Open Collections

Isolation of two trypsin-like proteases associated with the outer membranes of Bacteroides gingivalis By HELEN G. SCOTT D.D.S., University of Alberta, 1976 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE In THE FACULTY OF GRADUATE STUDIES THE FACULTY OF DENTISTRY DEPARTMENT OF CLINICAL DENTAL SCIENCES We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November 1988 © Helen G. Scott, 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of n i n i r a l n p n f a l S r i p n p P R The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date September 2, 1988 DE-6(3/81) ABSTRACT Two proteases, P-1 and P-4, present in the outer membranes of Bacteroides gingivalis ATCC 33277, were isolated and partially purified by SDS-PAGE. The purification procedures resulted in a specific activity 45-fold (P-1) and 40-fold (P-4) greater than that of the crude fractions. Electrophoresis of proteolytically active P-1 produced three bands corresponding to molecular weights of 235, 220 and 200 kD whereas electrophoresis of proteolytically active P-4 produced one major band corresponding to 74 kD. The optimum temperature and pH for activity were determined using N-5-benzoyl-D-arginine-p-nitroanilide as substrate. The proteases were most active at 37 °C and at pH values between 6.0 and 6.5. Both proteases required a reducing agent for activity and were inhibited by a variety of serine and thiol protease inhibitors. Arginine-containing peptides were readily hydrolyzed whereas the proteases were less active against glycine-containing peptides. The proteases hydrolyzed IgA, IgG, gelatin, azocoll and azoalbumin; acid-soluble collagen was not degraded. The results of this investigation suggest that these are trypsin-like proteases which have a thiol component as part of the active site. ii TABLE OF CONTENTS Abstract i i List of tables v List of figures v i Acknowledgements v i i Introduction 1 Classification of black-pigmented Bacteroides 1 Antigenic and serologic characteristics of B. gingivalis 3 Ultrastructural studies 3 Ecology of Bacteroides gingivalis 4 Growth and nutrition 4 Metabolic end-products of Bacteroides gingivalis 7 Formation and characterization of vesicles 7 Adherence properties of Bacteroides gingivalis 8 Fimbriae 9 Surface-binding proteins 9 Hemagglutination properties 1 0 Experimental infections involving BPBs 1 1 Periodontal diseases 1 2 Virulence factors of Bacteroides gingivalis 1 4 Adherence properties 1 4 Capsule 15 Outer membrane vesicles 1 6 Lipopolysaccharide 1 7 Metabolic end-products 1 7 Enzyme activities 1 8 Types of proteases 2 1 Proteases of Bacteroides gingivalis 2 2 Protelytic activity of other pathogenic organisms 2 7 Aim of the investigation 2 9 iii Material and methods Bacterial strain and culture conditions Isolation and purification of proteases Preparation of outer membranes Preparative polyacrylamide gel electrophoresis Protein determination Protein profiles of OM-1 and OM-2 Protein profiles of P-1 and P-4 Proteolytic activity of OM-1, OM-2, P-1, P-4 Peptidase activity Optimum pH Optimum temperature and heat stability Inhibition of protease activity Hydrolysis of protein substrates Immunoreactivity Chemicals Results Purif ication Peptidase activity Optimum pH Optimum temperature and heat stability Inhibition of enzyme activity Activity against protein substrates Immunoreactivity Discussion Bibliography iv LIST OF TABLES TABLE PAGE 1 Purification of P-1 and P-4 4 7 2 Peptidase activity of P-1 and P-4 5 2 3 Effect of protease inhibitors and metal ions 5 9 4 Activity against protein substrates 6 0 v LIST OF FIGURES FIGURE > PAGE 1 SDS-PAGE of outer membranes-1 4 3 2 SDS-PAGE of outer membranes-2 4 4 3 Zymogram of proteases after preparative SDS-PAGE 4 5 4 SDS-PAGE of P-1 4 9 5 SDS-PAGE of P-4 5 0 6 SDS-PAGE showing proteolytic purity of OM-2, P-1, P-4 5 1 7 Optimum pH for the hydrolysis of BAPNA by P-1 5 5 8 Optimum pH for the hydrolysis of BAPNA by P-4 5 6 9 Optimum temperature for the hydrolysis of BAPNA 57 1 0 Heat stability of P-4 58 1 1 Hydrolysis of IgA and IgG by P-1 and P-4 6 2 1 2 Hydrolysis of gelatin and collagen by P-1 and P-4 6 3 1 3 Hydrolysis of collagen by trypsin and P-1 6 4 vi I ACKNOWLEDGEMENTS I am deeply grateful to Dr. B. C. McBride for his understanding and guidance, and for his continued support and encouragement of this project, and for the financial support of the University of British Columbia. My sincere thanks to Drs. John Silver and Jukka Uitto for serving on my committee and for their guidance and support, and I gratefully acknowledge the constant encouragement of the members of the Division of Periodontics. I wish to give special thanks to Dr. Daniel Grenier who so generously shared his knowledge and experience with me, and who was always a source of encouragement. I give my sincere thanks to Meja Song, Heather Merilees and Pauline Hannam who asssisted me by sharing their technical skills and experience, and for their constant support and encouragement. I extend to my fellow students, Angela Joe, Nadarajah Ganeshkumar, Umadatt Singh and Blair Heffelfinger, my sincere gratitude for consistently sharing their experience and knowledge, and for their fellowship and humour. I also wish to thank Bruce McCaughey and Harold Traeger for their photography in the preparation of this work. I take this opportunity to express my gratitude to my family and friends for their unfailing support and confidence. vii INTRODUCTION Bacteroides gingivalis is considered to be a periodontal pathogen based upon the f indings of numerous invest igat ions which have identif ied this o rgan i sm in associat ion with a variety of periodontal d i seases (reviewed by van Winkelhoff et al, 1988). Furthermore, B. gingivalis possesses a number of virulence factors which may be potentially damaging to the periodontal t issues. The periodontal pocket prov ides an eco log ica l environment which is ideal for co lon izat ion by these organisms. Classification of black-pigmented Bacteroides B lack-p igmented Bacteroides (BPB) are Gram-negat ive, obl igately anaerobic , non-moti le, nonsporeforming, rod-shaped organ isms which produce brown to black co lonies on sol id media containing blood. This group of micro-organisms was first desc r ibed by Ol iver and Wherry (1921) who named the organ ism Bacterium melaninogenicum because they be l ieved the dark pigment to be melanin. These bacteria were later ass igned to the genus Bacteroides and named Bacteroides melaninogenicus (Roy and Kelly, 1939). This was the only B P B spec ies recognized until 1970, although Sawyer et al (1962) and Cou ran t and G i b bon s (1967) had shown b i o chem i ca l and immunologica l heterogeneity among strains of B. melaninogenicus. Holdeman and Moore (1970) divided B. melaninogenicus into three subspec ies based upon their fermentat ive abi l i t ies: B. melaninogenicus subsp . melaninogenicus (strongly fermentat ive) , B. melaninogenicus subsp . intermedius (weakly fermentat ive) and B. melaninogenicus s u b s p . asaccharolyticus (non-fermentative). Suf f ic ient b i o chem i ca l and genet i c d i f fe rences were found be tween the saccharolyt ic and asaccharolyt ic subspec ies to elevate B. melaninogenicus subsp. asaccharolyticus to the spec ies level ([B. asaccharolyticus], F inegold and 1 Barnes, 1977). Further research (Coykendall et al, 1980; van Steenbergen et al, 1979) indicated that the DNA of asaccharolytic B P B strains isolated from the human oral cavity contained between 47 and 49 mol% guanine plus cytosine (G+C) whereas the D N A isolated from non-oral strains contained between 52 and 53 mol% (G+C). Furthermore, no DNA homology was found between the DNAs of the oral and the non-oral strains. It was therefore proposed, that a second asaccharolyt ic spec ies , Bacteroides gingivalis, be recognized which would designate oral strains (Coykendall et al, 1980; van Steenbergen et al, 1982). A new species of asaccharolytic Bacteroides, which can be distinguished from B. gingivalis and from B. asaccharolyticus, has been proposed by van Steenbergen et al (1984): Bacteroides endodontalis. This proposal came after the evaluation of the two asaccharolytic B P B strains by Van Steenbergen et al, (1982). T h e s e two strains were originally isolated by Sundqvist (G. Sundqvist, Ph. D. thesis, University of Umea, Umea, Sweden, 1976, as reviewed by Mayrand & Holt, 1988). B. endodontalis can be differentiated from the other asaccharolytic species by its lower G + C content and by the absence of antigens common to other asaccharolytic B P B s . Furthermore, all three species can be identified by their differing protein profiles following sodium dodecyl sulfate-polyacry lamide gel e lectrophores is (van Steenbergen et al , 1984). B. endodontalis has been isolated from mixed oral infections, predominantly from pyogenic infections of odontogenic origin (van Winkelhoff et al, 1985) and from root canal infections (Haapasalo et al, 1986). Bacteroides gingivalis is isolated primarily from the human gingival sulcus or periodontal pocket and differs from the non-oral B. asaccharolyticus by the production of phenylacetic acid (Kaczmarek and Coykendal l , 1980; Mayrand, 1979), by hemagglutination of sheep erythrocytes (Slots and Genco, 1979), by a trypsin-like enzyme activity (Slots, 1981), and by serology (Mansheim et al, 1978; Mouton et al, 1981). 2 Antigenic and serologic characteristics of Bacteroides gingivalis Reed et al (1980) demonstrated the antigenic specificity of B. asaccharolyticus and B. gingivalis using Immunoelectrophoresis and immunodiffusion techniques. The results of their work showed that none of the strains from non-oral sites was antigenically similar to B. gingivalis and furthermore, no common antigens were found between the asaccharolytic and saccharolytic BPB strains. Parent et al (1986) observed that B. gingivalis consisted of at least two serogroups, and found that the human biotype exhibited 25 surface antigens, two of which were specific for the biotype. Fisher et al also demonstrated two serogroups of B. gingivalis (J.G. Fisher, J.J. Zambon, P. Chen and R.J. Genco, J . Dent. Res. 65:816, abstr. 817, 1986) and found that these groups were associated with the virulence of the strain. The virulent strains produced a major protein band at 54 kD which was absent in the avirulent strains. Furthermore the electrophoretic profile produced by cell envelope proteins of virulent and avirulent strains differed (Bramanti and Holt, J . Dent. Res. 66:223, abstr. 930, 1987), and this difference was apparently represented in the pathogenicity of the organism. Ultrastructural studies Electron microscopy observations (Takazoe et al, 1971) revealed the presence of an external capsule and fimbria-like structures in a pathogenic strain of B. gingivalis. The electron microscopy studies of Mansheim and Kasper (1977), revealed a typical Gram-negative morphology: the cells had an inner and outer cell membrane separated by a thin peptidoglycan layer. Furthermore, the oral strain, but not the non-oral strain, was found to have capsular material external to, and associated with, the outer membrane. A recent electron microscopic survey by Handley and Tipler (1986) also revealed the presence of fimbriae and of a layer of material external to the outer membrane. In addition, B. gingivalis produces outer membrane vesicles (blebs) which are 3 formed by pinching or budding of the outer membrane (Grenier and Mayrand, 1987; Listgarten and Lai, 1979; Woo et al, 1979). The number of vesicles released may be related to growth conditions (McKee et al, 1986). The vesicles have been shown to be proteolytic and collagenolytic, to hemagglutinate erythrocytes and to mediate attachment between noncoaggregating bacterial species (Grenier and Mayrand, 1987). E c o l o g y of Bacteroides gingivalis Bacteroides gingivalis colonizes the human gingival sulcus and periodontal pocket (Slots and Genco, 1979) and has occasionally been isolated from oral mucosal surfaces (van Winkelhoff, 1986b; Zambon et al, 1981). The organism is usually absent from a healthy gingival sulcus, however, it may constitute <1-5% of the cultivable subgingival flora from individuals with gingivitis (Slots, 1982; Slots et al, 1978; White and Mayrand, 1981; Zambon et al, 1981), and the proportion of B. gingivalis increases significantly in cases of adult periodontitis (Loesche et al 1985; Slots, 1977b; Spiegel et al, 1979; Tanner et al, 1979; White and Mayrand, 1981). In addition, this organism is routinely found in those sites which are deemed to be actively progressing (Slots et al, 1986; Tanner et al 1984). G r o w t h a n d nutr i t ion Morphologically, BPB are uniform or pleomorphic rods; cells in broth culture are coccobacillary, 0.5 urn wide by 1-2 urn long. Growth on blood agar produces colonies which are 1-2 mm in diameter, convex, and form black pigment in 7-10 days. This black pigmentation, originally thought to be melanin by Oliver and Wherry (1921), was found by Shah et al (1979) to be protohemin, a water-soluble, intracellular or cell-associated derivative of hemoglobin (Duerdan, 1975). Bacteroides gingivalis will grow on solid medium or in broth culture containing small peptides plus hemin and vitamin K as growth factors (Gibbons and MacDonald, 1960; Wahren and Gibbons, 1970). Vitamin K can be replaced by 4 menadione (Gibbons and MacDonald, 1960), and succinate can take the place of hemin as a growth factor (Mayrand and McBride, 1980). Hunt et al (1986) proposed a selective medium for B. gingivalis {Bacteroides gingivalis agar, BGA) containing bacitracin, colistin and nalidixic acid as the selective agents. They found that this medium could be used to differentiate between B. gingivalis and B. asaccharolyticus and could isolate B. gingivalis from other oral bacteria. Colistin and nalidixic acid inhibit aerobic and facultatively anaerobic, Gram-negative bacteria, and bacitracin inhibits Gram-positive bacteria and is active against many other oral species. Seddon et al (1988) described chemically defined and minimal media for the growth of Bacteroides gingivalis. The chemically defined medium contained only twelve components and was able to support the growth of cells in culture which were morphologically identical to those grown in other media. Furthermore, the metabolic end-products of cells grown in this medium were reproducible and yielded patterns similar to those produced in complex media. The growth rates were approximately 50% slower than those of cells grown in a complex medium, however, when the defined medium was supplemented with protein hydrolysates, the rate of cell growth was increased. The effect of hemin concentration upon the physiology and virulence of B. gingivalis strain W50 was studied by McKee et al (1986). They found that no growth occurred in the absence of hemin, although a complex proteinaceous medium supplemented with vitamin was used. Cells grown under conditions of hemin limitation produced few fimbriae, but large numbers of extracellular vesicles were observed surrounding the cells and free within the medium. Conversely, under conditions of hemin excess, the cells were heavily fimbriated and few vesicles were noted. Furthermore, cells grown in medium lacking hemin were avirulent when injected into mice whereas the injection of cells grown in medium supplemented with hemin produced up to 50% mortality in mice. Their conclusion was that the hemin concentration of the medium modulated not only 5 cell growth, but also the virulence of the organism. Based upon this apparently obligate requirement for hemin, it is not surprising that B. gingivalis is highly effective in degrading heme-containing plasma proteins (Carlsson et al, 1984) or that there is a proportionate increase in the number of Bacteroides species when bleeding develops as a result of disease progression (Loesche and Syed, 1978) . Shah and Williams (1987) found that B. gingivalis grew best in medium in which glucose was absent, and if glucose was present, little was metabolized (3%). Prolific growth was produced in medium containing protein hydrolysates. Nutritional relationships can exist between certain bacteria the result of which is an enhancement of growth for one or both organisms. For example, BPB species exhibit enhanced growth in the presence of naphthoquinone, a vitamin K-related compound (MacDonald et al, 1963) and the hemin required by B. gingivalis can be replaced by the succinate produced by facultative organisms which ferment glucose (Mayrand & McBride, 1980) or by capnophilic bacteria (Grenier & Mayrand, 1985). More recently, Grenier & Mayrand (1986) found that protoheme, produced by W. recta in coculture with B. gingivalis, enhanced the growth of this Bacteroides species. Furthermore, once established, B. gingivalis can promote its own growth by suppression of other bacterial species by the production of hematin or bacteriocin (Takazoe et al, 1984). The growth of a pathogenic strain of B. gingivalis (W50) under conditions of varying pH was examined by McDermid et al (1988). Using chemostat conditions, stable growth occurred over the pH range 6.7 to 8.3, but could not be maintained at pH 6.5 or pH 8.5. A maximum yield of cells was obtained when the cultures were maintained between pH values of 7.0 and 8.0. The enzymatic activity of the cells was also found to vary according to the growth pH. Trypsin-like activity increased with the growth pH and was maximal at pH 8.0, whereas the specific enzyme activity of collagenase and hyaluronidase was greatest at, or below a neutral growth pH. 6 Metabolic end-products of Bacteroides gingivalis Some products resulting from the metabolism of nutrients by B. gingivalis are considered to be virulence factors of the organism (van Steenbergen et al, 1982). These include organic acids: acetic, proprionic, butyric, isobutyric, isovaleric (Lambe et al, 1982; Reed et al, 1980; Shah et al, 1976) which are produced in common with other Bacteroides species. However, phenylacetic acid is produced only by B. gingivalis (Kaczmarek and Coykendall, 1980; Mayrand, 1979) and was found to be directly proportional to the Trypticase content of the medium; L-phenylalanine and peptides containing this amino acid also enhanced the production of phenylacetic acid (Bourgeau and Mayrand, 1983). Ammonia and indole (MacDonald and Gibbons, 1962; van Steenbergen et al, 1986b) and volatile sulphur compounds, including hydrogen sulphide, dimethyl sulphide and methylmercaptan (Tonzetich and McBride, 1981) are produced by B. gingivalis and are considered to be virulence factors detrimental to the host. The large quantities of ammonia produced (Shah et al, 1987) contribute to the rise in pH during growth of B. gingivalis in broth culture (McDermid, 1988) and may have a similar effect within the periodontal pocket where the pH has been shown to increase with an increase in pocket depth and with the severity of the host inflammatory response (Bickel and Cimasoni, 1985). Formation and characterization of extracellular vesic les The formation of extracellular vesicles by B. gingivalis has been reported by Williams and Holt (1985) and by McKee et al (1986). Williams and Holt (1985) recovered outer membrane fragments (vesicles) by ultracentrifugation of culture supernatant. They found that although there was a distribution of sizes, vesicles of approximately 50 nm predominated. The polypeptide profile of the vesicles was similar, but not identical, to that of the outer membranes, as analyzed by SDS-PAGE. There were at least two polypeptide bands, at 16 and 18 kD, found in the outer membrane fraction of B. gingivalis which were not detectable in the vesicle fraction. In addition, there were minor differences 7 between the fractions in the intensity of the protein staining of some of the bands. Grenier and Mayrand (1987) characterized some of the biological activities of extracellular vesicles isolated from B. gingivalis. They, too, found a variation in vesicle size from 20-150 nm, but a predominance of vesicles of 50 nm diameter. However, in contrast to the study of Williams and Holt (1985), Grenier and Mayrand (1987) found two supplementary polypeptide bands between 45 and 66 kD in the vesicle fraction which were not apparent in the polypeptide profile of the outer membranes. This difference between the two studies could be the result of having used two different strains of B. gingivalis for comparison of the polypeptide patterns; Williams and Holt (1985) used B. gingivalis strain W whereas Grenier and Mayrand (1987) used a nonpathogenic strain, A T C C 33277. The vesicles were found to be both proteolytic and collagenolytic and were able to agglutinate sheep erythrocytes, activities which are identical with those of the whole cell. Furthermore, the vesicles mediated coaggregation between bacterial cells of two species, Eubacterium saburreum and Capnocytophaga ochracea, by forming a physical link or bridge between the cells, as seen by electron microscopy. This coaggregation was stable over a pH range from 4.5 to 8.5 (Grenier & Mayrand, 1987). Adherence properties of Bacteroides gingivalis A primary requirement of an organism, in order to colonize and grow within a host, is an ability to attach to host surfaces. Specific adhesins on the surface of bacterial cells are responsible for the attachment to specific host receptors (Holt, 1982). The adhesins on the cell surface of Gram-negative bacteria include type-specific pili or fimbriae, hemagglutinins and other surface-binding proteins (Ofek and Perry, 1985). It is characteristic of B. gingivalis to colonize the dentition at discrete sites. Attachment potential, host and bacterial interactions, nutritional, physical and chemical factors are determinants which will affect the colonization pattern. 8 1. Fimbriae Fimbriae from B. gingivalis have been isolated, purified and characterized by Yoshimura et al, (1984) and Yoshimura et al, (1985). Their studies indicate that the fimbriae from B. gingivalis are thin, curly, heat-stable filaments approximately 5 nm long and with a diameter of 4 nm. The component subunit, fimbrillin, has an apparent molecular weight of 43 kD and a primary structure different from that of other Gram-negative bacteria (Yoshimura et al, 1985). Furthermore, native fimbriae and denatured fimbrillin have differing immunological characteristics and show little cross-reactivity. Work done by Okuda & Takazoe (1974) seemed to indicate that the fimbriae conferred hemagglutinating (HA) activity, however, a pure preparation of fimbriae (Yoshimura et al, 1984) showed neither HA activity, nor inhibition of HA activity. Slots and Gibbons (1978) reported that B. gingivalis was able to attach to both buccal and crevicular epithelial cells, and to the surfaces of Gram-positive bacteria. However, the attachment of B. gingivalis to epithelial cells and to erythrocytes was inhibited by both saliva and serum. These fluids did not inhibit the attachment to Gram-positive bacteria. What this suggests is that there may be two types of adhesins on the surface of B. gingivalis. This hypothesis is supported by the fact that both those strains which have HA activity, and those which do not, exhibit fimbriae on their cell surface. Furthermore, heating for 15 minutes at 60 °C abolished the HA activity of a partially purified preparation of fimbriae (Slots and Gibbons, 1978). 2. Surface-binding properties The ability to bind to specific host proteins may assist the organism in colonization, in protection from host defences or by providing key nutrients. B. gingivalis has been found to bind fibrinogen (Lantz et al, 1986). The binding was rapid, highly specific and saturable. Furthermore, this group indicated that B. gingivalis possessed a cell-associated thiol protease which 9 could degrade fibrinogen. Winkler et al (1987) investigated the ability of some Gram-positive and Gram-negative oral organisms, including B. gingivalis, to attach to a basement-membrane-like matrix and to purified basement membrane proteins: fibronectin, laminin and Type IV collagen. Their results showed that, generally, the Gram-negative organisms bound in greater numbers to the intact matrix, whereas the Gram-positive organisms attached preferentially to the isolated proteins. Of all the Gram-negative organisms tested, B. gingivalis bound to the intact matrix in the greatest numbers and although it did not bind to a great degree to the isolated proteins, it attached in greater numbers to the isolated Type IV collagen than did either of the other Gram-negative organisms or the Gram-positive organisms. The apparent affinity of B. gingivalis for the intact matrix, rather than for the isolated proteins, may be related to involvement of a three-dimensional structure of the matrix in adherence. The importance of the binding of B. gingivalis to the isolated Type IV collagen should not be overlooked in that the basement membrane, which is composed of Type IV collagen, offers the last potential barrier to bacterial translocation from the pocket to the underlying connective tissues. 3. Hemagglutination properties The ability of B. gingivalis to hemagglutinate erythrocytes (Boyd and McBride, 1984; Slots and Genco, 1979) is an important taxonomic character which can differentiate B. gingivalis from other Bacteroides species. However, the receptor responsible for HA activity has not been completely elucidated. Surface components (capsular polysaccharide and lipopolysaccharide [LPS]) extracted from B, gingivalis neither exhibited nor inhibited HA activity, however, pili extracted from B. gingivalis did posess HA ability (Okuda et al, 1981). Although, as mentioned earlier, others have had difficulty reproducing this work. 10 Boyd and McBride (1984) showed that the HA activity was associated with low-molecular-weight LPS, protein, and loosely bound lipid of the outer membrane and furthermore, that procedures which removed fimbriae from the surface of B. gingivalis had no effect upon the ability of the whole cells to cause hemagglutination. In their study, the hemagglutinating activity and the bacterial aggregating activity associated with a crude outer membrane fraction were separated. The fraction responsible for bacterial aggregation comprised a large quantity of protein and carbohydrate with little lipid material, whereas the fraction responsible for hemagglutination contained loosely-bound lipid, carbohydrate and only a small amount of protein. The work of Yoshimura et al (1984) also showed that a preparation of purified fimbriae from B. gingivalis had neither a positive correlation with HA activity nor did it inhibit HA activity. Further investigation into the hemagglutination factor of B. gingivalis by Okuda et al (1986) resulted in the isolation and characterization of a hemagglutination factor from culture supernatant of B. gingivalis strain 381. Their results indicated that the factor was composed mainly of protein (73%) with smaller quantities of sugar (12%) and phosphorous (6%). The apparent molecular weight was 40 kD, as determined by SDS-PAGE. Experimental infections involving asaccharolyt ic B P B s Asaccharolytic BPB species have often been cited as pathogens in experimental infections (Grenier & Mayrand,1985; Mayrand & McBride, 1980; Socransky & Gibbons, 1965). In these studies, combinations of bacteria were isolated from various oral sources and were tested for their ability to induce abscess formation and transmissible infections when inoculated subcutaneously into guinea pigs. The general consensus from these studies was that individual bacterial species or isolates were not able to induce an infection, however, mixtures of two or more bacteria produced a pathogenic effect. A common finding of these studies was that when BPBs were included in the 11 infectious mixture, a transmissible infection could be produced. However, when the BPB species, and especially asaccharolytic BPBs, were deleted from the mixture, transmissibility was not exhibited. It is possible that the synergistic infective mechanism is related to growth factors produced by the associated bacteria. However, a recent study by Grenier & Mayrand (1987) has shown that six out of fourteen strains of Bacteroides gingivalis tested for virulence were pathogenic in pure culture. These pathogenic strains tended to be highly collagenolytic and proteolytic whereas strains which were not pathogenic had lower collagenolytic activity but possessed a high proteolytic activity. Periodontal d iseases Specific periodontal diseases have been described and are distinguished by such factors as the age of the patient, the type and distribution of bone and attachment loss and by the species of bacteria associated with the disease (Page & Schroeder, 1982). For example, adult periodontitis has been associated with BPBs (B. intermedius and B. gingivalis) and Actinobacillus actinomycetemcomitans, (Slots,1979; Slots, 1986a; Slots & Listgarten, 1988; Socransky, 1977; Zambon, 1985; Zambon et al, 1981), and spirochetes and other motile organisms (Listgarten & Hellden,1978). Bacteroides gingivalis, Capnocytophaga and Actinobacillus actinomycetemcomitans have been isolated from patients diagnosed as having juvenile periodontitis (Kornman & Robertson, 1985; Liljenberg & Lindhe, 1980; Moore et al 1985; Newman et al, 1976; Slots, 1976). At least 300 different bacterial species have been identified within the periodontal pocket. In health, bacteria comprising the microflora of the gingival sulcus are primarily Gram-positive, facultative organisms (Slots, 1977a), whereas in adults, a proportionate increase in the number of Gram-negative, anaerobic species occurs as the disease increases in severity (Slots, 1977b; Slots et al, 1978). This is the case as well for juvenile periodonditis where the subgingival microflora has been found to comprise Gram-negative facultative or 12 anaerobic rods (Baehni et al, 1979; Kornman & Robertson, 1985; Moore et al, 1985; Slots, 1976; Tanner et al, 1979). Bacteroides gingivalis has been repeatedly implicated in the establishment and progression of periodontal diseases. Many recent studies have confirmed the finding of Burdon (1928) that Bacteroides species were present in disease sites (reviewed by Slots, 1982; Slots & Listgarten, 1988; van Winkelhoff et al, 1988). B. gingivalis has been recovered not only from sites in adults diagnosed as having generalized advanced periodontitis, (Loesche et al, 1985; Slots, 1977b; Spiegel et al, 1979; Tanner et al, 1979; White & Mayrand, 1981) but also from sites in adult patients which were classified as actively progressing (Slots et al, 1986; Tanner et al, 1984). That periodontal diseases are the result of mixed infections has been the indication from all of the investigations into the microflora related to both periodontal health and disease. The composition of the subgingival flora found in health varies significantly from that found in the various forms of periodontal disease. In a state of health, 95% of the total microflora, as determined by phase contrast or darkfield microscopy, is composed of nonmotile rods and cocci (Listgarten and Hellden, 1978). As disease progressses, there is a marked decrease in the numbers of cocci and a proportionate increase in the number of motile rods and spirochetes (Lindhe et al, 1980; Temporo et al 1983). Di Murro et al (1987), in a study evaluating the subgingival microflora isolated from patients diagnosed as having rapidly progressive periodontitis, found B. gingivalis to be consistently involved. Furthermore, a recent report by van Dyke et al (1988), indicated that B. gingivalis was an etiologic agent in severe, recurrent adult periodontitis. B. gingivalis has been demonstrated in both generalized juvenile periodontitis (Loesche et al, 1985; Wilson et al , 1985) and in localized juvenile periodontitis (Kornman & Robertson,1985; Moore et al, 1985). 13 Additional evidence for the role of B. gingivalis in the various types of periodontal diseases comes from immunological studies. Virtually all investigators agree that in patients diagnosed with generalized juvenile periodontitis or adult periodontitis the serum antibody levels to B. gingivalis are higher than in other groups of individuals ( Altman et al, 1982; Ebersole et al, 1982; Ebersole et al, 1986; Farida et al, 1986; Mouton et al, 1981; Taubman et al, 1982; Vincent et al, 1985). Elevated levels of antibody to B. gingivalis are also noted in gingival crevicular fluid as compared to serum levels (Ebersole et al, 1985; Schonfeld & Kagan, 1982; Tew et al, 1985b). A recent study by Holt et al (1988) indicated that B. gingivalis could be successfully implanted into the periodontal microbiota of monkeys. B. gingivalis strain 3079.03 was isolated from a ligature-induced periodontal site in a cynomolgus monkey and was used to derive a rifampin-resistant B. gingivalis strain. Subsequent to the implantation of the derived strain, not only had serum antibody levels to the B. gingivalis strain increased as compared to controls, but in addition, significant bone loss was observed radiographically at implanted sites, ostensibly due to a burst of activity. Thus, this work appears to demonstrate a direct connection between the implantation of B. gingivalis and clinical alterations in serum antibody levels and in crestal bone levels. Virulence factors of Bacteroides gingivalis Potentially pathogenic organisms require a combination of properties in order to manifest their effect. Much research has been directed toward the properties of B. gingivalis which contribute to its virulence. 1. Adherence properties One of the first requirements of an organism, in order to initiate disease, is an ability to attach to its host, for only in so doing will the organism be able to colonize and multiply within the host. The features of B. gingivalis which effect adherence have been discussed previously under adherence properties and include 14 fimbriae, hemagglutinins and surface-binding proteins. The recent work by Ohmori et al (1987) and Hanazawa et al (1988) helps to elucidate the virulence properties of the fimbriae of B. gingivalis. Ohmori et al (1987) reported that human gingival fibroblasts spontaneously produce thymocyte-activating factor (FTAF), which in turn stimulates mitogen-induced thymoctye proliferation. The importance of FTAF is that it may function to activate the immune response, which has a potentially damaging effect upon the periodontal tissues of the host. In the follow-up study, Hanazawa et al (1988) examined culture supernatants from gingival fibroblasts for their ability to stimulate thymocyte proliferation. Once this was established, the effect of purified fimbriae on FTAF production was determined by culturing gingival fibroblasts with various doses of purified fimbriae. Lipopolysaccharide was not detected on a silver-stained gel following SDS-PAGE of the fimbriae. Their results indicated that the fimbriae stimulated FTAF production in a dose-dependent manner. Thus fimbriae may have a duel role in the pathogensis of disease: attachment and the activation of factors which elicit a response from the host defense system. 2. Capsu le of Bacteroides gingivalis The capsule of B. gingivalis has a number of functions: as a physicochemical barrier between the cell and the external environment, as protection against dessication by binding water molecules, as a defense against the host immune system by avoiding the phagocytic action of polymorphonuclear leukocytes (PMNLs), and if entrapment by PMNLs does occur, the capsule aids in preventing hydrolytic degradation of the organism. The capsule consists of a layer of electron-dense material, external to the outer membrane, approximately 15 nm thick (Handley & Tipler, 1986; Listgarten & Lai, 1979; Woo et al, 1979) and is composed of a polysaccharide heteropolymer (Mayrand & Holt , 1988). Studies relating virulence to the presence of a capsule have shown that those strains which are encapsulated have greater resistance to phagocytosis (Okuda & 15 Takazoe, 1973). Furthermore, when capsular material extracted from a BPB species was added to a system utilizing Staphylococcus aureus, phagocytosis was inhibited. Mayrand & Holt (1988), reviewing the work of van Steenbergen et al (in publication), noted that virulent strains of B. gingivalis were more resistant to killing by human serum and by PMNLs plus serum. Furthermore, the virulent organisms did not autoagglutinate, had a thicker capsule and were more hydrophilic than less virulent strains. The conclusion from their results was that observed differences in virulence can be attributed, at least in part, to differences in capsular structure. 3. Outer membrane vesicles The outer membrane vesicles (previously described) are similar, if not identical, to the outer membrane of B. gingivalis. Their virulence is related to proteolytic and collagenolytic activities, as well as to an ability to hemagglutinate erythrocytes and to promote adherence between noncoaggregating species (Grenier & Mayrand, 1987). Although their role in the pathogenesis of disease is unclear, it is possible that, because of their small size, the vesicles could easily cross epithelial barriers which would be impermeable to whole cells. In this way, the lytic capacity of the cell could be expanded, allowing it to obtain nutrients from a larger area, and creating a pathway for invasion of the whole cell into the tissues. Furthermore, the vesicles could compete for antibodies of the host defence system, thereby impeding the specific antibacterial immune defense (Grenier & Mayrand, 1987). In a study by Smalley & Birss (1987), vesicles from B. gingivalis strain W50 were found to have tryspin-like enzyme activity, supporting the finding of Grenier & Mayrand (1987) who found not only a trypsin-like, but also a collagenolytic acitvity, associated with the extracellular vesicles. In a more 16 recent study by Smalley, Birss & Shuttleworth (1988), a trypsin-like enzyme was purified from a cell- and particle-free culture supernatant of B. gingivalis. The activity was associated with both a 58 kD peptide (by SDS-PAGE) and a higher molecular-weight complex. Both prepared fractions had the ability to degrade human plasma fibronectin in the presence and the absence of a reducing agent (dithiothreitol). 4. Lipopolysaccharide of Bacteroides gingivalis The lipopolysaccharide (LPS) of B. gingivalis is associated with the outer leaflet of the outer membrane and comprises three covalently-linked components: a polysaccharide O antigen which extends into the surrounding medium from the surface of the outer membrane, a core polysaccharide found at the surface of the outer membrane, and the lipid A moiety which is embedded in the outer membrane. The studies of Mansheim & Kasper (1977) and Mansheim et al (1978) indicated that the LPS of BPB species was biochemically different from that of other Gram-negative bacteria in that it lacked heptose and 2-keto-3-deoxyoctonate. The absence of these compounds in lipopolysaccharide from BPB species has been disputed (B. Johne, I Olsen, & K. Bryn, J . Dent. Res. 67:368, abstr. 2046,1988). The virulence of LPS from B. gingivalis is manifested in a number of ways: (1) incubation of gingival fibroblasts in culture with LPS inhibits growth of the fibroblasts (Larjava et al, 1987; Layman & Diedrich, 1987), (2) LPS has been shown to cause bone resorption in vitro ( Millar et al, 1986; Nair et al, 1983) and is able to inhibit bone collagen formation (Millar et al, 1986), (3) through induction of interleukin-1, LPS is able to activate the inflammatory response which is postulated to have a role in disease pathogenesis (Hanazawa et al, 1985). 5. Metabolic end-products of Bacteroides gingivalis Other substances which may contribute to the virulence of B. gingivalis include the characteristic end-products of asaccharolytic metabolism. For example, 17 butyric and proprionic acid have been shown to have a cytotoxic effect upon various human or animal cells in culture (Goldstein et al, 1984; Grenier & Mayrand, 1985; Singer & Buchner, 1981; Touw et al, 1982; van Steenbergen et al, 1982). Volatile sulphur compounds, including hydrogen sulphide, methylmercaptan and dimethyl disulphide, produced by B. gingivalis (Tonzetich & McBride, 1981) have a potentially damaging effect upon tissues of the host because they can influence the permeability of the oral mucosal tissues (Ng & Tonzetich, 1983) and can reduce collagen synthesis (Tonzetich & McBride, 1981). Indole and ammonia (MacDonald & Gibbons, 1961; van Steenbergen et al, 1986b) are also potentially toxic metabolic products of B. gingivalis. 6. E n z y m e a c t i v i t i e s o f Bacteroides gingivalis a . P e r t u r b a t i o n o f h o s t d e f e n s e m e c h a n i s m s B. gingivalis is able to elaborate specific enzymes which may contribute to its virulence by destroying or impeding the defensive response of the host immune system. Carlsson et al (1984a) found the presence of enzymes which were able to degrade the plasma proteinase inhibitors, 3-1-antitrypsin and 3-2-macroglobulin. In vivo, these proteinase inhibitors probably function to regulate the activity of proteases released from PMNLs (Starkey & Barrett, 1977), however, the destruction of such inhibitors by B. gingivalis could favor greater tissue degradation and more rapid disease progression. Enzymes which can specifically degrade the plasma proteins albumin, haemopexin, haptoglobulin and transferrin have been reported (Carlsson et al, 1984). The degradation of the iron-transporting proteins may reflect the requirement of B. gingivalis for hemin. Nilsson et al (1985) also found a virulent strain of B. gingivalis (W30) able to recognize and inactivate other important plasma proteinase inhibitors including complement (Cl) inhibitor which is a key modulator of both the classical and alternative pathways of complement activation. 18 Furthermore, antithrombin, plasminogen, prekallikrein, prothrombinase complex, clotting factor X and most of the 2-antiplasmin were functionally eliminated after 30 minutes of incubation with the bacterial suspension. Bacteroides gingivalis is able to alter the host defense system by degrading IgG, IgM and complement factors C3 and C5 (Sundqvist et al, 1985) and secretory IgA (Sato et al, 1987). Kilian (1981) demonstrated the degradation of immunoglobulins A-), A2 and G and in a follow-up investigation (Mortensen 81 Kilian, 1984) purified and characterized an immunoglobulin A1 protease from a BPB species, B. melaninogenicus subsp. melaninogenicus. It may not be possible to compare this work directly with that of either Sundqvist et al (1985) or Sato et al (1987) because the bacterial strain used in the study of Mortensen & Kilian (1984) was a sacchrolytic strain whereas the studies of Sundqvist et al (1985) and Sato et al (1987) utilized oral asaccharolytic strains (B. gingivalis). b. Provision of nutrients The extensive proteolytic enzyme system of B. gingivalis is able to degrade a wide variety of proteinaceous substances. The large number of small peptides produced by protein hydrolysis may then be transported into the cell and used to satisfy its nutritional requirements. c. Destruction of host tissues A strong fibrinolytic activity is evidenced by B. gingivalis (P.A. Mashimo & J. Slots, J . Dent. Res. 62: 663, abstr. 123, 1983; Lantz et al, 1986; Nitzan et al, 1978; Wikstrdm et al, 1983) which provides it with the potential for tissue invasion. Other enzymatic activity may be responsible for the destruction of tissue matrix. The early studies by Courant et al (1965) and Socransky (1970), in which whole plaque samples were analyzed, indicated that the plaque contained lytic enzymes capable of destroying the major components of the gingival connective tissue. More recently, Grenier & Mayrand (1983) investigated the production 19 of enzymes by B. gingivalis which had the potential to degrade various components of the gingival ground substance. Their findings showed that B. gingivalis elaborated gelatinase, chondroitin sulfatase, hyaluronidase, deoxyribonuclease and fibrinolysis all of which may have an effect upon tissue destruction in the pathogenesis of periodontal disease. Gibbons & MacDonald (1961) and Robertson et al (1982) noted the cell-associated collagenase activity of B. gingivalis. This organism has subsequently been shown to be the only BPB species having specific collagenase activity which will degrade native Type 1 collagen (Mayrand & Grenier, 1985; Sundqvist et al, 1987; van Steenbergen & de Graaff, 1986). This type of collagen is the major constituent of gingival connective tissue (From & Schultz-Haudt, 1963) and although it is resistant to many proteolytic enzymes, it can be degraded by both bacterial and mammalian collagenase (Loesche et al, 1982). B. gingivalis also possesses a trypsin-like activity (Laughon et al, 1982; Slots, 1981; van Winkelhoff et al, 1986c) which appears to be an important part of collagen degradation. The trypsin-like activity of B. gingivalis strain W50 was found to be enhanced in cultures grown at increasingly higher pH values, maximal at pH 8.0 (McDermid et al, 1988). Furthermore, the finding of McDermid et al (1988) that collagen degradation required the action of collagenase and a trypsin-like activity emphasizes the importance of these two enzymes in tissue degradation. The trypsin-like activity of B. gingivalis (Laughon et al, 1982; Slots, 1981; van Winkelhoff et al, 1986c) may also contribute to the virulence of this organism by the conversion of latent host collagenase into active collagenase (Golub et al, 1985) and by the activation of the complement system, thereby stimulating prostaglandin-mediated osteoclastic bone resorption (Schenkein, 1982). Unlike eucaryotic collagenases, which cleave native collagen at a single site, collagenase from B. gingivalis hydrolyzes collagen into small peptides 20 (Robertson et al, 1982; Toda et al, 1984), perhaps making the fragments more susceptible to degradation by other, nonspecific proteases. The trypsin-like activity may also contribute to the virulence of B. gingivalis by the induction and activation of host procollagenase (H. Birkedal-Hansen, J. Dent. Res. 62: 101, abstr. S51, 1987). Types of proteases Proteases may be characterized as one or more of the following: 1. serine protease: characterized by the presence of a unique serine residue at the active site. This group includes mammalian trypsin, chymotrypsin and elastase and several bacterial proteases. The usual reaction catalyzed by these enzymes is the hydrolysis of peptide bonds in proteins and peptides. Trypsin cleaves bonds only after lysine and arginine whereas chymotrypsin cleaves bonds only after large hydrophobic residues, eg. tyrosine. Elastase is a specific protease which is able to degrade elastin. Trypsin and chymotrypsin can be inhibited by specific inhibitors which bind to the active site. 2. thiol protease: characterized by the presence of a cysteine side chain at the active site, comparable to the serine residue of a serine protease. Papain is an example of a thiol protease. These proteases can be inhibited by reagents which react with thiol groups or by a variety of metal ions which form complexes with thiol groups. 3. carboxyl (acidic) protease: characterized by the presence of two aspartic acid residues at the active site. These proteases are active at acid pH. Pepsin is a representative of this group and will hydrolyze peptides with hydrophobic residues on either side of the scissile bond. Pepsin can be inhibited by the specific inhibitor, pepstatin. 4. metalloprotease: characterized by the presence of a bound metal (usually 21 zinc) at the active site. These proteases require divalent metal ions for activity. This group of enzymes includes carboxy peptidases A and B (exopeptidases) and thermolysin (endopeptidase). Inhibitors of these proteases include ethylenediaminetetracetic acid and ascorbic acid. Destruction of both hard and soft tissues can occur as a result of the enzyme activities of Bacteroides gingivalis. The major component of the tissues of the periodontium is Type I collagen which, in disease, can be rapidly destroyed (Page & Schroeder, 1976) possibly by both specific collagenase and nonspecific protease activity. The proteases of Bacteroides gingivalis have been the subject of much research. Proteases of Bacteroides gingivalis Uitto & Raeste (1978) determined collagenase activity against Type I collagen using gingival fluid from both healthy and disease sites. This activity was found to be 7 times greater in inflamed versus healthy tissue, however in both groups, collagenolysis was increased by brief exposure to trypsin, suggesting the presence of latent collagenase which could be activated by neutral protease. Collagenase activity was enhanced in both groups by bacterial plaque and by incubation of leukocytes with bacterial plaque extract. In a later study, Uitto (1983) examined the effect upon basement membrane collagen (Type IV) using proteinases from human gingiva, leukocytes and bacterial plaque. As determined by the release of hydroxyproline, the highest activity was found in leukocyte and plaque extracts. Type IV collagen, because of discontinuities in the triple helical structure (Schuppan et al,1980) is susceptible to degradation by nonspecific proteases, for example, gelatinase and elastase-like enzymes. Robertson et al (1982) assessed the collagenolytic activity of a number of oral organisms, including BPB species, Actinobacillus actinomycetemcomitans (A.a.), Fusobacterium, Capnocytophaga, and Selenomonas. Degradation of 22 collagen was determined by SDS-PAGE following incubation of collagen in solution and collagen fibrils with cell sonicates and culture supernatants. These investigators found that the BPB species had a cell-associated collagenolytic activity, whereas collagenolytic activity was observed in both the cell sonicates and media preparations of A.a. Enhancement of the proteolytic activity was found when the cells were grown in peptide-depleted medium. Toda et al (1984), however, did find some collagenolytic activity in culture supernatant from B. gingivalis. The activity was enhanced by the addition of reducing agents such as dithiothreitol (DTT) and L-cysteine, suggesting the presence of a thiol-dependent collagenolytic activity, the production of which increased with bacterial growth (up to 40 hours). The origin of this activity was not precisely defined. The work of Mayrand & Grenier (1985) further elucidated the ability of oral organisms to degrade collagen. In this study, twelve species of oral bacteria, including six BPB species, were evaluated. Their results indicated that, of the BPB examined, B. gingivalis was the only species which could degrade collagen previously sterilized with ethylene oxide. A rapid collagenolysis occurred which was determined to be the result of the combined activity of a specific collagenase and nonspecific proteases. Although other strains were capable of degrading collagen, the hydrolysis was much slower (24 vs 2 hours of incubation) and appeared to be the result of the activity of nonspecific proteases only. The findings of Sundqvist et al (1987) and van Steenbergen & de Graaf (1987) support the results of Mayrand & Grenier (1985). These studies also found B. gingivalis to be the only BPB species, of those examined, to demonstrate specific collagenolytic activity. The results of Sundqvist et al (1987) indicated that the activity was cell-associated. As well as having a specific collagenolytic activity, B. gingivalis has been shown to possess other proteolytic activities, both cell-free and cell-associated, which 23 may be important in the pathogenesis of disease. These nonspecific proteases may produce destruction of tissue directly or may induce factors of the host defense system which can have a deleterious effect upon the periodontal tissues. Fujimura & Nakamura (1981) characterized two partially purified proteases from cell extracts of a strain of B. melaninogenicus. Since the supernatant from an ultracentrifugation following cell sonication was used as the starting material, it is unlikely that the proteases isolated were cell-bound. These proteases were inhibited by thiol compounds, but not by diisopropylfluorophosphate (DFP), an inhibitor of serine proteases. Neither collagenase nor elastase activity was detected. A partially purified protease from B. gingivalis strain 381 was characterized by Yoshimura et al (1984) using the chromogenic synthetic substrate, benzoyl-D-L-arginine-p-nitroanilide (BAPNA). The activity was found in a whole cell preparation containing both the inner and outer membranes and the peptidoglycan, therefore the exact localization of the enzyme was not possible, although it was classified as being membrane-bound. The trypsin-like activity was inhibited by thiol protease inhibitors and stimulated by reducing agents such as DTT and L-cysteine. Ono et al (1987) also found a trypsin-like activity in a partially purified protease from culture supernatant of B. gingivalis strain 381. These investigators suggested that the activity was due to the same enzyme as previously studied by Laughon et al (1982) and Yoshimura et al (1984). The work by Ono et al (1987) extended the characterization of this protease by determination of the apparent molecular weight by SDS-PAGE to be 49,000 D. It should be noted, however, that the protease studied by Yoshimura et al (1984) was partially purified from a whole cell envelope fraction, whereas Ono et al (1987) used culture supernatant as a starting material. Tsutsui et al (1987) used the supernatant from the ultracentrifugation of 24 sonicated B. gingivalis cells. A protease was isolated and purified by ammonium sulphate precipitaiton and sequential column chromatography. The apparent molecular weight by SDS-PAGE was found to be 50,000 D. This membrane-free protease was inhibited by serine and trypsin inhibitors and was most active at an alkaline pH (8.5). The activity was enhanced by the reducing agents L-cysteine and 2-mercaptoethanol. Although there are similarities to the trypsin-like activity found by Yoshimura et al (1984), this protease was slightly inhibited by M g + + and EDTA, whereas significant inhibition from EDTA and activation by M g + + was found by Yoshimura et al (1984). In addition, the activity found by Tsutsui et al (1987) was not significantly affected by p-chloromercuribenzoic acid (PCMB), although it did affect the activity of the protease reported by Yoshimura et al (1984). In a later study, Fujimura & Nakamura (1987) examined the characteristics of a trypsin-like protease isolated from a Triton X-100 (nonionic detergent) soluble extract of sonicated cells of B. gingivalis strain 33277 and purified by sequential column chromatography. This enzyme was active at neutral pH (7.5), inhibited by thiol inhibitors and EDTA, with an apparent molecular weight of 65,000 D as determined by SDS-PAGE. The properties of this enzyme are similar to those of the enzyme reported by Yoshimura et al (1984), however, different strains of B. gingivalis were used in each of these studies. Sorsa et al (1987) found both a trypsin-like protease and collagenase in a supernatant fraction resulting from a low speed centrifugation of sonicated cells of B. gingivalis strain 33277. The trypsin-iike protease was able to degrade native Type IV collagen and denatured Type I collagen whereas the collagenase cleaved native Type I collagen at multiple cleavage sites over time. The trypsin-like protease was inhibited by serine protease inhibitors, EDTA, and ascorbic acid; activity was enhanced by reducing agents. The importance of this is that basement membrane is composed of Type IV collagen, thus such a nonspecific, trypsin-like activity could facilitate the penetration of other bacterial toxins and 25 enzymes from the periodontal pocket into the underlying connective tissue. Grenier & McBride (1987) were able to isolate, purify and characterize a glycylprolyl protease from the outer membranes of B. gingivalis strain 33277. The purification was completed by preparative SDS-PAGE and showed a single band with an apparent molecular weight of 29,000 D for the active form of the protease. This protease was inhibited by reducing agents, in contrast to other proteases isolated from B. gingivalis. Proteolytic activity was exhibited against both a synthetic peptide containing the glycylprolyl peptide and denatured collagen. The protease was found in all B. gingivalis strains tested, but not in other BPB species examined. Two groups have investigated the effect of specific peptidases of B. gingivalis. Abiko et al (1985) characterized a partially purified glycylprolyl dipeptidylaminopeptidase isolated from spent culture medium, supernatant from washed cells and supernatant from cell extracts of B. gingivalis strain 381. The enzyme was found to enhance the hydrolysis of partially degraded Type I collagen. Suido et al (1987) isolated two peptidases from the culture supernatant of B. gingivalis strain 381. One, a glycylprolyl peptidase thought to be the same as that found by Abiko et al (1985), was inhibited by serine protease inhibitors. The other enzyme, N-CBz-glycyl-glycyl-arginyl peptidase, was first associated with the cell fraction, but after 48 hours in culture, could be found in the supernatant. This enzyme was characterized as a thiol protease, however, it was not determined whether or not it represented any of the previously described thiol proteases. Finally, Otsuka et al (1987) isolated a protease(s) from culture supernatant from B. gingivalis strain 381. This protease was separated into three isoenzymes, all of which were dependent upon thiol group reagents for activity (DTT). The previous studies of Yoshimura et al (1984), Robertson et al (1982) and Toda et al (1984) suggest the presence of a thiol-dependent 26 trypsin-like protease associated with the proteolytic activity of B. gingivalis. On the basis of their results with specific substrates and inhibitors and the requirement for reducing agents for activity, Otsuka et al (1987) suggest that the designation of this enzyme as a trypsin-like protease be changed to one of a thiol protease. The highly varied and extensive proteolytic nature of B. gingivalis has been shown in the work of Grenier, Chao, McBride (in press). Eight proteases were isolated by these researchers using BSA-polyacrylamide gel electrophoresis. The proteases were isolated, and subsequently characterized, from culture supernatants, cell extracts, purified outer membranes, and outer membrance vesicles. Although only a portion of the proteolytic nature of B. gingivalis can be appreciated from the literature cited, the importance of proteolysis in disease pathogenesis cannot be underestimated, especially as the organism appears to have a direct effect upon the periodontal tissues, and is able to not only induce a response from the host which can enhance disease progression, but can also inhibit the defense system of the host. Proteolytic activity of other pathogenic organisms A number of organisms considered to be pathogenic have been shown to possess cell-associated or cell-free products which are an integral part of their virulence and contribute to their ability to cause disease. One such organism is Pseudomonas aeruginosa, an opportunistic pathogen capable of producing a broad spectrum of disease under a variety of conditions (reviewed by Pollack, 1984). More recently, Heck et al (1986) found Pseudomonas elastase was capable of degrading human collagen, Types III and IV, and that Type I collagen was degraded by both P. aeruginosa elastase and an alkaline protease. Degradation products resulting from the incubation of collagen with the proteases at 25 °C were visualized by staining with Coomassie blue following SDS-PAGE . Their 27 conclusions were that the tissue destruction observed in P. aeruginosa infections may be the result of collagen degradation by nonspecific extracellular proteases. Collagenolytic activity was observed by Mufioz et al (1984) to be related to the virulence of Entamoeba histolytica. This group had previously shown that E. histolytica produced collagenase which was specific for Type I collagen. Their subsequent study in 1984 examined pathogenic and nonpathogenic strains of E. histolytica. The results indicated that collagenolytic activity correlated positively with the virulence of the organism as determined by the ability of the organisms to produce liver lesions in hamsters. This evidence for the relationship between collagenolysis and virulence is similar to that documented by Grenier & Mayrand (1987) in their study of pathogenic and nonpathogenic strains of Bacteroides gingivalis. The pathogenicity of Legionella pneumophila has been associated with a tissue-destructive protease (TDP) by Williams et al (1987) and Conlan et al (1988). Conlan et al (1988) was able to demonstrate the in vivo production of TDP in the lungs of infected guinea pigs, while the immunocytochemical techniques of Williams et al (1987) demonstrated the presence of TDP in situ. These findings reveal an important role for TDP in the pathogenesis of Legionnaires' disease. An extracellular protease produced by Aeromonas salmonicida is thought to be a virulence factor in the pathogenesis of fish furunculosis. A study by Sakai (1985) demonstrated the importance of the protease in disease by inducing a protease-deficient mutant which, when injected into salmon and trout, produced neither disease nor mortality. Serratia marcesens produces a number of extracellular proteases which are capable of degrading defense-oriented humoral proteins and tissue components. Molla et al (1986) determined that a virulent strain of Serratia produced an extracellular protease which cleaved IgG and lgA1. The protease was not 28 inhibited by endogenous human proteinase inhibitors, 3-1-protease inhibitor and 3-2-macroglobulin, and the inhibitors themselves were completely degraded. Extensive pulmonary edema and hemorrhage was noted by Lyerly & Keger (1983) following administration of a purified protease from Serratia marcescens to the lungs of guinea pigs and mice. The tissue destruction observed was similar to that produced during acute Serratia pneumonia. Aims of the study All of the organisms described have in common a pathogenic role in specific diseases. The ability of these organisms to produce disease appears to be the result of the combined effects of proteolytic activity, either cell-free or cell-associated. The organism with which this study was concerned was Bacteroides gingivalis and the specific aims of the investigation were to isolate, purify and characterize two membrane-associated proteases. The importance of identifying and characterizing the proteolytic activity of B. gingivalis relates to its presumed role in the pathogenesis and progression of various periodontal diseases. The findings of many investigators indicate potentially destructive effects of B. gingivalis proteases in the form of perturbation of the host defense system and direct tissue destruction. The degradation of host proteins is a potential means by which the organism can satisfy its metabolic requirements. The clinical aspect of the identification of Bacteroides gingivalis proteases is that it may be possible to use specific protease inhibitors in the prevention or treatment of periodontal diseases. 29 MATERIALS and METHODS Bacterial strain and culture conditions Bacteroides gingivalis, strain 33277, was grown in Brain Heart Infusion broth (3.7%; Difco) containing hemin (10 ug/ml; Sigma) and vitamin K (1 ug/ml; Sigma). The cultures were incubated in an anaerobic chamber ( N 2 - H 2 - C 0 2 [85:10:5]) at 37 °C. Isolation and purification of proteases Protease was prepared from the two sources of outer membranes outlined below: Preparation of outer membranes 1 . Outer membranes-1 (OM-1): Outer membranes were prepared following the method of Boyd and McBride (1984) and Grenier and McBride (1987). Bacterial cells from Bacteroides gingivalis, strain 33277, were harvested by centrifugation at 8,000 x g from an early-stationary-phase culture. The cells were washed twice in 0.15 M NaCI and suspended in 50 mM phosphate buffer (pH 7.4) containing 0.15 M NaCI. To remove the outer membranes, the cells were sheared through needles of progressively smaller gauge to a final needle size of 26 1/2 gauge. This was followed by mixing in a Waring blender for seven 30-second periods. The blender jar and contents were placed in an ice bath for 2 minutes following each 30-second mixing. The mixture was centrifuged for 20 minutes at 8,000 x g to remove whole cells and debris. The resulting supernatant was centrifuged for 2 hours at 80,000 x g and the gel-like translucent pellets of outer membranes were suspended in distilled water and lyophilized. 2. Outer membranes-2 (OM-2): A cell extract was prepared according to the method of Grenier, Chao, and McBride (in press). Sixty grams (wet weight) of cells treated as described in the previous section (OM-1) were sonicated (discontinous sonication, five 3-minute periods, with continuous cooling in an ice bath, and 2 minute resting periods between each sonication; 30% duty cycle, output 5; Sonifier Cell Disrupter 350; Branson Sonic Power Co.). The sonicated 30 mixture was centrifuged twice (20 minutes at 8,000 x g) to remove cell debris, and the resulting cell extract was concentrated to approximately 1/3 of its original volume by lyophilization. The findings of Grenier, Chao, and McBride (in press) show that the proteolytic activity of the cell extract is identical to that of purified outer membranes and furthermore, this activity can be removed from the cell extract by ultracentrifugation. These findings are good evidence that the activity in the cell extract is due to the presence of outer membrane fragments, and not to soluble proteases unique to this cellular fraction. Preparative polyacrylamide gel e lectrophoresis (PAGE) 1. outer membranes-1: Preparative electrophoresis, with the buffer system of Laemmli (1970), was used to accomplish the initial separation of the proteases. One hundred mg of lyophilized outer membranes were suspended in 3.5 ml of 50 mM Tris hydrochloride buffer (pH 7.2) and sonicated ( 1 x 1 0 s). The membranes were then solubilized by adding 2.5 ml of sodium dodecyl sulphate (SDS) solubilization buffer (0.125 M Tris hydrochloride, 4% SDS, 20% glycerol, 1% bromphenol blue) and incubating the mixture for 30 minutes at 37 °C. This concentration and volume of outer membrane solution was loaded onto 6 mm thick, 12% (wt/vol) polyacrylamide resolving gels with 4.5% (wt/vol) polyacrylamide stacking gels. Electrophoresis was carried out at a constant current of 80 mA for two gels, with cooling, for 16 hours (overnight). Following electrophoresis, a narrow vertical strip (approximately 2 cm wide) was cut from one side of the gel and the remainder of the gel stored at 4 °C. The strip was treated according to the method of Foltmann et al (1985 ) in order to detect proteolytic activity. The gel strip was washed with 2.5% Triton-X 100 in 0.3 M sodium acetate buffer (pH 5.3) for 1 hour to remove SDS. The strip was then equilibrated in the acetate buffer without Triton X-100 for 1 hour. 31 Proteolytic activity was detected by placing the gel strip upon a previously prepared 1% agarose gel containing 1% skim milk and 25 mM dithiothreitol (DTT). After incubation for 5-6 hours at 37 °C , proteolytic activity was detected as white bands in a clear matrix. Two of these bands, representing protease material, were selected for further purification and characterization (P-1 and P-4). The designations P-1 and P-4 are taken from the work of Grenier, Chao, and McBride (in press) which describes the proteolytic profile of Bacteroides gingivalis. The profiles of the proteolytic activities found in Bacteroides gingivalis culture supernatants, outer membranes, vesicles and cell extracts were analyzed in SDS-polyacrylamide gels containing covalently bound bovine serum albumin. Proteolytic activity of P-1 (M r 200,000 D) and P-4 (M r 80,000 D) was found in culture supernatants, in cell extracts, in purified outer membranes and in the vesicle preparation of Bacteroides gingivalis . The section of the gel corresponding to proteolytic activity was excised from the remaining gel slab, cut into small pieces and suspended in 10 mM Tris hydrochloride buffer (pH 7.0). The proteins were eluted overnight at 4 °C with constant agitation. The eluate from P-1 and from P-4 was dialyzed against distilled water for 8 hours at 4 °C and then lyophilized. The lyophilized eluates (P-1 and P-4) from the preparative electrophoresis procedure were suspended separately in 1.0 ml of 10 mM Tris hydrochloride buffer (pH 7.2). Solubilization of the protein contained in the fractions was accomplished by adding 0.5 ml of SDS-solubilization buffer and incubating the solutions for 30 minutes at 37 °C . The solutions were then loaded onto 1.5 mm thick, 10% (wt/vol) polyacrylamide resolving gels with 4.5% (wt/vol) polyacrylamide stacking gels and subjected to a second electrophoresis. The electrophoretic purification procedure was carried out at a constant current of 20 mA for two gels, with cooling, for 16 hours (overnight). The areas of proteolytic activity 32 were detected and prepared as described above. The electrophoretic purification procedure was repeated once more using the lyophilized eluate from the previous run as the sample. The partially purified proteases were suspended in 50 mM Tris hydrochloride buffer (pH 7.0) and kept at -20 ° C . 2. outer membranes 2: Aliquots of 5 ml of the cell extract were incubated with 2 ml of the SDS-solubilization buffer for 30 minutes at 37 °C . The electrophoretic purification procedure was carried out on 6.0 mm gels and 1.5 mm gels as outlined above for outer membranes-1, and the purified protease material suspended and stored in the same manner as that purified from the outer membrane-1 material. The proteases isolated and purified from both OM-1 and OM-2 were used in the characterization studies. All of the results reported, except those for peptidase activity, were obtained from proteases purified from OM-2. Protein determination The amount of protein in the purified fractions and in the crude fractions of OM-1 and OM-2 was determined according to the Bradford (Bio-Rad) assay for protein. A standard curve was prepared using bovine serum albumin (BSA) standard stock solution (100 mg/ml; Sigma) diluted with distilled water to 25 pg/ml. Aliquots of this diluted standard solution, containing 0 to 20 u,g of protein, were reacted with 0.2 ml of the Bio-Rad colour reagent and the absorbance measured at A 5 g 5 Protein profiles of OM-1 and OM-2 Electrophoresis of OM-1 and OM-2 was carried out on mini-slab gels (0.075 cm thicknesss; Bio-Rad Laboratories, Richmond, Calif.) containing 12% (wt/vol) polyacrylamide in the resolving gel and 4.5% (wt/vol) polyacrylamide in the stacking gel. Two sets of samples were prepared: (i) lyophilized OM-1 and OM-2 were incubated separately with the SDS-solubilization buffer for 30 minutes at 37 °C (nonboiled fractions), and 33 (ii) lyophilized OM-1 and OM-2 were incubated separately with the SDS-solubilization buffer for 5 minutes at 100 °C (boiled fractions). Aliquots from both preparations, containing 30 u.g of protein, were loaded onto the gels and electrophoresis carried out at room temperature at a constant current of 200 volts (for two gels). An aliquot containing standard molecular weight proteins was loaded onto the gels as a reference. The reference proteins included myosin ([H chain], 200,000 D), phosphorylase b (97,000 D), bovine serum albumin (68,000 D), ovalbumin (43,000 D), <?-chymotrypsinogen (25,700 D) and /Mactoglobulin (18,400 D). Following electrophoresis, the protein profiles were visualized by staining with silver nitrate following a modification of the procedure of Oakley et al (1980). Protein profiles of P-1 and P-4 The molecular weight of the proteases was determined by SDS-PAGE using 12% (wt/vol) polyacrylamide mini-slab gels. The migration of the proteases was compared to the migration of known protein standards (as described above). Aliquots of each of the proteases were incubated with SDS-solubilization buffer for either 5 minutes at 100 °C, or for 30 minutes at 37 °C prior to loading onto the gels. Following electrophoresis, the gels were stained with silver nitrate (Oakley et al, 1980). Lipopolysaccharide (LPS) was determined according to the method of Tsai and Frasch (1980). Proteolytic activity of OM-1, OM-2, P-1 and P-4 In addition to the silver stained samples, OM-1 and OM-2 were compared for proteolytic activity, along with an outer membrane vesicle preparation (kindly provided by Dr. Daniel Grenier, University of British Columbia), and the partially purified protease fractions, using the method of Kelleher and Juliano (1984) and D. Grenier & B.C. McBride (J. Dent. Res.67:368, abstr. 2045, 1988). Bovine serum albumin (BSA) was incorporated into the SDS-polyacrylamide gels to provide the substrate for proteolysis. 34 The BSA-acrylamide conjugate was prepared according to the method of Kelleher and Juliano (1984). Four hundred mg of linear polyacrylamide were dissolved in 20 ml of 0.2 M sodium phosphate buffer (PBS), pH 6.8, and then mixed with 8 ml of 25% gluteraldehyde (in water). The mixture was incubated for 24 hours at 37 °C. The non-conjugated gluteraldehyde was removed by dialyzing the mixture against 10 L of distilled water at 4 °C for four days with three changes of water per day. To the dialyzed mixture 2 ml of 10% BSA (0.2 M PBS, pH 7.2) were added and the mixture allowed to react for 24 hours at 25 °C. Following incubation, the reaction was halted by adding 250 mg of glycine and 3 mg of sodium azide. The mixture was then incubated for 24 hours at 4 °C and again dialyzed against 10 L of distilled water containing 0.05% sodium azide at 4 °C for 2 days. The BSA-acrylamide conjugate was stored at 4 °C. Proteolytic activity was determined by analyzing the samples in SDS-polyacrylamide gels containing covalently bound bovine serum albumin (BSA). Briefly, the electrophoresis was carried out as described above, using 10% (wt/vol) polyacrylamide in the resolving gel. Prior to casting, the polyacrylamide was mixed with BSA-acrylamide conjugate to give a concentration of 200 ug of protein per ml of gel (5% [vol/vol] BSA-conjugate in the total casting volume). The crude fractions and partially purified proteases were prepared by incubation with the SDS-solubilization buffer for 30 minutes at 37 ° C . Aliquots of the prepared samples were loaded onto the BSA-polyacrylamide mini-slab gels and subjected to electrophoresis at a constant current of 200 volts (for two gels). Following electrophoresis, the gel was shaken gently for 30 minutes in 100 mM Tris hydrochloride buffer (pH 7.0) containing 2% Triton X-100, washed twice in distilled water and then equilibrated in the Tris buffer without Triton X-100 for 30 minutes. After equilibration, the gel was incubated for 2 hours at 37 °C in a development buffer of 100 mM Tris hydrochloride (pH 7.0) containing 2.5 mM calcium chloride and 50 mM L-cysteine as reducing agent. The gel was subsequently stained for protein with Coomassie brilliant blue R-250. After 35 destaining, the proteolytic activity could be visualized as clear bands against a dark blue background. D e t e r m i n a t i o n of p e p t i d a s e a c t i v i t y Peptidase activity was determined according to the method of Berdal and Olsvik (1983). Each of the proteases was incubated with each of the chromogenic peptides in the same manner as outlined below for the BAPNA (N-<?-Benzoyl-D-arginine-p-nitroanilide) assay: Briefly, 75 pi of each of the proteases was incubated with 75 uJ of a 2 mM solution of each of the chromogenic peptides in a reaction mixture containing 225 u_l of 50 mM Tris hydrochloride buffer (pH7.5 at 37 °C) and 2 mM DTT . Incubation was carried out overnight at 37 °C, following which, the release of p-nitroaniline was measured after diazotization. Trichloroacetic acid (40% in water, 375 pi) was added to the mixture followed by 112.5 u.l of freshly prepared sodium nitrite (0.1% in water). After 5 minutes at room temperature, 112.5 pi of sodium sulphamate (0.5% in water) was added and the reaction mixture incubated for 5 minutes at room temperature. Lastly, 112.5 uJ of freshly prepared N-1-naphthylethylene diamine dihydrochloride (0.1% in water) was added and following 5 minutes incubation at room temperature, the absorbance at A 5 4 5 was recorded. The diazotization procedure alters the absorption maxima and increases the sensitivity of the test. Determination of peptidase activity was semi-quantitave based upon the following scale: no colour reaction (0), faint red colour (+/-), clearly visible red colour (+), strong red colour (++). D e t e r m i n a t i o n of o p t i m u m pH A reaction mixture was prepared containing 75 u.l of the partially purified proteases (P-1, 1.7 u,g total protein; P-4, 1.1 pg total protein), 75 u.l of 2 mM N-<?-Benzoyl-D-arginine-p-nitroanilide (BAPNA) and 225 u.l of one of 36 the following buffers: 0.2 M acetate buffer, pH 4.8 and 5.4; 0.2 M phosphate buffer, pH 5.6 to 8.1; 0.2 M Tris hydrochloride buffer, pH 7.0 to 8.6. Each of the buffers contained sufficient DTT such that the reaction mixture was 2 mM in DTT. The reaction mixture was incubated overnight at 37 °C and the release of p-nitroaniline was measured after diazotization as outlined above for the BAPNA assay. Determination of optimum temperature and heat stability Using the BAPNA assay outlined above, the optimum temperature for activity was determined by incubating the substrate overnight with the proteases in a sample mixture containing 50 mM Tris hydrochloride buffer, pH 7.5, and 2 mM DTT. The incubation was carried out at 25, 30, 37, 45 °C. The heat stability of the proteases was determined as above, except that the protease used (P-4) was incubated at temperatures of 4, 37, 55, 80, and 100 ° C for 30 minutes, 2 hours or 24 hours prior to determination of enzyme activity using the BAPNA assay. Inhibition of protease activity Protease activity was determined in the presence of various protease inhibitors and metal ions: EDTA, phenylmethylsulfonyl fluoride (PMSF), N-9-p-tosyl-L-lysine chloromethyl ketone hydrochloride (TLCK), L-1-tosylamine-2-phenyIethyIchloromethylketone (TPCK), SDS, iodoacetate, p-chloromercuribenzoic acid (PCMB), aprotinin, CaCI 2 , HgCI 2, MgCI 2 , ZnCI 2 . Each of the purified proteases was incubated with each of the inhibitors for 30 minutes at 37 °C prior to determination of protease activity using the BAPNA assay . A positive control containing protease, substrate and buffer with DTT, but no inhibitor, was taken to be 100% reactive. The relative reactivity of the proteases following incubation with each of the selected inhibitors was calculated on this basis. 37 Hydrolysis of protein substrates by P-1 and P-4 Proteolytic activity against azocoll and azoalbumin was determined according to the method of Mayrand and McBride (1980). The reaction mixture for the azocoll assay contained 15 u.l of purified protease, 50 u.l of azocoll (30%), 450 pi of 50 mM Tris hydrochloride buffer (pH 7.5 at 37 °C) and 2 mM DTT. The mixture was incubated overnight at 37 °C, subsequently centrifuged at 12,000 x g for 3 minutes and the absorbance of the supernatant read at A 5 8 0 . A sample mixture containing azocoll, buffer and DTT served as the control. The reaction mixture for the azoalbumin assay contained 100 u.l of purified protease, 500 u1 of azoalbumin (3% in 50 mM Tris hydrochloride buffer, pH 7.5 at 37 °C), and 25 mM DTT. The mixture was incubated overnight at 37 °C. Following the incubation, 500 u.1 of 20% trichloroacetic acid (TCA) was added to stop the reaction and the resulting precipitate was removed by centrifuging at 12,000 x g for 3 minutes. To the resulting supernatant (TCA-soluble material) 150 pi of 6 N NaOH was added, and the resulting absorbance read at A 3 7 5 . A sample mixture containing azoalbumin, buffer and DTT served as the control. Proteolysis of IgA, IgG, gelatin, and acid-soluble collagen was determined by assaying for lower-molecular-weight degradation products by SDS-PAGE. The proteins in question were incubated overnight at 37 °C in a reaction mixture containing 25 uJ of partially purified protease, 25 u,l of protein substrate (1 mg/ml), 25 u.l of 50 mM Tris hydrochloride buffer (pH 7.5 at 37 °C), and 2 mM DTT. Subsequently , the samples were incubated for 5 minutes at 100 °C with 20 u.l of SDS-solubilization buffer. Aliquots of each of the samples were loaded onto mini-slab gels containing an appropriate polyacrylamide concentration to assay for degradation products. After electrophoresis, the gels were stained with Coomassie brilliant blue R-250. Control samples were prepared containing the same reaction mixture but these were not incubated at 37 °C, instead solubilization buffer was immediately added to the samples and the 38 mixtures incubated for 5 minutes at 100 °C. D e t e r m i n a t i o n of i m m u n o r e a c t i v i t y The antibody used in this determination was prepared and kindly provided by Dr. Daniel Grenier, University of British Columbia. The antibody was against a membrane-associated Bacteroides gingivalis glycylprolyl protease (Grenier and McBride, 1987) and did not contain antibody against lipopolysaccharide. Cross-reactivity between the partially purified proteases (P-1 and P-4) and the prepared antibody was determined by Western blotting (Burnette, 1981). Electrophoresis was carried out as described above on a 12% (wt/vol) polyacrylamide resolving gel onto which had been loaded aliquots of P-1 (0.68 u.g of total protein), P-4 (0.45 u.g of total protein), crude OM-1 (20 u.g of total protein), crude OM-2 (20 u,g of total protein) and outer membrane vesicles (10 u.g of total protein). After the run, electrophoretic blot transfer of material was conducted according to the method of Burnette (1981). The gel was placed upon three layers of 3M Whatman filter paper previously wetted in the blotting buffer (glycine [14.4 g/L], Tris hydrochloride [3.02 g/L], methanol [200 ml/L], pH 8.3). Nitrocellulose paper, previously wetted in Tris buffered saline (TBS; 20 mM Tris hydrochloride, 0.5 M NaCI; pH 7.5) was placed upon the gel and another three layers of wetted filter paper placed over the nitrocellulose paper. The whole stack was placed into the holding cassette of the Bio-Rad Blot cell (Bio-Rad Laboratories, Richmond, Calif.) and the cell filled with blotting buffer. Electrophoretic transfer was carried out for 2 hours at room temperature with a constant current of 60 volts. Following electrophoresis, the nitrocellulose was first shaken gently with blocking buffer (3% BSA) for 1 hour, and then with the first antibody (anti-glycylproyl protease, 1:400 in 1% BSA) for 1 hour. After incubation with the first antibody, the nitrocellulose was washed with distilled water followed by washing for 20 minutes in 2-100 ml changes of TBS 39 containing 0.05% Tween-20 (TTBS). Subsequently, the nitrocellulose was shaken gently with the second antibody (goat anti-mouse IgG horseradish peroxidase conjugate [Bio-Rad], 1:3000 in 1% BSA) for 1 hour, followed by washing in distilled water and TTBS as outlined above. Staining of the antigen-antibody complexes with antibody raised in rabbit to the glycylprolyl protease was carried out with the Bio-Rad Immun Blot (GAR-HRP) assay kit. C h e m i c a l s Synthetic peptides, IgG, IgA, azocoll, azoalbumin , acid-soluble collagen and the protease inhibitors PMSF, TPCK, TLCK, PCMB were purchased from Sigma. Electrophoresis chemicals (glycine, SDS , acrylamide, bis-acrylamide, ammonium persulfate and TEMED) were purchased from Bio-Rad. 40 RESULTS P u r i f i c a t i o n Compar ison of OM-1 and OM-2. The yield of purified outer membranes was 233 mg, dry weight, (7.8 mg per liter of culture) in one preparation, from an original wet weight of cells of 104 gm (3.5 gm per liter of culture). In a second preparation, 560 mg, dry weight, (14.0 mg per liter of culture), of outer membranes were purified from an original wet weight of cells of 155 gm (3.9 gm per liter of culture). The protein content of the outer membrane fractions was determined to be 1.5 mg/ml; the cell extract contained 2.0 mg/ml of protein. The protein profile of the outer membrane preparation (OM-1) obtained by silver staining (Fig. 1) was similar to that previously reported in Boyd & McBride (1984) and Grenier & McBride (1987). The pattern was similar, but not identical, to that of a cell extract prepared by sonication, as outlined in Materials and Methods for outer membranes-2 (OM-2, Fig. 2). The major bands revealed in the profile of OM-2 had the same M r as those of OM-1, the differences being in the number and intensity of staining of the minor bands. An apparent difference is the absence of several low molecular weight bands in the protein profile of OM-2 (Fig. 2, lanes E, F, G, H ) as compared to the protein profile of OM-1 (Fig. 1, lanes E, F, G, H). The low molecular weight bands are not visible in either the boiled or nonboiled samples of OM-2 (Fig. 2, lanes E, F, G, H). A possible explanation is that the integrity of the proteins in OM-1 are preserved whereas the proteins in OM-2 are subjected to proteolytic activity from proteases released during cell sonication. Electrophoresis and silver staining of boiled samples of both OM-1 and OM-2 produced major bands corresponding to M r of 68 kD , 45 kD and 40 kD. A number of minor bands corresponding to M r both higher and lower than these major bands were apparent. 41 Electrophoresis followed by silver staining of the nonboiled samples revealed a major band corresponding to an M r of 45 kD in the OM-1 preparation (Fig. 1, lanes B and F) and a similar major band corresponding to 40 kD in the OM-2 preparation (Fig. 2, lanes B and F). These major bands were also found in the respective protein profiles of the boiled samples (Figs. 1 and 2, lanes A and E). In Figures 1 and 2, lanes C, D, G, and H show the protein profiles of the boiled and nonboiled samples following solubilization in buffer containing reducing agent (5-mercaptoethanol). When compared with the protein profiles of the boiled and nonboiled samples (Figs.1 and 2, lanes A, B, E, F) solubilized in buffer without reducing agent, it apparent that there are fewer bands in those lanes which contain samples solubilized in buffer containing reducing agent (Figs. 1 and 2, lanes C, D, G, H), most noticeable when comparing lanes F and H. It is likely that the reducing agent is able to protect the sulfhydryl bonds in the proteases, thus resulting in fewer visible bands. The skim milk zymogram obtained following preparative gel electrophoresis of OM-1 and OM-2 revealed three proteolytic bands with apparent molecular weights above 68,000 D (Fig. 3). The bands were not similar in their intensity of activity, the lower molecular weight band (P-4) and the intermediate band having more activity against the casein substrate than the higher molecular weight band (P-1). The relative intensity of the bands was consistent from one preparation to another. P-1 and P-4 were selected for further purification and characterization. 42 Figure 1. S D S - P A G E of outer membrane-1 preparation. Lane A: boiled OM-1, 0.2 ug dry weight, SDS-sol buffer* Lane B: nonboiled OM-1, 0.2 ug dry weight, SDS-sol buffer Lane C: boiled OM-1, 0.2 u.g dry weight, /J-ME-sol buffer** Lane D: nonboiled OM-1, 0.2 ug, dry weight, /3-ME-sol buffer Lane E: boiled OM-1, 2.0 ug dry weight, SDS-sol buffer Lane F: nonboiled OM-1, 2.0 ug dry weight, SDS-sol buffer Lane G: boiled OM-1, 2.0 ug dry weight, tf-ME-sol buffer Lane H: nonboiled OM-1, 2.0 ug dry weight, /3-ME-sol buffer * sodium dodecyl sulfate solubilization buffer ** fl-mercaptoethanol solubilization buffer Arrows indicate molecular weight markers, from top to bottom, respectively: 97,000 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) 25,700 D (<9-chymotrypsin) 4 3 43a Figure 2. S D S - P A G E of outer membrane-2 preparation. Lane A: boiled OM-2, 0.1 x 10"3 ug total protein, SDS-sol buffer* Lane B: nonboiled OM-2, 0.1 x 10' 3 ug total protein, SDS-sol buffer** Lane C: boiled OM-2, 0.1 x 10"3 ug total protein, fl-ME-sol buffer Lane D: nonboiled OM-2, 01. x 10 - 3 ug total protein, /?-ME-sol buffer Lane E: boiled OM-2, 1.0 x 10~2 ug totall protein, SDS-sol buffer Lane F: nonboiled OM-2, 1.0 x 10 - 2 ug total protein, SDS-sol buffer Lane G: boiled OM-2, 1.0 x 10"2 ug total protein, fl-ME-sol buffer Lane H: nonboiled OM-2, 1.0 x 10 - 2 u.g total protein, tf-ME-sol buffer * sodium dodecyl sulfate solubilization buffer tf-mercaptoethanol solubilization buffer Arrows indicate molecular weight markers, from top to bottom, respectively: 97,400 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) 25,700 D (<9-chymotrypsin) 44 44a Figure 3. Zymogram of proteases after preparative SDS-PAGE. 45 45a Proteolytic activity of samples against bovine serum albumin (BSA) The protocol outlined in Materials and Methods for the detection of proteolytic activity on BSA-polyacrylamide gels was selected following a number of experiments to determine the optimum length of time for treatment of the gels in Triton-X-100 and for the incubation of the gels in the development buffer. The test gels were subjected to treatment times in Triton X-100 of 15, 30, 60 or 120 minutes whereas the incubation times in development buffer were varied from 0.5,1.5, 2.0 to 5.0 hours. The optimal concentration of BSA in the casting volume was determined to be 5% after testing the protocol using gels containing 4, 5 or 15% BSA-conjugate (vol/vol) in the total casting volume. The optimal treatment time in Triton-X-100 was 30 minutes whereas the optimal incubation time in development buffer was 2 hours. Determination of total protein After the third preparative PAGE, the amount of protein contained in the proteases purified from the OM-1 preparation was determined according to the Bradford (Bio-Rad) protein assay, and found to be 1.2 mg/ml (P-1) and 1.6 mg/ml (P-4), whereas the same proteases purified from the OM-2 preparation contained 67.5 u,g/ml (P-1) and 45.0 pg/ml (P-4). The lower protein content found in P-1 and P-4 purified from the OM-2 preparation is not unexpected in that the cells used to obtain the extract had been previously stripped, to a large degree, of the outer membranes with which the proteases are associated. Specific activity of P-1 and P-4 The purification procedures resulted in an increase of specific activity 45-fold and 40-fold greater than that of the OM-2 preparation, P-1 and P-4, respectively (Table 1). 46 TABLE 1. Purification of P-1 and P-4 from the OM-2 preparation of Bacteroides gingivalis 33277 Fraction Protein Total Sp act Pur i f i - Yield (mg) units* U/mg of cation (%) protein OM-2 20 6.0 0.3 1 100 P-1 0.067 0.9 13.4 45 15 ** P-4 0.067 0.8 1 1.9 40 13 ** One unit of enzyme activity is defined as the amount of enzyme which released I (xmol of p-nitroaniline from 2 mM BAPNA in 120 minutes at 37 °C . This is not necessarily the true yield since the original preparation contained a number of BAPNA-degrading proteases. Following the second SDS-PAGE purification, the protein profile of boiled P-1 revealed one major band corresponding to an M r 68 kD and six minor bands corresponding to M r from approximately 50 to 18 kD (Fig. 4, lane E) whereas nonboiled P-1 revealed three major bands corresponding to M r in the range of 200 kD and one minor band with an M r of approximately 50 kD (Fig. 4, lane F). The 50 kD band was present in the profiles of both the nonboiled and boiled samples. The bands were visualized by silver staining the gels after electrophoresis. Following the third and final SDS-PAGE purification, electrophoresis of boiled P-1 produced five bands which corresponded to molecular weights ranging from 68 kD (major band) to 18 kD (Fig.4, lane G). Nonboiled P-1 produced three major bands following electrophoresis. The bands obtained after silver staining corresponded to molecular weights of 235 kD, 220 kD, and 200 kD (Fig. 4, lane H). 47 After the second SDS-PAGE purification, electrophoresis of boiled P-4 produced eight bands corresponding to M r of 68 to 18 kD (Fig. 5, lane E) whereas nonboiled P-4 revealed one major band corresponding to an M r of 200 kD (Fig. 5, lane F). After the third and final SDS-PAGE purification, electrophoresis of boiled P-4 produced two major bands which corresponded to molecular weights of 68 and 50 kD, and a minor band corresponding to 40 kD (Fig. 5, lane G). Nonboiled P-4 produced one major band following electrophoresis which corresponded to a molecular weight of 74 kD (Fig. 5, lane H). The three bands produced following electrophoresis of the boiled P-4 (68, 50, 40 kD) were also present in the protein profile of boiled P-1 (Fig. 4, lane G). The M r of the partially purified proteases determined by silver staining correlated well with the bands of proteolytic activity found in gels in which BSA had been incorporated as a substrate for the proteases (Fig. 6). Proteolytic activity can be visualized (Fig. 6, lanes C and D) in the area of 200 kD (P-1) and at 74 kD (P-4; Fig. 6, lane B). Lane D contains protease (P-1) eluted from the upper one-half of a gel section excised from a preparative gel whereas Lane C contains protease (P-1) eluted from the lower one-half of the same preparative gel section. If the two were combined, it would be equal to the whole P-1 protease as visualized following silver staining in Figure 4. No bands were detected following electrophoresis of the purified proteases and subsequent staining of the gels for lipopolysaccharide. 48 F i g u r e 4. S D S - P A G E of P-1. Lane A: boiled OM-1, 2.0 ug dry weight, SDS-sol buffer* Lane B: nonboiled OM-1, 2.0 u.g dry weight, SDS-sol buffer Lane C: boiled P-1, after preparative SDS-PAGE, SDS-sol buffer Lane D: nonboiled P-1, after preparative SDS-PAGE, SDS-sol buffer Lane E: boiled P-1, after 2nd purification by SDS-PAGE, SDS-sol buffer Lane F: nonboiled P-1, after 2nd purification by SD-PAGE, SDS-sol buffer Lane G: boiled P-1, 0.2 u.g total protein, after 3rd purification by SDS-PAGE, SDS-sol buffer Lane H: nonboiled P-1, 0.2 u.g total protein, after 3rd purification by SDS-PAGE, SDS-sol buffer sodium dodecyl sulfate solubilization buffer Arrows indicate molecular weight markers, from top to bottom, respectively: 200,000 D (myosin, H chain) 97,000 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) 25,700 D (^-chymotrypsin) 18,400 D (/3-lactoglobulin) 49 A B C D E F G H HI jm 49a Figure 5. SDS-PAGE of P-4. Lane A: boiled OM-1, 2.0 u.g dry weight, SDS-sol buffer* Lane B: nonboiled OM-1, 2.0 u.g dry weight, SDS-sol buffer Lane C: boiled P-4, after preparative SDS-PAGE, SDS-sol buffer Lane D: nonboiled P-4, after preparative SDS-PAGE, SDS-sol buffer Lane E: boiled P-4, after 2nd purification by SDS-PAGE, SDS-sol buffer Lane F: nonboiled P-4, after 2nd purification by SDS-PAGE, SDS-sol buffer Lane G: boiled P-4, 0.2 ug total protein, after 3rd purification by SDS-PAGE, SDS-sol buffer Lane H: nonboiled P-4, 0.2 u.g total protein, after 3rd purification by SDS-PAGE, SDS-sol buffer * sodium dodecyl sulfate solubilization buffer Arrows indicate molecular weight markers, from top to bottom, respectively: 200,000 D (myosin, H chain) 97,000 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) 25,700 D (<9-chymotrypsinogen) 18,400 D (/Mactoglobulin) 5 0 A B C D E F G H 50a Figure 6. Proteolytic purity of OM-2, P-1 and P-4. Lane A: nonboiled OM-2, 30 u.g total protein, SDS-sol buffer * Lane B: nonboiled P-4, 0.22 u.g total protein, SDS-sol buffer Lane C: nonboiled P-1 (lower 1/2), 0.45 ug total protein, SDS-sol buffer Lane D: nonboiled P-1 (upper 1/2), 0.45 ug total protein, SDS-sol buffer Lane E: molecular weight markers, from top to bottom, respectively: 200,000 D (myosin H chain) 97,400 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) 25,700 D (<9-chymotrypsinogen) * sodium dodecyl solubilization buffer 51 A B C D E 51a P e p t i d a s e a c t i v i t y Peptidase activity is shown in Table 2. There was no difference between the proteases regarding peptidase activity. Both were able to readily hydrolyze substrates containing benzoyl-arginine groups: the dipeptide, N-3-Benzoyl-D-a r g i n i n e - p - N A and the tetrapeptide, N-Benzoyl-L- isoleucyl-L-glutamylglycyl-L-arginine-p-NA. The two dipeptides containing glycine, Gly-L-Pro-p-NA and Gly-L-Arg-p-NA , were less actively hydrolyzed, and the aminopeptide L- -glutamyl-p-NA was only weakly degraded. TABLE 2. Pept idase activity of P-1 and P-4 Substrate P-1 Activity* P-4 Activity* L-Alanine-p-N A 0 0 L-Arginine-p-N A 0 0 L- o'-Glutamyl-p-N A + / - +/-L-Leucine-p-N A 0 0 L-Lysine-p-N A 0 0 L-Methionine-p-N A 0 0 L-Proline-p-N A 0 0 L-Valine-p-N A 0 0 N-3-Benzoyl-D-arginine-p-N A + + + + Glycyl-L-proline-p- N A + / - + / -Glycyl-L-arginine-p-N A + + L-Alany l -L-a lany l -L-phenylalanine-p-NA 0 0 N-Benzoyl-L-isoleucy I-L-glutamylglycyl-L-arginine-p-N A + + + + Methoxysuccinyl-L-alanyl-L-alanyl-L- proline-L-valine-p-N A 0 0 * The determinat ion of pept idase activity on synthet ic subst ra tes was semi-quant i tat ive ba sed upon the fo l lowing s ca l e : no co lour react ion (0), faint red co lour (+/-), c lear ly v is ib le red colour (+), strong red colour (++). 52 O p t i m u m pH The optimum pH for activity of both proteases was found to be between pH 6.0 and 6.5, although there was significant activity over the pH range of 5.5 to 7.0 (Figs. 7 and 8). Both proteases appeared to have significant activity in the acetate buffer whereas activity appeared to be inhibited in Tris hydrochloride buffer. O p t i m u m t e m p e r a t u r e As determined by the BAPNA assay, the optimum temperature for activity for both proteases was found to be 37 °C (Fig. 9). Activity declined at temperatures either higher or lower than this value. P-4 appeared to be much more sensitive to changes in temperature than did P-1. At 45 °C, P-1 retained approximately 90% of the relative reactivity whereas P-4 had only 45% of the relative reactivity at the same temperature. At 25 °C, P-4 exhibited 42% of the relative reactivity whereas P-1 retained 47% of the relative activity at the same temperature. H e a t s t a b i l i t y P-4 was used as the representative sample to determine the heat stability of the partially purified proteases, based upon its apparent sensitivity to changes in temperature. The proteolytic activity of P-4 was completely abolished after 2 hours of incubation at 80 °C or after boiling for 5 minutes. Incubation at temperatures below 55 °C for 30 minutes or for 2 hours had relatively little effect upon proteolytic activity and 41% of the relative activity was retained after 16 hours of incubation at 55 °C. Electrophoresis of P-4 which had been incubated at 37 °C for 48 hours revealed autodegradation. This could be visualized, following silver staining, as a loss of 53 resolution of the protein bands when compared to those of the non-incubated control (Fig. 10). The major band (Fig. 10, lanes A, C, E) corresponds to the 74 kD band seen in Figure 5 (lane H) following electrophoresis of nonboiled P-4. The major bands (Fig. 10, lanes B, D, F) correspond to the 68, 50, and 40 kD bands seen in Figure 5 (lane G) following electrophoresis of boiled P-4. Comparison of the bands in Lanes B, D, and F (Fig. 10) reveals the loss of the 50 kD band in Lane D (boiled P-4, incubated 48 hours at 37 °C). The loss of this band indicates autodegradation of the P-4 protease after this incubation period. The protease remained stable for at least 3 months at -20 °C; repeated freezing and thawing had no apparent effect upon proteolytic activity. 54 Figure 7. Optimum pH for the hydrolysis of BAPNA by P-1. The reaction mixture contained 75 u,l of purified P-1, 75 u.l of 2 mM BAPNA, 225 U.I of buffer, and 2 mM DTT. The following buffers were used: 0.2 M acetate buffer, pH 4.8 and 5.4; 0.2 M phosphate buffer, pH 5.6 to 8.1; 0.2 M Tris hydrochloride buffer, pH 7.0 to 8.6. The reaction mixture was incubated overnight at 37 °C and the release of p-nitroaniline was measured after diazotization as outlined for the BAPNA assay in Materials and Methods. 55 55a Figure 8. Optimum pH for the hydrolysis of BAPNA by P-4. The reaction mixture contained 75 u.1 of purified P-4, 75 u.1 of 2 mM BAPNA, 225 u.1 of buffer and 2 mM DTT. The following buffers were used: 0.2 M acetate buffer, pH 4.8 and 5.4; 0.2 phosphate buffer, pH 5.6 to 8.1; 0.2 Tris hydrochloride buffer, pH 7.0 to 8.6. The reaction mixture was incubated overnight at 37 °C and the release of p-nitroaniline was measured after diazotization, as outlined for the BAPNA assay in Materials and Methods. 56 56a F igure 9. Op t imum temperature for the hyd ro l y s i s of B A P N A by P-1 and P-4. The reaction mixtures contained 75 u.l of either P-1 or P-4, 75 u.l of 2 mM BAPNA, 225 u.1 of Tris hydrochloride buffer, pH 7.5 and 2 mM DTT. The reaction mixtures were incubated overnight at the following temperatures: 25, 30, 37, and 45 °C. The release of p-nitroaniline was measured after diazotization as outlined for the BAPNA assay in Materials and Methods. 57 110 -i Temp. 57a Figure 10. Heat stability of P-4. Lane A: nonboiled P-4, incubated 48 hours at 4 °C Lane B: boiled P-4, incubated 48 hours at 4 °C Lane C: nonboiled P-4, incubated 48 hours at 37 °C Lane D: boiled P-4, incubated 48 hours at 37 °C Lane E: nonboiled P-4, control maintained at -20 °C Lane F: boiled P-4, control maintained at -20 °C Lane G: molecular weight markers: 97,400 D (phosphorylase b) 43,000 D (ovalbumin) 25,700 D (3-chymotrypsinogen) 58 58a Inhibition of enzyme activity The effect of various inhibitors and metal ions on the hydrolysis of BAPNA is shown in Table 3. The proteases were strongly inhibited by the chymotrypsin inhibitor TPCK, and by the thiol inhibitors PCMB, iodoacetate and Hg ions. Inhibition occurred to a lesser extent from the serine protease inhibitor PMSF, from the trypsin inhibitor TLCK, from the anionic detergent SDS, and from Zn ions, which are inhibitory to thiol proteases. The activity of the proteases against BAPNA was enhanced by EDTA, aprotinin, and Ca and Mg ions. T A B L E 3. Effect of protease inhibitors and metal ions on enzyme activity Compound Concn (mM) P-1 % Res idua l a c t i v i t y P-4 % Res idua l a c t i v i t y None 100 100 EDTA 20 304 241 PMSF 4 44 85 TLCK 2 44 41 TPCK 2 8 7 SDS 20 50 29 PCMB 4 2 0 Iodoacetate 2 0 9 Aprotinin 8 ]ig/ml 318 234 CaCI 2 2 217 203 HgCI 2 2 2 1 MgCI 2 2 323 3 4 7 ZnCI 2 2 74 60 * Activity wa s measured by the B A P N A assay as outl ined in Material and Methods. 59 Activity against protein substrates The activity of the proteases against protein substrates is shown in Table 4. Both of the proteases were able to degrade gelatin, IgA, IgG and the two azo substrates, azocoll and azoalbumin, under the conditions of the experiments. Acid-soluble collagen was not hydrolyzed by either protease during overnight incubation at 25 °C . T A B L E 4. Proteolytic activity of P-1 and P-4 Subs t r a t e P-1 Activity* P-4 Activity* Azocoll + + Azoalbumin + + Gelatin* + + Acid-soluble collagen* IgG* + + IgA* + + * The determinat ion of activity against proteins was based upon the abil ity of the proteases to produce fragments of lower molecular weight as compared to the original protein profile seen in polyacrylamide gels stained with Coomass i e blue. The degradation of IgA and IgG is apparent (Fig. 11) as a loss of bands following overnight incubation (37 ° C ) of the substrates with the proteases (Fig. 11, lanes B, D, F, H) as compared with the IgA and IgG samples which were not incubated with the proteases (Fig. 11, lanes A, C , E, G). Gelatin was completely degraded by both P-1 and P-4 (Fig. 12). The degradation is indicated by the loss of bands (Fig. 12, lane B, gelatin with P-1 and lane F, gelatin with P-4) following overnight incubation (37 ° C ) of the substrate with the proteases as compared to the profiles seen in lanes A and E (Fig. 12) which contain samples of non-incubated substrate and protease. The 68 kD band seen in 60 lanes A, B, C, and D (Fig. 12) may be a contaminant related to the P-1 protease rather than to the substrate, since it appears to remain unaltered following the overnight incubation. It was initially thought that P-1 had collagenolytic activity since acid-soluble collagen was completely degraded following overnight incubation with the protease at 37 °C (Fig. 12, lane D). Therefore, a subsequent experiment, having a reaction mixture which comprised acid-soluble collagen, boiled P-1, a non-specific protease (trypsin), Tris hydrochloride buffer and DTT, was completed following identical experimental conditions. After electrophoresis of the samples, it was evident that collagenolysis had occurred in the presence of trypsin and boiled Fraction 1 (Fig. 13, lane B). Since boiling was previously shown to completely abolish the proteolytic activity of the purified protease, the degradation which occurred could not be the result of proteolytic activity due to P-1. The results suggest that the non-specific protease (trypsin) was responsible for the collagenolytic activity in concert with the disruption of the helical structure of collagen which occurs at 37 °C (Fig. 13, lane B). The collagenolytic activity of the nonboiled P-1 could be the result of a similar effect upon acid-soluble collagen during incubation at 37 °C . Alternatively, this could result from a small amount of residual SDS in the P-1 protease which may disrupt the tertiary structure of the collagen molecule, making it accessible to a non-specific protease. 61 Figure 1 1 . Hydrolysis of immunoglobulins A and G by P-1 and P-4 . Lane A: P-1 with IgA, T 0 Lane B: P-1 with IgA, T 1 8 * Lane C: P-4 with IgA, T 0 Lane D: P-4 with IgA, T 1 8 Lane E: P-1 with IgG, T 0 Lane F: P-1 with IgG, T 1 8 Lane G: P-4 with IgG, T 0 Lane H: P-4 with IgG, T 1 8 * no incubation ** overnight incubation of protease fraction with substrate at 37 °C Arrows indicate molecular weight markers, from top to bottom, respectively: 97,400 D (phosphorylase b) 43,000 D (ovalbumin) 25,700 D (^-chymotrypsin) The reaction mixtures contained 25 pi protease, 25 pi IgA or IgG (1 mg/ml water), 25 u.1 Tris hydrochloride buffer, pH 7.5 at 37 °C , 2mM DTT. The mixtures were either immediately boiled for 5 minutes with 20 u.1 of SDS-solubilization buffer, or incubated overnight at 37 °C followed by boiling for 5 minutes with 20 u.1 of SDS-solubilization buffer. Aliquots of 20 pi were loaded onto a 12% (wt/vol) mini-slab gel. The gel was stained with Coomassie blue dye after electrophoresis. 62 A B C D E F G H 62a Figure 12. Hydrolysis of gelatin and collagen by P-1 and P-4. Lane A: P-1 with gelatin, T 0* Lane B: P-1 with gelatin, T 1 8 * Lane C: P-1 with collagen, T 0 Lane D: P-1 with collagen, T-|8 Lane E: P-4 with gelatin, T 0 Lane F: P-4 with gelatin, T 1 8 Lane G: P-4 with collagen, T 0 Lane H: P-4 with collagen, T 1 8 no incubation overnight incubation of protease fraction and substrate at 37 °C Arrows indicate molecular weight markers, from top to bottom, respectively: 200,000 D (myosin H chain) 97,400 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) The reaction mixtures contained 25 u.l protease fraction, 25 uJ gelatin (1 mg/ml) or acid-soluble collagen (1 mg/ml 0.1% acetic acid), 25 u.l Tris hydrochloride buffer, pH 7.5 at 37 °C, 2 mM DTT. The mixtures were either immediately boiled for 5 minutes with 20 u,l of SDS-solubilization buffer, or incubated overnight at 37 °C followed by boiling for 5 minutes with 20 ju.l of SDS-solubilization buffer. Aliquots of 20 uJ were loaded onto a 7.5% (wt/vol) polyacrylamide resolving mini-slab gel. The gel was stained with Coomassie blue dye after electrophoresis. 63 A B C D E F G H 63a Figure 13. Hydrolysis of collagen by trypsin and P-1. Lane A: boiled P-1, trypsin and collagen, To* Lane B: boiled P-1, trypsin and collagen, T18* Lane C: nonboiled P-1, trypsin and collagen, To Lane D: nonboiled P-1, trypsin and collagen, T18 Lane E: trypsin and collagen, To Lane F: trypsin and collagen, T18 * no incubation overnight incubation of P-1 and /or tryspin with substrate at 37 °C Arrows indicate molecular weight markers, from top to bottom, respectively: 200,000 D (myosin H chain) 97,400 D (phosphorylase b) 68,000 D (bovine serum albumin) 43,000 D (ovalbumin) Reaction mixture for Lanes A-D: 25 u.l boiled or nonboiled P-1, 25 u.l acid-soluble collagen (1 mg/ml 0.1% acetic acid), 25 u.l trypsin (0.4 mg/ml water), 25 u.l Tris hydrochloride buffer (pH 7.5 at 37 °C), 2mM DTT. Reaction mixture for Lanes E and F: 25 uJ trypsin (0.4 mg/ml water), 25 u.1 acid-soluble collagen (1 mg/ml 0.1% acetic acid), 25 u.l Tris hydrochloride buffer (pH 7.5 at 37 °C), 2 mM DTT. The reactions mixtures were either boiled immediately (T 0 ) with 20 u.l of the SDS-solubilization buffer for 5 minutes or incubated overnight ( T 1 8 ) at 37 ° C followed by boiling for 5 minutes with 20 u.l of the SDS-solubilization buffer. Aliquots of 20 u,l were loaded onto a 7.5% (wt/vol) polyacrylamide resolving mini-slab gel. The gel was stained with Coomassie blue dye after electrophoresis. 64 A B C D E F 64a Reaction with antibody to glycylprolyl protease No cross-reactivity was found when the partially purified proteases were reacted with antibody to a glycylprolyl protease. This was determined by Western blotting as outlined in Material and Methods. It is not surprising to find a lack of cross-reactivity since the glycylprolyl protease was deemed to be pure, based upon the electrophoretic pattern produced and subsequent staining with silver which is a very sensitive method for detection of protein (Oakley et al, 1 980) 6 5 DISCUSSION The proteolytic activity of Bacteroides gingivalis gives this organism many potential advantages in terms of pathogenicity. A number of proteases have been characterized with regard to molecular weight, temperature and pH optima, specific inhibitors and ability to hydrolyze specific substrates. The proteases, P-1 and P-4, were isolated from the outer membrane preparation (OM-1) obtained by shearing whole cells and from the outer membrane preparation (OM-2) derived from the sonication of whole cells (Grenier & McBride, in press). Since the shearing method for removal of the outer membranes is very gentle, it is likely that little cell breakage occurred during this procedure. However, sonication is a destructive process, and probably resulted in cell damage, releasing intracellular proteins as well as the outer membranes. Grenier & McBride (in press) have described the proteolytic activity of culture supernatants, purified outer membranes, outer membrane vesicles and cell extracts of Bacteroides gingivalis using BSA-polyacrylamide gel electrophoresis. Their findings indicate that the proteolytic activity of the purified outer membranes is identical to that of the cell extracts. This activity was found to be absent from the cell extracts following ultracentrifugation which would result in the removal of the outer membrane fragments from the cell extract. The loss of proteolytic activity from the cell extract after ultracentifugation is confirmation that the proteolytic activity is associated with the outer membranes rather than with soluble proteases unique to the cell extract. The BSA-polyacrylamide gel electrophoretic procedure proved to be a rapid and very sensitive method for monitoring proteolytic activity during the purification of P-1 and P-4. Gel electrophoresis was chosen as the method of purification because it is rapid, sensitive, and reproducible. Furthermore, SDS-polyacrylamide gel 66 electrophoresis offers the ability to easily separate proteins according to their molecular weight. Separation of high molecular weight proteins is often not possible by more traditional methods because of the formation of protein aggregates. The higher molecular weight protease (P-1) was fractionable to proteolytic purity, as determined by electrophoresis using SDS-polyacrylamide gels conjugated to bovine serum albumin. The area of proteolysis was visible as a single band on the stained gel which suggests that P-1 was pure in terms of its proteolytic activity. Three narrow bands, corresponding to molecular weights of 235 kD, 220 kD and 200 kD, were visible following SDS-PAGE of nonboiled P-1 and subsequent silver staining for protein. Several bands were visible following electrophoresis and silver staining of boiled P-1 (68-18 kD). These bands may be due to denaturation of the protease after boiling with SDS-solubilization buffer. It is possible that at 37 °C, in the presence of SDS, the protease and associated molecules migrate as one band. However, during incubation at 100 °C, in the presence of SDS, the associated molecules may separate from the protease and each migrate as a separate band. The outer membrane is associated with insoluble hydrophobic molecules, protein and lipopolysaccharide. It is very likely that some of these molecules remain associated with the protease following purification procedures. The position of P-1 protelytic activity visible on a BSA-polyacrylamide gel (corresponding to 200 kD) correlated well with the position of the three bands seen on a silver stained gel after electrophoresis of nonboiled P-1 (235-200 kD). The presence of three bands visible after silver staining may simply reflect the conformation of the protease, as suggested by Heukeshoven & Dernick (1985) and is not an indication that each band has individual proteolytic activity. The results of the study by Heukeshoven & Dernick (1985) indicated 67 that differences in staining occurred when the protein three-dimensional structure was changed by treatment with detergents, e.g. SDS. Furthermore, the three-dimensional protein structure, and therefore the presentation of the reactive moiety in space, appeared to be the single most important factor in stainability. The appearance of a single band of proteolytic activity on the BSA-polyacrylamide gel would tend to support this, and is additional evidence of the proteolytic and protein purity of P-1. Furthermore, the appearance of a single major band, corresponding to 68 kD, after electrophoresis of boiled P-1 is another indication that the three bands represent a single protease. P-4, fractionnated to proteolytic purity (as above for P-1), in its active (nonboiled) form, had a molecular weight corresponding to 74 kD whereas the boiled protease produced two major bands after SDS-PAGE corresponding to 68 and 50 kD. It may be that these two bands again represent molecules associated with the nonboiled protease and are produced as a result of the denaturation of the protease during treatment with SDS-solubilization buffer at 100 ° C . Alternatively, the bands could be the result of a conformational change in the protease following denaturation of the protease by boiling with SDS-solubilization buffer. The position of proteolytic activity (74 kD) on the BSA-polyacrylamide gel correlates very well with the single major band corresponding to 74 kD, visualized by silver staining after SDS-PAGE of nonboiled P-4, and is an indication of a high degree of proteolytic and protein purity of the P-4 protease. Lipopolysaccharide was not found in either P-1 or P-4, after SDS-PAGE and silver staining for lipopolysaccharide (results not shown). Although the proteases had different molecular weights, they were identical in their activity. The lower molecular weight protease (P-4) appeared to be more reactive than the higher molecular weight protease (P-1). The proteases were identical in their requirement for reducing agent for activity and in pH and temperature optima. The small difference in heat stability between P-1 and P-4 6 8 may be due to a protein complex which stabilizes P-1, making it less susceptible to changes in temperature. The proteases were identical in their activity against synthetic peptides and protein substrates. Both proteases were especially active against synthetic peptides containing arginine which, on the basis of this substrate specificity, suggests that the proteases are trypsin-like rather than chymotrypsin-like, even though inhibition resulted from the specific chymotrypsin inhibitor, TPCK. It is interesting that although the glycylprolyl dipeptide was slowly hydrolyzed, both denatured collagen and gelatin were readily degraded. Further investigation of the collagenolytic activity of the higher molecular weight protease (P-1) appears warranted, based upon the potential ability of P-1 to degrade acid-soluble collagen at 37 °C. Both proteases completely degraded IgG and IgA. This ability is important in terms of virulence, giving the organism the potential to interfere with the host defense system and thus enhance its own survival. Previous studies (Mortensen & Kilian, 1984; Sato et al, 1987; Sunqvist et al, 1985) have reported similar results. The inhibition studies indicated that both P-1 and P-4 are serine proteases. Inhibition was produced by several serine protease inhibitors, including those active against both trypsin- (PMSF and TLCK) and chymotrypsin-like enzymes (TPCK). In addition, the sensitivity of these proteases to thiol inhibitors (PCMB, Hg and Zn ions) suggests that these groups are present at the active site. These findings are similar to those of other investigations describing trypsin-like proteases of B. gingivalis (Fujimura & Nakamura, 1987; Ono et al, 1987; Sorsa et al, 1987; Yoshimura et al, 1984). The enhancement of activity from EDTA and aprotinin may be the result of chelation of an inhibitory cation associated with the proteases. Supportive evidence is the increase in activity when cations, Ca and Mg, are added to the reaction mixture. 6 9 The immunoreactivity studies indicate that the glycylprolyl protease purified from outer membranes of B. gingivalis 33277 (Grenier and McBride,1987) is different from the proteases purified in this work. The specificity of the antibody to the glycylprolyl protease was reinforced by the fact that only the band representing the M r of the glycylprolyl protease (29,000 D) was detected when the antibody was reacted with crude preparations of outer membrane, cell extract and with the outer membrane vesicle preparation, as outlined in Materials and Methods. Whether these fractions represent a single protease or two different proteases is not entirely clear from this investigation, since in all characteristics except molecular weight, the proteases were identical. The initial detection of three proteolytic bands on the skim milk zymogram could represent three different proteases or a single protease having three proteolytically active moieties. Further work using immunological and genetic techniques may be able to resolve this. Preparation of specific antibody to each of the purified proteases, and investigation of any cross-reactivity, should determine whether these are unique proteases or part of single large protein. Furthermore, isolation and purification of the intermediate protease seen on the zymogram would likely provide more information concerning this group of proteases. It is quite likely, based upon the findings of this investigation, that the three bands apparent on the initial zymogram represent different forms of a single protease. Evidence for this conclusion is the appearance of the major band corresponding to 68 kD which is common to both P-1 and P-4 following electrophoresis of boiled P-1 and boiled P-4. The trypsin-like protease described by Fujimura and Nakamura (1987) was also isolated from a cell extract of sonicated cells of B. gingivalis. However, their protease, soluble in Triton X-100, was found to be active at neutral pH (pH 7.5) and had a lower molecular weight by SDS-PAGE (65 kD) than either the P-1 or P-4 protease in this study. The earlier investigation of Fujimura and 70 Nakamura (1981) revealed the presence of two proteases, probably membrane-free, which were also inhibited by thiol inhibitors; however, no molecular weights were determined. Yoshimura et al (1984) and Ono et al (1987) described trypsin-like proteases with similar activities to the P-1 and P-4 proteases, however, the proteases were isolated from different crude fractions in each of these investigations. The protease isolated by Yoshimura et al (1984) could have been associated with either the inner or the outer membrane of the cell whereas Ono et al (1987) used culture supernatant as a starting material. Yoshimura et al (1984) did not determine the molecular weight of the partially purified protease whereas the M r of the protease purified by Ono et al (1987) was determined to be 49 kD by SDS-PAGE. Differences other than M r exist among the proteases found in these two investigations and the proteases purified in this study. The maximum activity of the protease was determined to be at pH 7.6 (Ono et al, 1987), however, information regarding pH optimum is not given by Yoshimura et al (1984). In contrast, both P-1 and P-4 had their optimum activity between the pH values between 6.0 and 6.5. Furthermore, P-1 and P-4 appear to be associated only with the outer membrane of the cell. The trypsin-like protease isolated from culture supernatant by Sorsa et al (1987) appeared similar to P-1 and P-4 in activity. The molecular weight given for their protease was 35 kD (by gel filtration). Therefore, it is not possible to make a direct comparison to the proteases isolated in this study. Activity similar to that found for P-1 and P-4 was apparent in extracellular vesicles (Grenier & Mayrand, 1987). The vesicle fraction was found to degrade azocoll and to possess trypsin-like activity, evidenced by the hydrolysis of BAPNA. The vesicles were also active against collagen. Furthermore, the polypeptide pattern (by SDS-PAGE) of the vesicle preparation was found to be similar, although not identical to that of an outer membrane fraction. It is reasonable to postulate that the proteases associated with the outer membrane 71 per se are also associated with the outer membrane vesicles. Thus both P-1 and P-4 may be bound to the vesicles and able to exert their effect within the host at some distance from the cell itself. Although it is generally recognized that B. gingivalis is involved in the pathogenesis of a number of periodontal diseases, the specific role of the proteases associated with this organism is not entirely understood. Certainly all of the information available concerning the proteolytic activity of B. gingivalis indicates an extensive and varied ability of the organism to cause destruction to the host tissues by both direct and indirect means. As such, further understanding of the characteristics of this virulent organism can only assist in the understanding and treatment of periodontal disease. 72 BIBLIOGRAPHY 1. Abiko, Y., M. Hayakawa, S. Murai and H. Takiguchi. 1985. Glycylprolyl dipeptidylaminopeptidase from Bacteroides gingivalis. J. Dent. Res. 64: 106-111. 2. Altman, L.C., R.C. Page, J.L. Ebersole and G.E. Vandesteen. 1982. Assessment of host defenses and serum antibodies to suspected periodontal pathogens in patients with various types of periodontitis. J . Perio. Res. 17: 495-497. 3. Baehni, P.C, C. Tsai, W.P. McArthur, B.F. Hammond and N.S. Taichman. 1979. Interaction of inflammatory cells and oral micro-organisms. VIII. Detection of leukocytic activity of a plaque-derived micro-organism. Infect. Immun. 24: 233-243. 4. Berdal, B. P. and O. Olsvik. 1983. Legionella extracellular protease activity on chromogenic peptides: a simplified procedure for biochemical enzyme identification. Acta Path. Microbiol. Immunol. Scand. 91: 89-91. 5. Bickel, M and G. Cimasoni. 1985. The pH of human crevicular fluid measured by a new microanalytical technique. J . Perio. Res. 20: 3 5 - 4 0 . 6. Bourgeau, G. and D. Mayrand. 1983. Phenylacetic acid production by Bacteroides gingivalis from phenylalanine and phenylalanine-containing peptides. Can. J . Microbiol. 29: 1184-1189. 7. Boyd, J. and B.C. McBride. 1984. Fractionation of hemagglutinating and bacterial binding adhesions of Bacteroides gingivalis. Infect. Immun. 45: 403-409. 8. Brunette, W. N. 1981. "Western blotting": electrophoretic transfer of proteins from sodium dodecyl sulphate-polyacryulamide gels to unmodified nitrocellulose and radiographic detection with antibodies and radioiodinated protein A. Anal. Biochem. 112: 195-203. 7 3 9. Burdon, K.L. 1928. Bacterium melaninogenicum from normal and pathogenic tissues. J . Infect. Dis. 42: 161-162. 10. Carlsson, J., B. F. Herrmann, J. F. Hofling and G. K. Sundqvist. 1984. Degradation of the human proteinase inhibitors alpha-1-antitrypsin and alpha-2-macroglobulin by Bacteroides gingivalis. Infect. Immun. 43: 6 4 4 - 6 4 8 . 11. Carlsson, J., J. F. Holfling and G. K. Sundqvist. 1984. Degradation of albumin, haemopexin, haptoglobin and transferrin, by black-pigmented Bacteroides species. J. Med. Microbiol. 18: 39-46. 12. Conlan, J.W., A. Williams and L.A.E. Ashworth. 1988. In vivo production of a tissue-destructive protease by Legionella pneumophila in the lungs of experimentally infected guinea-pigs. J. Gen. Microbiol. 134: 1 4 3 - 1 4 9 . 13. Courant, P.R. and R.J. Gibbons. 1967. Biochemical and immunological heterogeneity of Bacteroides melaninogenicus. Arch. Oral Biol. 12: 1 6 0 5 - 1 6 1 3 . 14. Courant, P. R., I. Paunio and R. J. Gibbons. 1965. Infectivity and hyaluronidase activity of debris from healthy and diseased gingiva. Arch. Oral Biol. 10: 119-125. 15. Coykendall, A.L., F.S. Kaczmarek and J . Slots. 1980. Genetic heterogeneity in Bacteroides asaccharolyticus (Holdeman and Moore 1970) Finegold and Barnes (Approved Lists, 1980) and proposal of Bacteroides gingivalis sp. nov. and Bacteroides macacae (Slots, J . and Genco, R.) comb. nov. Int. Syst. Bacteriol. 30: 559-564. 16. Deurden, B. I. 1975. Pigment production by Bacteroides species with reference to sub-classification. J. Med. Microbiol. 8: 13-125. 17. Di Murro, C , R. Nisini, M. Cattabriga, A. Simonetti-D'Arca, S. Le Mole, M. Paolantonio, L. Sebastiani and R. D'Amelio. 1987. Rapidly progressive periodontitis-Neutrophil chemotaxis inhibitory factors associated with the presence of Bacteroides gingivalis in crevicular fluid. J. Periodontol. 58: 868-872. 74 18. Ebersole, J.L., M.A. Taubman and D.J. Smith. 1985. Gingival crevicular fluid antibody to oral microorganisms. II. Distribution and specificity of local antibody responses. J . Perio. Res. 20: 349-356. 19. Ebersole, J.L., M.A. Taubman, D.J. Smith and D.E. Frey. 1986. Human immune responses to oral microorganisms: patterns of systemic antibody levels to Bacteroides species. Infect. Immun. 51: 507-513. 20. Ebersole, J.L., M.A. Taubman, D.J. Smith and S.S. Socransky. 1982. Humoral immune responses and diagnosis of human periodontal disease. J . Perio. Res. 17: 478-480.106-111. 21. Farida, R., P.D. Marsh, H. N. Newman, D.C. Rule and L. Ivanyi. 1986. Serological investigation of various forms of inflammatory periodontitis. J. Perio. Res. 21: 365-374. 22. Finegold, S.M. and E.M. Barnes. 1977. Report of the ICSB Taxonomic Subcommittee on Gram-Negative Anaerobic Rods. Proposal that the saccharolytic and asaccharolytic strains at present classified in the species Bacteroides melaninogenicus (Oliver and Wherry) be classified in two species as Bacteroides melaninogenicus and Bacteroides asaccharolyticus. Int. J . Bacteriol. 27: 388-391. 23. Foltman, B., P.B. Szecsi and N.I. Tarasova. 1985. Detection of proteases by clotting of casein after gel electrophoresis. Anal. Biochem. 146: 353-360 . 24. From, S.H. and S. D. Schultz-Haudt. 1963. Comparative histological and microchemical evaluations of the collagen content of human gingiva. J. Periodontol. 34: 216-222. 25. Fujimura, S. and T. Nakamura. 1981. Isolation and characterization of proteases from Bacteroides melaninogenicus. Infect. Immun. 33: 738-742. 26. Fujimura, S. and T. Nakamura. 1987. Isolation and characterization of a protease from Bacteroides gingivalis. Infect. Immun. 55: 716-720. 75 27. Gibbons, R.J. and J.B. MacDonald. 1960. Hemin and vitamin K compounds as required factors for the cultivation of certain strains of Bacteroides melaninogenicus. J. Bacteriol. 80: 164-170. 28. Goldstein, E.J.C., D. M. Citron and S.M. Finegold. 1984. Role of anaerobic bacteria in bite wound infections. Rev. Infect. Dis. 6: 177-183. 29. Golub, L.M., H.M. Lee, T.F. McNamara, N.S. Ramamuthy, J . Zambon and S. Ciano. 1985. Further evidence that tetracyclines inhibit collagenase activity in human crevicular fluid and from other mammalian sources. J . Perio. Res. 20: 12-23. 30. Grenier, D., G. Chao and B. C. McBride. 1988. Characterization of SDS-stable Bacteroides gingivalis proteases by polyacrylamide gel electrophoresis. Infect. Immun. (In press). 31. Grenier, D. and D. Mayrand. 1983. Etudes d'infections mixtes anaerobies comportant Bacteroides gingivalis. Can. J . Microbiol. 29: 6 1 2 - 6 1 8 . 32. Grenier, D. and D.Mayrand. 1985. Cytotoxic effects of culture supernatants of oral bacteria and various organic acids on Vero cells. Can. J . Microbiol. 31: 302-304. 33. Grenier, D. and D. Mayrand. 1986. Nutritional relationships between oral bacteria. Infect. Immun. 53: 616-620. 34. Grenier, D. and D. Mayrand. 1987. Selected characteristics of pathogenic and nonpathogenic strains of Bacteroides gingivalis. Can. J . Microbiol. 25: 738-740. 35. Grenier, D. and D. Mayrand. 1987. Functional characterization of extracellular vesicles produced by Bacteroides gingivalis. Infect. Immun. 55: 111-117. 36. Grenier, D. and B.C. McBride. 1987. Isolation of a membrane associated Bacteroides gingivalis glycylprolyl protease. Infect. Immun. 55: 3131-3136. 76 37. Haapsalo, M., H. Ranta, K. Ranta and H. Shah. 1986. Black-pigmented Bacteroides spp. in human apical periodontitis. Infect. Immun. 53: 149-153. 38. Hanazawa, S., K. Hirose, Y. Ohmori, S. Amano and S. Kitano. 1988. Bacteroides gingivalis fimbriae stimulate production of thymocyte-activating factor by human gingival fibroblasts. Infect. Immun. 56: 272-274. 39. Hanazawa, S., K. Nakada, Y. Ohmori, T. Miyoshi, S. Amano and S. Kitano.1985. Functional role of interleukin 1 in periodontal disease: induction of interleukin -1 production by Bacteroides gingivalis lipopolysaccharide in peritoneal macrophages from C3H/HeN and C3H/HeJ mice. Infect. Immun. 50: 262-270. 40. Handley, P.S. and L.S. Tipler. 1986. An electronmicroscope survey of the surface structures and hydrophobicity of oral and non-oral species of the bacterial genus Bacteroides. Arch. Oral Biol. 31: 325-335. 41. Heck, L.W., K. Morihara, W.B. McRae and E. Miller. 1986. Specific cleavage of human Type III and IV collagens by Pseudomonas aeruginosa elastase. Infect. Immun. 51:115-118. 42. Heukeshoven, J. and R. Dernick. 1985. Simplified method for silver staining of proteins in polyacrylamide gels and the mechanism of silver staining. Electrophoresis. 6: 103-112. 43. Holdeman, L.V. and W.E.C. Moore. 1970. Outline of clinical methods in anaerobic bacteriology, 2nd ed. Virginia Polytechnic Institute and State University Anaerobe Laboratory, Blacksburg. 44. Holt, S.C. 1982. Bacterial adhesion in pathogenesis: an introductory statement, p. 261-265. In D. Schlessinger (ed.), Microbiology-1982. American Society for Microbiology, Washington, D.C. 45. Holt, S .C, J. Ebersole, J. Felton, M. Brunsvold and K. Kornman. 1988. Implantation of Bacteroides gingivalis in nonhuman primates initiates progression of periodontitis. Science. 239: 55-57. 46. Hunt, D.E., J.V. Jones and V.R. Dowell, Jr. 1986. Selective 7 7 medium for the isolation of Bacteroides gingivalis. J . Clin. Microbiol. 23: 4 4 1 - 4 4 5 . 47. Kaczmarek, F.S. and A.L. Coykendall. 1980. Production of phenylacetic by strains of Bacteroides asaccharolyticus and Bacteroides gingivalis (sp. nov.). J. Clin. Microbiol. 12: 288-290. 48. Kelleher, E. J . and R. L. Juliano. 1984. Detection of proteases in polyacrylamide gels containing covalently bound substrates. Anal. Biochem. 136: 470-475. 49. Kilian, M. 1981. Degradation of immunoglobulins A1, A2 and G by suspected principal periodontal pathogens. Infect. Immun. 34: 757-765. 50. Kornman, K.S. and P.B. Robertson. 1985. Clinical and microbiological evaluation of therapy for juvenile periodontitis. J . Periodontol. 56: 443-446. 51 . Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. (London). 277: 680-685. 52. Lambe, D.W., Jr., K.P. Ferguson and W.R. Mayberry. 1982. Characterization of Bacteroides gingivalis by direct fluorescent antibody staining and cellular fatty acid profiles. Can. J . Microbiol. 28: 367-374. 53. Lantz, M., R.W. Rowland, L.M. Switalski and M. Hook. 1986. Interactions of Bacteroides gingivalis with fibrinogen. Infect. Immun. 55: 201-205. 54. Larjava, H., V-J Uitto, E. Eerola and M. Haapsalo. 1987. Inhibition of gingival fibroblast growth by Bacteroides gingivalis. Infect. Immun. 55: 387-392. 55. Laughon, B.E., S.A. Syed amd W. J . Loesche. 1982. API ZYM system for identification of Bacteroides spp., Capnocytophaga spp. and spirochetes of oral origin. J. Clin. Microbiol. 15: 97-102. 78 56. Layman,D.L. and D.L. Diedrich. 1987. Growth inhibitory effects of endotoxins from Bacteroides gingivalis and intermedius on human gingival fibroblasts in vitro. J . Periodontol. 58: 389-392. 57. Liljenberg, B. and J. Lindhe. 1980. Juvenile periodontitis. Some microbiological, histopathological and clinical characteristics. J . Clin. Perio. 7: 48-61. 58. Lindhe, J. , Liljenberg, B. and Listgarten, M. A. 1980. Some microbiological, histopathological features of periodontal disease in man. J . Periodontol. 51: 264-269. 59. Listgarten, M.A. and L. Hellden. 1978. Relative distribution of bacteria at clinically healthy and periodontal^ diseased sites in humans. J. Clin. Periodont. 5: 115-132. 60. Listgarten, M. A. and C. H. Lai. 1979. Comparative ultrastructure of Bacteroides melaninogenicus subspecies. J. Perio. Res. 14: 332-340. 61. Listgarten, M.A. and CH. Lai. 1979. Comparative ultrastructure of Bacteroides melaninogenicus subspecies. J . Periodontal. Res. 14: 3 3 2 - 3 4 0 . 62. Loesche, W.J., V. Paunio, M.P. Woolfolk and R.N. Hockett. 1974. Collagenolytic activity of dental plaque associated with periodontal pathology. Infect. Immun. 9: 329-336. 63. Loesche, W.J. and S.A. Syed. 1978. Bacteriology of human experimental gingivitis: effect of plaque and gingivitis score. Infect. Immun. 21: 830-839. 64. Loesche, W., S.A. Syed, E. Schmidt, E.C Morrison. 1985. Bacterial profiles of subgingival plaques in periodontitis. J. Periodontol. 56: 447-456. 65. Lylerly, D.M. and A.S. Kreger. 1983. Importance of serratia protease in the pathogenesis of experimental Serratia marcescens pneumonia. Infect. Immun. 40: 113-119. 66. MacDonald, J.B. and R .J. Gibbons. 1962. The relationship of 79 indigenous bacteria to periodontal disease. J. Dent. Res. 41: 320-326. 67. MacDonald J.B., S.S. Socransky, R.J. Gibbons. 1963. Aspects of the pathogenesis of mixed anaerobic infections of mucous membranes. J . Dent. Res. 42: 529-544. 68. Mansheim, B.J. and D.L. Kasper. 1977. Purification and immunochemical characterization of the outer membrane complex of Bacteroides melaninogenicus subspecies asaccharolyticus .J. Infect. Dis. 135:787-799 . 69. Mansheim, B.J., A.B. Onderdonk and D.L.Kasper. 1978. Immunochemical and biologic studies of the lipopolysaccharide of Bacteroides melaninogenicus subspecies asaccharolyticus. J . Immunol. 120: 72-78. 70. Mayrand, D. 1979. Identification of clinical isolates of selected species of Bacteroides: production of phenylacetic acid. Can. J. Microbiol. 45. 25: 927-928. 71. Mayrand, D. and D. Grenier. 1985. Detection of collagenase activity in oral bacteria. Can. J . Microbiol. 31: 134-138. 72. Mayrand, D. and S.C. Holt. 1988. Biology of asaccharolytic black-pigmented Bacteroides species. Microbiol. Rev. 52: 134-152. 73. Mayrand, D. and B.C. McBride. 1980. Ecological relationships of bacteria involved in a simple mixed anaerobic infection. Infect. Immun. 27: 44-50. 74. McDermid, A.S., A.S. McKee and P.D. Marsh. 1988. Effect of environmental pH on enzyme activity and growth of Bacteroides gingivalis W50. Infect. Immun. 56: 1096-1100. 75. McKee, A.S., A.S. McDermid, A. Baskerville, A.B. Dowsett, D.C. Ellwood and P.D. Marsh. 1986. Effect of hemin on the physiology and virulence of Bacteroides gingivalis W50. Infect. Immun. 52:349-355. 76. Millar, S.J., E.G. Goldstein, M.J. Levine and E Hausmann. 1986. 80 Modulation of bone metabolism by two chemically distinct lipopolysaccharide fractions from Bacteroides gingivalis. Infect. Immun. 51: 302-306. 77. Molla, A., K. Matsumoto, I. Oyamada, T. Katsuki and H. Maeda. 1986. Degradation of protease inhibitors, immunoglobulins, and other serum proteins by Serratia protease and its toxicity to fibroblasts in culture. Infect. Immun. 53: 522-529. i 78. Moore, W.E.C., L.V. Holdeman, E.P. Cato, R.M. Smibert, J.A. Burmeister, K.G. Palcanis and R.R. Ranney. 1985. Comparative bacteriology of juvenile periodontitis. Infect. Immun. 48: 507-519. 79. Mortensen, S. B. and M. Kilian. 1984. Purification and characterization of an immunoglobulin A1 protease from Bacteroides melaninogenicus. Infect. Immun. 45: 550-557. 80. Mouton, C , P.G. Hammond, J. Slots and R.J. Genco. 1981. Serum antibodies to oral Bacteroides asaccharolyticus (Bacteroides gingivalis): Relationship to age and periodontal disease. Infect. Immun. 31: 182-192. 81. Mouton, C , P.G. Hammond, J. Slots, M.J. Reed and R.J. Genco. 1981. Identification of Bacteroides gingivalis by fluorescent antibody staining. Ann. Microbiol. (Paris). 132B: 69-83. 82. Mufioz, M.L., M. Rojkind, J. Calderon, M. Tanimoto, S. Arias-Negrete and A. Martinez-Palomo. 1984. Entamoeba histolytica: collagenolytic activity and virulence. J . Protozool. 31: 468-470. 83. Nair, B. C , W. R. Mayberry, R. Dziak, P. B. Chen, M. J. Levine and E. Hausmann. 1983. Biological effects of a purified lipopolysaccharide from Bacteroides gingivalis. J. Perio. Res. 18: 40-40. 84. Newman, M. G., S. S. Socransky, E. D. Savitt, D. A. Propas and A. Crawford. 1976. Studies of the microbiology of periodontosis. J . Periodontol. 47: 373-379. 85. Ng, W. and J. Tonzetich. 1983. Effect of H 2 on permeability of oral 81 mucosa. American Association for Dental Research, Program and abstracts 62, no. 953. 86. Nilsson, T., J. Carlsson and G. Sundqvist. 1985. Inactivation of key factors of the plasma proteinase cascade systems by Bacteroides gingivalis. Infect. Immun. 50: 467-471. 87. Nitzan, D., J.F. Sperry and T.D. Wilkins. 1987. Inactivation of key factors of the plasma proteinase cascade systems by Bacteroides gingivalis. Arch. Oral Biol. 23: 465-470. 88. Oakley, B. R., D. R. Kirsch and N. R. Morris. 1980. A simplified ultra-sensitive silver stain for detecting proteins in polyacrylamide gels. Anal. Biochem. 105: 361-363. 89. Ofek, I. and A. Perry. 1985. Molecular basis of bacterial adherence to tissues, p. 7-13. In S.E. Margenhagen and B. Rosan (ed.), Molecular basis of oral microbial adhesion. American Society for Microbiology, Washington, D.C. 90. Ohmori, Y., S. Hanazawa, S. Amano, T. Miyoshi, K. Hirose and S. Kitano. 1987. Spontaneous production of thymocyte-activating factor by human gingival fibroblasts and its autoregulatory effect on their proliferation. Infect. Immun. 55: 947-954. 91. Okuda, K., J. Slots and R.J. Genco. 1981. Bacteroides gingivalis, Bacteroides asaccharolyticus, and Bacteroides melaninogenicus subspecies: cell surface morphology and adherence to erythrocytes and human buccal epithelial cells. Curr. Microbiol. 6:7-12. 92. Okuda, K. and I. Takazoe. 1973. Antiphagocytic effects of the capsular structure of a pathogenic strain of Bacteroides melaninogenicus. Bull. Tokyo Dent. Coll. 14: 99-104. 93. Okuda, K. and I. Takazoe. 1974. Hemagglutinating activity of Bacteroides melaninogenicus. Arch Oral Biol. 19: 415-416. 94. Okuda, K., A. Yamamoto, Y. Naito, I. Takazoe, J. Slots, and R.J. Genco. 1986. Purification and properties of hemagglutinin from culture supernatant of Bacteroides gingivalis. Infect. Immun. 54: 659-665. 82 95. Oliver, W.W., and W.B. Wherry. 1921. Notes on some bacterial parasites of the human mucous membranes. J . Infect. Dis. 28: 341-345. 96. Ono, M., K. Okuda and I. Takazoe. 1987. Purification and characterization of a thiol-protease from Bacteroides gingivalis strain 381. Oral Microbiol. Immunol. 2: 77-81. 97. Otsuka, M., J. Endo, D. Hinode, A. Nagata, R. Maehara, M. Sato and R. Nakamura. 1987. Isolation and characterization of protease from culture supernatant of Bacteroides gingivalis. J. Perio. Res. 22: 491-498. 98. Page, R.C. and H. E. Schroeder. 1976. Pathogenesis of inflammatory peridontal disease. A summary of current work. Lab. Invest. 34: 235-249. 99. Page, R. and H. E. 1982. Features in man and animals, p. 222-229. In R. Page and H. E. Schroeder, Periodontitis in man and other animals. A comparative review. Karger, New York. 100. Parent, R., C. Mouton, L. Lamonde and D. Bouchard. 1986. Human and animal serotypes of Bacteroides gingivalis defined by crossed immunoelectrophoresis. Infect. Immun. 51: 909-918. 101. Pollack, M. 1984. The virulence of Pseudomonas aeruginosa. Rev. Infect. Dis. 6 (S): S617-S626. 102. Reed, M.J., J. Slots, C. Mouton and R.J. Genco. 1980. Antigenic studies of oral and nonoral black-pigmented Bacteroides strains. Infect. Immun. 29: 564-574. 103. Robertson, P.B., M. Lantz, P.T. Marucha, K.S. Kornman, C L . Trummel and S.C. Holt. 1982. Collagenolytic activity associated with Bacteroides species and Actinobacillus actinomycetemcomitans. J. Perio. Res. 17: 275-283. 104. Roy, T.E., and CD. Kelly. 1939. Genus VIII. Bacteroides Castellini and Chalmers, p. 556-559. In D.H. Bergy, R.S. Breed, 83 E.G.D. Murray and A.P. Hitchens (ed.), Bergy's manual of determinative bacteriology, 5th ed. The Williams & Wilkins Co., Baltimore. 105. Sakai, D.K. 1985. Loss of virulence in a protease-deficient mutant of Aeromonas salmonicida. Infect. Immun. 48: 146-152. 106. Sato, M., M. Otsuka, R. Maehara, J. Endo and R. Nakamura. 1987. Degradation of human secretory immunoglobulin A by protease isolated from the anaerobic periodontopathogenic bacterium, Bacteroides gingivalis. Arch. Oral Biol. 32: 235-238. 107. Sawyer, S.J., J.B. MacDonald and R.J. Gibbons. 1962. Biochemical characteristics of Bacteroides melaninogenicus. A study of thirty-one strains. Arch. Oral Biol. 7: 685-691. 108. Schenkein, H.A. 1982. The complement system in periodontal diseases, p. 299-308. In R. J . Genco and S.E. Mergenhagen (ed.), Host-parasite interactions in periodontal disease. American Society for Microbiology, Washington, D. C. 109. Schonfeld, S.E. and J. M. Kagan. 1982. Specificity of gingival plasma cells for bacterial somatic antigens. J . Perio. Res. 17: 60-69. 110. Schuppan, D., R. Timpl and R. W. Glanville. 1980. Discontinuities in the triple helical sequence Gly-X-Y of basement membrane (Type IV) collagen. FEBS Lett. 115: 297 111. Seddon, S. V., H. N. Shah, J. M. Hardie and J. P. Robinson. 1988. Chemically defined and minimal media for Bacteroides gingivalis. Curr. Microbiol. 17: 147-149. 112. Shah, H. N., R. Bonnet, B. Matteen and R. A. Williams. 1979. The porphyrin pigmentation of subspecies of Bacteroides melaninogenicus. Biochem. J . 180: 45-50. 113. Shah, H.N. and R.A.D. Williams. 1987. Utilization of glucose and amino acids by Bacteroides intermedius and Bacteroides gingivalis. Curr. Microbiol. 15: 241-246. 114. Shah, H.N., R.A.D. Williams, G.D. Bowden and J.M. Hardie. 84 1976. Comparison of the biochemical properties of Bacteroides melaninogenicus from human dental plaque and other sites. J . Appl. Bacteriol. 41: 473-492. 115. Singer, R.E. and B.A. Buchner. 1981. Butyrate and propionate: important components of toxic dental plaque extracts. Infect. Immun. 32: 4 5 8 - 4 6 3 . 116. Slots, J. 1976. The predominant cultivable organisms in juvenile periodontitis. Scand. J Dent. Res. 84: 1-10. 117. .Slots, J. 1977a. Microflora in the healthy gingival sulcus in man. Scand. J. Dent. Res. 85: 247-254. 118. Slots, J. 1977b. The predominant cultivable microflora of advanced periodontitis. Scand. J . Dent. Res. 85: 114-121. 119. Slots, J. 1979. Subgingival microflora and periodontal disease. J . Clin. Perio. 6: 351-382. 120. Slots, J. 1981. Enzymatic characterization of some oral and nonoral gram-negative bacteria with the API ZYM system. J . Clin. Microbiol.14: 288-294. 121. Slots, J. 1982. Importance of black-pigmented Bacteroides in human periodontal disease, p. 27-45. In R.J. Genco and S.E. Mergenhagen (ed.), Host-parasite interactions in periodontal diseases. American Society for Microbiology, Washington, D.C. 122. Slots, J. 1986a. Bacterial specificity in adult periodontitis. A summary of recent work. J . Clin. Perio. 13: 570-577. 123. Slots, J. , L. Bragd, M. Wikstrom and G. Dahlen. 1986. The occurrence of Actinobacillus actinomycetemcomitans, Bacteroides gingivalis and Bacteroides intermedius in destructive periodontal disease in adults. J. Clin. Periodontol.13: 570-577. 124. Slots, J. and R.J. Genco. 1979. Direct hemagglutination technique for differentiating Bacteroides asaccharolyticus oral strains from nonoral 85 strains. J . Clin. Microbiol. 10: 371-373. 125. Slots, J. and R .J. Gibbons. 1978. Attachment of Bacteroides melaninogenicus subsp. asaccharolyticus to oral surfaces and its possible role in colonization of the mouth and of periodontal pockets. Infect. Immun. 19: 254-264. 126. Slots, J. and M. A. Listgarten. 1988. Bacteriodes gingivalis, Bacteroides intermedius and Actinobacillus actinomycetemcomitans in human periodontal diseases. J. Clin. Perio. 15: 85-93. 127. Slots, J. , D. Moenbo, J. Langeback and A. Frandsen. 1978. Microbiota of gingivitis in man. Scand. J. Dent. Res. 86: 174-181. 128. Smalley, J.W. and A.J. Birss. 1987. Trypsin-like activity of the extracellular membrane vesicles of Bacteroides gingivalis W50. J . Gen. Microbiol. 133: 2883-2894. 129. Smalley, J.W., A.J. Birss and C.A. Shuttleworth. 1988. The degradation of Type I collagen and human' plasma fibronectin by the trypsin-like enzyme and extracellular membrane vesicles of Bacteroides gingivalis W50. Arch. Oral Biol. 33: 323-329. 130. Socransky, S. S. 1970. Relationship of vacteria to the etiology of periodontal disease. J. Dent. Res. 49: 203-222. 131. Socransky, S. S. 1977. Microbiology of periodontal disease-Present status and future considerations. J . Periodontol. 48: 497-504. 132. Socransky, S.S. and R .J. Gibbons. 1965. Required role of Bacteroides melaninogenicus in mixed anaerobic infections. J . Infect. Dis. 115: 247-253. 133. Sorsa, T., V-J. Uitto, K. Suomalainen, H. Turto and S. Lindy. 1987. A trypsin-like protease from Bacteroides gingivalis: partial purification and characterization. J . Perio. Res. 22: 1-6. 134. Speigel, C.A., S.E. Hayduk, G.E. Minah and G.N. Krywolap. 1979. B lack-p igmented Bacteroides from clinically characterized periodontal sites. J . Perio. Res. 14: 376-382. 8 6 135. Starkey, P. M. and A. J. Barrett. 1977. 32-Macroglobulin, a physiological regulator of proteinase activity, p. 663-696. In A.J. Barrett (ed.), Proteinases in mammalian cells and tissues. North-Holland Publishing Co., Amsterdam. 136. Suido, H., M.E. Neiders, P.K. Barua, M. Nakamura, P.A. Mashimo and R.J. Genco. 1987. Characterization of N-CBz-glycyl-glycyl-arginyl peptidase and glycyl-prolyl peptidase of Bacteroides gingivalis. J . Perio. Res. 22: 412-418. 137. Sundqvist, G.K., J. Carlsson and L. Hanstrom. 1987. Collagenolytic activity of black-pigmented Bacteroides species. J. Perio. Res. 22: 300-306. 138. Sundqvist, G., J. Carlsson, B. Herrmann and A. Tarnvik. 1985. Degradation of human immunoglobulins G and M and complement factors C3 and C5 by black-pigmented Bacteroides. J . Med. Microbiol. 19: 85-94. 139. Takazoe, I. and T. Nakamura. 1971. Experimental mixed infection by human gingival crevice material. Bull. Tokyo Dent. Coll. 12: 85-93. 140. Takazoe, I., T. Nakamura and K. Okuda. 1984. Colonization of the subgingival area by Bacteroides gingivalis. J. Dent. Res. 63: 42-46. 141. Tanner, A.C.R., C. Haffer, G.T. Bratthall, R.A. Visconti and S.S. Socransky. 1979. A study of the bacteria associated with advancing periodontitis in man. J . Clin. Periodontol. 6: 278-307. 142. Tanner, A.C.R., S.S. Socransky and J.M. Goodson. 1984. Microbiota of periodontal pockets losing crestal alveolar bone. J. Perio. Res. 19: 279-291. 143. Taubman, M.A., J.L. Ebersole and D.J. Smith. 1982. Association between systemic and local antibody in periodontal disease. In R.J. Genco and S.E. Mergenhagen (ed.),Host-parasite interactions in periodontal disease. American Society for Microbiology, Washington, D.C. 144. Temporo, P.J., H. S. Reynolds, J . Slots. 1983. Microbial morphotypes in periodontal health and disease. J . Dent. Res. 62:178. 87 145. Tew, J.G., D.R. Marshall, J.A. Burmeister and R.R. Ranney. 1985b. Relationship between gingival crevicular fluid and serum antibody titers in young adults with generalized and localized periodontitis. Infect. Immun. 49: 487-493. 146. Toda, K., M. Otsuka, Y. Ishikawa, M. Sato, Y. Yamamoto and R. Nakamura. 1984. Thiol-dependent collagenolytic activity in culture media of Bacteroides gingivalis. J . Perio. Res. 19: 372-381. 147. Tonzetich, J. and B.C. McBride. 1981. Characterization of volatile sulphur production by pathogenic and non-pathogenic strains of oral Bacteroides. Arch. Oral Biol. 26: 963-969. 148. Touw, J.J.A., T.J.M. van Steenbergen and J. de Graaff. 1982. Butyrate: a cytotoxin for vero cells produced by Bacteroides gingivalis and Bacteroides asaccharolyticus. Antonie van Leeuwenhoek J . Microbiol. Serol. 48: 1315-1325. 149. Tsai, C. M. and C. E. Frasch. 1980. A sensitive silver stain for detecting lipopolysaccharide in polyacrylamide gels. Anal. Biochem. 119: 1 1 5-119 . 150. Tsutsui, H., T. Kinouchi, Y. Wakano and Y. Ohnishi. 1987. Purification and characterization of a protease from Bacteroides gingivalis 381. Infect. Immun. 55: 420-427. 151. Uitto, V-J. 1983. Degradation of basement membrane collagen by proteinases from human gingiva, leukocytes and bacterial plaque. J. Periodontol. 54: 740-745. 152. Uitto, V-J. and A. M. Raeste. 1978. Activation of latent collagenase of human leukocytes and gingival fluid by bacterial plaque. J. Dent. Res. 57: 8 4 4 - 8 5 1 . 153. van Dyke, T. E., S. Offenbacher, D. Place, V. R. Dowell and J. Jones.1988. Refractory periodontitis: Mixed infection with Bacteroides gingivalisand other unusual Bacteroides species. A case report. J. Periodont. 59: 184-189. 88 154. van Steenbergen, T.J.M. and J . de Graaff. 1986. Proteolytic activity of black-pigmented Bacteroides. FEMS Microbiol. Lett. 33: 219-222. 155. van Steenbergen, T.J.M., J .J . De Soet and J. de Graaff. 1979. DNA base composition of various strains of Bacteroides melaninogenicus. FEMS Microbiol. Lett. 5: 127-130. 156. van Steenbergen, T.J.M., L.M.S. Van der Mispel, and J. De Graaff. 1986b. Effects of ammonia and volatile fatty acids produced by oral bacteria on tissue culture cells. J. Dent. Res. 65: 909-912. 157. van Steenbergen, T.J.M., A.J. van Winkelhoff, D. Mayrand, D. Grenier and J. de Graaff. 1984. Bacteroides endodontalis sp. nov., an asaccharolytic black-pigmented Bacteroides species from infected dental root canals. Int. J. Syst. Bacteriol. 34: 118-120. 158. van Steenbergen, T.J.M., C. A. Vlaanderen and J. de Graaff. 1982. Confirmation of Bacteroides gingivalis as a species distinct from Bacteroides asaccharolyticus. Int. J . Syst. Bacteriol. 31:236-241. 159. van Winkelhoff, A. J. , A. W. Carlee and J. de Graaff. 1985. Bacteroides endodontalis and other black-pigmented Bacteroides species in odontogenic abscesses. Infect. Immun. 49: 494-497. 160. van Winkelhoff, A.J., U. Van der Velden, E.G. Winkel and J. De Graaff. 1986b. Black pigmented Bacteroides and motile organisms on oral mucosal surfaces in individuals with and without periodontal breakdown. J. Perio. Res. 21: 434-439. 161. van Winkelhoff, A.J., T.J.M. van Steenbergen and J. de Graaff. 1988. The role of black-pigmented Bacteroides in human oral infections. J . Clin. Periodontol. 15: 145-155. 162. van Winkelhoff, A.J., T.J.M. van Steenbergen, N. Kippuw and J. de Graaff. 1986c. Enzyme characterization of oral and nonoral black-pigmented Bacteroides species. Antonie van Leeuwenhoek. 50: 789-798. 163. Vincent, J.W., J.B. Sizuki, W.A. Falkler, Jr., and W.C. Cornett. 1985. Reaction of human sera from juvenile periodontitis, rapidly 89 progressive periodontitis, and adult periodontitis patients with selected periodontopathogens. J . Periodontol. 56: 464-469. 164. Wahren, A. and R.J. Gibbons. 1970. Amino acid fermentation by Bacteroides melaninogenicus. Antonie van Leeuwenhoek J . Microbiol. Serol. 36: 149-159. 165. White, D. and D. Mayrand. 1981. Association of oral Bacteroides with gingivitis and adult periodontitis. J . Periodontal. Res. 16: 259-265. 166. Wikstrom, M.B., G. Dahlen and A. Lindhe. 1983. Fibrinogenolytic and fibrinolytic activity in oral microorganisms. J . Clin. Microbiol. 17: 7 5 9 - 7 6 7 . 167. Williams, A., A. Baskerville, A.B. Dowsett and J.W. Conlan. 1987. Immunocytochemical demonstration of the association between Legionella pneumophila, its tissue-destructive protease, and pulmonary lesions in experimental Legionnaires' disease. J . Pathol. 153: 257-264. 168. Williams, G.D. and S.C. Holt. 1985. Characteristics of the outer membrane of selected oral Bacteroides species. Can. J . Microbiol. 31: 2 3 8 - 2 5 0 . 169. Wilson, M.E., J.J. Zambon, J.B. Suzuki and R.J. Genco. 1985. Generalized juvenile periodontitis, defective neutrophil chemotaxis and Bacteroides gingivalis in a 13-year-old female. A case report. J . Periodontol. 1985. 56: 457-463. 170. Winkler, J.R., S.R. John, R.H. Kramer, C.I. Hoover and P.A. Murray. 1987. Attachment of oral bacteria to a basement-membrane-like matrix and to purified matrix proteins. Infect. Immun. 55: 2721-2726. 171. Woo, D.D.L.. S.C. Holt and E.R. Leadbetter. 1979. Ultrastructure of Bacteroides species: Bacteroides asaccharolyticus, Bacteroides fragilis, Bacteroides melaninogenicus and B. melaninogenicus subspecies intermedius. J. Infect. Dis. 319: 534-546. 172. Yoshimura, F., M. Nishikata, T. Suzuki, E.I. Hoover and E. 90 Newbrun. 1984. Characterization of a trypsin-like protease from the bacterium Bacteroides gingivalis isolated from human dental plaque. Arch. Oral Biol. 7: 559-564. 173. Yoshimura, F., K. Takahashi, Y. Nodasaka and T. Suzuki. 1984. Purification and characterization of a novel type of fimbriae from the oral anaerobe Bacteroides gingivalis. J . Bacteriol. 160: 949-957. 174. Yoshimura, F., T. Takasawa, M. Yoneyama, T. Yamaguchi, H. Shiokawa and T. Suzuki. 1985. Fimbriae from the oral anaerobe Bacteroides gingivalis: physical, chemical and immunological properties. J Bacteriol.163: 730-734. 175. Zambon, J.J. 1985. Actinobacillus actinomycetemcomitans in human periodontal disease. J . Clin. Perio. 12: 1-20. 176. Zambon, J.J., H.S. Reynolds and J. Slots. 1981. Black pigmented Bacteroides sp. in the human oral cavity. Infect. Immun. 32: 198-203. 91