamentous bacteria isolated from human root surface caries. Archs orul Biol. 17, 1333-1342.
Jordan H. V. and Keyes P. H. (1964) Aerobic, grampositive, filamentous bacteria as etiologic agents of experimental periodontal disease in hamsters. Archs oral Biol. 9, 401-41 I. Kawamoto Y., Hayakawa M. and Abiko Y. (1991) Purification and immunochemical characterization of a recombinant outer membrane protein from Bacteroides gingicalis. Inr. J. Biochem. 23, 1053-1061. Kinder S. A. and Holt S. C. (1989) Characterization of coaggregation between Bacteroides gingicalis T22 and Fusobacterium nucleatum TIS. Infect. Immun. 57, 3425-3433.
REFERENCES Abiko Y., Hayakawa M.. Aoki H.. Kikuchi T., Shimatake H. and Takiguchi H. (1990) Cloning of a Bucferoides gingicalis outer membrane protein gene in Escherichia Coli. Archs oral Biol. 35, 689-695.
Bourgeau G. and Mayrand D. (1990) Aggregation of Actinomyces strains by extracellular vesicles produced by Bacteroides gingivalis. Can. J. Microbial. 36, 362-W.
Bradford M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analyt. Biochem. 72, 248-254.
Childs W. C. and Gibbons R. J. (1988) Use of percoll density gradients for studying the attachment of bacteria to oral eoithelial cells. J. dent. Res. 67. 826-830. Cisar J. 0.: Kolenbrander P. E. and McIntire F. C. (1979) Specificity of coaggregation reactions between human oral streptococci and strains of Acfinomyces riscosus or Acfinomyces naeslundii. Infect. Immun. 24, 742-752.
Cisar J. 0.. Barsumian E. L., Curl S. H., Vatter A. E., Sandberg A. L. and Siraganian R. P. (1981) Detection and localization of a lectin on Acfinomyces viscosus Tl4V by monoclonal antibodies. J. Immunol. 127, 1318-1322. Cisar J. 0.. Sandberg A. L. and Mergenhagen S. E. (1984) The function and distribution of different fimbriae on strains of Acfinomyces ciscosus and Actinomyces naeslundii. J. dent. Res. 63, 393-396.
Costello A. H., Cisar J. 0.. Kolenbrander P. E. and Gabriel 0. (1979) Neuraminidase-dependent hemagglutination of human erythrocytes by human strains of Acrinomyces ciscosus and Acfinomyces naeslundii. Infect. Immun. 26, 563-572.
Ebisu S.. Nakae H. and Okada H. (1988) Coaggregation of Eikenella corrodens with oral bacteria mediated by bacterial lectin-like substance. Adv. dent. Res. 2, 323-327. Ellen R. P. and Grove D. A. (1989) Bacteroides gingicalis vesicles bind to and aggregate Acfinomyces viscosus. Infect. Immun. 57, 1618-1620. Ellen R. P., Fillery E. D., Chan K. H. and Grove D. A. (1980) Sialidase-enhanced lectin-like mechanism for Acrinomyces riscosus and Actinomyces naeslundii hemagglutination. Znfict. Immun. 27, 335-343. Ellen R. P., Schwarz-Faulkner S. and Grove D. A. (1988) Coaggregation among peridontal pathogens, emphasizing Bacteroides gingivalis-Actinomyces viscosus cohesion on a saliva-coated mineral surface. Can. J. Microbial. 34, 299-306.
Gibbons R. J. and Nygaard M. (1970) Interbacterial aggregation of plaque bacteria. Archs oral Biol. 15, 1397-1400. Grenier D. and Mayrand D. (1987) Functional characterization of extracellular vesicles produced by Bacferoides giflgicalis. fnfecf. Immun. 55, I I l-l 17. Inoshita E., Amano A., Hanioka T., Tamagawa H., Shizukuishi S. and Tsunemitsu A. (1986) Isolation and some properties of exohemagglutinin from the culture medium of Bacferoides gingivalis 381. Infect. Immun. 52, 421-427.
Kolenbrander P. E. and Williams B. L. (1981) Lactosereversible coaggregation between oral actinomycetes and Streptococcus sanguis. Infect. Immun. 33, 95-102.
Li J., Ellen R. P., Hoover C. I. and Felton J. R. (1991) Association of proteases of Porphyromonas (Bacteroides) gingiralis with its adhesion to Acfinomyces ciscosus. J. dent. Res. 70, 82-86.
McIntire F. C., Vatter A. E., Baros J. and Arnold J. (1978) Mechanism of coaggregation between Acfinomyces ciscosus Tl4V and Streptococcus sanguis 34. In_fect. Immun. 21, 978-988.
McKee A. S., McDermid A. S., Baskerville A.. Dowsett A. B., Ellwood D. C. and Marsh P. D. (1986) Effect of hemin on physiology and virulence of Bacferoides gingivalis WSO. Infect. Immun. 52, 349-355. Nagata H., Murakami Y., Inoshita E., Shizukuishi S. and Tsunemitsu A. (1990) Inhibitory effect of human plasma and saliva on coaggregation between Bacreroides gingicalis and Streptococcus mitis. J. dent. Res. 69, 1476-1479.
Naito Y. and Gibbons R. J. (1988) Attachment of Bacteroides gingicalis to collagenous substrata. J. dent. Res. 67, 1075-1080.
Okuda K., Slots J. and Genco R. J. (1981) Bacteroides gingiralis, Bacferoides asaccharolyticus and Bacteroides melaninogenicus subspecies: cell surface morphology and adherence to erythrocytes and human buccal epithelial cells. C’urr. Microbial. 6, 7-12. Okuda K., Yamamoto A., Naito Y.. Takazoe I., Slots J. and Genco R. J. (1986) Purification and properties of hemagglutinin from culture supernatant of Bacteroides gingivalis. Infect. Immun. 54, 659-665.
Revis G. J., Vatter A. E., Crowle A. J. and Cisar J. 0. (1982) Antibodies against the Ag2 fimbriae of Actinomyces ciscosus Tl4V inhibit lactose-sensitive bacterial adherence. Infect. Immun. 36, 1217-1222. Schwarz S.. Ellen R. P. and Grove D. A. (1987) Bacteroides gingiralis-Acfinomyces uiscosus cohesive interactions as measured by a quantitative binding assay. Infect. h~nnm. 55, 2391-2397. Singh U.. Grenier D. and McBride 8. C. (1989) Bacferoides gingicalis vesicles mediate attachment of streptococci to serum-coated hydroxyapatite. Oral Microbial. Immun. 4, 199-203. Slots J. (1982) Importance of black-pigmented Bacteroides in human periodontal disease. In Host-Parasite Interactions in Peridontal Disease (Eds Genco R. J. and Margenhagen S. E.), pp. l-12. Am. Sot. Microbial., Washington, DC. Slots J. and Gibbons R. J. (1978) Attachment of Bacteroides melaninogenicus subsp. asaccharolyticus to oral surfaces and its possible role in colonization of the mouth and of peridontal pockets. hfect. fmmun. 19, 254-264.
Socransky S. S., Manganiello A. D., Propas D., Oram V. and van Houte J. (1977) Bacteriological studies of developing supragingival dental plaque. J. periodont. Res. 12, 90-106.
724
K. HIRATXJKAet al.
Theilade E., Fejerskov 0.. Prachyabrued W. and Kilian M. (1974) Microbiologic study on developing plaque in human fissures. Stand. J. dent. Rex 82, 420-427.
Tinanoff N., Gross A. and Brady J. M. (1976) Development of plaque on enamel. Parallel investigations. J. peridonf. Res. 11, 197-209.
Wheeler T. T. and Clark W. B. (19gO) Fibril-mediated adherence of Acfinom,vces uiscosus to saliva-treated hydroxyapatite. Infecr. Immun. 28, 577-584. Wheeler T. T., Clark W. B. and Birdsell D. C. (1979) Adherence of Acrinomws ciscosus Tl4V and T14AV to hydroxyapatite surfaces in vitro and human teeth in oivo. Infect. Immun. 25, 1066-1074.
