Activation of Blood Coagulation Factor X by Arginine-specific Cysteine Proteinases (Gingipain-Rs) from Porphyromonas gingivalis *

The effect of two arginine-specific cysteine proteinases (gingipain Rs) from Porphyromonas gingivalis, a causative bacterium of adult periodontitis, on human blood coagulation was investigated. Activated partial thromboplastin time and prothrombin time were shortened by these proteinases, with a 95-kDa gingipain R containing adhesin domains being 5-fold more efficient in comparison to a 50-kDa gingipain R containing the catalytic domain alone. The 50-kDa enzyme reduced each coagulation time in several plasmas deficient in various coagulation factors, while it was ineffective in factor X-deficient plasma unless reconstituted with this protein. Each proteinase activated factor X in a dose- and time-dependent manner, with Michaelis constants (K m ) being found to be lower than the normal plasma factor X concentration, strongly suggesting that factor X activation by gingipain Rs, especially the 95-kDa form which is strongly activated by phospholipids, could occur in plasma. This is the first report of factor X activation by bacterial proteinases and indicates that the gingipain Rs could be responsible for the production of thrombin and, indirectly, with the generation of prostaglandins, interleukin-1, etc., which have been found to be associated with the development of periodontitis induced by P. gingivalis infections. Furthermore, the data support the hypothesis that induction of blood coagulation by bacterial proteinases may be a causative agent in the pathogenesis of disseminated intravascular coagulation in sepsis.

The effect of two arginine-specific cysteine proteinases (gingipain Rs) from Porphyromonas gingivalis, a causative bacterium of adult periodontitis, on human blood coagulation was investigated. Activated partial thromboplastin time and prothrombin time were shortened by these proteinases, with a 95-kDa gingipain R containing adhesin domains being 5-fold more efficient in comparison to a 50-kDa gingipain R containing the catalytic domain alone. The 50-kDa enzyme reduced each coagulation time in several plasmas deficient in various coagulation factors, while it was ineffective in factor X-deficient plasma unless reconstituted with this protein. Each proteinase activated factor X in a doseand time-dependent manner, with Michaelis constants (K m ) being found to be lower than the normal plasma factor X concentration, strongly suggesting that factor X activation by gingipain Rs, especially the 95-kDa form which is strongly activated by phospholipids, could occur in plasma. This is the first report of factor X activation by bacterial proteinases and indicates that the gingipain Rs could be responsible for the production of thrombin and, indirectly, with the generation of prostaglandins, interleukin-1, etc., which have been found to be associated with the development of periodontitis induced by P. gingivalis infections. Furthermore, the data support the hypothesis that induction of blood coagulation by bacterial proteinases may be a causative agent in the pathogenesis of disseminated intravascular coagulation in sepsis.
Periodontitis is an infectious disease associated with a loss of connective tissue, resorption of alveolar bone, and formation of periodontal pockets. It is the most common cause of tooth loss in adults, primarily because of the declining incidence of dental caries in the general population (1,2). Although the pathogenesis of periodontitis is not completely understood, prostaglandins (3,4) and interleukin-1 (5,6), which increase in gingival crevicular fluid in periodontal pockets, are considered to be predominant factors in the tissue destruction process associated with this disease. However, the mechanism of the production of these inflammatory mediators is still unclear.
Thrombin, which is primarily associated with the cleavage of fibrinogen to generate fibrin clots (7)(8)(9)(10)(11), is a key proteinase in the blood coagulation system. However, besides its central role in hemostasis, this proteinase is also a potent stimulator of prostaglandin synthesis in osteoblasts (12), with in vitro bone resorption appearing to be dependent, at least in part, on thrombin-stimulated prostaglandin synthesis (13,14). In addition, thrombin also potentiates lipopolysaccharide-stimulated interleukin-1 production by macrophages (15). These facts suggest that thrombin may play a major role in the development of periodontitis by indirectly causing tissue breakdown including alveolar bone resorption. However, whether thrombin is produced at periodontitis sites is still unknown.
A close relationship between Porphyromonas gingivalis (formally Bacteroides gingivalis) and adult periodontitis has been reported (16 -18), with proteolytic enzymes which are produced in large quantity by this bacterium acting as important pathogenic agents (19 -21). We have previously purified two arginine-specific cysteine proteinases, referred to as 50-kDa and 95-kDa gingipain 1 Rs, from P. gingivalis (22,23), and found each to be capable of generating bradykinin through prekallikrein activation (24) and degrading fibrinogen in human plasma (25). Both of these events may be potentially associated with crevicular fluid production and bleeding tendency, respectively, at periodontitis sites. The 95-kDa proteinase is composed of both a catalytic and an adhesin domain (26), the former being 81% identical in primary structure with that of the 50-kDa form. 2 However, little is known with regard to any other specific functional differences between the two enzymes. Since coagulation factors are activated by cleavage of peptide bonds at the carboxyl-terminal side of specific arginine residues (27), it is possible that each of the gingipains Rs could activate the coagulation cascade system to produce thrombin, an enzyme known to have multiple functions. Therefore, to study other possible pathogenic roles for the two gingipain Rs, we investigated their effect on activating this system. It was found that each proteinase could induce blood coagulation through factor X activation. However, there were significant differences between 95-kDa gingipain R and the 50-kDa form in performing this function. mas deficient in factors VII, VIII, IX, X, XI, XII, or prekallikrein were purchased from Sigma. SIMPLASTIN ® , AUTOMATED APTT, and Platelin ® (rabbit brain phospholipids) were purchased from Organon Teknika, Corp. (Durham, NC) while D-Phe-L-Pro-L-Arg-chloromethylketone (FPR-ck) was obtained form Bachem Biosci. Inc. (King of Prussia, PA). Purified human factor X was purchased from Enzyme Research Laboratories, Inc. (South Bend, IN) and purified human factor Xa was acquired from Kaketsuken (Kumamoto, Japan). t-Butyloxycarbonyl-L-isoleucyl-L-glutamyl-glycyl-L-arginine-4-methyl-coumaryl-7-amide (Boc-Ile-Glu-Gly-Arg-MCA) was purchased from the Peptide Institute (Minoh, Japan). p-Nitrophenyl-pЈ-guanidinobenzoate was from Nacalai tesque (Kyoto, Japan). Normal human plasma was obtained by centrifugation of a mixture of 9 volumes of freshly drawn blood from healthy volunteers and 1 volume of 3.8% (w/v) sodium citrate.
Proteinase Purification-50-kDa gingipain R and 95-kDa gingipain R were isolated according to the method described by Pike et al. (23). The amount of active enzyme in each purified proteinase was determined by active site titration using FPR-ck. 3 The concentration of active gingipain R was calculated from the amount of inhibitor needed for complete inactivation of the proteinase.
Activation of Proteinases-Each gingipain R form was activated with 10 mM cysteine in 0.2 M Hepes buffer, pH 8.0, containing 5 mM CaCl 2 at 37°C for 10 min. The activated proteinase (2 M) was then diluted with 10 mM Tris-HCl, pH 7.3, containing 150 mM NaCl (TBS) and 5 mM CaCl 2 prior to use.
Proteinase Activity Assay-The amidolytic activity of the gingipain Rs was determined using benzoyl-L-arginine-p-nitroaniline. Samples were preincubated at 37°C for 5 min in 0.1 M Tris-HCl buffer, pH 7.6, containing 200 mM Gly-Gly, 5 mM CaCl 2 , and 10 mM cysteine, pH 7.6, and then assayed for amidolytic activity using 1 mM substrate. The formation of p-nitroaniline was monitored spectrophotometrically at 405 nm.
Clotting Assay-Clotting time was measured with COAG-A-MATE ® XC (General Diagnostics, Morris Plains, NJ) following the manufacturer's instructions. For prothrombin time (PT) assay, 90 l of plasma and 10 l of a given proteinase were incubated in a plastic cell at 37°C for 3 min, followed by addition 200 l of SIMPLASTIN ® to initiate coagulation. For activated partial thromboplastin time (APTT) assay, 90 l of plasma and 10 l of proteinase were incubated in a plastic cell at 37°C for 1 min, followed by addition of 100 l of AUTOMATED AAPTT ® , incubation at 37°C for 3 min, and addition of 100 l of 25 mM CaCl 2 to initiate coagulation.
Kinetic Analysis of Factor X Activation-Factor X, dissolved in 450 l of 0.1 M Tris-HCl, pH 7.6, containing 0.15 M NaCl, 5 mM CaCl 2 , and 45 g/ml phospholipids, was incubated with 50 l of either of the gingipain Rs (0.5 nM) at 37°C for 30, 60, 90, 120, or 150 s. One-hundred ml of the solution was then added to 500 l of the same buffer supplemented with 1.5 M antipain, to completely inhibit cysteine proteinase activity but not the amidolytic activity of factor Xa at the concentration used. This was followed by addition of 20 l of a factor Xa-specific substrate, Boc-Ile-Glu-Gly-Arg-MCA (10 mM). The amount of 7-amino-4-methyl coumarin (AMC) released by factor Xa at 37°C was measured fluorometrically with a fluorescence spectrophotometer (Model 650 -40, Hitachi), the fluorescence at 440 nm with excitation at 380 nm being monitored with a recorder. The factor Xa concentration produced by either proteinase was calculated by using as a standard the amidolytic activity of purified factor Xa which had been active site-titrated with p-nitro-pЈ-guanidinobenzoate (28). The initial velocity of factor Xa production at various factor X concentrations was determined by the best fit line for each factor Xa concentration at the five incubation periods mentioned above. Several factor X concentrations in a range from 35 to 425 nM were used for the kinetic study.
The values of the Michaelis constant (K m ) and the maximum velocity (V max ) in the Michaelis-Menten equation were obtained using three different plots, [S] 0 /v versus [S] 0 , 1/v versus 1/[S] 0 and v versus v/[S] 0 (v and [S] 0 denote the catalytic rate and the initial substrate concentration, respectively), where the best fit values were determined by the method of least squares with Taylor expansion, described by Sakoda and Hiromi (29).
Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis-Analysis by SDS-polyacrylamide gel electrophoresis was performed with 15% slab gels according to the method of Laemmli (30), with 0.8% Coomassie Brilliant Blue R-250 being used for protein staining.
Amino-terminal Sequence Analysis-Automatic sequence analysis was performed with a pulse-liquid phase sequencer (model 477A Pro-tein Sequencer, Perkin-Elmer/Applied Biosystems Inc.). To determine the amino-terminal sequence of factor X-derived fragments, each protein fragment was separated by SDS-polyacrylamide gel electrophoresis and transferred to the Immobilon TM polyvinylidene difluoride transfer membrane (Millipore Co., Ltd., Bedford, MA). The proteins transferred to the polyvinylidene difluoride membrane were visualized by staining with Coomassie Brilliant Blue R-250. The bands excised were placed on a Polybrene-treated glass filter and sequence analysis was performed.
Determination of Factor X Concentration-The molar concentrations of purified factor X were calculated using A 280 nm 1% [/stack ϭ 13.3 and a molecular mass of 59 kDa (31).

Activation of Blood Coagulation by Gingipain
Rs-To investigate the effect of gingipain Rs on blood coagulation, normal human plasma was incubated with either of the proteinases and examined for APTT. Both proteinases decreased APTT in a dose-dependent manner at concentrations as low as 0.1 nM, with the 95-kDa gingipain R being approximately 5-fold more effective than that of the 50-kDa gingipain R (Fig. 1). Since TLCK-treated proteinases did not shorten APTT (Fig. 1), the effect of gingipain Rs is dependent on their proteolytic activity. To investigate the mechanism for reducing APTT by gingipain Rs, the effect observed was further studied using plasmas deficient in a given factor found in the intrinsic coagulation pathway. The 50-kDa gingipain R shortened APTTs of plasmas deficient in factor XII, prekallikrein, XI, IX, or VIII in a dosedependent manner (Table I). However, this enzyme did not affect the APTT of factor X-deficient plasma, unless it was reconstituted with the missing factor (Table I). It is suggested from these results that gingipain Rs shortened APTT through factor X activation. Normal human plasma incubated with either of the proteinases was also examined for PT to investigate the effect of gingipain Rs on the extrinsic pathway factors. It was again found that both proteinases reduced PT in a dose-dependent manner at concentrations above 1 nM, with the effect requiring active enzyme (Fig. 2). The 95-kDa gingipain R shortened the PT 5-fold more effectively than the 50-kDa form (Fig. 2), as previously observed in the case of APTT. The 50-kDa gingipain R also reduced PT of factor VII-deficient plasma, whereas there was little effect on the PT of factor X-deficient 3 J. Potempa, R. Pike, and J. Travis, manuscript in preparation. plasma, unless it was reconstituted with factor X (Table II). These data again suggest that gingipain R decreased PT through factor X activation and that accelerated blood coagulation induced by gingipain Rs is specifically due to activation of this factor.
Activation of Factor X by Gingipains Rs-To investigate factor X activation by gingipain Rs, we incubated these cysteine proteinases with purified factor X and measured the factor Xa activity produced. Both proteinases generated factor Xa linearly in a dose-and time-dependent manner (Fig. 3, A and B). Consistent with the data shown in Figs. 1 and 2, the 95-kDa gingipain R produced more factor Xa than the 50-kDa gingipain R (Fig. 3, A and B). TLCK-treated gingipain Rs did not induce factor X activation (Fig. 3A).
Cleavage of Factor X by the 50-kDa Gingipain R-To investigate the mechanism of factor X activation by gingipain R in more detail, we examined the pattern of cleavage during this process on an SDS-polyacrylamide gel, followed by amino-terminal sequence analysis of the fragments generated. Treatment with proteinase did not shift the 21-kDa molecule, which had an amino-terminal sequence ANSFLXXMKK (X is presumed to be ␦-carboxyglutamic acid) identical to that of the factor X light chain (Fig. 4). This indicates that cleavage had only occurred within the heavy chain which contains the catalytic domain blocked by the activation peptide. Our data also suggests that the first cleavage by gingipain R resulted in a shift in molecular mass from 47 to 45 kDa (lane b) without affecting the amino-terminal sequence (SVAQATSSSG-), presumably because of hydrolysis between Arg 278 -Gly 288 and release of a 19-amino acid peptide. This is consistent with the fact that the molecular mass of the final form of the heavy chain is slightly smaller than that of the factor Xa heavy chain (lanes d and e). The amino-terminal sequences of the 35-and 32-kDa molecules were GDNNLTRIVG-and IVGGQECKDG-, respectively. From the amino acid sequence of human factor Xa (32), the generation of the 35-kDa fragment must have occurred through cleavage between Arg 45 and Gly 46 within the modified heavy chain to release a 45-amino acid activation glycopeptide. Finally, cleavage must occur between Arg 52 and Ile 53 to release the remainder of the activation peptide, also glycosylated. All of these proteolysis sites are in agreement with the fact that gingipain Rs only cleave polypeptide chains after arginine residues (22,23) and at positions which are consistent with the activation of Factor X to Factor Xa.
Effect of Phospholipids on Gingipain R Factor X Activation-Phospholipids are important cofactors in blood coagulation and accelerate the cascade reaction of coagulation factors. We, therefore, studied the effect of phospholipids on factor X activation by gingipain Rs. 95-kDa gingipain R factor Xa production increased in a phospholipid concentration-dependent manner with the effect reaching a plateau at concentrations above 40 g/ml, and yielding a 7-8-fold increase in factor X activation at enzyme concentrations of both 0.02 and 0.05 nM (Fig. 5). Interestingly, phospholipids did not stimulate the activation of factor X by the 50-kDa gingipain R (Fig. 5). Phospholipids did not augment 95-kDa gingipain R factor X activation in the absence of calcium, while factor X activation by the two proteinases was not affected by calcium in the absence of phospholipids (Fig. 5). It is likely that the calcium ion-mediated binding of the adhesin domains to phospholipids, in addition to factor X binding, is involved in augmentation of the factor X activation.
Kinetics of Factor X Activation by Gingipain Rs-To investigate the kinetics of factor X activation by gingipain Rs, the values of K m and k cat were determined for the interaction of purified factor X with either enzyme at a phospholipid concen-

plasmas deficient in a factor in the intrinsic coagulation pathway
The effect of 50-kDa gingipain R on activated partial thromboplastin time was examined in plasmas deficient in a factor utilized in the intrinsic pathway, according to the method shown in Fig. 1 c Clotted within 20 s, the detection limit. d Factor X-deficient plasma reconstituted with 9 g/ml factor X.  a Factor X-deficient plasma reconstituted with 9 g/ml purified factor X. tration of 40 g/ml. The values of K m and k cat were 98 nM and 0.4/s for 95-kDa gingipain R and 136 nM and 0.012/s for 50-kDa gingipain R (Figs. 6, A and B), the catalytic efficiency (k cat /K m ) of the former enzyme being 46-fold higher than that of the 50-kDa gingipain R. This result does not appear to be consistent with the fact that the 95-kDa gingipain R activated blood coagulation 5ϳ6-fold more efficiently than the 50-kDa gingipain R (Figs. 1 and 2). This is probably because Ca 2ϩ are absent during the incubation of plasma with each proteinase and added just before coagulation initiation, resulting in a lessened effect of phospholipids on 95-kDa gingipain R-stimulated factor X activation. The kinetic constants of gingipain Rs for factor X conversion were compared with those of activated factor VII (VIIa) complexed with tissue factor (factor VIIa-TF) (33), factor IX complexed with activated factor VIII (VIIIa) (34), or Russell's viper venom factor X coagulant protein (RVV-XCP), a factor X-specific activator (34). The K m values of the cysteine proteinases were higher than those of physiological factor X activators but lower than the K m value of RVV-XCP (Table III). The k cat values of gingipain Rs were lower than any of other factor X activators, but the k cat value of 95-kDa gingipain R was close to the values of factor VIIa-TF and RVV-XCP (Table III). The k cat /K m value was much lower than the values of physiological factor X activators but was about half of the value of RVV-XCP. These data suggest that 95-kDa gingipain R is a more potent factor X activator than the 50-kDa form and is comparable to RVV-XCP in factor X activation. DISCUSSION In this study we have provided experimental evidence to support the hypothesis that gingipain Rs can induce blood coagulation through activation of factor X. Although snake venom enzymes are known to activate factor X (35,36), no bacterial proteinase has been reported so far to perform this same function. Indeed, staphylocoagulase is the only bacterial coagulase which can induce human plasma coagulation, appar-FIG. 3. Activation of factor X by the 50-or 95-kDa gingipain R. Ten microliters of a proteinase was added to 90 l of factor X (1 M dissolved in 0.1 M Tris-HCl, pH 7.6, 150 mM NaCl, and 5 mM CaCl 2 ) and incubated at 37°C. Then, 500 l of the above buffer supplemented with 1.5 M antipain was added to stop the activation. Twenty microliters of Boc-Ile-Glu-Gly-Arg-MCA (10 mM) was added to the solution and the velocity of AMC release was measured at 37°C. A, proteinases of various concentrations were incubated with factor X for 5 min. The final concentrations of each proteinase are shown. ⅜, 50-kDa gingipain R; Ç, 95-kDa gingipain R; q, TLCK-inactivated 50-kDa gingipain R; å, TLCK-inactivated 95-kDa gingipain R. B, a proteinase (0.05 nM, the final concentration) was incubated with factor X at 37°C for various periods. , 50-kDa gingipain R; q, 95-kDa gingipain R; E, TBS instead of a proteinase solution. ently through formation of an active molecular complex with prothrombin (37). The present work, therefore, demonstrates the existence of factor X activators of bacterial origin for the first time. Previously, it has been shown that gingipain Rs can activate prekallikrein and convert it to kallikrein (24), an activator of Hageman factor (38) which initiates the intrinsic coagulation pathway (39). Accordingly, it is possible that gingipain Rs trigger blood coagulation through this mechanism. However, this pathway is unlikely to be involved in gingipain R-induced blood coagulation, since the APTT of prekallikreindeficient plasma, like other deficient plasmas (except factor X-deficient plasma), is also shortened by 50-kDa gingipain R (Table I). Furthermore, this bacterial proteinase also shortened PT of factor VII-deficient plasma but not of factor X-deficient plasma (Table II). Thus, it seems likely that the factor X activation by gingipain Rs is the primary site through which blood coagulation is induced by these proteinases.
The K m values of gingipain Rs (Table III) are lower than the factor X concentration in normal plasma (around 10 g/ml, 170 nM) (27), which, together with their effects on APTT and PT ( Figs. 1 and 2), supports the hypothesis that factor X activation by gingipain Rs can occur in plasma. Since phospholipids and calcium ions are ubiquitous cell membrane components and ions, respectively, their stimulatory effect on 95-kDa gingipain R-induced factor X activation (Fig. 5) can occur in vivo. Together with the fact that the k cat /K m value is close to the value of RVV-XCP (Table III), a strong factor X activator, it is likely that the high molecular weight gingipain R is physiologically a more important factor X activator than the smaller form.
We previously reported the binding activity of the adhesin domains in high molecular mass forms of gingipains to various proteins (40). The fact that the phopholipids augment factor X activation only by 95-kDa gingipain R (Fig. 5), complexed with adhesin domains, suggest the binding of the adhesin domains to phospholipids. Exoenzyme S, recently confirmed as an adhesin for Pseudomonas aeruginosa (41), binds to phosphatidylethanolamine specifically in the presence of 5 mM CaCl 2 (42) and lipid receptors have been shown to be involved in the binding of Helicobacter pylori and Burkholderia (Basonym Pseudomonas) cepacis to epithelial cells (42,43). Thus, it may be natural for 95-kDa gingipain R adhesin domains to bind to phospholipids. However, phospholipids augmented neither Boc-Ile-Glu-Gly-Arg-MCA hydrolysis, prekallikrein activation, nor fibrinogen degradation by the complexed form, 4 hence the phospholipids effect appears to be specific for factor X activation. The augmentation of 95-kDa gingipain R factor X activation by phospholipids, together with the lower K m and much higher k cat values of 95-kDa gingipain R than those of the 50-kDa form (Table III), indicates a further function for the adhesin domains in increasing the affinity to factor X as well as facilitating substrate cleavage.
In human plasma, gingipain Rs generate bradykinin (24), a potent vascular permeability enhancing peptide. In parallel, and through factor X activation thrombin is also generated ( Fig. 1 and 2). Thus, in blood there is likely to be the simultaneous production of each by gingipain Rs, and for this reason it is believed that both may also be involved in the production of gingival crevicular fluid (24) and of prostaglandins and interleukin-1 (13)(14)(15), respectively, during infectious episodes. Hence, the elevated levels of prostaglandins and interleukin-1 in gingival crevicular fluid of adult periodontitis patients (3,4,6) may be connected to the thrombin produced by P. gingivalis proteinases. In addition, thrombin and fibrinopeptide B, released from fibrinogen by thrombin are each neutrophil chemotactic factors (44,45) and may contribute to infiltration of these phagocytic cells at periodontitis sites (46). Finally, factor Xa also enhances macrophage interleukin-1 production (15) and is mitogenic for endothelial cells (47), smooth muscle cells (48), and lymphocytes (49). Therefore, the data described here would imply that gingipain R-induced activation of blood coagulation through conversion of factor X to Xa can easily be involved in the development of periodontitis.
In septic patients, disseminated intravascular coagulation (DIC) occurs frequently. Recent studies indicate that the initial activation of coagulation in sepsis is primarily dependent on activation of the extrinsic pathway, initiated by tissue factor expressed on monocytes (50). Since endotoxin, itself, or endo-4 T. Imamura, unpublished data.  toxin-induced tumor necrosis factor-␣ can stimulate monocytes to express tissue factor (51,52), the tendency has been to recognize such factors as the primary bacterial agents involved in the induction of DIC coagulation in sepsis. However, our finding that gingipain Rs can also readily induce blood coagulation suggest the possibility of bacterial proteinase-induced DIC coagulation in sepsis. It has been reported that a patient with an acute dentoalveolar abscess developed septicemia of Bacteroides melaninogenicus, a bacterium similar to P. gingivalis, and died with DIC (53). The fact that no plasma proteinase inhibitor effectively inactivates gingipain Rs (data not shown) supports this possibility. If so, inhibitors of such enzymes could be a therapy for DIC.