Trypsin-like Enzyme Activity of the Extracellular Membrane Vesicles of Bacteroides gingivalis W50

Journal of General Microbiology (1987), 133, 2883-2894. Printed in Great Britain 2883 Trypsin-like Enzyme Activity of the Extracellular Membrane Ve...
Author: Noreen Maxwell
0 downloads 2 Views 1MB Size
Journal of General Microbiology (1987), 133, 2883-2894.

Printed in Great Britain

2883

Trypsin-like Enzyme Activity of the Extracellular Membrane Vesicles of Bacteroides gingivalis W50 By J . W . S M A L L E Y * A N D A. J . BIRSS Department of Dental Sciences, School of Dental Surgery, University of Liverpool, PO Box 147, Liverpool L69 3BX, UK (Received 27 January 1987; revised 26 May 1987) Trypsin-like enzyme activity in spent culture media from 3-d-old batch cultures of Bacteroides gingivalis W50 was measured by using the hydrolysis of Na-benzoyl-L-arginine-p-nitroanilide. The cell-free culture medium was fractionated by differential centrifugation at 10000g and 75000 g, yielding two particulate fractions and a soluble supernatant fraction. About 80%of the total recoverable activity was associated with the particulate fractions, the remainder being in the supernatant. Electron microscopy of ruthenium-red/osmium stained ultrathin sections of the pellet fractions showed them to be composed of vesicular particles (extracellular vesicles), between 50 and 250 nm in diameter. Enzyme activity in all three fractions was enhanced by dithiothreitol. Gel-permeation chromatography of the soluble fraction yielded one peak of activity which contained 64 kDa and 58 kDa polypeptides. Enzyme activity from the vesicular fractions could be solubilized by sonication, giving a similar chromatographic profile to the supernatant fraction. The main peak of activity was composed of 64 kDa and 58 kDa polypeptides. In addition, there was a higher molecular mass enzyme activity peak composed of the 64 kDa and 58 kDa components along with 111 kDa, 93 kDa and 70 kDa polypeptides. We conclude that the trypsin-like enzyme of B. gingivalis is released as a soluble protein and is also associated with extracellular vesicles, in which it may exist as a soluble component and also as a protein complex.

INTRODUCTION

Proteolytic activities associated with black-pigmented Bacteroides species have long been considered as virulence factors in the pathogenesis of periodontal disease (MacDonald et al., 1963; Slots, 1981 ; Slots & Genco 1984). Bacteroidesgingivalis, which is frequently isolated from periodontal lesions in adults with advanced periodontitis (Slots, 1982), possesses a spectrum of proteases including a trypsin-like enzyme (Mayrand et al., 1980). The presence of this trypsinlike enzyme in whole bacterial cultures may be used to differentiate B. gingivalis from other black-pigmented Bacteroides species (Laughton et al., 1982) and it increases the potential of this organism to mediate destruction of periodontal tissues. The trypsin-like enzyme from whole B. gingivalis cells was purified and characterized by Yoshimura et al. (1984). It is active against a number of synthetic and natural substrates including benzoyl-L-arginine-p-nitroanilide,benzoyl-DL-arginine-p-naphthylamide, tosyl-L-arginine methyl ester, casein, ovalbumin and bovine serum albumin. Enzyme activity appears to be associated with the whole cell envelope fraction. A trypsin-like enzyme was also recovered from spent growth medium of B. gingivalis by ammonium sulphate precipitation (Fujimura & Nakamura, 1986); it has a molecular mass of 65 kDa as assessed by SDS-PAGE. Inhibitor studies with N-ethylmaleimide have shown the enzyme to be thiol dependent. It is active against Abbreviations: WCF, whole culture filtrate; ECV, extracellular vesicles; CFS, culture filtrate supernatant; BAPNA, benzoyl-L-arginine-p-nitroanilide; p-NA, p-nitroaniline. 0001-3965 0 1987 SGM Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

L-

2884

J . W . SMALLEY A N D A . J . BlRSS

azoalbumin and azocoll as well as synthetic substrates for trypsin. However, these studies have not indicated whether the enzyme is released solely as a soluble peptide or whether it is associated with an insoluble membrane fraction. The release of 'blebs' or membrane vesicles by Bacteroides species has been described (Wooet al., 1979; Okuda et af., 1981 ;Slots & Gibbons, 1978; Handley & Tipler, 1986). Chemostat grown cultures of B. gingivafis W50 display surface blebs and extracellular vesicles (McKee et af., 1986) and we have recovered extracellular vesicles from batch grown cultures of this organism. However, little is known about the pathogenic significance of extracellular vesicles. We have investigated the possibility that these released extracellular vesicular structures may serve as a source of the extracellular trypsin-like enzyme of B. gingivafis.(A preliminary report of this work was presented at the 1986 XIV International Congress of Microbiology.) METHODS

Organism and growth conditions. Bacteroides gingivalis W50 was obtained from Dr P. D. Marsh, PHLS, Porton Down, Wiltshire, UK, and was maintained by subculture on laked-horse-blood agar plates comprising W ilkensChalgren agar (Oxoid) and 5% (v/v) citrated horse blood. Starter cultures (3-d-old, 50 ml) of Schaedler anaerobe broth (Oxoid) were used to inoculate 950 ml batches of pre-reduced Schaedler broth supplemented with 0.1 % KN03, 0.1 % NaHC03 and 0.5 pg menadione ml-l. Cultures were stirred in an anaerobic atmosphere of 80% N2, 10% C 0 2 and 10% H,; the bacteria were harvested after 3 d by centrifugation at IOOOOg for 75 rnin at 4 "C and washed three times in 0.01 M-phosphate-buffered isotonic saline (PBS; 0.14 M-NaCl, 0.0026 M-KCl, 0.0078 MNa2HP04, 04016 M - N ~ H ~ P OpH , ) 7.3. After each wash, the organisms were recovered by centrifugation at 20000g for 15 rnin at 4 "C. Treatment of spent culture medium. After removal of the bacteria, the supernatant was filtered through a glassfibre pre-filter (SM 13400, thickness 0.55 mm; Sartorius) and then through a 0.45 pm cellulose filter (Millipore) to remove any contaminating cells. The filtered medium was then dialysed for 48 h against repeated changes of distilled water at 4 "C; the retained material was freeze-dried, and termed whole culture filtrate (WCF). The freeze-dried material was suspended in water at a concentration of 100 mg ml-' and centrifuged at 10000 g for 1 h to yield a pellet (lO00Og pellet). The supernatant from this step was further centrifuged at 75000g for 3 h at 4 "C to yield a pellet termed extracellular vesicles (ECV) and a supernatant fraction termed culture filtrate supernatant (CFS). Preparation of the outer membrane fraction. An outer membrane fraction was prepared from whole bacteria by the method of Mansheim & Kasper (1977). Briefly, harvested cells were washed twice with PBS, pH 7.3; the cells were recovered each time by centrifugation at 20000 g for 15 rnin at 4 "C. Pelleted bacteria were resuspended in an equal volume of PBS and this was added to an equal volume of 0.02 M - N ~ ~ E D TThe A . resulting suspension (pH 7.4) was incubated at 37 "C for 30 rnin and syringed twice through a 25 gauge hypodermic needle with manual pressure and then centrifuged at 20000g for 30 rnin at 4 "C. The supernatant was carefully removed and dialysed for 24 h against repeated changes of distilled water at 4 "C. The non-dialysable material representing the outer membrane fraction was recovered by freeze-drying. Owing to the small amounts of material recovered, a differential centrifugation step was not done. Enzyme assays. Trypsin-like enzyme activity was measured using the synthetic chromogenic substrate benzoylL-arginine-p-nitroanilide(L-BAPNA) (Sigma). The enzymic hydrolysis was assessed in a total reaction volume of 2 0 0 ~ 1in 50m~-Tris/HCl,pH 7.7, containing L-BAPNA at a final concentration of 0 . 4 m ~at 37 "C. The absorbance of p-nitroaniline (p-NA) released from the substrate was measured at 405 nm. Appropriate controls were included to account for spontaneous hydrolysis of the substrate. One unit (U) of enzyme activity corresponds to 1 nmol p-NA released min-' at 37 "C. SDS-PAGE. The discontinuous method of Laemmli (1970) was used. Samples for electrophoresis were solubilized by boiling for 5 rnin in sample application buffer containing 50 mM-dithiothreitol (DTT) and 2 M-urea. Polyacrylamide slab gels (lo%, w/v) 1.5 x 140 x I10 mm wide were used with a 3% stacking gel. Electrophoresis was done at 40 mA per gel using bromophenol blue as tracker dye and separated peptide bands were visualized after staining with 0.1 % Kenacid blue in 50% (v/v) methanol and 7% (v/v) acetic acid. Destaining was achieved by diffusion in methanol/acetic acid/water (5 :7 :88, by vol.). The molecular masses of separated polypeptide bands were calculated by reference to the plot of log,, molecular mass versus relative mobility for the standard proteins b-galactosidase (1 16 kDa), phosphorylase B (97 kDa), bovine serum albumin (68 kDa), ovine albumin (45 kDa) and carbonic anhydrase (29 kDa). Analytical methods. Protein was assayed by the Lowry method with bovine serum albumin as standard. Electron microscopy. Negative staining was done using aqueous methylamine tungstate. Bacterial suspensions or CFS fractions (dispersed in distilled water at a concentration of 5 mg ml-') were mixed with an equal volume of 2% (w/v) methylamine tungstate (pH 7.2) for 1 min on carbon coated, collodion covered copper grids (300 mesh;

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Trypsin-like enzyme of Bacteroides gingivalis

2885

Agar Aids, Stanstead, Essex, UK). Excess liquid was removed with blotting paper and the grids were dried under reduced pressure over PzOs and examined in a JEOL 100 CX electron microscope operated at an accelerating voltage of 60 kV. Similar samples to those above were also examined in ultrathin section by transmission electron microscopy. Samples were first fixed for a period of 30 min in 2.5% (w/v) glutaraldehyde in 0.1 M-Sodium cacodylate buffer, pH 7.4, containing 0.05% ruthenium red. Post-fixation staining with osmium was done by exposing samples to 1 % (w/v) osmium tetroxide in 0.1 Msodium cacodylate buffer, pH 7-4, for 30 min in the presence of 0.05% ruthenium red. Stained specimens were dehydrated through methanol and embedded in Spurr resin (Spurr, 1969). Ultrathin sections of approximately 70 nm were cut using a diamond knife on a LKB Ultratome 111, applied to 300 mesh copper grids and examined in the above microscope at an operating potential of 60 kV. Gel-permeationchromatography. This was done on a column of Sephacryl S-200 (Pharmacia) eluted with 0.2 MNaCl buffered with 0.1 M-Tris/HCI, pH 7.7, at a flow rate of 10 ml h-l. Samples (30 mg dry wt) of the 10000g and 75000g pellet fractions were sonicated on ice for 1 min in 1 ml of eluant buffer using a probe sonicator (MSE) operated at maximum amplitude. The solubilized material recovered as the supernatant from the sonicate after a centrifugation step (15 min at lOOOOg) was applied directly to the column. A sample (50 mg) of the CFS fraction was similarly chromatographed but without prior sonication treatment. Eluate fractions (2 ml) were monitored for protein by absorption of UV light (280nm) and for trypsin-like activity. Fractions in regions of the chromatograms displaying enzyme activity were pooled, dialysed against repeated changes of distilled water at 4 "C and the non-dialysable material was concentrated by freeze-drying. The void volume of the column (V,) and total volume ( V,)were assessed by chromatography of Blue Dextran 2000 (Pharmacia) and L-tyrosine respectively. RESULTS

Trypsin-like enzyme activity was detected in cell-free WCF fractions of 3-d-old batch cultures of B. gingivalis W50, by using the hydrolysis of the synthetic substrate L-BAPNA at 37 "C and pH 7-7. Enzyme activity was also observed in the outer membrane fraction prepared from EDTA-treated PBS-washed whole cells of B. gingivalis W50. The spent culture supernatant, after vacuum filtration through 0.45 pm cellulose filters and dialysis, was centrifuged differentially to yield two particulate fractions, the 10000 g and 75000g pellets, and the particle-free CFS. The recoveries and the distribution of the measurable enzyme activities of the supernatant fractions in the absence of added reducing agents for a typical 1 litre culture are shown in Table 1. On a dry wt basis, the lOOOOg and 75000g pellets represented 1.9 and 1.7% of the WCF respectively. A further four batch cultures of B. gingivalis yielded average recoveries of 1.78% (range 2.09 to 1.64) and 1.5% (range 1.63 to 1.24) for the above fractions respectively. Similarly, the average protein content for the CFS fraction was 52.5% (range 54 to 51.5). However, variability was observed in the protein content for the two pellet fractions: the mean for the lOOOOg pellet was 29% (range 23 to 36), and for the 75000g pellet 36% (range 24 to 48). The reason for this variability is not known. For the five outer membrane preparations examined, the mean protein content was 7.25% & 2.1 (sD). The majority of the enzyme activity was recovered in the 10000g and 75 000 g pellet fractions, each representing approximately 40% of the total recoverable activity. However, the greatest enzyme specific activity [39.31 U (mg protein)-' 3 was displayed by the 75 000 g pellet fraction. Although the CFS fraction represented 20% of the total activity, its specific activity was 0.5% that of the 75 000 g pellet fraction. Enzyme activity released from whole bacterial cells by EDTA treatment represented only 7.5 % of the total recoverable activity, but had a specific activity of 2.16 U (mg protein)-' , 20-fold less than the 75000 g pellet fraction. The enzyme activity of the whole culture filtrate was 72% of the totalled individual 75000g, lOOOOg and CFS fractions. The effect on enzyme activity of adding DTT was investigated in the 75000g and 1OOOOg pellets. Samples (0.1 ml) of the pellet fractions at a concentration of 0.1 mg dry wt ml-1 in 100 mM-Tris/HCl, pH 7.7, were incubated with L-BAPNA (final concentration 0.4 mM) and DTT at concentrations between 0-625 and 20 mM. The results are shown in Fig. 1. A maximum enzyme specific activity of 200 U (mg protein)-' was obtained for the 75000g pellet in the presence of 10 mM-DTT. This activity represented a 5.1-fold increase over the enzyme specific activity observed in the absence of reducing agent. A similar activity profile against DTT concentration was obtained for the l O O O O g pellet. The enzyme specific activity at 10 ~ M - D T T

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Table 1. Distribution of trypsin-like activity from the culture supernatant of a 3-d-old batch culture of B. gingivalis W50

w

E

Values are means of triplicate estimates. Recoveries are expressed as dry wt (1 culture)-'.

Fraction (a) Whole culture

filtrate (WCF) (b) 1OOOOg pellet (c) 75OOOg pellet (d) 75 OOO g supernatant (CW (e) Outer membrane

Recovery wt (mg I-' )

Percentage dry wt of WCF

1574 30 28 1511

67

1-9 1.7

Total protein* (mg)

Percentage by wt of protein

Total enzyme activity (U)

Percentage of total enzyme activity in (h), (c) and (d)

810

51-5

507

-

10.4 7.2 778

34.8 26.0 51.5

274 283 145

39.03 40-31 20.66

5.3

8-0

10-7

-

* Measured using the

Folin-phenol reagent.

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Specific activity [U (mg protein)-'] 0-62 26-34 39.3 1 0-18 2.03

Trypsin-like enzyme of Bacteroides gingivalis

~~

~~

2887

~~

10 20 DTT concn (mM) Fig. 1 . Effect of DTT on trypsin-like enzyme activity associated with 10000 g pellet fractions. 0

(a) the

75000g and (b) the

for this fraction was 148 U (mg protein)-’, representing a 5-7-fold increase over the non-DTT stimulated activity. The CFS fraction displayed a 15-fold increase in enzyme specific activity in the presence of 10 mM-DTT. Gel-permeation chromatographic profiles of the 75 000 g and 10000g pellets and CFS fractions are shown in Fig. 2. Enzyme activity was ‘solubilized’ from the 75000g and 100OOg pellets by sonication for 1 min in 0.2 M-NaCl, 0-1 M-Tris/HCl, pH 7.7. The CFS fraction containing trypsin-like activity was chromatographed without prior sonication treatment. Low levels of enzyme activity were observed when eluate fractions were assayed in the absence of DTT. Enzyme assays were therefore done in the presence of 10 mM-DTT over an incubation period of 18 h. Enzyme activities are displayed as absorbance at 405 nm. Broad areas of material absorbing at 280 nm were observed in all three chromatograms. Only the lOOOOg and 75000g pellets displayed any high molecular mass material absorbing at 280 nm and eluting at the column void volume. No enzyme activity was associated with either of these high molecular mass fractions. The enzyme activity elution profiles for the two ‘solubilized’ 75000g and lOOOOg pellet fractions were essentially the same. Each displayed two regions of enzyme activity within the chromatogram. In contrast, the CFS fraction displayed only one peak of enzyme activity (Fig. 2 c ) which corresponded in elution position to the lowest molecular mass enzyme peak obtained from the two particulate fractions. The regions of the chromatogram displaying enzyme activity were pooled and dialysed against distilled water at 4 “C. The non-dialysable fractions were recovered by freeze-drying and subjected to SDS-PAGE under reducing conditions. The separated peptide components, stained with Kenacid blue, were scanned densitometrically at 550 nm (Fig. 3). The higher molecular mass peak (Fig. 3a) of enzyme activity from the 75000g pellet displayed a spectrum of peptides of molecular masses 111, 94, 70, 64 and 58 kDa, along with some lower molecular mass bands. The main peak of enzyme activity (Fig. 3c), however, displayed two polypeptides of 64 and 58 kDa as the major components, but with evidence of some lower molecular mass components. SDS-PAGE profiles similar to the 75000 g pellet column fractions I and I1 were obtained from the l O O O O g pellet. The components present in peak I (Fig. 3b) were calculated as having molecular masses of 11 1, 94, 86 and 70 kDa. Two peptides of molecular masses 64 and 58 kDa were observed in the lower molecular mass enzyme peak as the major components (Fig. 34.These data indicated strong similarities between the material released by sonication from the lOOOOg and 75000g pellets. The peak of enzyme activity from the CFS fraction (Fig. 3e) displayed two polypeptide bands with molecular masses of 64 and 58 kDa. There was also evidence of a small amount of lower molecular mass peptides.

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

2888

J . W . S M A L L E Y A N D A . J . BIRSS

0.16

1.6

0.08

0.8

" ,

8

3

v

0.16

x Y

..-> d

v

;

2 0.08

w

0.12

0.3

0.04

0.1

Eluate vol. (ml)

I

Fig. 2. Gel-permeation chromatography of material from (a) the CFS fraction; (b) released from 30 mg 75000 g pellet and (c) released from 30 mg 100oOg pellet. The sonication-solubilized materials (b and c) were recovered as the supernatant fractions after centrifugation, applied to a 90 x 1.5 cm column of Sephacryl S-200 and eluted with 0.2 M-NaCI, 0-1 M-Tris/HCl, pH 7.7, at a flow rate of 10 ml h-I. The CFS fraction was chromatographed directly without prior sonication. Eluate fractions (2 ml) were in the monitored for protein by absorbance at 280 nm ( 0 )and for trypsin-like enzyme activity presence of 10 mM-DTT, expressed as the change in absorbance at 405 nm after incubation for 18 h at 37 "C. Vo (72 ml) and V, (190 ml) are indicated.

u)

As a control, samples of unused Schaedler growth medium were dialysed exhaustively against water. The non-dialysable material was concentrated by freeze-drying and analysed by SDSPAGE on a 10% acrylamide gel. Electrophoresis of this material at 800 pg per gel track (data not presented here) showed that it did not contain any Kenacid blue stainable peptides.

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Trypsin-like enzyme of Bacteroides gingivalis 11697

45

68

2889

29

0.25

3

0.2

n

P

c

Y

0 11697

68

29

45

0

0

E

A '

64 58

P

11697 68'

0.15

"t

t"

Origin 11697

0.25 - 4 4

-

5

P

c w7

P

Dye front 6R

4

45

4

29

-

4

(e)

58-4

C FS

0

45

29

f

&gin

f "

Dye hont

64

0

t

Origin

t"

Dye front

Fig. 3. Densitometric scans at 550 nm of SDS-PAGE peptide profiles of enzyme containing fractions obtained after Sephacryl S-200 chromatography of the 75000 g pellet (a and c), the 10000 g pellet ( b and d ) and CFS (e). Electrophoresis was done under reducing conditions on 10% (w/v) gels. Protein bands were visualized after staining with Kenacid blue. The molecular masses (kDa) of individual polypeptides and protein standard markers are indicated.

Electron microscopy The surface morphological features of B. gingivalis W50 cells from 3-d-old cultures were examined by negative staining using methylamine tungstate (Fig. 4a). Whole bacterial cells displayed 'blebs' or vesicles of between 100 and 150 nm in diameter, which appeared to be associated with the cell surface. Morphologically similar structures were also observed free of the cell surface. Negative staining of the 75000g pellet fraction showed it to be composed of vesicle-like structures ranging in size between 50 and 200 nm in diameter (Fig. 46). The 10000g pellet fraction (not shown) was composed of structures of similar morphology and size range to those in the 75 000 g pellet. The culture filtrate supernatant was examined by negative staining at a concentration of 10 mg ml-l, but was found to be devoid of vesicular structures. In all negatively stained pellet fractions and cell suspensions examined, fimbriae were absent. Bacteria were stained with osmium tetroxide and ruthenium red in order to visualize acid polyanionic polymers (Luft, 1971; Woo et al., 1979). Cell wall structures typical of those described for B . gingivalis (Okuda et al., 1981) were observed (Fig. 5 0 ) . A prominent granular ruthenium-red-positive extramural layer was also observed indicative of a polysaccharide

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Trypsin-like enzyme of Bacteroides gingivalis

Fig. 5 . Electron micrographs of ultrathin sections of (a) B. gingiualis W50 cells and (b) 75000g pellets fixed and stained with ruthenium red and osmium tetroxide. In (a) the inner cytoplasmic membrane (IM), outer membrane (OM) and periplasmic space (PS) are indicated. The bacteria are covered with a loosely fitting electron dense external granular layer (EGL), and are accompanied by extracellular vesicles (ECV) and cell-surface associated vesicles (SAV). Some of the vesicles have an external granular layer (EGL). In (h) the vesicles have a double tracked membrane (M).Bars, 200 nm.

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

289 1

2892

J . W . SMALLEY A N D A . J . B I R S S

capsule (Kasper, 1976). Blebs or vesicle-like structures were observed associated with the cell surface. In all fields examined, vesicular particles free of the cell surface were observed; they varied between 50 and 250 nm in diameter. Some of these ECVs displayed a ruthenium-redpositive layer similar to that surrounding whole cells. The ruthenium red/osmium stained 75000g and lOOOOg pellets were also examined (Fig. 5b). Both these fractions (lOOOOg not shown) were composed of vesicular particles, morphologically similar to those observed as cellsurface associated or cell-free structures in the whole bacterial cell preparations. The size of these vesicles ranged between 50 and 250 nm, and the lumina were devoid of any electron dense material. Each displayed a single double-tracked membrane but did not appear to be stained with ruthenium red. The reason for the loss of ruthenium red staining is not known. A loosely attached ruthenium-red-positive capsule was described for the extracellular vesicles of B. asaccharolyticus by Woo et al. (1979). It is possible that ruthenium red stainability of vesicles in this study was lost during manipulation. DISCUSSION

In this study extracellular trypsin-like enzyme activity was recovered from the cell- and particle-free spent growth supernatant of B. gingivalis W50.Trypsin-like enzyme activity was also recovered from the culture media of B. gingivalis by Fujimura & Nakamura (1986). These workers purified an enzyme of molecular mass 65 kDa, as estimated by SDS-PAGE, from Triton solubilized ammonium sulphate precipitates of spent media. It was not reported whether enzyme activity was present as an insoluble fraction. However, our data show that a significant proportion (80%) of trypsin-like activity in culture filtrates is associated with an insoluble vesicle fraction. This, however, may have been an underestimate of total vesicle production and hence, enzyme production, since both negatively stained and ultrathin sections of bacterial cells harvested were contaminated with vesicular particles. The cellular origins of the extracellular trypsin-like enzyme or the location in the extracellular vesicles are as yet unclear. In whole cells of B. gingivalis the enzyme can be solubilized from the cell envelope fraction by Triton X-100 and is thought to be located on the periplasmic side of the inner membrane (Yoshimura et al., 1984). Bleb formation and release of vesicles from the bacterial cell surface may be accompanied by envacuolation of the trypsin-like enzyme from the periplasmic region. Alternatively, the enzyme may be integrated into the vesicle membrane. This is likely since alkaline phosphatase, which occurs in the periplasmic space (Cheng et al., 197l), may be recovered extracellularly as a lipopolysaccharide-enzyme complex (Lindsay et al., 1973). Extracellular membrane-bound vesicles are thought to be derived by blebbing of the outer membrane and are considered to be the source of cell-free endotoxin (Russell, 1976). Indeed, we have found (unpublished data) that both the l O O O O g and the 75000g vesicle fractions display similar SDS-PAGE polypeptide and lipopolysaccharide banding patterns to the outer membrane fraction of B . gingivalis W50. However, it is still unclear whether the trypsin-like enzyme is derived from the outer membrane. Enzyme specific activity in the outer membrane fraction is far lower than that of the vesicle fractions, although this may be due to the presence of residual EDTA which inhibits the trypsin-like enzyme (Yoshimura et al., 1984). The stimulation of enzyme activity by DTT is in accord with the results of Yoshimura et al. (1984). A 10-fold optimum stimulation of the enzyme from whole cells or from the two pellet fractions was achieved at 10 mM-DTT, although the level of stimulation was half that observed by Yoshimura et al. (1984) for the purified enzyme. The enhancement of activity in our whole CFS fraction was, however, 15-fold, three times greater than that of the pellet fractions. The reason for these differences is at present unknown. In keeping with the work of Fujimura & Nakamura (1986), a 64 kDa polypeptide was found in both the soluble and the vesicle enzyme-containing fractions after column chromatography and SDS-PAGE. Release of soluble enzyme by mild sonication suggests its loose association with the vesicle structures. The 64 kDa polypeptide was found in column chromatographic fractions I and I1 released from both l O O O O g and 75000g vesicle fractions. In fraction I this component was associated with other polypeptides of molecular mass 1 1 1, 93 and 70 kDa. It is

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Trypsin-like enzyme of’Bacteroides gingivalis

2893

tempting to speculate that this fraction represents a protein complex to which the enzyme is linked by disulphide bonding. This would explain the difference in elution position from fraction I1 under non-reducing conditions. We would further speculate that the soluble CFS enzyme that does not appear to be associated with any of the above higher molecular mass polypeptides is released from the cell surface directly or liberated from ECV-protein complexes by proteolytic cleavage or the action of reducing agents. It must not be overlooked, however, that other polypeptides within these fractions are also responsible for trypsin-like enzyme activity . Whilst cell-surface vesicles and ECVs have been described previously in batch grown cultures of B. gingivalis (Okuda et al., 1981 ; Shah et al., 1976; Slots & Gibbons, 1978) their pathogenic significance is unclear. Chemostat growth of B. gingivalis W50 in the presence of haemin has a marked effect on virulence and surface morphology : haemin limitation results in the production of greater amounts of surface vesicle and ECV material (McKee et al., 1986). Of the blackpigmented Bacteroides, B. gingicalis is the most efficient at degrading haem-containing plasma proteins in vitro (Carlsson et al., 1984). Release of proteolytic enzyme containing vesicles may enable this organism to obtain amino acids and haem from such substrates. Indeed, ECVs are released by Escherichia coli during amino acid limitation (Knox et al., 1966, 1967)and it has been suggested that such structures function as packaging for degradative enzymes (Russell, 1976) in which they are stabilized by lipopolysaccharide (Costerton et al., 1974). Release of such a proteolytic enzyme may not be exclusive to B . gingivalis. Indeed, a 62 kDa protease capable of degrading IgA has been recovered from the culture supernatants of B . melaninogenicus (Mortensen & Kilian, 1984). It remains to be seen whether this enzyme is related to that produced by B. gingivalis and if it is released in association with vesicular structures. There is some debate whether bacterial cells can penetrate crevicular epithelium or connective tissues in periodontal disease (Saglie et al. 1981 ; Allenspach-Petrzilka & Guggenheim, 1983). However, both soluble enzyme and ECV structures may gain access to and mediate tissue destruction. Indeed, this possibility is strengthened by our unpublished observations that both CFS soluble enzyme and ECV preparations can degrade type I collagen and fibronectin in soluble and polymeric form. These substrates are connective tissue matrix components in periodontal ligament (Shuttleworth & Smalley, 1983) and gingival tissues (Cho et al., 1984) and their degradation may lead to loss of tissue integrity. In addition, conditions of low Eh in periodontal lesions may promote increased enzyme activity and potentiate the degradative capacity of vesicular structures. Released ECVs may, therefore, play an important role in the pathogenesis of periodontal disease. The authors wish to thank Mrs Janette Chesters for help with the electron microscopy. REFERENCES

ALLENSPACH-PETRZILKA, G. E. & GUGGENHEIM, B. (1983). Bacterial invasion of the periodontium; an important factor in the pathogenesis of periodontitis. Journal of Clinical Periodontology 10, 609-6 17. G . K. CARLSSON, J., HOFLING,J. F. & SUNDQVIST, (1 984). Degradation of albumin, haemopexin, haptoglobin and transferin, by black-pigmented Bacteroides species. Journal of Medical Microbiology 18,3946. J. W. CHENG,K.-J., INGRAM,J. M. & COSTERTON, (1971). Interactions of alkaline phosphatase and the cell wall of Pseudomonas aeruginosa. Journal of Bacteriologj? 107, 325-336. CHO, M. J., GARANT,P. R. & LEE, Y. L. (1984). Immunological localisation of collagen (I and 111) and fibronectin in inflamed and non-inflamed gingival connective tissue and sulcular fluid of beagle dogs. Journal of Periodontal Research 19,638641.

COSTERTON, J. W., INGRAM,J. M. & CHENG,K.-J. (1974). Structure and function of the cell envelope of Gram-negative bacteria. Bacteriological Reviews 38, 87-1 10. FUJIMURA, S. & NAKAMURA, T . (1986). Purification and characterisation of trypsin-like protease from Bacteroides gingiualis. Journal of Dental Research 65, abs. 107, 737. HANDLEY, P. S. & TIPLER,L. S. (1986). An electron microscope survey of the surface structures and hydrophobicity of oral and non-oral species of the bacterial genus Bacteroides. Archives of Oral Biology 31, 325-335. KASPER,D. L. (1976). The polysaccharide capsule of Bacteroides fragilis subspecies fragilis : immunochemica1 and morphologic definition. Journal of Znfectious Diseases 133, 79-87. KNOX,K. W., VESK,M. & WORK,E. (1966). Relation between excreted lipopolysaccharide complexes and

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

2894

J. W . S M A L L E Y A N D A . J. B I R S S

surface structures of a lysine-limited culture of Escherichia coli. Journal of Bacteriology 92, 12061217. KNOX,K. W., CULLEN,J. & WORK,E. (1967). An extracellular lipopolysaccharide-phospholipid-protein complex produced by Escherichia coli grown under lysine-limiting conditions. Biochemical Journal 103, 192-20 1. LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, London 227, 68&685. W. J. (1982). LAUGHTON, B. E., SYED,S. A. & LOESCHE, API ZYM system for identification of Bacteroides spp., Capnocytophaga spp., and spirochaetes of oral origin. Journal of Clinical Microbiology 15, 97- 102. S. S., WHEELER, B., SANDERSON, K. E. & LINDSAY, COSTERTON, J . W. (1973). The release of alkaline phosphatase and of lipopolysaccharide during the growth of rough and smooth strains of Salmonella typhimurium. Canadian Journal of Microbiology 19, 335-343. LUFT, J. H. (1971). Ruthenium red and violet. I. Chemistry, purification, methods of use for electron microscopy and mechanism of action. Anatomical Record 171, 347-368. MACDONALD, J. B., SOCRANSKY, S. S. & GIBBONS, R. J. (1963). Aspects of the pathogenesis of mixed anaerobic infections of mucous membranes. Journal of Dental Research 42, 529-544. MCKEE,A. S., MCDERMID, A. S., BASKERVILLE, A., D. C. &MARSH,P. D. DOWSETT, A. B., ELLWOOD, (1986). Effect of hemin on the physiology and virulence of Bacteroides gingivalis W50. Infection and Immunity 52, 349-355. B. J. & KASPER,D. L. (1977). Purification MANSHEIM, and immunochemical characterization of the outer membrane complex of Bacteroides melaninogenicus subspecies asaccharolyticus. Journal of Infectious Diseases 135, 787-799. MAYRAND, D., MCBRIDE,B. C., EDWARDS,T. & JENSEN,S. (1 980). Characterisation of Bacteroides asaccharolyticus and B. melaninogenicusoral isolates. Canadian Journal of Microbiology 26, 1 178-1 183. MORTENSEN, S. B. & KILIAN,M. (1984). Purification and characterisation of an immunoglobulin A 1 protease from Bacteroides melaninogenicus. Infection and Immunity 45, 550-557. OKUDA,K., SLOTS, J. & GENCO, R. J. (1981). Bacteroides gingiilalis, Bacteroides asaccharolyticus and Bacteroides melaninogenicus subspecies : cell surface morphology and adherence to erythrocytes

and human buccal epithelial cells. Current Microbiology 6, 7-12. RUSSELL, R. R. B. (1976). Free endotoxin - a review. Microbios Letters 2, 125-1 35. SAGLIE,R., NEWMAN, M. G., CARRANZA, F. A. & PATTISON, G. L. (1981). Bacterial invasion of gingiva in advanced periodontitis in humans. Journal of Periodontology 53, 2 17-222. SHAH,H. N., WILLIAMS, R. A. D., BOWDEN, G. W. & HARDIE, J . M. (1976). Comparison of the biochemical properties of Bacteroides melaninogenicus irom human dental plaque and other sites. Journal of Applied Bacteriology 41, 473-492. SHUITLEWORTH, C. A. & SMALLEY, J. W. (1983). Periodontal ligament. International Review of Connective Tissue Research 10, 21 1-247. SLOTS,J. (1981). Enzymatic characterization of some oral and non-oral Gram-negative bacteria with the API ZY M system. Journal of Clinical Microbiology 14, 288-294. SLOTS, J. (1982). Importance of black pigmented Bacteroides in human periodontal disease. In Host Parasite Interactions in Periodontal Diseases, pp. 2745. Edited by R. J. Genco & S. E. Mergenhagen. Washington, DC : American Society for Microbiology. SLOTS, J. & GENCO,R. J. (1984). Black pigmented Bacteroides species, Capnocytophaga species and Actinobacillus actinomycetemcomitans in human periodontal disease : virulence factors in colonization survival and tissue destruction. Journal of Dental Research 63, 412-42 1. SLOTS, J. & GIBBONS,R. J. (1978). Attachment of Bacteroides melaninogenicus subsp. asaccharolyticus to oral surfaces and its possible role in colonization of the mouth and of periodontal pockets. Infection and Immunity 19, 254-264. SPURR,A. J. (1969). A low viscosity epoxy resin embedding medium for electron microscopy. Journal of Whrastructural Research 26, 3 1-43. Woo, D. D. L., HOLT, S. C. & L E A D B E ~ EE. R ,R. (1 979). Ultrastructure of Bacteroides species : Bacteroides asaccharolyticus, Bacteroides fragilis, Bacteroides melaninogenicussubspecies melaninogenicusand Bacteroides melaninogenicus subspecies intermedius. Journal of Infectious Diseases 139, 534-546. F., NISHIKATA, M., SUZUKI,T., HOOVER, YOSHIMURA, E. (1984). Characterisation of a C. I. & NEWBURN, trypsin-like protease from the bacterium Bacteroides gingivalis isolated from human dental plaque. Archives of Oral Biology 29, 559-564.

Downloaded from www.microbiologyresearch.org by IP: 37.44.207.200 On: Wed, 18 Jan 2017 07:50:54

Suggest Documents