Purification and partial characterization of A D-like fragment from human fibrinogen, produced by human leukocyte elastase

300

Bto(htmtca et Bt,,pht ~t,a t~ta. 755(lq,~3p3()(I 31)¢~ kl,,cvler Biomcdk_'al Pro,,,

BBA21359

PURIFICATION AND PARTIAL CHARACTERIZATION OF A D-LIKE FRAGMENT FROM HUMAN FIBRINOGEN, PRODUCED BY HUMAN LEUKOCYTE ELASTASE LYDI S T E R R E N B E R G a . , W I L L E M N I E U W E N H U I Z E N ~ and JO H E R M A N S h " Gaubtus Instttute, Health Research Dwlston TARO, Herenstraat 5d, 2313 AD Letden and /' Department ~f Medtcal Stattstte~. State ~hut~erstty of Letden, Wassenaarseweg 80. 2333 A L Letden (The Netherlands) (Received November 1 lth, 1982)

Key words" Fibrmogen fragment, Fragment D, Elastase dlgestton

Digestion of human fibrinogen with human leukocyte elastase in the presence of Ca z+ yields a D-like fragment of M r 93000. This fragment was purified by gel filtration on Sephacryl S-200 followed by chromatofocusing. The purified fragment was partially characterized and compared with a fragment termed D-cate, which is produced by plasmin digestion of fibrinogen in the presence of Ca 2+ . The molecular weights of the constituent chains of the D-like fragment and D-care were similar. The D-like fragment precipitated with antisera directed against D-cate, but not with antisera against fragment E. The name D-elastase for the fragment is suggested. Differences between the D-elastase and D-care fragments were found in amino-terminal amino acids, in isoelectric point and in the expression of D antigenic determinants. Two major functional differences were demonstrated: fragment D-elastase had a much stronger anticlotting potency than D-cate and the binding of Ca 2+ by D-elastase and D-care differed qualitatively and quantitatively. Since it has been suggested that the calcium-binding and anticlotting properties of D-cate are related to a carboxyl-terminal 13000 stretch of the ,/-chain, the present findings for D-elastase indicate that the differences in these properties between D-care and D-elastase are due to differences in this area of the molecule.

Introduction

In the symmetric fibrinogen molecule three core fragments are present, one called fragment E and two called fragment D **. These core fragments are separated but remain intact at late stages of fibrinogen digestion by plasmin. They contain dif* Present address. Department of Pharmacology, Faculty of Pharmacy, Umversity of Utrecht, Catharijnesmgel 60, Utrecht. The Netherlands. ** For nomenclature, see Ref. 1. Abbreviations: D-cate, D-fragment of M r 93000 produced by digestion of fibrmogen with plasmin m the presence of Ca2÷: D-elastase, D-fragment of M r 93000 produced by digestion of fibrinogen with leukocyte elastase m the presence of Ca `'+. DFP, dlisopropylfluorophosphate, EGTA, ethylene glycol bts( ]~-aminoethyl ether)-N, N'-tetraacetlc acid 0304-4165/83/0000-0000/$03.00 ,f' 1983 Elsevier Biomedical Press

ferent antigenic determinants. Fragment E is the amino-terminal part of fibrinogen, the D fragments are derived from the midchain and the carboxyl-terminal regions of fibrinogen [2,3]. As plasmin is thought to be primarily responsible for fibrin(ogen) digestion in vivo, most effort has been made to purify and characterize plasmin degradation products. In 1975 Plow and Edgington [4] suggested an alternative pathway for fibrino (geno)lysis, i.e., by enzymes of leukocytes. Since then, fibrin(ogen) degradation by leukocyte enzymes in vitro has been studied by several investigators [5-8]. Although the pattern of degradation of fibrinogen by leukocyte enzymes resembles that by plasmin, it seems that products arise that differ functionally and immunologically from the plasmin products [7-9]. Until now, however, no

301 leukocyte enzyme fibrin(ogen) degradation fragments have been purified and characterized. In this paper we describe the purification by gel filtration and chromatofocusing of a D-like fragment of fibrinogen produced by leukocyte elastase. The fragment is compared with the plasmic fragment D-cate. Materials and Methods

Trasylol ~ (aprotinin solution, I0000 kallikrein inhibitor units (KIU)/ml) was purchased from Bayer, Mijdrecht, The Netherlands. EGTA was from ICN Pharmaceuticals Inc., Plainview, NY, U.S.A. DFP was obtained from Serva, Heidelberg, F.R.G. and dissolved in dry 2-propanol to give a 0.1 M solution. Sephacryl S-200, CM-Sephadex C50, polybuffer exchanger PBE 94 and polybuffer PB 74 were bought from Pharmacia, Uppsala, Sweden. Human fibrinogen was obtained from Kabi, Mrlndal, Sweden (human fibrinogen grade L) or purified from plasma of healthy donors as described [10]. Fibrinogen fragment D-cate was prepared by digestion of fibrinogen by plasmin in the presence of Ca 2÷ [1]. Purification of the fragment was achieved according to the method described in Ref. 11. Human leukocyte elastase was partially purified from polymorphonuclear granulocytes from healthy donors essentially as described in Ref. 12 and was further purified by ion-exchange chromatography on CM Sephadex C50 at pH 8.5 [13]. The purified elastase showed one band of M r 30 000 on SDS-polyacrylamide gel electrophoresis. Antiserum against plasmic fragments D and E was raised in rabbits by subcutaneous injection of the purified fragments analogous to the production of antiserum to fibrinogen as described in Ref. 10. Antiserum against fragment D-cate did not react with plasmic fragment E; antiserum against fragment E did not react with fragment D-cate. SDS-polyacrylamide gel electrophoresis was performed according to the method described in Ref 14; proteins were stained for carbohydrate as described in Ref. 15: Amino-terminal amino acids were determined by the dansylchloride method [161. One-dimensional immunoelectrophoresis and

tandem crossed immunoelectrophoresis were performed essentially as described in Refs. 17 and 18. 1% Agarose (BDH Chemicals Ltd, Poole, U.K.) in 0.03 M diethyl barbituric acid, pH 8.6, and 0.005 M EGTA was used. The electrophoresis buffer consisted of 0.075 M diethyl barbituric acid, pH 8.6. Proteins were run in the first dimension for l h at 30 V/cm, 12°C, and in the second dimension overnight at 6 V/cm, 12°C. The isoelectric point of the purified fragments was determined by isoelectric focusing [11]. Calcium-binding was measured as described [19] and calculated by computer analysis as described in Ref. 20. Anticlotting properties of the fragments were measured as follows: 50/tl normal pooled plasma, collected in sodium citrate, was mixed with 100 txl of a solution of the fragment to be tested in phosphate-buffered saline. After 2 min at 37°C, 50 /xl bovine thrombin solution was added and the clotting time assessed using a steel hook which was moved up and down. The thrombin concentration was such as to provide a clotting time of 20 + 1 s when buffer without fragment was used. Preparanon of a digest rich in a (D-like)fragment with M r 93 000

Purified human leukocyte elastase was added to a fibrinogen solution (1 m g / m l in 0.04 M diethyl barbituric acid/0.10 M NaCI, pH 7.75) containing 0.01 M calcium chloride, and 10 KIU Trasylol/ml to prevent degradation of fibrinogen by plasmin. Digestion at 37°C was continued until the degradation mixture contained no fibrinogen, small amounts of X- and Y-like fragments and relatively large amounts of a fragment with an M r of 93 000. The digestion was stopped by the addition of DFP to a final concentration of 0.002 M. The degradation products were precipitated at an ammonium sulphate concentration of 50% saturation. The precipitate was dissolved in 0.02 M sodium hydrogen carbonate buffer pH 8.9. Purtfl¢atlon

The digestion mixture of 100 mg fibrinogen was applied to a column of Sephacryl S-200 (3 × 100 cm) which was equilibrated with 0.02 M sodium hydrogen carbonate buffer, pH 8.9. Elution was performed at a flow rate of 60 ml/h. Fractions

302

c o n t a i n i n g the M r 93000 fragment were p o o l e d a n d p r e c i p i t a t e d with a m m o n i u m sulphate. The p r e c i p i t a t e was d i a l y z e d against 0.025 M imidazole-HC1, p H 7.4, a n d a p p l i e d to a c o l u m n of p o l y b u f f e r e x c h a n g e r PBE 94 (0.8 × 50 cm) equil i b r a t e d with 0.025 M i m i d a z o l e - H C l , p H 7.4, for c h r o m a t o f o c u s i n g . Bound material was fractiona t e d by elution with a l : 10 dilution of p o l y b u f f e r PB 74, p H 3.2. F r a c t i o n s were a n a l y z e d b y SDSp o l y a c r y l a m i d e gel electrophoresis. The fractions c o n t a i n i n g the M r 93000 fragment were pooled, c o n c e n t r a t e d by a m m o n i u m sulphate precipitation, dialyzed against p h o s p h a t e - b u f f e r e d saline a n d stored at - 2 0 ° C .

Results Fig. l a shows the SDS-gel electrophoresis pattern of the fibrinogen elastase digest used as the starting m a t e r i a l for the purification of the M r 93000 fragment. A c o r r e s p o n d i n g stage of fibrinogen d e g r a d a t i o n b y p l a s m i n is shown in Fig. lb. I n the elastase digest p r o d u c t s are present with a m o l e c u l a r weight c o r r e s p o n d i n g to p l a s m i n frag-

m e n t X ( M r 250000), Y ( M r 150000) and D ( M r 93 000). Based on the c o r r e s p o n d e n c e in molecular weight with plasmic fragment D, the M r 93000 fragment in the elastase d e g r a d a t i o n mixture has tentatively been n a m e d ' D - l i k e ' . On SDS-gel electrophoresis no fragment c o r r e s p o n d i n g to plasmin fragment E was found. However, p r o d u c t s are present m the d e g r a d a t i o n mixture whtch react with a n t i - E antisera but not with a n t i - D antisera, as is found on i m m u n o e l e c t r o p h o r e s i s . Plasmin fragment D-cate is a terminal p r o d u c t and, therefore, a late stage of digestion ts used for its purification (Fig. lc). In contrast, we used an early stage of d e g r a d a t i o n for the purification of the D-like M r 93 000 fragment. The reason was the o b s e r v e d d e g r a d a t i o n of the D-like M r 93 000 fragm e n t in late stages of fibrinogen digestion with elastase (Fig. ld). The M r 93000 fragment was purified by gel filtration on Sephacryl S-200 and by c h r o m a t o f o c u s i n g . Fig. 2a shows elution profile of the digestion mixture on the Sephacryl S-200 column. T h e X- a n d Y-like p r o d u c t s were sepa r a t e d from the M~ 93000 D-like fragments, as was shown on SDS-gel electrophoresis of the fractions. Small a m o u n t s of M r 85 000 fragments were present in the fractions c o n t a i n i n g the M r 93000 fragment. The fractions with the M r 93000 fragm e n t gave a faint p r e c i p i t a t e with anti-E antisera. Fig. 2b shows the elution p a t t e r n of the pool of the AIr 93 000 fragment from the Sephacryl S-200 colu m n in c h r o m a t o f o c u s i n g . The M r 85 000 fragment was present in the shoulder o f the main peak. F r a c t i o n s c o n t a i n i n g this fragment were excluded from the pool. The p o o l e d M r 93000 fragment TABLE I NUMBER OF CALCIUM-BINDING SITES AND DISS O C I A T I O N CONSTANTS IN FRAGMENT D-ELASTASE

Fig. 1. S D S - p o l y a c r y l a m i d e gel electrophoresis of fibrmogen

degradation products according to the method of Weber and Osborn [14]. 7% gels were run at 8 mA/gel for 2h. (a) Early-stage elastase digest used for the purification of the D-hke Mr 93000 fragment. (b) Early stage of fibrmogen degradauon by plasmm. (c) Late-stage plasmin degradation mixture used for the punhcatlon of plasmin fragment D-cate. (d) Late-stage elastase degradation mixture. (e) Purified D-hke Mr 93000 fragment produced by elastase, not reduced. (f) Purified D-hke Mr 93000 fragment produced by elastase, after reduction.

Number and dissoc~ation constants are obtained from analy~s of the Scatchard plot and of the r versus concentration curve [19]. Values are from separate analysis of data obtained at (A) low calcium concentrations (up to 1 10-5 M) and (B) higher calcmm concentrauons (over 10 5 M) A

Number of s~tes

1.36

0 86

D~ssoclatton constant N u m b e r of sites

0.84 10 -5 M 4.07

1.04.10 4 73

Dlssociatton constant

3.85.10 -5 M

4.83.10 5 M

5M

303

O6

[

Ii

O5

04

2 03

02

°' j " f 0

I alo

100

"0

~&CYlO~

~30

NUMBED

0 0

/

' ii

",\

, 2'0

'

4'0

/ -6G FRACTION NUMBE~

\\ 8G

~00

Fig. 2a. Gel filtratmn on Sephacryl S-200 of a degradation mixture of h u m a n fibrmogen produced by leukocyte elastase. Experimental conditions are given m the text. The shaded area contains a fragment similar to plasmin D-cate (D-elastase, see text), b. Elution pattern of partmlly purified D-elastase from the Sephacryl S-200 column on chromatofocusing. Experimental conditions are given in Materials and Methods. The shaded area contains D-elastase.

180

1.50

I---

_z 12o m

I[ o I---

9O

6O

3O

0

0

~

I

.L

,

1

2

3

4

CONCENTRATION

-

(pM)

Fig. 3. Anticoagulant properties of D-cate (©) and D-elastase (I). For experimental conditions, see Materials and Methods.

precipitated with anti-D-cate antisera, but not with anti-E antisera. The adjective 'D-like' for the M r 93 000 fragment thus proved to be correct. Figs. le and 1f show the gel pattern after SDS-gel electrophoresis of the purified fragment without (Fig. le) or after (Fig. I f) reduction. After complete reduction of the fragment protein bands were found with M r values of 43000, 38000 and 12000. These values yield a calculative molecular weight of 93 000 for the D-like fragment. This value was also found on SDS-gel electrophoresis of the D-like fragment under non-reducing conditions. Carbohydrate, present in the B/3- and the y-chain of the fibrinogen molecule, was found in the M r 43 000chain of the D-like fragment. The M r 38 000 band moved faster when a short reduction of the M r 93 000 fragment was carried out in the presence of Ca 2÷ . Such an increased mobility has been shown by Lawrie and Kemp [21] for the y-chain of fibrinogen under similar conditions. We conclude on the basis of the carbohydrate staining and of the Ca2+-dependent mobility of the M r 38~300 band that the M r 43000, 38000 and 12000 chains represent the remnants of the Bfl-, y- and Aa-chain, respectively. The M, 93 000 fragment thus resembles D-cate in M r of the constituent Aa-, B/3- and y-chains. We propose to designate the D-like M, 93 000 fragment as D-elastase. Amino-terminal amino acids of D-elastase were

304

2~ ////

2C ///

•

o,

°

,;

2'o

A

2o

CONCENTRATION (pM)

Fig. 4. Calcium binding by fragment D-care (O) and D-elasta~e (O). r, number of Ca "~+ bound per molecule D. Concentration, equihbnum concentration of calcium. Sohd hnes, experimental curves. Only the part of the curve for calcium concentrations up to 40 # M is depicted in order to demonstrate clearly the change in binding characteristics at 10 5M calcium for Delastase. Data were obtained for calcium concentrations up to 100 p.M. For the analysis of the curve for D-elastase between 10 and 100 /~M, 12 data points were avadable, only three of which are depicted. Dotted hne, theoreucal curve for D-elastase, generated by adding the curves separalely fitted to the experimental data, for calcmm concentrauons up to 10 .5 M and calcmm concentrations above 10- 5 M. See text for further explanation.

Asx, Lys and Pro. The amino-terminal acids of D-cate were Asx. Val, and Ser. The isoelectric point of D-elastase was pH 6.9 as determined by isoelectric focusing. The isoelectric point of D-cate is p H 6.5 [ll]. Fig. 3 shows the mean results of the anticlotting assay obtained with three independently prepared batches of D-elastase. Fragments D-elastase prolonged the thrombin time more than D-cate fragments. Table I summarizes the results of the calciumbinding studies. The number of calcium-binding sites and the dissociation constant Kj were calculated assuming that there are only two sets of binding sites [20]. This calculation was performed separately for low (up to 10 -5 M) and higher (above 10 -5 M) calcium concentrations, as the

shape of the experimentally obtained r versu~ concentration plot (Fig. 4) ~uggested a change m bradmg characteristics at about 1. 10 5 M. At low calcium concentrations (up to 10 ~ M), one ~et of binding sites was found with a K~ of 0.9. 10 5 M (A). At higher calcium concentrations two sets were found: 4 - 5 binding sites with a K d of (3.8-4.8)-10 5 M (B) plus several sites with a much higher K a. Fig. 4 also shows the theoretical curve, obtained by adding the fitted curves for the calcium binding up to 1 0 - S M and that for the calcium binding above 10 5 M generated by using the numbers of binding sites and the dissociation constants a~ derived from the experimental cttrve~ ( n ~ = | , K d = 1.0" l0 5 M : tl B = 4 . 5 , K d =4.3. 10-5 M). D-cate and D-elastase were compared ,mmunologically by tandem crossed immunoelectrophoresis using several batches of antisera against D-cate. Some of the antiserum batches revealed partial identity between D-cate and D-elastase. Discussion

Degradation of fibrinogen by leukocyte elastase in the presence of Ca 2 + leads to the production of a D-like fragment of AIr 93000. The presence of Ca z* in the degradation mixture is essential for the production of the Mr 93000 fragment: omission of Ca 2 + or addition of EGTA results in the formation of a heterogeneous population of degradation fragments in a M r range of 90 000-80 000 (results not shown). It has been demonstrated by Purves et al. [22] and by Haverkate and Timan [1] that Ca-" ~ protects D-care completely against further degradation by plasmin. Our late-steps elastase digest (Fig. ld) shows that the protectton by Ca z ~ of the D-like M~ 93000 fragment against degradation by elastase ts not absolute. For the purification of the Mr 93 000 fragment we stopped the degradatton of fibrinogen at an early stage, The absence from the degradation mixture of fragments comparable in M~ to plasmm fragment E suggests degradation of the amino-terminal part (E-part) of fibrinogen to small products, This has also been reported by Halter et al. [6]. E-hke material was, however, found in the degradation mixture on immunoelectrophoresis: the mobility of this material was greater than that of plasmin fragment E.

305

A two-step purification method was used to obtain a pure D-like M r 93 000 fragment. In the first step (gel filtration) X- and Y-like fragments were separated from the D-like fragment. Separation of the amino-terminal product(s) and the Dlike fragment was achieved on chromatofocusing. Since the purified fragment was similar to D-cate in molecular weight of its constituent chains and precipitated with anti-D-cate antisera but not with anti-E antisera, the name D-elastase is suggested. Major differences between D-elastase and Dcate were found in properties associated with the ),-carboxyl-terminal part of the fragments. Firstly, D-elastase and D-cate differ in anticlotting properties. Fragment D-cate has been shown to prolong the clotting-time of fibrinogen in a thrombin time test [23]. It has been suggested that D-cate should possess a site which interferes with the polymerization of fibrin monomers [24,25]. The presence or expression of this polymerization site depends on the presence of the carboxyl-terminal 13000 stretches of the ),-chain [23,26]. The much stronger anticlotting properties of D-elastase, as compared with D-cate, might suggest a different, more favourable conformation of this site in Delastase. Secondly, the binding of Ca 2÷ by D-elastase differs from binding by D-cate., It has been shown [19,20,22] that human fibrinogen contains three calcium-binding sites. Two sites, one in each Dmoiety, are preserved when fibrinogen is digested by plasmin to fragments D-cate [19,20]. Laudano and Doolittle [27] have suggested that the calcium-binding depends on the carboxyl-terminal stretches of the Bfl- and ),-chains. The binding site in D-cate is lost when the carboxyl-terminal stretch of the "t-chain is split off on digestion of D-cate by plasmin in the presence of EGTA [19]. As fragment D-elastase is related to the Dmoiety of the fibrinogen molecule, at most one calcium-binding site would be expected. Surprisingly, analysis of the data of the calcium-binding studies for D-elastase, in a way similar to that used for D-cate [20], revealed 4-5 binding sites on D-elastase. The divergence in calcium binding of D-elastase and D-cate starts at calcium concentrations above l0 -5 M, as is evident from the r versus concentration curves of D-cate and D-elastase (Fig. 4). Separate analysis of the data obtained at

calcium concentrations up to 10 -5 M reveals one calcium-binding site with a dissociation constant, K d, of l • 10 -5 M. This result closely corresponds with the observed dissociation constant for the binding site of D-cate (1.3.10 -5 M [20]). If our approach of separate analysis of the data below and above l0 -5 M calcium is correct, the bend in the r versus concentration curve indicates a transition in binding from l to 4-5 Ca 2÷ . This might suggest a conformational change in D-elastase due to the binding of the first Ca 2+ . The molecular basis for such a phenomenon is obscure. The calcium-binding studies and the anticlotting properties suggest that essential structural or conformational differences between D-elastase and D-cate exist in the carboxyl-terminal part of the fragments. The observed partial identity with Dcate, using anti-D-cate antisera, might reflect these differences. Our own observations (results not shown) and those of Hafter et al. [6] on the degradation of fibrin by leukocyte elastase demonstrate that elastase is able to split off carboxylterminal stretches of the )'-chain in the presence of C a 2+ . The removed part of the )'-chain contains the cross-linking site, and thus monomeric D-fragments are produced from cross-linked fibrin. We do not known if the fibrinogen D-elastase fragment which we used in this study lacks the crosslinking sites of the ),-chain. If so, an explanation for the observed major differences between Delastase and D-cate could be a modified carboxylterminal conformation in D-elastase due to the absence of the carboxyl-terminal stretches of the ),-chain. This will be studied further. Acknowledgements The authors wish to thank Mr. G.J. Van Liempt and Ms. A. Blonk for their technical assistance, Mr. A. Vermond for his help in the calcium-binding experiments and Dr. G.L. Haberland and Dr. E. Wischh/Sfer from Bayer AG for their gift of freeze-dried Trasylol ® for the purification of leukocyte elastase. References l Haverkate, F. and Timan, G. (1977) Thromb. Res. 10, 803-812

306 2 Nussenzwetg, V., Sehgman, M and Grabar, P. (1961) Ann. Inst. Pasteur 100, 490-508 3 Marder, V.J. and Budzynskl, A.Z. (1974) in Progress m Hemostasts and Thrombosis (Spaet, T.H., ed.), Vol 2, 141-174, Grune & Stratton, New York 4 Plow, E.F. and Edgington, T.S (1975) J C h n Invest. 56. 30-38 5 Bingenheimer, C., Gramse, M., Egbrmg, R. and Havemann. K. (1981) Hoppe-Seyler's Z. Physiol. Chem. 362, 853-863 6 Hafter, R., Petrl, K., Schiessler, H. and Graeff, H. (1980) Blut 40,. 37 7 Bilezeklan, S.B. and Nossel, H L. (1977) Blood 50, 21-28 8 Gramse, M., Bingenheimer, C., Schmtdt, W., Egbrmg, R. and Havemann, K. (1978) J. Clin. Invest. 61, 1027-1032 9 Plow, E.F. and Edgmgton, T.S. (1978) Thromb. Res. 12, 653-663 10 Van Ruilven-Vermeer, I.A.M. and Nleuwenhuizen, W. (1978) Blochem. J. 169, 653-658 11 Van Rui.lven-Vermeer, I.A.M.. Nleuwenhmzen, W., Haverkate, F. and Ttman, G. (1979) Hoppe-Seyler's Phystol Chem. 360, 633-637 12 Baugh, R.J. and Travls, J. (1976) Biochemistry 15. 836-841 13 Mennmger, H., Kruze, D., Fehr, K. and Born, A (1978) Verh. Dtsch. Ges. Rheumatol 5, 200-203 14 Weber, K. and Osborn, M (1969) J. Biol. Chem. 224, 4406- 4412

15 Zacchanus, R.M., Zell. T.E., Morrlson, J.M • Woodlock. J J (1969) Anal. Biochem. 30, 148-152 16 Gray, W.R. (1977) Methods Enzymol. 25B, 121 138 17 Grabar, P. ~1964) m Immunoelectrophorettc Analysv, (Grabar, P and Burtm, P.. eds.) p. 3, Elsevier, Amsterdam 18 Kr211, J. (1973) m A Manual of Quant~tattve Immunoelectrophorests (Axelson, N . H , Kroll, J and Weeke, B.. eds.), pp. 57-59, Umversltaet~ Forlaggt, Oslo 19 Nleuwenhuizen, W . Vermond, A.. Noo~jen. W.J. and Haverkate, F. (1979) FEBS Lett 98, 257-259 20 N~euwenhmzen, W., Van Rmjven-Vermeer, I.A.M, Nootjen, W.J, Vermond, A and Haverkate. F ~1981) Thromb Res. 22, 653-657 21 Lawne, J.S and Kemp, G {1979) Biochem. Btophss Acta 577, 415-423 22 Purves, L . R , Lmdsey, G.G and Franks, J.J. 11978) S. Afr J Scl. 74. 202-200 23 Haverkate, F., Tlman, G. and Nleuwenhmzen, W. (19791 Eur. J Chn. Invest. 9, 253-255 24 Olexa, S A. and Budzynskl. A.Z. (1980) Proc. Natl. Acad ScL U.S.A. 77, 1374 1378 25 Kudryk, B.J., Collen, D., Woods, K.R. and Blomback, B (19741J. Blol Chem 249. 3322-3325 26 Olexa, S.A. and Budzynsk~, A.Z (1981) J Bml.Chem 256, 3544 3549 27 Laudano, A.P and Doohttle, R.F. (1981) Science 212. 457-459