- Email: [email protected]
Preparation of thrombomodulin from human placenta
THROMBOSIS RESEARCH 37; 353-364, 1985 0049-3848/85 $3.00 + .OO Printed in the USA. Copyright (c) 1985 Pergamon Press Ltd. All rights reserved.
PREPARATION
OF THROMBOMODULIN
Shinichiro Institute
Kurosawa
of Hematology
Jichi Medical
FROM HUMAN PLACENTA
and Nobuo
Aoki
and Department of Medicine,
School, Tochigi-Ken
329-04, Japan
(Received 11.10.1984; Accepted in revised form 30.11.1984 by Editor S. Iwanaga)
ABSTRACT Thrombomodulin, a cell surface cofactor for thrombincatalyzed activation of protein C, was isolated from human placenta by a combination of affinity chromatography and gel filtration. Molecular weight of purified human thrombomodulin, estimated by sodium dodecyl sulfate electrophoresis after reduction, was 88,000. Its isoelectric point was around pH 4. The thrombomodulin was soluble only in the presence of detergent. The purified thrombomodulin was inactivated by disulfide bond reduction, but was stable to heat, pH The thrombomodulin extremes and protein denaturants. served as a cofactor of thrombin-catalyzed protein C activation in human as well as bovine enzyme systems.
INTRODUCTION Protein C plays an important role in regulating the blood coagulation process in vivo (l), and its congenital deficiency results in thrombophilic states (2). Protein C must be activated to exert its physiological role as an anticoagulant, and its activation is catalyzed by thrombin in the presence of a cofactor called thrombomodulin (3). Thrombomodulin is presumably present on the vascular endothelium as a membrane-bound protein (4). This cofactor has been isolated from rabbit lungs, and its effect on thrombin-catalyzed activation of protein C has
contents were presented at an annual meeting of Japan Society of Hematology on April 12, 1984, in Kyoto, Japan. Key words: Thrombomodulin, Protein C, Placenta, MCA-substrate The
353
354
HUMAN THROMBOMODULIN
Vo1.37, No.3
been extensively studied using materials of bovine origin (5). However, preparation of human thrombomodulin and its effect on the activation system of protein C in humans have not been reported. In this communication, we present some results of our study on isolation of thrombomodulin from human placentae, its characterization and its effect on activation of human protein C.
MATERIALS
AND METHODS
Extract of human placenta In a few hours after delivery of the placenta, it was perfused via the umbilical artery with buffer A (0.02 M Tris-HCl, 0.25 M sucrose, 1 mM benzamidine-HCl, pH The placenta was then minced and homogenized with 7.5) at 4'C. buffer A by the use of a Waring blender. The homogenate was centrifuged at 30,000 g for 40 min. The pellet was suspended in buffer A and recentrifuged. The same procedure was repeated for atotalof 6 times. All procedures were carried outat4'C or on ice. Finally, thrombomodulin was extracted from the pellet by homogenization and centrifugation as described above with 150 ml of bufferA containing 0.5% (v/v) Triton X-100 (Sigma Chemical co. U.S.A.). Thrombin Thrombin was purified from a commercial bovine (Mochida Pharm. Co., Tokyo) or human (Midori-Juji Corp., Osaka) thrombin preparation by the method of Lundblad (6). DIP-thrombin-agarose Sepharose-4B (Pharmacia, Sweden) was activated with CNBr bv the method of Cuatrecasas (7). . and thrombin was linked to the agarose to give a final thrombin concentration The thrombin was then around 1 mg of thrombin/ml of agarose. inactivated by incubation with 1 mM diisopropylphosphofluoridate (Sigma) in 0.1 M NaCl, 0.02 M Tris-HCl, pH 7.5. Protein C Bovine andhumanproteinc was purifiedbythe method of Stenflo (8) and the method of Suzuki et al. (9), respectively. Antithrombin III Antithrombin III was purified from human plasma as described by Miller-Anderson et al. (10). MCA-substrate for activated protein C Boc-Leu-Ser-Thr-Arg-MCA (4-methyl-coumaryl-7-amide) (11) was purchased from Protein Research Foundation, Osaka, Japan. Agarose beads for gel filtration from LKB (Sweden).
Ultro Gel ACA 34 was purchased
Assay of thromboxnodulin Ten ~1 of bovine protein C (500 pg/ml) was mixed with 5 plof bovine thrombin (40 u/ml) and 10 ).11of the sample to be assayed, and the total volume was adjusted to 100 ~1 with 0.02 M TriS-HCl, 0.1 M NaCl, 3.5 mM CaC12, pH 7.5. The mixture was incubated at 37'C for 30 min. After the incubation, 150 ~1 of antithrombin III solution (300 pg/ml) in buffer B (0.05 M Tris-HCl, 0.1 M NaCl, 1 mM CaC12, pH 8.5) was added to the mixture and incubated further for 15 min to neutralize
Vo1.37, No.3
355
HUMAN THROMBOMODULIN
thrombin. Subsequently, 250 ~1 of MCA-substrate solution (100 PM) in buffer B was added to the mixture and incubated at 37'C for 10 min. The reaction was immediately stopped by adding 500 ~1 of 20% (v/v) acetic acid. AMC (7-amino-4-methylcoumarine) liberated was then measured by a spectrofluorophotometer (Shimadzu RF-510, Kyoto, Japan) with excitation at 380 nm and emission at 460 nm, using a standard curve constructed with a standard AMC (Protein Research Foundation, Osaka, Japan). The activity of active protein C generated was linearly related to the amount of thrombomodulin used (see Results), and the amount of thrombomodulin that ultimately liberated one nmole of AMC/min.ml in the assay system was defined as one unit of thrombomodulin. Buffers used for chromatography 0.1 M NaCl, 0.02 M Tris-HCl, 0.5 mM CaC12, 1 mM benzamidine-HCl, 0.5% Lubrol PX, pH 7.5 (0.1 M NaCl-Lubrol buffer): 0.2 M NaCl with other constituents being the same as above (0.2 M NaCl-Lubrol buffer); 0.4 M NaCl with other constituents being the same as above (0.4 M NaCl-Lubrol buffer); 0.4 M NaCl-Lubrol buffer with 1 mM EDTA replacing CaC12 (0.4 M NaCl-EDTA-Lubrol buffer); and 1 M NaCl with other constituentsbeingthe same as 0.4 M NaCl-EDTA-Lubrolbuffer (1 M NaCl-EDTA-Lubrol buffer). Polyacrylamide gel electrophoresis The electrophoresis was carried out according to the sodium dodecyl sulfate (SDS) system of Laemmli (12), using a 10% polyacrylamide slab gel with and without dithiothreitol. After electrophoresis, the gels were stained by the silver-staining procedure described by Morrissey (13). For estimation of molecular weight, a mixture of molecular weight protein standards (Bio-Rad Lab., Calif.) was run on the same gel. Isoelectric focusing Isoelectric focusing was carried out according to methods described by Versterberg (14), in an LKB jacketed column (llO-ml capacity) using 4% ampholytes, pH range 3-10, at 300 v for 70 h at 2OC. Activated partial thromboplastin time Platelet-poor plasma was obtained by centrifugation of blood anticoagulated with 0.1 vol of 3.8% sodium citrate. One hundred ~1 of plasma was mixed at 37'C with 100 ~1 of activated cephaloplastin reagent (Actin, Dade Diagnostic, Inc.). The clotting reaction was started by adding 100 1.11 of 0.025 M calcium chloride at 37'C, and the time required for clot formation after the addition of calcium was recorded. Other reagents Pepsin was purchased from Worthington. serum albumin and Lubrol PX were purchased from Sigma. other chemicals used are all analytical grades.
Bovine The
RESULTS Chromatography of the placental extract on DIP-thrombin-agarose. The conditions used for the chromatography were essentially the
356
HUMAN THROMBOMOOULIW
Vo1.37, No.3
same as those described by Esmon et al. (5). Calcium chloride was added to the extract to a concentration of 0.5 mM before chromatography. Typical results are shown in Fig. 1. The majority of contaminating proteins in the extract did not bind to the agarose equilibriated with 0.1-0.2 M NaCl. Thrombomodulinboundtothe agarose was elutedwith 1 M NaCl. The yield was variable and ranged from 25 to 90% of the activity applied on the column. This chromatography increased the specific activity by 5-lo-fold.
n
FRACTIONS
FIG. 1 Chromatography of the placental extract on DIP-thrombin-agarose. A DIP-thrombin-agarose column (2.6 x 16.5 cm) was equilibrated with O.lM NaCl-Lubrolbuffer, and then 200 ml of the extract After sample (1.044 Aztju, 250 u/ml) was applied on the column. application, the columnwas washed with 2 liters of 0.2 M NaClLubrol buffer over a period of 72 hours, and then thrombomodulin was eluted with 1 M NaCl-EDTA-Lubrol buffer: 3.5-ml fractions were collected. The eluate was pooled as indicated by the bar, and the pooled eluate was concentrated by ultrafiltration.
Gel filtration The eluate from the DIP-thrombin agarose was pooled and concentrated approximately 20 times with a Millipore CX-30 ultrafiltration unit fitted to a CX agitator (Japan Millipore, Ltd., Tokyo). The concentrated eluate was then applied on the column of Ultrogel ACA 34 for gel filtration. Typical results are shown in Fig. 2. There were usually two major protein peaks and an additional two minor peaks. A large single peak of thrombomodulin activity was found between the two major protein peaks. The recovery of thrombomodulin was 50-908 of the activity applied on the column, and the gel filtration increased the specific activity by 7-15-fold.
vo1.37, No.3
HUMAN THROMBOMODULIN
357
FRACTIONS
FIG. 2 Gel filtration of partially purified thrombomodulin on Ultrogel ACA 34. The concentrated eluate (5 ml) from the first DIPthrombin-agarose column (1.214 A280, 1448 u/ml) was applied on the column (2.5 x 89 cm) equilibrated with 0.1 M NaCl-Lubrol buffer. The gel filtration was carried out with the same buffer with a flow rate at 25 ml/hour, and the effluent was collected in 2.46-ml fractions. The effluent was pooled as indicated by the bar, and it was concentrated by ultrafiltration.
Rechromatography on DIP-thrombin-agarose The effluent containing thrombomodulin from the gel filtration was pooled and concentrated approximately 10 times with a Millipore CX-30. The concentrate was added with calcium chloride to obtain a calcium concentration of 0.5 mM, and then it was applied on the column of DIP-thrombin-agarose and eluted with a linear gradient. Typical results are shown in Fig. 3. Thrombomodulin was eluted with approximately 0.8 M NaCl. The eluate possessing maximal thrombomodulin activity was dialyzed against 0.02 M Tris-HCl, 0.1 M NaCl, 0.5% Lubrol PX, pH 7.5, and concentrated by means of The recovery of thrombomodulin Millipore CX-30 ultrafiltration. was around 95% of the activity applied on the column, and this procedure increased the specific activity by about 3-fold. SDS polyacrylamide gel electrophoresis The purified thrombomodulin migrated as a single band on SDS-gel electrophoresis, both before (Mr=71,000) and after (Mr=88,000) disulfide bond reduction (Fig. 4). Isoelectric focusing One hundred ~1 of purified thrombomodulin (30,000 u/ml) was electrofocused in the llO-ml column, and 1 ml fractions were collected. The pH and thrombomodulin activity of
358
HUMAN THROMBOMODULIN
0.1
Vo1.37, No.3
r
I
s N
0.0s
4
0
FRACTIONS
FIG. 3 Rechromatography of partially purified thrombomodulin on DIPthrombin-agarose. The concentrated effluent from the gel filtration containing thrombomodulin (Fig. 2) was adjusted to a CaC12 concentrationof 0.5 mM andwas appliedonthe column (2.6 x 16 cm) equilibrated with 0.1 M NaCl-Lubrol buffer. After sample application, the columnwaswashedwith 30 mlof 0.4 M NaCl-Lubrol buffer and then with 30 ml of 0.4 M NaCl-EDTA-Lubrol buffer. The column was developed with a linear gradient, from 0.4 M to 1 M NaCl, containing EDTA-Lubrol buffer. At the completion of the gradient, additional 1 M NaCl-EDTA-Lubrol buffer was applied. Fractions of 2.35 ml were collected. The eluate was pooled as indicatedbythebar, and it was concentrated by ultrafiltration.
each fractions were determined. Thrombomodulin activity was found in the fractions of pH 3.6-4.3, with the maximum at pH 4.2. Stability Purified thrombomodulin (0.073 A2s0/ml) was tested for Stability under various conditions. Purified thrombomodulin in 0.1 M Nacl, 0.02 M glycineacetate, pH 2.5, was incubated with pepsin (5 pg/ml) at 37'C for
TABLE Stability
1
of human thrombomodulin
Conditions Pepsin B-Mercaptoethanol SDS, 1% Urea, 8M Guanidinium Cl, 6M PH 2 pH 10 100°C
8 activity 43
25 80-90 85-90 70-80 80-90 85-90 t1/2 min, 20
vo1.37, No.3
- R nR NRNR
HUMAN THROMBOMODULIN
359
FIG.
4.
SDS-polyacrylamide slab gel electrophoresis of purified thrombomodulin. Purified thrombomodulin (0.011 A2s0) was mixed and heated at 1OO'C for 3 min with an equal volume of 2% SDSwith or without reducing agent (2% dithiothreitol), and a 20 pl-aliquotofthetreated sample was subjected to the electroThe acrylamide concentration phoresis. Proteins in the separating gel was 10%. were visualized by silver staining. samples. R: NR and nR: non-reduced One the non-reduced reduced sample. samples (nR) migrated as an oblique band under the influence of the reducing agent diffusing out from an adjacent reduced sample during the electrophoresis.
FIG.
5
Activation of bovine protein C by thrombinthrombomodulin. Bovine protein C (50 pg/ml) was incubated at 37'C with purified bovine thrombin c (2 u/ml) and purified > 300thrombomodulin (10 times dilution of 0.0197 A280 in the activation mixture) in 0.1 M NaCl, 0.02 M Tris-HCl, 3.5 mM CaC12, pH 7.5. An aliquot of loo-p1 was withdrawn at intervals 0 6 10 20 30 and was added to 150 ~1 of antithrombin III (300 Activation time IminI Ccgjml) in 0.1 M NaCl, 0.05 M Tris-HCl, 1 mM CaC12, pH 8.5. The mixture was incubated at 37°C for 15 min. Subsequently, 250 ~1 of MCA-substrate (200 PM) in the same buffer, pH 8.5, was added, and the initial rate of the release of AMC was measured as described under Materials and Methods. z
soo-
120 min. Samples were removed and diluted 100 times with 0.1 M NaCl, 0.02 M Tris-HCl, pH 7.5, was incubated with various reagents at 20°C for 150 min. Purified thrombomodulin in the same buffered saline was heated in a boiling water bath for various lengths of time. Purified thrombomodulin was also left
360
HUMAN THROMBOMODULIN
vo1.37, No.3
at pH 2 or pH 10 at 20°C for 150 min. The activities remained after these treatments were assayed and expressed as % of the original activity. The results are shown in Table 1. ProteinC activation ProteinC was incubated at 37OC with the mixture of thrsrn-nd purified thrombomodulin. At various time intervals, an aliquot was withdrawn and mixed with antithrombin III to neutralize thrombin. The activated protein C was then measured for the amidase activity on MCA-substrate as described in Materials and Methods. As seen in Fig. 5, activation of bovine protein C progressed linearly on incubation. 'When thrombomodulin was omitted from the activation mixture, the
FIG. 6 500
T ;i k
400
g
300
.r’ .g 200 2 100
0
0
10
6
Thrombomodulin
lpll
FIG. 7
600 = <
400
2 z 2
300
Linear relationship between the extent of bovine protein C activation and the amount of thrombomodulin. In the experiments in Fig. 5, the amount of thrombomodulin in the 100 ~1 activation mixture was varied, and the activity of activated protein C developed after plotted against the amount of thrombomodulin used. The activity that releases one nmole of AMC/min.ml was defined as one unit.
.z 2 200 **
100
0 0
10
6
Thrombomodulin
IpI)
Activation of human protein C by human thrombin-thrombomodulin. Experimental conditions such as the concentrations of protein C and thrombin were all the same as in Fig. 5, except that the all materials were of human origin. The activity is expressed as in Fig. 6.
Vo1.37, No.3
HUMAN THROMBOMODULIN
36'
activation of protein C was very slow and was negligible in the experimental conditions used. When the amount of thrombomodulin was varied, the activated protein C activity developed was linearly proportional to the amount of thrombomodulin used (Fig. 6). Thrombomodulin alone without thrombin did not activate protein C. The same experiments were performed using human The protein C and human thrombin instead of bovine materials. results corresponding to Fig. 6 are presented in Fig. 7. Anticoagulant activity TABLE 2 Influence of the presence of thromboAnticoagulant effect of human modulin on the activated thrombomodulin on partial partialthromboplasti thromboplastin time of time of plasma was human plasma examined. Various amounts (O-10 ~1) of purified thrombomodulin Clotting time (set)* Sample (pl) (0.055 Azbo/ml) were Test Control added to 100 ~1 of plasma, and the acti26.1 0 vated partial thrombo30.4 68.4 plastin time test was 2.5 32.9 103.9 performed. The results 5.0 37.3 >200 are shown in Table 2. 10.0 The clotting time was prolonged by the presence of thrombomodulin, and the prolongation was dependent on the amount of thrombomodulin added. The control experiments performed with buffered solution used for dissolving thrombomodulin showed only slight prolongation of the clotting time.
DISCUSSION Thrombomodulin was isolated from human placentae by a method which is essentially the same as the one described by Esmon et al. (5) for rabbit lungs, except for the use of gel filtration between the two steps of affinity chromatography. The overall yield was quite variable and was 2-90% of the original Triton X-100 extract of the placental homogenate, and the increase of the specific activity in terms of activity per unit of protein was roughly 200 times. The latter estimate, however, may be subject to errors since the protein absorbance was so low that exact measurement was difficult and might have been influenced by the possible presence of a trace amount of benzamidine which was used during the chromatography. The molecular weight of the purified thrombomodulin estimated by SDS-gelelectrophoresis was 71,000 or 88,000 before or after disulfide bond reduction, respectively. These values are higher than those estimated for rabbit thrombomodulin, 68,000 and 74,000, respectively (5). Esmon et al. (5) considered that the molecular weight estimate before disulfide bond reduction may not represent the true value since SDS may not completely
362
HUMAN THROMBOHODULIN
Vo1.37, No.3
bind to the molecule without disulfide reduction, and the best fit molecular weight based on the amino acid composition (Mr 73,600) was very close to that (Mr 74,000) obtained by SDS-gel electrophoresis after disulfide bond reduction. Therefore, a better estimate of the molecular weight of human thrombomodulin may be 88,000, which is significantly higher than the 74,000 molecular weight of rabbit thrombomodulin. The thrombomodulin is considered to be an acidic protein since the isoelectric point of the purified molecule was estimated to be around pH 4. The purified thrombomodulin was exceptionally stable to heat, pH extremes and protein denaturants, but was inactivated by disulfide bond reduction or by pepsin treatment (Table 1). These properties are very similar to those of rabbit thrombomodulin (5). ProteinCofhumanaswellasofbovineoriginswas readily activated in the presence of calcium ions by thrombin and the thrombomodulin. The activation of protein C by thrombin in the absence of thrombomodulin was very slow, and an addition of the thrombomodulin accelerated the activation. The rate of activation depends on the amount of the thrombomodulin added (Figs. 6 and 7). Thus, the thrombomodulin isolated from human placentae plays a role as a cofactor of thrombin-catalyzed protein C activationinthehumanaswellasinthebovine enzyme system. When bovine thrombin was used as a catalyst, the activation rate was directly proportional to the amount of the thrombomodulin When human added up to 500 units/ml of protein C activation. thrombin was used as a catalyst, the rate of activation increased parabolically to the amount of the thrombomodulin added (Fig. 7). The difference may be accounted for by the difference of Kd of thrombin-thrombomodulin complex formation, since thrombin-thrombomodulin complex rather than thrombin itself is a responsive enzyme for protein C activation and thrombin preparations used in both systems have a similar degree of purity (homogeneous in SDS-polyacrylamide gel electrophoresis and =G 1500 u/mg intermsof clottingactivityper unit of protein). Thrombin-thrombomodulin complex formation is a reversible equilibrium reaction, therefore the rate of protein C activation plotted against the amount of thrombomodulin should theoretically give a parabolic curve. However, bovine thrombinhuman thrombomodulin complex formation may have a smaller Kd, giving a seemingly linear relationship between protein C activation and the amount of thrombomodulin in a wide range. An additional important property of thrombomodulin is to neutralize thrombin activity by binding to thrombin. The purified thrombomodulin was tested for this anticoagulant activity in the partial thromboplastin time test of plasma. An addition of the purified thrombomodulin readily prolonged the clotting time (Table 2). Although involvement of activated protein C in prolongation of the partial thromboplastin time can not be ruled out, theprolongationof the clotting timemaybemainlydue to neutralization of thrombin evolved since the amount of protein C activatedin sucha short time as 30 seconds would not be enough to inactivate factors V and VIII to such a degree as prolongs the partial thromboplastin time. These properties of the thrombomodulin isolated from human
vo1.37, No.3
363
HUMAN THROMBOMODULIN
placentae are all in common with those revealed by the study on rabbit thrombomodulin (5). Differences may exist in some physicochemical properties such as molecular weight as presented These in the present study and in its immunological properties. are subjects for further study. Addendum : After submission of this paper, a paper dealing with the same subject has appeared (H.H. Salem et al. J. Biol. Chem. 259:12246, 1984).
REFERENCES 1. STENFLO, J. Structure and function of protein C. Semin. Thromb. Hemostas. 10, 109-121, 1984. 2. GRIFFIN, J.H. Clinical studies of protein C. Semin. Thromb. Hemostas. 10, 162-166, 1984. 3. ESMON, C.T. and ESMON, N.L. Protein C activation. Thromb. Hemostas. 10, 122-130, 1984.
Semin.
4. OWEN, W.G. and ESMON, C.T. Functional properties of an endothelial cell cofactor for thrombin-catalyzed activation of protein C.J. Biol. Chem. 256, 5532-5535, 1981. 5. ESMON, N.L., OWEN, W.G. and ESMON, C.T. Isolation of a membrane-bound cofactor for thrombin-catalyzed activation protein C. J. Biol. Chem. 257, 859-864, 1982.
of
6. LUNDBLAD, R.L., UHTEG, L.C., VOGEL, C.N., KINGDON, H.S. and MANN, K.G. Preparation and partial characterization of two fonns of bovine thrombin. Biochem. Biophys. Res. Commun. 66 482-489, 1975. _' 7. CUATRECASAS, P. Protein purification by affinity chromatography. J. Biol. Chem. 245, 3059-3065, 1970. 8. STENFLO, J. A new vitamin Chem. 251, 355-363, 1976.
K-dependent
protein. J. Biol.
9. SUZUKI, K., STENFLO, J., DAHLBACK, 8. and TEODORSSON, B. Inactivation of human coagulation factor V by activated protein C. J. Biol. Chem. 258, 1914-1920, 1983. 10. MILLER-ANDERSSON, M., BORG, H., and ANDERSSON, L-O. Purification of antithrombin III by affinity chromatography. Thromb. Res. 5, 439-452, 1974. 11. OHNO, Y., KATO, H., MORITA, T., IWANAGA, S., TAKADA, K., SAKAKIBARA, S. and STENFLO, J. A new fluorogenic peptide substrate for vitamin K-dependent blood coagulation factor, bovine protein C. J. Biochem. 90, 1387-1395, 1981. 12. LAEMMLI, U.K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680685, 1970.
364
HUMAN THROMBOMODULIN
Vo1.37, No.3
13. MORRISSEY, J.H. Silver stain for proteins in polyacrylamide gels. A modified procedure with enhanced uniform sensitivity. Anal. Biochem. 117, 307-310, 1981. 14. VESTERBERG, 0. and SVENSSON, H. Isoelectric fractionation, analysis, and characterization of ampholytes in natural pH gradients. IV. Further studies on the resolving power in connection with separation of myoglobins. ACta Chem. Stand. 20, 820-834, 1966.
PREPARATION
OF THROMBOMODULIN
Shinichiro Institute
Kurosawa
of Hematology
Jichi Medical
FROM HUMAN PLACENTA
and Nobuo
Aoki
and Department of Medicine,
School, Tochigi-Ken
329-04, Japan
(Received 11.10.1984; Accepted in revised form 30.11.1984 by Editor S. Iwanaga)
ABSTRACT Thrombomodulin, a cell surface cofactor for thrombincatalyzed activation of protein C, was isolated from human placenta by a combination of affinity chromatography and gel filtration. Molecular weight of purified human thrombomodulin, estimated by sodium dodecyl sulfate electrophoresis after reduction, was 88,000. Its isoelectric point was around pH 4. The thrombomodulin was soluble only in the presence of detergent. The purified thrombomodulin was inactivated by disulfide bond reduction, but was stable to heat, pH The thrombomodulin extremes and protein denaturants. served as a cofactor of thrombin-catalyzed protein C activation in human as well as bovine enzyme systems.
INTRODUCTION Protein C plays an important role in regulating the blood coagulation process in vivo (l), and its congenital deficiency results in thrombophilic states (2). Protein C must be activated to exert its physiological role as an anticoagulant, and its activation is catalyzed by thrombin in the presence of a cofactor called thrombomodulin (3). Thrombomodulin is presumably present on the vascular endothelium as a membrane-bound protein (4). This cofactor has been isolated from rabbit lungs, and its effect on thrombin-catalyzed activation of protein C has
contents were presented at an annual meeting of Japan Society of Hematology on April 12, 1984, in Kyoto, Japan. Key words: Thrombomodulin, Protein C, Placenta, MCA-substrate The
353
354
HUMAN THROMBOMODULIN
Vo1.37, No.3
been extensively studied using materials of bovine origin (5). However, preparation of human thrombomodulin and its effect on the activation system of protein C in humans have not been reported. In this communication, we present some results of our study on isolation of thrombomodulin from human placentae, its characterization and its effect on activation of human protein C.
MATERIALS
AND METHODS
Extract of human placenta In a few hours after delivery of the placenta, it was perfused via the umbilical artery with buffer A (0.02 M Tris-HCl, 0.25 M sucrose, 1 mM benzamidine-HCl, pH The placenta was then minced and homogenized with 7.5) at 4'C. buffer A by the use of a Waring blender. The homogenate was centrifuged at 30,000 g for 40 min. The pellet was suspended in buffer A and recentrifuged. The same procedure was repeated for atotalof 6 times. All procedures were carried outat4'C or on ice. Finally, thrombomodulin was extracted from the pellet by homogenization and centrifugation as described above with 150 ml of bufferA containing 0.5% (v/v) Triton X-100 (Sigma Chemical co. U.S.A.). Thrombin Thrombin was purified from a commercial bovine (Mochida Pharm. Co., Tokyo) or human (Midori-Juji Corp., Osaka) thrombin preparation by the method of Lundblad (6). DIP-thrombin-agarose Sepharose-4B (Pharmacia, Sweden) was activated with CNBr bv the method of Cuatrecasas (7). . and thrombin was linked to the agarose to give a final thrombin concentration The thrombin was then around 1 mg of thrombin/ml of agarose. inactivated by incubation with 1 mM diisopropylphosphofluoridate (Sigma) in 0.1 M NaCl, 0.02 M Tris-HCl, pH 7.5. Protein C Bovine andhumanproteinc was purifiedbythe method of Stenflo (8) and the method of Suzuki et al. (9), respectively. Antithrombin III Antithrombin III was purified from human plasma as described by Miller-Anderson et al. (10). MCA-substrate for activated protein C Boc-Leu-Ser-Thr-Arg-MCA (4-methyl-coumaryl-7-amide) (11) was purchased from Protein Research Foundation, Osaka, Japan. Agarose beads for gel filtration from LKB (Sweden).
Ultro Gel ACA 34 was purchased
Assay of thromboxnodulin Ten ~1 of bovine protein C (500 pg/ml) was mixed with 5 plof bovine thrombin (40 u/ml) and 10 ).11of the sample to be assayed, and the total volume was adjusted to 100 ~1 with 0.02 M TriS-HCl, 0.1 M NaCl, 3.5 mM CaC12, pH 7.5. The mixture was incubated at 37'C for 30 min. After the incubation, 150 ~1 of antithrombin III solution (300 pg/ml) in buffer B (0.05 M Tris-HCl, 0.1 M NaCl, 1 mM CaC12, pH 8.5) was added to the mixture and incubated further for 15 min to neutralize
Vo1.37, No.3
355
HUMAN THROMBOMODULIN
thrombin. Subsequently, 250 ~1 of MCA-substrate solution (100 PM) in buffer B was added to the mixture and incubated at 37'C for 10 min. The reaction was immediately stopped by adding 500 ~1 of 20% (v/v) acetic acid. AMC (7-amino-4-methylcoumarine) liberated was then measured by a spectrofluorophotometer (Shimadzu RF-510, Kyoto, Japan) with excitation at 380 nm and emission at 460 nm, using a standard curve constructed with a standard AMC (Protein Research Foundation, Osaka, Japan). The activity of active protein C generated was linearly related to the amount of thrombomodulin used (see Results), and the amount of thrombomodulin that ultimately liberated one nmole of AMC/min.ml in the assay system was defined as one unit of thrombomodulin. Buffers used for chromatography 0.1 M NaCl, 0.02 M Tris-HCl, 0.5 mM CaC12, 1 mM benzamidine-HCl, 0.5% Lubrol PX, pH 7.5 (0.1 M NaCl-Lubrol buffer): 0.2 M NaCl with other constituents being the same as above (0.2 M NaCl-Lubrol buffer); 0.4 M NaCl with other constituents being the same as above (0.4 M NaCl-Lubrol buffer); 0.4 M NaCl-Lubrol buffer with 1 mM EDTA replacing CaC12 (0.4 M NaCl-EDTA-Lubrol buffer); and 1 M NaCl with other constituentsbeingthe same as 0.4 M NaCl-EDTA-Lubrolbuffer (1 M NaCl-EDTA-Lubrol buffer). Polyacrylamide gel electrophoresis The electrophoresis was carried out according to the sodium dodecyl sulfate (SDS) system of Laemmli (12), using a 10% polyacrylamide slab gel with and without dithiothreitol. After electrophoresis, the gels were stained by the silver-staining procedure described by Morrissey (13). For estimation of molecular weight, a mixture of molecular weight protein standards (Bio-Rad Lab., Calif.) was run on the same gel. Isoelectric focusing Isoelectric focusing was carried out according to methods described by Versterberg (14), in an LKB jacketed column (llO-ml capacity) using 4% ampholytes, pH range 3-10, at 300 v for 70 h at 2OC. Activated partial thromboplastin time Platelet-poor plasma was obtained by centrifugation of blood anticoagulated with 0.1 vol of 3.8% sodium citrate. One hundred ~1 of plasma was mixed at 37'C with 100 ~1 of activated cephaloplastin reagent (Actin, Dade Diagnostic, Inc.). The clotting reaction was started by adding 100 1.11 of 0.025 M calcium chloride at 37'C, and the time required for clot formation after the addition of calcium was recorded. Other reagents Pepsin was purchased from Worthington. serum albumin and Lubrol PX were purchased from Sigma. other chemicals used are all analytical grades.
Bovine The
RESULTS Chromatography of the placental extract on DIP-thrombin-agarose. The conditions used for the chromatography were essentially the
356
HUMAN THROMBOMOOULIW
Vo1.37, No.3
same as those described by Esmon et al. (5). Calcium chloride was added to the extract to a concentration of 0.5 mM before chromatography. Typical results are shown in Fig. 1. The majority of contaminating proteins in the extract did not bind to the agarose equilibriated with 0.1-0.2 M NaCl. Thrombomodulinboundtothe agarose was elutedwith 1 M NaCl. The yield was variable and ranged from 25 to 90% of the activity applied on the column. This chromatography increased the specific activity by 5-lo-fold.
n
FRACTIONS
FIG. 1 Chromatography of the placental extract on DIP-thrombin-agarose. A DIP-thrombin-agarose column (2.6 x 16.5 cm) was equilibrated with O.lM NaCl-Lubrolbuffer, and then 200 ml of the extract After sample (1.044 Aztju, 250 u/ml) was applied on the column. application, the columnwas washed with 2 liters of 0.2 M NaClLubrol buffer over a period of 72 hours, and then thrombomodulin was eluted with 1 M NaCl-EDTA-Lubrol buffer: 3.5-ml fractions were collected. The eluate was pooled as indicated by the bar, and the pooled eluate was concentrated by ultrafiltration.
Gel filtration The eluate from the DIP-thrombin agarose was pooled and concentrated approximately 20 times with a Millipore CX-30 ultrafiltration unit fitted to a CX agitator (Japan Millipore, Ltd., Tokyo). The concentrated eluate was then applied on the column of Ultrogel ACA 34 for gel filtration. Typical results are shown in Fig. 2. There were usually two major protein peaks and an additional two minor peaks. A large single peak of thrombomodulin activity was found between the two major protein peaks. The recovery of thrombomodulin was 50-908 of the activity applied on the column, and the gel filtration increased the specific activity by 7-15-fold.
vo1.37, No.3
HUMAN THROMBOMODULIN
357
FRACTIONS
FIG. 2 Gel filtration of partially purified thrombomodulin on Ultrogel ACA 34. The concentrated eluate (5 ml) from the first DIPthrombin-agarose column (1.214 A280, 1448 u/ml) was applied on the column (2.5 x 89 cm) equilibrated with 0.1 M NaCl-Lubrol buffer. The gel filtration was carried out with the same buffer with a flow rate at 25 ml/hour, and the effluent was collected in 2.46-ml fractions. The effluent was pooled as indicated by the bar, and it was concentrated by ultrafiltration.
Rechromatography on DIP-thrombin-agarose The effluent containing thrombomodulin from the gel filtration was pooled and concentrated approximately 10 times with a Millipore CX-30. The concentrate was added with calcium chloride to obtain a calcium concentration of 0.5 mM, and then it was applied on the column of DIP-thrombin-agarose and eluted with a linear gradient. Typical results are shown in Fig. 3. Thrombomodulin was eluted with approximately 0.8 M NaCl. The eluate possessing maximal thrombomodulin activity was dialyzed against 0.02 M Tris-HCl, 0.1 M NaCl, 0.5% Lubrol PX, pH 7.5, and concentrated by means of The recovery of thrombomodulin Millipore CX-30 ultrafiltration. was around 95% of the activity applied on the column, and this procedure increased the specific activity by about 3-fold. SDS polyacrylamide gel electrophoresis The purified thrombomodulin migrated as a single band on SDS-gel electrophoresis, both before (Mr=71,000) and after (Mr=88,000) disulfide bond reduction (Fig. 4). Isoelectric focusing One hundred ~1 of purified thrombomodulin (30,000 u/ml) was electrofocused in the llO-ml column, and 1 ml fractions were collected. The pH and thrombomodulin activity of
358
HUMAN THROMBOMODULIN
0.1
Vo1.37, No.3
r
I
s N
0.0s
4
0
FRACTIONS
FIG. 3 Rechromatography of partially purified thrombomodulin on DIPthrombin-agarose. The concentrated effluent from the gel filtration containing thrombomodulin (Fig. 2) was adjusted to a CaC12 concentrationof 0.5 mM andwas appliedonthe column (2.6 x 16 cm) equilibrated with 0.1 M NaCl-Lubrol buffer. After sample application, the columnwaswashedwith 30 mlof 0.4 M NaCl-Lubrol buffer and then with 30 ml of 0.4 M NaCl-EDTA-Lubrol buffer. The column was developed with a linear gradient, from 0.4 M to 1 M NaCl, containing EDTA-Lubrol buffer. At the completion of the gradient, additional 1 M NaCl-EDTA-Lubrol buffer was applied. Fractions of 2.35 ml were collected. The eluate was pooled as indicatedbythebar, and it was concentrated by ultrafiltration.
each fractions were determined. Thrombomodulin activity was found in the fractions of pH 3.6-4.3, with the maximum at pH 4.2. Stability Purified thrombomodulin (0.073 A2s0/ml) was tested for Stability under various conditions. Purified thrombomodulin in 0.1 M Nacl, 0.02 M glycineacetate, pH 2.5, was incubated with pepsin (5 pg/ml) at 37'C for
TABLE Stability
1
of human thrombomodulin
Conditions Pepsin B-Mercaptoethanol SDS, 1% Urea, 8M Guanidinium Cl, 6M PH 2 pH 10 100°C
8 activity 43
25 80-90 85-90 70-80 80-90 85-90 t1/2 min, 20
vo1.37, No.3
- R nR NRNR
HUMAN THROMBOMODULIN
359
FIG.
4.
SDS-polyacrylamide slab gel electrophoresis of purified thrombomodulin. Purified thrombomodulin (0.011 A2s0) was mixed and heated at 1OO'C for 3 min with an equal volume of 2% SDSwith or without reducing agent (2% dithiothreitol), and a 20 pl-aliquotofthetreated sample was subjected to the electroThe acrylamide concentration phoresis. Proteins in the separating gel was 10%. were visualized by silver staining. samples. R: NR and nR: non-reduced One the non-reduced reduced sample. samples (nR) migrated as an oblique band under the influence of the reducing agent diffusing out from an adjacent reduced sample during the electrophoresis.
FIG.
5
Activation of bovine protein C by thrombinthrombomodulin. Bovine protein C (50 pg/ml) was incubated at 37'C with purified bovine thrombin c (2 u/ml) and purified > 300thrombomodulin (10 times dilution of 0.0197 A280 in the activation mixture) in 0.1 M NaCl, 0.02 M Tris-HCl, 3.5 mM CaC12, pH 7.5. An aliquot of loo-p1 was withdrawn at intervals 0 6 10 20 30 and was added to 150 ~1 of antithrombin III (300 Activation time IminI Ccgjml) in 0.1 M NaCl, 0.05 M Tris-HCl, 1 mM CaC12, pH 8.5. The mixture was incubated at 37°C for 15 min. Subsequently, 250 ~1 of MCA-substrate (200 PM) in the same buffer, pH 8.5, was added, and the initial rate of the release of AMC was measured as described under Materials and Methods. z
soo-
120 min. Samples were removed and diluted 100 times with 0.1 M NaCl, 0.02 M Tris-HCl, pH 7.5, was incubated with various reagents at 20°C for 150 min. Purified thrombomodulin in the same buffered saline was heated in a boiling water bath for various lengths of time. Purified thrombomodulin was also left
360
HUMAN THROMBOMODULIN
vo1.37, No.3
at pH 2 or pH 10 at 20°C for 150 min. The activities remained after these treatments were assayed and expressed as % of the original activity. The results are shown in Table 1. ProteinC activation ProteinC was incubated at 37OC with the mixture of thrsrn-nd purified thrombomodulin. At various time intervals, an aliquot was withdrawn and mixed with antithrombin III to neutralize thrombin. The activated protein C was then measured for the amidase activity on MCA-substrate as described in Materials and Methods. As seen in Fig. 5, activation of bovine protein C progressed linearly on incubation. 'When thrombomodulin was omitted from the activation mixture, the
FIG. 6 500
T ;i k
400
g
300
.r’ .g 200 2 100
0
0
10
6
Thrombomodulin
lpll
FIG. 7
600 = <
400
2 z 2
300
Linear relationship between the extent of bovine protein C activation and the amount of thrombomodulin. In the experiments in Fig. 5, the amount of thrombomodulin in the 100 ~1 activation mixture was varied, and the activity of activated protein C developed after plotted against the amount of thrombomodulin used. The activity that releases one nmole of AMC/min.ml was defined as one unit.
.z 2 200 **
100
0 0
10
6
Thrombomodulin
IpI)
Activation of human protein C by human thrombin-thrombomodulin. Experimental conditions such as the concentrations of protein C and thrombin were all the same as in Fig. 5, except that the all materials were of human origin. The activity is expressed as in Fig. 6.
Vo1.37, No.3
HUMAN THROMBOMODULIN
36'
activation of protein C was very slow and was negligible in the experimental conditions used. When the amount of thrombomodulin was varied, the activated protein C activity developed was linearly proportional to the amount of thrombomodulin used (Fig. 6). Thrombomodulin alone without thrombin did not activate protein C. The same experiments were performed using human The protein C and human thrombin instead of bovine materials. results corresponding to Fig. 6 are presented in Fig. 7. Anticoagulant activity TABLE 2 Influence of the presence of thromboAnticoagulant effect of human modulin on the activated thrombomodulin on partial partialthromboplasti thromboplastin time of time of plasma was human plasma examined. Various amounts (O-10 ~1) of purified thrombomodulin Clotting time (set)* Sample (pl) (0.055 Azbo/ml) were Test Control added to 100 ~1 of plasma, and the acti26.1 0 vated partial thrombo30.4 68.4 plastin time test was 2.5 32.9 103.9 performed. The results 5.0 37.3 >200 are shown in Table 2. 10.0 The clotting time was prolonged by the presence of thrombomodulin, and the prolongation was dependent on the amount of thrombomodulin added. The control experiments performed with buffered solution used for dissolving thrombomodulin showed only slight prolongation of the clotting time.
DISCUSSION Thrombomodulin was isolated from human placentae by a method which is essentially the same as the one described by Esmon et al. (5) for rabbit lungs, except for the use of gel filtration between the two steps of affinity chromatography. The overall yield was quite variable and was 2-90% of the original Triton X-100 extract of the placental homogenate, and the increase of the specific activity in terms of activity per unit of protein was roughly 200 times. The latter estimate, however, may be subject to errors since the protein absorbance was so low that exact measurement was difficult and might have been influenced by the possible presence of a trace amount of benzamidine which was used during the chromatography. The molecular weight of the purified thrombomodulin estimated by SDS-gelelectrophoresis was 71,000 or 88,000 before or after disulfide bond reduction, respectively. These values are higher than those estimated for rabbit thrombomodulin, 68,000 and 74,000, respectively (5). Esmon et al. (5) considered that the molecular weight estimate before disulfide bond reduction may not represent the true value since SDS may not completely
362
HUMAN THROMBOHODULIN
Vo1.37, No.3
bind to the molecule without disulfide reduction, and the best fit molecular weight based on the amino acid composition (Mr 73,600) was very close to that (Mr 74,000) obtained by SDS-gel electrophoresis after disulfide bond reduction. Therefore, a better estimate of the molecular weight of human thrombomodulin may be 88,000, which is significantly higher than the 74,000 molecular weight of rabbit thrombomodulin. The thrombomodulin is considered to be an acidic protein since the isoelectric point of the purified molecule was estimated to be around pH 4. The purified thrombomodulin was exceptionally stable to heat, pH extremes and protein denaturants, but was inactivated by disulfide bond reduction or by pepsin treatment (Table 1). These properties are very similar to those of rabbit thrombomodulin (5). ProteinCofhumanaswellasofbovineoriginswas readily activated in the presence of calcium ions by thrombin and the thrombomodulin. The activation of protein C by thrombin in the absence of thrombomodulin was very slow, and an addition of the thrombomodulin accelerated the activation. The rate of activation depends on the amount of the thrombomodulin added (Figs. 6 and 7). Thus, the thrombomodulin isolated from human placentae plays a role as a cofactor of thrombin-catalyzed protein C activationinthehumanaswellasinthebovine enzyme system. When bovine thrombin was used as a catalyst, the activation rate was directly proportional to the amount of the thrombomodulin When human added up to 500 units/ml of protein C activation. thrombin was used as a catalyst, the rate of activation increased parabolically to the amount of the thrombomodulin added (Fig. 7). The difference may be accounted for by the difference of Kd of thrombin-thrombomodulin complex formation, since thrombin-thrombomodulin complex rather than thrombin itself is a responsive enzyme for protein C activation and thrombin preparations used in both systems have a similar degree of purity (homogeneous in SDS-polyacrylamide gel electrophoresis and =G 1500 u/mg intermsof clottingactivityper unit of protein). Thrombin-thrombomodulin complex formation is a reversible equilibrium reaction, therefore the rate of protein C activation plotted against the amount of thrombomodulin should theoretically give a parabolic curve. However, bovine thrombinhuman thrombomodulin complex formation may have a smaller Kd, giving a seemingly linear relationship between protein C activation and the amount of thrombomodulin in a wide range. An additional important property of thrombomodulin is to neutralize thrombin activity by binding to thrombin. The purified thrombomodulin was tested for this anticoagulant activity in the partial thromboplastin time test of plasma. An addition of the purified thrombomodulin readily prolonged the clotting time (Table 2). Although involvement of activated protein C in prolongation of the partial thromboplastin time can not be ruled out, theprolongationof the clotting timemaybemainlydue to neutralization of thrombin evolved since the amount of protein C activatedin sucha short time as 30 seconds would not be enough to inactivate factors V and VIII to such a degree as prolongs the partial thromboplastin time. These properties of the thrombomodulin isolated from human
vo1.37, No.3
363
HUMAN THROMBOMODULIN
placentae are all in common with those revealed by the study on rabbit thrombomodulin (5). Differences may exist in some physicochemical properties such as molecular weight as presented These in the present study and in its immunological properties. are subjects for further study. Addendum : After submission of this paper, a paper dealing with the same subject has appeared (H.H. Salem et al. J. Biol. Chem. 259:12246, 1984).
REFERENCES 1. STENFLO, J. Structure and function of protein C. Semin. Thromb. Hemostas. 10, 109-121, 1984. 2. GRIFFIN, J.H. Clinical studies of protein C. Semin. Thromb. Hemostas. 10, 162-166, 1984. 3. ESMON, C.T. and ESMON, N.L. Protein C activation. Thromb. Hemostas. 10, 122-130, 1984.
Semin.
4. OWEN, W.G. and ESMON, C.T. Functional properties of an endothelial cell cofactor for thrombin-catalyzed activation of protein C.J. Biol. Chem. 256, 5532-5535, 1981. 5. ESMON, N.L., OWEN, W.G. and ESMON, C.T. Isolation of a membrane-bound cofactor for thrombin-catalyzed activation protein C. J. Biol. Chem. 257, 859-864, 1982.
of
6. LUNDBLAD, R.L., UHTEG, L.C., VOGEL, C.N., KINGDON, H.S. and MANN, K.G. Preparation and partial characterization of two fonns of bovine thrombin. Biochem. Biophys. Res. Commun. 66 482-489, 1975. _' 7. CUATRECASAS, P. Protein purification by affinity chromatography. J. Biol. Chem. 245, 3059-3065, 1970. 8. STENFLO, J. A new vitamin Chem. 251, 355-363, 1976.
K-dependent
protein. J. Biol.
9. SUZUKI, K., STENFLO, J., DAHLBACK, 8. and TEODORSSON, B. Inactivation of human coagulation factor V by activated protein C. J. Biol. Chem. 258, 1914-1920, 1983. 10. MILLER-ANDERSSON, M., BORG, H., and ANDERSSON, L-O. Purification of antithrombin III by affinity chromatography. Thromb. Res. 5, 439-452, 1974. 11. OHNO, Y., KATO, H., MORITA, T., IWANAGA, S., TAKADA, K., SAKAKIBARA, S. and STENFLO, J. A new fluorogenic peptide substrate for vitamin K-dependent blood coagulation factor, bovine protein C. J. Biochem. 90, 1387-1395, 1981. 12. LAEMMLI, U.K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680685, 1970.
364
HUMAN THROMBOMODULIN
Vo1.37, No.3
13. MORRISSEY, J.H. Silver stain for proteins in polyacrylamide gels. A modified procedure with enhanced uniform sensitivity. Anal. Biochem. 117, 307-310, 1981. 14. VESTERBERG, 0. and SVENSSON, H. Isoelectric fractionation, analysis, and characterization of ampholytes in natural pH gradients. IV. Further studies on the resolving power in connection with separation of myoglobins. ACta Chem. Stand. 20, 820-834, 1966.