Eur. J. Biochem. 234, 350-357 (1995) 0 FEBS 1995

Functional properties of a recombinant chimeric protein with combined thrombin inhibitory and plasminogen-activating potential H. Roger LIJNEN', Stephan WNENDT', Johannes SCHNEIDER', Elke JANOCHA', Berthe VAN HOEF', Desire COLLEN I and Gerd J. STEFFENS'

' Center for Molecular and Vascular Biology, University of Leuven, Belgium ' Grunenthal GmbH, Aachen, Germany

(Received 23 June 1995) - EJB 95 1014/4

A chimeric protein (rscu-PA-40-kDa/Hir), consisting of the C-terminal amino acids 53 -65 of hirudin (Hir), fused via a 14-amino-acid linker sequence to the C-terminal of a 40-kDa fragment (Ser47-Leu411) of recombinant (r) single-chain (sc) urokinase-type plasminogen activator (rscu-PA), was produced by expression of the corresponding chimeric cDNA i n Escherichia coli cells. The thrombin inhibitory poten- tial of purified rscu-PA-40-kDdHir was confirmed by complete inhibition of the coagulant activity of thrombin at 20-30-fold molar excess of the chimera, and by the resistance of rscu-PA-40-kDa/Hir to proteolytic cleavage by thrombin. rscu-PA-40-kDa/Hir prolonged the thrombin time of normal human plasma in a dose-dependent way (reduction of the apparent thrombin concentration to 50% with 95 nM chimeric protein as compared to 4.7 nM hirudin), and inhibited thrombin-mediated platelet aggregation (reduction of the apparent thrombin concentration to 50% with 40 nM chimeric protein).

The chimera had a specific activity on fibrin films of 57000 IU/mg as compared to 95000 IU/mg for rscu-PA. The urokinase-like amidolytic activity of the single-chain protein was only 220 IU/mg but increased to 169000 IU/mg after treatment with plasmin, which resulted in quantitative conversion to a two-chain (tc) derivative (rtcu-PA-40-kDa/Hir). Corresponding values for rscu-PA were 270 and 226 000 IU/mg. The catalytic efficiencies for plasmin-mediated conversion to two-chain molecules were compara- ble for rscu-PA-40-kDaMr and rscu-PA (0.63 and 0.65 pM-' . s-', respectively). The plasminogen- activating potential of the single-chain chimera was comparable to that of rscu-PA ; the catalytic efficien- cies for plasminogen activation by their two-chain counterparts were also similar (0.55 and 0.73 pM- I . s-', respectively). In 2 h, 50% lysis of 'z51-fibrin-labeled clots prepared from platelet-poor human plasma and immersed in normal plasma was obtained with 1.3 pg/ml rscu-PA-40-kDa/Hir and with 0.67 pg/ml rscu-PA, with corresponding residual fibrinogen levels of 74% and 87%, respectively. In the absence of fibrin, 50% fibrinogenolysis in 2 h i n normal human plasma required 2.1 pgiml rscu-PA, but 7.9 pgiml rscu-PA-40-kDa/Hir.

Thus, the chimera rscu-PA-40-kDUHir has maintained the specific fibrinolytic and plasminogen acti- vating activity of rscu-PA as well as its fibrinolytic potency in plasma, whereas it displayed a similar or somewhat better fibrin specificity. In addition, the fibrinolytically active concentration in a plasma me- dium is severalfold lower than the concentration required for thrombin inhibition, which may limit sys- temic anticoagulant activity. Therefore, further evaluation of the thrombolytic and antithrombotic potential of such chimeric molecules seems to be warranted.

Keywurds: urokinase-type plasminogen activator; hirudin; thrombolytic agent; antithrombotic agent.

Clinical experience, mainly in the treatment of patients with acute myocardial infarction, has revealed that the presently

Correspondence to H. R. Lijnen, Center for Molecular and Vascular Biology, University of Leuven, Campus Gasthuisberg, 0 & N, Herestraat 49, B-3000 Leuven, Belgium

Fux: +32 16 345990. Abbreviations. u-PA, urokinase-type plasminogen activator; scu-PA,

single-chain u-PA ; rscu-PA, recombinant 46-kDa scu-PA produced in Esckerickia coli; tcu-PA, two-chain u-PA obtained by treatment of scuPA with plasmin; rtcu-PA, recombinant tcu-PA ; rscu-PA-ilO-kDa, 40- kDa derivative of rscu-PA, consisting of amino acids Ser47-Leu411; rtcu-PA-40-kDa, two-chain molecule obtained by treatment of rscu-PA- 40 kDa with plasmin; Hir, C-terminal fragment of recombinant hirudin, consisting of amino acids 53 -65 ; rscu-PA-40-kDdHir, chimeric mole- cule consisting of rscu-PA-40-kDa with Hir fused to the C-terminal Leu via a 14-amino-acid linker sequence ; rtcu-PA-40-kDdHir, two-chain molecule obtained by treatment of rscu-PA-40-kDdHir with plasmin ; S-2403, ~-5-oxoprolyl-~-phenyIalanyl-~-lysine p-nitroanilide; S-2444, L-5-oxoprolyl-glycyl-L-arginine p-nitroanilide ; S-2238, D-phenylalanyl- L-pipecolyl-L-arginine p-nitroanilide; DPhe-Pro-ArgCH,CI, D-phenyl- alanyl-prolyl-arginine chloromethane.

Enzymes. Plasmin (EC 3.4.21.7); thrombin (EC 3.4.21.5); urokinase (EC 3.4.21.73).

available thrombolytic agents are efficient in terms of reduction of mortality, although significant shortcomings have been ob- served in terms of the occurrence of bleeding complications and platelet-mediated reocclusion [ I , 21. Therefore, it has been sug- gested that thrombolytic therapy cannot be considered as a mo- notherapy but that potent and fibrin-specific plasminogen activa- tors should be combined with antiplatelet and/or antithrombin agents or strategies 131. In the present study we report the func- tional evaluation of a recombinant chimeric protein consisting of regions of single-chain urokinase-type plasminogen activator (rscu-PA) fused to regions of hirudin.

rscu-PA contains 411 amino acids and is converted to an active two-chain derivative (rtcu-PA) by cleavage of the Lys158- Ile159 peptide bond by plasmin [4] and to an inactive two-chain derivative by cleavage of the Argl56-Phel57 peptide bond by thrombin [5]. A low-M, single-chain derivative lacking the 143 NH,-terminal amino acids has been reported with intact enzymic activity, fibrin specificity and thrombolytic potential [6, 71. The fibrin specificity is, however, only expressed in the single-chain form. scu-PA appears to have some intrinsic plasminogen-acti- vating potential, which represents S0.5 % of the catalytic effi- ciency of tcu-PA, although some investigators have claimed that

Lijnen et al. ( E m J. Biochem. 234) 351

scu-PA has no measurable activity 181. The occurrence of a tran- sitional state of scu-PA with higher catalytic efficiency against native plasminogen than tcu-PA has been suggested [9], and it has been shown that fibrin fragment E-2 selectively promotes plasminogen activation by scu-PA [lo]. Subsequent studies con- firmed that the fibrin specificity of scu-PA does not require its conversion to tcu-PA but appears to be mediated by enhanced binding of plasminogen to partially digested fibrin [11].

Hirudin, a polypeptide with 65 or 66 amino acid residues, first isolated from the leech Hirudo medicinalis, is a very potent and specific thrombin inhibitor 1121. I t inhibits thrombin by forming a tight but non-covalent equimolar complex [13]. From the three-dimensional structure of hirudin as determined by NMK techniques [14-161, and from the crystal structure of the thrombin-hirudin complex [17, 181, a model has been proposed for the thrombin-hirudin interaction with the following fea- tures : (a) extended areas of both molecules are in close contact; (b) the specificity of hirudin is not due to interaction with the primary specificity pocket of thrombin, but rather through bind- ing at sites both close to and distant from the active site; and (c) the three N-terminal residues of hirudin bind in the catalytic site. This unique mechanism explains the high efficacy and spec- ificity of hirudin against thrombin.

In the present study, the C-terminal region comprising amino acids 53-65 of hirudin was fused via a 14-amino-acid linker sequence to the C-terminal of a rscu-PA moiety consisting of amino acids Ser47-Leu411, in order to combine thrombin in- hibitory and fibrinolytic activity in one molecule. The structure of this chimeric protein, rscu-PA-40-kDa/Hir, is schematically represented in Fig. 1. Its properties with respect to thrombin in- hibition and plasminogen activation were analyzed in purified systems and in a human plasma medium in vitro.

MATERIALS AND METHODS

Proteins and reagents. Recombinant single-chain uroki- nase-type plasminogen activator (rscu-PA, saruplase) was ob- tained from Grunenthal GmbH (Aachen, Germany). Stock solu- tions of two-chain (tcu-PA) moieties were prepared by maximal activation at 37°C of the single-chain u-PA moieties (final con- centration 90-120 pM) with plasmin (0.5%) for 8-10 min in 0.05 M Tris/HCl pH 7.4 containing 38 mM NaCl and 0.01 % Tween 80. Plasmin was neutralized by addition of 3 10-fold molar excess of aprotinin and the conversion of u-PA from a single-chain to a two-chain molecule was monitored by SDS/ PAGE after reduction with dithioerythritol. tcu-PA moieties were isolated by chromatography on benzamidine - Sepharose, as de- scribed previously 1191. Human plasminogen and plasmin were obtained and characterized as described elsewhere 120, 211. The chromogenic substrates S-2444 (Glp-GI y- Arg p-nitroanilide) for urokinase, S-2403 (D-Glp-Phe-Lys-p-nitroanilide) for plasmin and S-2238 (D-Phe-Pip-Arg-p-nitroanilide) for thrombin were purchased from Chromogenix (Antwerp, Belgium), and aproti- nin (TrasylolR) from Bayer (Leverkusen, FRG). Luciferase, ADP and ATP were purchased from Sigma (Bornem, Belgium) and collagen from Nycomed (Miinchen, Germany). Human a-throm- bin was a kind gift from Dr J. W. Fenton (Wadsworth Center, New York State Department of Health, Albany, NY). Hirudin and a polyclonal sheep antiserum against hirudin were a gift from Ciba-Geigy. '2s1-labeled fibrinogen was obtained from Amersham International. The international reference preparation for urokinase (66146) was obtained from the National Institute for Biological Standards and Control (Hertfordshire, UK). The thrombin inhibitor D-Phe-Pro-ArgCH,Cl was purchased from Calbiochem (Bierges, Belgium).

Normal human platelet-poor plasma was pooled fresh frozen plasma collected on acidlcitratekdextrose from healthy donors. Platelet-rich human plasma was prepared from fresh blood, COI- lected on citrate (final concentration 0.01 1 M), by centrifugation at room temperature for 20 min at 1000 rpm (150 g) . After deter- mination of the platelet count (Cell-Dyn 160, Sequoia-Turner, Mountain View CA), plasma samples with fixed platelet count were prepared by dilution of platelet-rich with platelet-poor plasma.

Laboratory techniques. SDS/PAGE without reduction or after reduction with 1 % dithioerythritol was performed on 10- 15 % gradient gels using the Phast SystemTM (Pharmacia) and staining with Coomassie brilliant blue. Immunoblotting of 20 % homogenous gels on nitrocellulose sheets was performed ac- cording to Towbin et al. 1221.

Specific fibrinolytic activities were measured on bovine fi- brin plates 1231. Specific amidolytic activities were determined directly with S-2444 (0.3 mM final concentration) and specific plasminogen-activating activities were determined indirectly with S-2403 (0.3 mM final concentration) in mixtures of plas- minogen (0.5 pM final concentration) and u-PA (final concentra- tion 5-50 ng/ml) after incubation at 37°C for 20 min in 0.05 M Tris/HCl pH 7.4 containing 38 mM NaCl and 0.01 % Tween 80. Protein concentrations were determined according to Bradford [24], using albumin as the standard, and specific activities were expressed in IU/mg by comparison with the international refer- ence preparation for urokinase. N-terminal amino acid sequence analysis was performed on an Applied Biosystems 477A protein sequencer with identification of amino acids by HPLC.

Fibrinogen levels in plasma were determined with a clotting rate assay [25] and a,-antiplasmin levels with a chromogenic substrate assay [26].

Production of rscuPA-40-kDa and rscu-PA-40-kDahIir. Expression plasmids for rscu-PA-40-kDa and rscu-PA-40-kDd Hir were constructed by modifying expression vector pBF160 [27] which carries a synthetic gene encoding rscu-PA under the control of a synthetic trp promoter. The region of the gene en- coding amino acids 1-46 of rscu-PA was removed by digestion of pBF160 with NdeI and NcoI and inserting a synthetic DNA linker comprising a ribosomal binding site and a transcriptional start codon. The resulting plasmid pSJ41 is the expression plas- mid for production of rscu-PA-40-kDa. The plasmid for ex- presson of the chimeric protein rscu-PA-40-kDdHir was con- structed by modifying pSJ41 as follows. The region between the BamHI and Hind111 restriction sites was removed and a fragment encoding amino acids 398-41 1 of u-PA, a 14-amino-acid linker sequence and amino acids 3 - 6 5 of hirudin was inserted. The resulting expression plasmid was designated pSJ95. Both plas- mids, pSJ41 and pSJ95, were checked by extensive restriction analyses and sequencing of the synthetic gene. A detailed de- scription of the plasmid construction is given by Steffens et al. 1281. The primary structure of the 392-arnino-acid protein rscu- PA-40-kDdHir is schematically represented in Fig. 1.

Recombinant expression and purification of rscu-PA-40-kDa and rscu-PA-40-kDaHir was performed as described for rscu- PA [271. Escherichia coli JM 103 cells carrying pSJ41 or pSJ95, respectively, were grown in 1 1 standard I medium (Merck) sup- plemented with 150 pg/ml ampicillin at 37°C up to an absor- bance at 578 nm of about 1. Expression of recombinant protein was induced by addition of indole acrylic acid (final concentra- tion 60mg/l) and cultures were grown up to an absorbance at 578 nm of about 5. Cells were harvested by centrifugation (15000Xg), resuspended in 200 ml water and homogenized with a french press. The suspension was then centrifuged again and the pellet, consisting of inclusion bodies, was solubilized in 500 ml 5 M guanidinium hydrochloride, 40 mM cysteine, 1 mM

352 Lijnen et al. (Eur: J. Biochem. 234)

Fig. 1. Schematic representation of the primary structure of rscu- PA-40-kDaMir. Amino acids Serl through Leu365 (corresponding to Ser47-Leu411 of rscu-PA [4]) are linked via a 14-amino-acid linker sequence to amino acids Asn380-Gln392 (corresponding to Asn53- Gln65 of hirudin [lS]). The amino acids are represented by their single- letter symbols and black bars indicate disulfide bonds. The active-site residues His158, Asp209 and Ser310 (corresponding to His204, Asp255 and Ser356 in rscu-PA) are indicated with an asterisk and the arrow indicates the cleavage site for conversion to rtcu-PA-40-kDdHir.

EDTA, pH 8. After dilution with 2000 ml 25 mM Tris/HCI pH 9, the solution was stirred overnight in order to allow a proper refolding of the recombinant polypeptides. rscu-PA-40-kDa and rscu-PA-40-kDa/Hir, were isolated by adding 8 g silica gel to the refolding solution. After sedimentation, the silica gel was transferred into a chromatographic column and washed with 3 vol. 0.1 M sodium acetate pH 4. The rscu-PA derivative was eluted with 0.5 M trimethylammonium chloride in 0.1 M sodium acetate pH 4. The eluate was dialyzed against 50 mM sodium citrate pH 4 and further purified by chromatography on copper chelate and carboxymethyl-cellulose matrices as described [28].

Inhibition of thrombin activity. Inhibition of thrombin (fi- nal concentration 40 nM) by the single-chain u-PA moieties (fi- nal concentration 0-20 pM) or by hirudin (final concentration 0-100 nM) in 0.1 M sodium phosphate pH 7.4 containing 0.01 % Tween 80 at 25°C was monitored as a function of time by measurement of residual thrombin activity with S-2238 (final concentration 0.1 mM) after 10-fold dilution of samples. In addi- tion, the single-chain u-PA moieties (final concentration 4 pM) were incubated with thrombin (final concentration 40 nM) in 0.05 M Tris/HCI pH 7.4 containing 38 mM NaCl and 0.01 % Tween 80 at 37°C. At timed intervals (0-30 min), aliquots were removed and urokinase activity was measured with S-2444 (final concentration 0.3 mM) after 25-fold dilution of samples. The molecular form of the u-PA moieties was monitored on SDS/ PAGE under reducing conditions, after neutralization of throm- bin activity by addition of DPhe-Pro-ArgCH,CI (final concentra- tion 10 pM).

The influence of addition of the single-chain u-PA moieties (final concentration 0-500 nM) or of hirudin (final concentra- tion 0- 10 nM) to normal or plasminogen-depleted human

plasma was determined on the thrombin time. Thrombin times were recorded as the clotting time in mixtures of 0.2 ml plasma and 0.1 ml of thrombin solution (20 nM final concentration). A calibration curve was constructed by measuring the thrombin time of plasma with different thrombin concentrations (final concentration 0-20 nM). In addition, thrombin (80 nM final concentration) was incubated with the single-chain u-PA moi- eties or with hirudin (final concentration 0-4 pM) for 1- 10 min at 37°C prior to its addition to plasma (final concentra- tion 20 nM) and recording of the thrombin time.

The influence of the single-chain u-PA moieties was also evaluated on thrombin-induced platelet aggregation. Therefore, 2.5 X 10' washed platelets/pl in Ernes buffer (0.8 g/l Hepes, 4 g/l NaH,PO, . H,O, 8 g/l NaCI, 0.2g/l KCI and 1 g/l glucose, pH 7.3) were incubated for 30 s at room temperature with the agents (final concentration 4-500 nM) and aggregation was in- duced with thrombin (threshold concentration of 0.35 nM). Platelet aggregation was monitored with an Elvi 840 dual-chan- nel aggregometer and expressed as a percentage of the control value (without addition of u-PA).

In addition, platelet aggregation and secretion assays were performed using platelet-rich plasma (450 pl, platelet count 2.5X105/p1) incubated for 1 min at room temperature with the single-chain u-PA moieties (final concentration 0-1 pM) in a total volume of 500 pl. As a control, platelet-rich plasma was incubated with 0.1 M sodium phosphate pH 7.4 containing 0.01 % Tween 80. Luciferase was added to a final concentration of 4 mg/ml and, after a I-min incubation at 37"C, collagen (final concentration 20 pg/ml) was added, and platelet aggregation and ATP secretion were monitored for 7 min at 37"C, using a Lumi- nometer aggregometer model 400 (Chrono-Log, Haventown PA). The extent of aggregation was determined as the change in light transmittance measured 7 min after addition of inducer, expressed as a percentage of the difference between platelet-rich plasma and platelet-poor plasma and calculated as a percentage of the control value (incubation with buffer). ATP secretion was calculated by comparison with the signal obtained following ad- dition of ATP (final concentration 5 pM) at the end of the ex- periment.

Treatment with plasmin. The single-chain u-PA moieties (final concentration approximately 1 pM) were incubated with plasmin (final concentration 0.6-2.5% on a molar basis) in 0.05 M Tris/HCI pH 7.4 containing 38 mM NaCl and 0.01 % Tween 80 at 37°C. At timed intervals (0-20 rnin), aliquots were removed and generated urokinase activity was measured with S-2444 (final concentration 0.3 mM) after 25-fold dilution of samples. Urokinase activity was expressed in international units (IU) by comparison with the reference preparation.

The kinetic parameters for the conversion of single-chain to two-chain u-PA moieties were determined by incubation of the single-chain molecules (final concentration 1 -20 pM) with plasmin (final concentration 2 nM) at 37°C in 0.05 M Tris/HCI pH 7.4 containing 38 mM NaCl and 0.01 % Tween 80. At timed intervals (0-2 min), aliquots were removed and generated uro- kinase activity was determined as described above.

Activation of plasminogen. Activation of plasminogen (fi- nal concentration 1 pM) at 37°C in 0.1 M sodium phosphate pH 7.4 was measured with the single-chain u-PA moieties (final concentration 2 nM). Therefore, at different time intervals (0- 30 min), aliquots were removed from the incubation mixtures and plasmin was quantified with S-2403 (final concentration 0.3 mM) after 25- 100-fold dilution of samples. Plasminogen activation is expressed as a percentage of the plasminogen con- centration used.

The kinetic parameters for the activation of plasminogen by the two-chain U-PA moieties were determined by incubation of

Lijnen et al. ( E m J. Biochern. 234) 353

A 1 2 3 4 5

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

Fig.2. SDSPAGE without (A) or after (B) reduction with dithio- erythritol using 10-15 70 gradient gels and staining with Coomassie brilliant blue. Lane 2, rscu-PA; lane 3, rscu-PA-40-kDa; lane 4, rscu- PA-40-kDdHir; lane 5, rtcu-PA; lane 6, rtcu-PA-40-kDa; lane 7, rtcu- PA-40-kDa/Hir; lanes 1 and 8 represent a protein calibration mixture consisting of phosphorylase b (94 kDa), bovine serum albumin (67 kDa), ovalbumin (45 kDa), carbonic anhydrase (30 kDa), soybean trypsin in- hibitor (20.1 kDa) and a-lactalbumin (14.4 kDa).

plasminogen (final concentration 1-20 pM) at 37°C in 0.1 M sodium phosphate pH 7.4 with the enzyme (final concentration 1.5 nM). Generated plasmin was measured at different time in- tervals (0-4 min) as described above, and initial activation rates were obtained from plots of the concentration of generated plas- min versus time. The kinetic parameters ( k Z and K,,,) were deter- mined by linear regression analysis from Lineweaver-Burk plots.

Fibrinolytic and fibrinogenolytic properties in human plasma in vitro. "'I-fibrin labeled platelet-poor plasma or plate- let-rich plasma clots were prepared from normal human plasma (3x10' platelets/pl) [29]. Lysis of 60 pl clots, submerged in 0.5 ml citrated normal human plasma, with different concentra- tions of u-PA moieties (final concentration 0- 10 pg/ml) was measured after a 2-h incubation at 37"C, from the release of radioactivity in the surrounding plasma [29]. Fibrinogen, plas- minogen and a,-antiplasmin levels were monitored as described above.

In addition, systemic activation of the fibrinolytic system in normal human plasma in the absence of fibrin by addition of single-chain or two-chain u-PA moieties (final concentration 0- 10 pg/ml) was monitored by measuring the residual fibrinogen levels after incubation at 37°C for 2 h.

RESULTS

Physicochemical characterization of rscu-PA, rscu-PA-40- kDa and rscu-PA 40-kDa/Hir. The proteins were obtained as homogeneous preparations, essentially in the single-chain form, as shown by SDS/PAGE under non-reducing or reducing condi- tions (Fig. 2). The apparent molecular mass is 46 kDa for rscu- PA, 40 kDa for rscu-PA-40-kDa and 43 kDa for rscu-PA-40- kDa/Hir. N-terminal amino acid sequence analysis of 50 pmol rscu-PA-40-kDalHir yielded one homogeneous sequence (with yields in pmol in parenthesis), corresponding to Ser(l9)- Lys(25)-Thr(44)-Xaa-Tyr(69)-Glu(25)-. Upon treatment with plasmin, quantitative conversion to tcu-PA moieties was ob- tained as shown by SDS/PACE under reducing conditions (Fig. 2B), whereas no degradation was apparent under non-re- ducing conditions (Fig. 2A). The B chain of rtcu-PA-40-kDalHir (Fig. 2B, lane 7) has a somewhat higher mass than that of rtcu- PA (lane 5 ) or rtcu-PA-40-kDa (lane 6), in agreement with the addition of the C-terminal fragment of hirudin. The A chain of both rtcu-PA-40-kDa and rtcu-PA-40-kDdHir has a lower mass

Fig. 3. Immunoblotting of SDSlPAGE under reducing conditions using 20 % high-density gels and staining with a polyclonal sheep antiserum against hirudin. Lane 1 , rtcu-PA-40-kDdHir; lane 2, rscu- PA-40-kDdHir; lane 3 , rtcu-PA-40-kDa; lane 4, rscu-PA-40-kDa; lane 5, hirudin.

than that of rtcu-PA, in agreement with the N-terminal deletion. N-terminal amino acid sequence analysis of approximately 100 pmol rtcu-PA-40-kDdHir yielded the sequence Ile(86)- Ile(97)-Gly(64)-Gly(88)-Glu(49)-Phe(73), in addition to the N-terminal sequence Ser(56)-Lys( 1 8)-Thr(49)-Xaa-Tyr(49)- Glu(SO), thus confirming cleavage of the Lysl12-IleI 13 pep- tide bond by plasmin (Fig. 1). Immunoblotting with a polyclonal anti-hirudin sheep serum on nitrocellulose sheets of SDS/PAGE under reducing conditions (Fig. 3), confirmed the presence of the C-terminal fragment of hirudin in rscu-PA-40-kDalHir (lane 2) and in the B chain of rtcu-PA-40-kDalHir (lane 1).

Inhibition of thrombin activity. Incubation of human a-throm- bin (final concentration 40 nM) with up to 500-fold molar excess of rscu-PA-40-kDa/Hir for 20 min at 25°C did not affect the amidolytic activity of thrombin as measured with S-2238. In contrast, hirudin at a twofold molar excess completely abolished the amidolytic activity of thrombin within 30 s (data not shown).

Addition of thrombin (1 % on a molar basis) to rscu-PA and to rscu-PA-40-kDa resulted in rapid and quantitative conversion to two-chain u-PA moieties, whereas no conversion of rscu-PA- 40-kDdHir was observed under the same experimental condi- tions (Fig. 4). No generation of u-PA activity could be detected with S-2238 in samples removed from the incubation mixtures up to 30 min. This is consistent with the conversion of rscu-PA and rscu-PA-40-kDa to inactive tcu-PA moieties (most likely by cleavage of the Arg156-Phe157 and Argl 10-Phel11 peptide bonds, respectively [S]), whereas rscu-PA-40-kDdHir is resis- tant to cleavage as a result of neutralization of thrombin by the hirudin moiety.

Addition of rscu-PA-40-kDdHir or of hirudin to normal hu- man plasma resulted in a dose-dependent prolongation of the thrombin time. The apparent thrombin concentration at each concentration of rscu-PA-40-kDalHir or hirudin added is ob- tained from a calibration curve representing the thrombin time as a function of the thrombin concentration used (0-20 nM final concentration ; not shown), and plotted against the concentration of agent (Fig. 5) . The apparent thrombin concentration in normal plasma is reduced to 50% with 95 nM (4.1 pg/ml) rscu-PA-40- kDa/Hir and with 4.7 nM hirudin. In plasminogen-depleted plasma, a 50 % reduction of the apparent thrombin concentration was obtained with 87 nM (3.8 pg/ml) rscu-PA-40-kDa/Hir (data

354 Lijnen et al. (Eur J . Biochem. 234)

A B C

s 1 2 3 4 1 2 3 4 1 2 3 4

Fig. 4. SDSmAGE of single-chain u-PA moieties treated with throm- bin, using 10-15 70 gradient gels under reducing conditions and staining with Coomassie brilliant blue. (A) 4 pM rscu-PA, (B) rscu- PA-40-kDa, or (C) rscu-PA-40-kDdHir were incubated with 40 nM thrombin. Samples were taken at time 0 (lanes 1) and at 7 min (lanes 2), 15 min (lanes 3) and 30 min (lanes 4) after addition of thrombin. Lane S represents a protein calibration mixture as in Fig. 2.

* m

0 ~ " " " " " ' " 0 2 4 6 8 1 0 1 2

[hirudin] (nM) i I I I I I 1 I I 1 5 1 I

0 40 80 120 160 200 240

[rscu-PA-40kDa/Hir] (nM)

Fig.5. Effect of rscu-PA-40-kDaMr (0) or hirudin (m) on the thrombin time of normal human plasma. The apparent thrombin con- centration, derived from a calibration curve (not shown) is plotted versus the concentration of rscu-PA-40-kDdHir or of hirudin.

not shown). No significant prolongation of the thrombin time in normal human plasma was observed upon addition of rscu-PA or rscu-PA-40-kDa to a final concentration of 1 pM.

Incubation of thrombin with up to 50-fold molar excess of rscu-PA-40-kDdHir for 1 min at 37°C prior to its addition to plasma also resulted in a dose-dependent prolongation of the thrombin time, consistent with neutralization of the thrombin activity. Similar results were obtained upon incubation for 5 or 10 min. Complete inhibition of thrombin was obtained with a 30-fold molar excess of rscu-PA-40-kDdHir (value extrapolated from a plot of residual thrombin activity versus molar excess of the chimera over thrombin). With hirudin, complete thrombin inhibition was obtained at equimolar concentration.

Aggregation of washed human platelets induced by a thresh- old thrombin concentration was reduced in a concentration-de- pendent way by addition of rscu-PA-40-kDa/Hir with a 50% reduction of the apparent thrombin concentration at approxi- mately 40 nM (Fig. 6). rscu-PA-40-kDa and rscu-PA at concen- trations of 200-400 nM did not affect thrombin-induced platelet ag.gregation. In platelet-rich plasma, rscu-PA, rscu-PA-40-kDa or rscu-PA-40-kDa/Hir (at concentrations from 1 - 50 pg/ml)

F -

0.4

0.3

0.2

8 O ' l i 0 a

10 100

[rscu-PA-40kDalHir] (nM)

Fig. 6. Effect of rscu-PA-40-kDaMir on thrombin-induced aggrega- tion of washed human platelets. The data show a representative experi- ment with the corresponding calibration curve in the inset.

had no significant effect on collagen-induced platelet aggrega- tion: aggregation was (mean ? SD of determinations with five different concentrations) 101 t 4 % of control with rscu-PA, 100 t 6 % of control with rscu-PA-40-kDa and 102 2 7 % of con- trol with rscu-PA-40-kDaRIir. In the same experiments, ATP secretion, measured upon collagen-induced platelet aggregation, was not affected by addition of the plasminogen activators: secretion was 1 0 4 ? 5 % of control with rscu-PA, 1 0 4 & 3 % of control with rscu-PA-40-kDa and 99 ? 9 % of control with rscu- PA-40-kDa/Hir.

Treatment with plasmin. Plasmin caused a time-dependent and concentration-dependent increase of the amidolytic activity against S-2444 of the single-chain u-PA moieties, resulting in a specific activity of 189000 IU/mg for rscu-PA-40-kDa/Hir, as compared to 105000 IU/mg for rscu-PA-40-kDa or 178000 IU/ mg for rscu-PA (Fig. 7). In the absence of plasmin, no increase in amidolytic activities was observed (not shown). SDSPAGE under reducing or non-reducing conditions (Fig. 2 A and B) showed that incubation with plasmin was associated with quanti- tative conversion of single-chain to two-chain u-PA moieties. The specific activities of single-chain and two-chain u-PA moi- eties, determined on fibrin plates, measured directly with S-2444 or indirectly with S-2403 in the presence of plasminogen, are summarized in Table 1 .

The kinetic parameters for the conversion of single-chain to two-chain u-PA moieties by plasmin, as determined by linear regression analysis from Lineweaver-Burk plots (not shown) were K,,, = 6.8 pM and k, = 4.4 s- ' for rscu-PA ( r = 0.997), K,, = 3.6 pM and k , = 0.93 s-' for rscu-PA-40-kDa ( r = 0.991) and K,,, = 6.3 pM and k , = 4.0 s for rscu-PA-40-kDa/Hir ( r = 0.996). The catalytic efficiency ( k J K J for plasmin-mediated conversion to two-chain u-PA thus is comparable for rscu-PA- 40-kDa/Hir and rscu-PA (0.63 pM-' . s- ' and 0.65 pM-' . s-l), and somewhat lower for rscu-PA-40-kDa (0.26 pM-' . s-I), mainly as a result of a lower k, value.

Activation of plasminogen. Fig. 8 shows time-dependent acti- vation of plasminogen (final concentration 1 pM) by the single- chain u-PA moieties (final concentration 2 nM). With rscu-PA and rscu-PA-40-kDa/Hir, about 80-90% of the plasminogen is activated within 20 min, as compared to about 60% for rscu-PA- 40-kDa.

Kinetic analysis revealed that plasminogen activation by the two-chain u-PA moieties follows Michaelis-Menten kinetics, as

Lijnen et al. ( E m J. Biochem. 234) 355

200

150

100

50

0

FY A I " 1

0 *- 0 5 10 15 20 .-

zoo Q fl-= *---A -A ffl Y

150

100

50

0 0 5 I 0 15 20

TIME (min)

Fig. 7. Treatment of single-chain u-PA moieties with plasmin. Gener- ated urokinase-like amidolytic activity is plottted versus time. (A) rscu- PA (final concentration 1.2 pM) with (mol/mol) 0.4% plasmin (O), 0.8% plasmin (+) or 1.7% plasmin (a). (B) rscu-PA-40-kDa (final con- centration 0.8 pM) with (mol/mol) 0.6% plasmin (.), 1.2% plasmin (+) or 2.5 % plasmin (a). (C) rscu-PA-40-kDa/Hir (final concentration 0.8 pM) with (mollmol) 0.6% plasmin (O), 1.2% plasmin (+) or 2.5% plasmin (m).

Table 1. Specific activities of u-PA moieties, determined by compari- son with the International Reference Preparation for Urokinase (661 46). Protein concentrations were determined according to Bradford [24]. Data are meanLSD of 3-6 determinations.

u-PA moiety Specific activity measured with

fibrin plate S-2444 plasminoged S-2403

~

IUlmg

rscu-PA 950002 3300 270% 27 106000223000 rscu-PA-40 kDa 43000% 3400 1402 34 40000? 5100 rscu-PA-40 kDa/Hir 57 000 ? 11 OOO 220 ? 24 54 000? 5 100 rtcu-PA 218000-+ 7100 226000t 9900 296000232000 rtcu-PA-40kDa 148000_f30000 113000fll 700 135000?14000 rtcu-PA-40 kDa/Hir 138 000+29 OOO 169 OOO? 12 900 267 OOO? 15 000

shown by linear double-reciprocal plots of the initial activation rate versus the plasminogen concentration (not shown). The ki- netic constants obtained by linear regression analysis, were K,, = 5.6 yM and k, = 4.1 s-' for rtcu-PA (Y = 0.996), K,,, = 5.2 yM and k Z = 1.5 s-' for rtcu-PA-40-kDa (Y = 0.998) and K,, = 5.5 pM and k, = 3.0 s-' for rtcu-PA-40-kDflir ( r = 0.996). The catalytic efficiency for plasminogen activation (kJ K,) thus is comparable for rtcu-PA and rtcu-PA-40-kDa/Hir (0.73 pM-' . s ' and 0.55 pM-' . s-') and somewhat lower for rtcu-PA-40-kDa (0.29 pM-' . s- I), mainly as a result of a lower k, value.

1000

600

400

200

0

c

0 5 10 15 20

TIME (min)

Fig. 8. Activation of 1 pM plasminogen by 2 nM rscu-PA (O), rscu- PA-40-kDa (+) or rscu-PA-40-kDa/Hir (a).

Table 2. Lysis of '251-fibrin-labeled human platelet-poor plasma clots immersed in citrated human plasma. EC,,, is the concentration of u- PA moiety required to obtain 50% clot lysis in 2 h. Residual fibrinogen and a,-antiplasmin levels were determined after 2 h at EC,,,. Data repre- sent mean 2 SEM of 3 independent experiments.

u-PA moiety ECm Residual Residual fibrinogen a,-anti-

plasmin

rscu-PA 0.67 2 0.05 8 7 2 3 70210 rscu-PA-40 kDa 2.4 50.19 6 9 2 4 5 7 2 I rscu-PA-40 kDa/Hir 1.3 20.04 74?3 5 7 z 4 rtcu-PA 0.59 2 0.02 1 7 2 4 242 9 rtcu-PA-40 kDa 1.2 20.07 2 4 2 2 2 1 2 2 rtcu-PA-40 kDa/Hir 0.69 2 0.06 2421 2 3 2 6

Fibrinolytic and fibrinogenolytic properties in human plasma in vitro. Dose-dependent lysis of '2'1-fibrin-labeled platelet-poor plasma clots submerged in normal human plasma was obtained with the single-chain and two-chain plasminogen activators (data not shown). A clot lysis of 50% in 2 h (EC,,) was obtained with (mean t SEM; n = 3) 0.67 f 0.05 yg/ml rscu- PA or 0.59 & 0.02 yg/rnl rtcu-PA, with 2.4 t 0.1 8 pg/ml or 1.2 t 0.07 pg/ml rscu-PA-40-kDa or rtcu-PA-40-kDa, and with 1.3 f 0.04 yg/ml or 0.69 t 0.06 pg/ml rscu-PA-40-kDdHir or rtcu-PA-40-kDa/Hir. Residual fibrinogen and a,-antiplasmin levels at EC,, after 2 h, determined graphically from dosehe- sponse curves (not shown) are summarized in Table 2.

lZ'I-labeled platelet-rich plasma clots submerged in normal human plasma appeared to be very resistant to lysis with all u- PA moieties tetsted (data not shown). The highest degree of lysis which was obtained, residual fibrinogen and a,-antiplasmin levels are summarized in Table 3.

In the absence of fibrin, dose-dependent systemic activation of the fibrinolytic system in normal human plasma was observed with the single-chain and two-chain plasminogen activators (not shown). The concentrations which, within 2 h, reduced the plasma concentration of fibrinogen to 50% of the baseline value were (mean 2 SEM; n = 3 ) 2.1 t 0.23 pg/ml or 0.32 f 0.01 pgl ml for rscuPA or rtcu-PA, as compared to 6.0k0.31 pg/ml or 0.82 2 0.01 yg/ml for rscu-PA-40-kDa or rtcu-PA-40-kDa and to 7.0 f 0.1 2 pg/ml or 0.78 t 0.07 pg/ml for rscu-PA-40-kDdHir or rtcu-PA-40- kDa/Hir.

356 Lijnen et al. ( E m J . Biochem. 234)

Table 3. Lysis of '*"I-fibrin-labeled platelet-rich plasma clots im- mersed in citrated human plasma. The maximal lysis is indicated, with the concentration at which it is achieved in parentheses. Residual fibrinogen and a,-antiplasmin levels were determined after 2 h at the concentration yielding maximal lysis. Data represent mean 2 SEM of 3 independent experiments.

u-Pa moiety Maximal lysis (concn) Residual Residual fibinogen u2-anti-

plasmin

rscu-PA 3 7 2 4 4.220.83 1 6 t 5 3 4 t 1 6 rscu-PA-40 kDa 3 8 Z 7 5.050 12' 2 4 6 2 1 8 rscu-PA-40kDa/Hir 37'6 3.320.83 20'10 57-1-17 rtcu-PA 2 0 2 2 2.120.42 2 1 2 9 1 7 2 4 rtcu-PA40 kDa 2 0 2 2 2 .520 1 7 2 7 2 2 2 6 rtcu-PA-40 kDa/Hir 2 0 2 1 2.1 20.42 1 8 t 5 33' 6

DISCUSSION

Despite their widespread use in patients with acute myocar- dial infarction, all currently available thrombolytic agents suffer from a number of significant limitations, including resistance to reperfusion, the occurrence of acute coronary reocclusion and bleeding complications [ l , 21. Several alternative and comple- mentary approaches to improve thrombolytic therapy have been explored: (a) earlier and accelerated treatment in order to reduce the duration of ischemia; (b) the development of alternative or engineered plasminogen activators with increased thrombolytic potency and/or specific thrombolytic activity, in order to en- hance coronary thrombolysis and (c) the use of more specific and potent anticoagulant and antiplatelet agents for conjunctive use with thrombolytic agents, with the aim to accelerate recana- lization and to prevent reocclusion [ I ] .

rscu-PA is a relatively fibrin-specific plasminogen activator of which the potency and specificity are presently under investi- gation in patients with acute myocardial infarction [30, 311. Hirudin is a potent and specific thrombin inhibitor which also inhibits clot-bound thrombin [32] that is resistant to inhibition by the heparin-antithrombin-I11 complex [33]. Hirugen, a syn- thetic peptide derived from residues 53-64 of hirudin, was shown to enhance thrombolysis induced by recombinant tissue- type plasminogen activator (rt-PA) and to delay reocclusion in a canine left anterior descending coronary artery model [34]. The antithrombotic potential of hirudin, in combination with throm- bolytic agents (rt-PA and streptokinase) is presently under study in large-scale clinical trials [35-371. In an alternative approach, Phaneuf et al. [38] have reported the covalent linkage of strepto- kinase to recombinant hirudin using chemical heterofunctional cross-linking agents [38]. Chemical conjugates may, however, have several disadvantages including low yields, random cou- pling and, possibly, interference with the functional properties of the constituting moieties [39].

In the present study, we report on the functional properties of a recombinant chimeric protein (rscu-PA-40-kDa/Hir) with combined thrombin-inhibitory and plasminogen-activating po- tential, which was obtained by fusion of amino acids 53-65 of hirudin to the C-terminal of a 40-kDa fragment of rscu-PA via a 14-amino-acid linker sequence. This molecule was selected out of a panel of chimeras in which the thrombin inhibitory activity could be modulated depending on the linker sequence used (Wnendt et al., unpublished data).

Functional analysis of rscu-PA-40-kDa/Hir and of its two- chain derivative obtained by treatment with plasmin, in compari-

son with rscu-PA-40-kDa and wild-type rscu-PA, revealed that the specific fibrinolytic activity and the plasminogen-activating potential of the chimera remained intact. Its fibrinolytic potency in a plasma medium was also maintained with intact or some- what enhanced fibrin specificity, as indicated by a similar rate of fibrinogenolysis in plasma in the presence of fibrin and a lower rate in the absence of fibrin. The thrombin inhibitory po- tential of the chimera was revealed by our findings of a dose- dependent prolongation of the thrombin time of normal human plasma and inhibition of thrombin-mediated platelet aggrega- tion. Interestingly, the amidolytic activity of thrombin against S-2238 was not inhibited in the presence of excess rscu-PA-40- kDa/Hir, in contrast to its activity against fibrinogen. This is compatible with the observation that the secondary binding site for fibrinogen is blocked because the C-terminal of hirudin wraps around thrombin along this binding site [17]. In contrast to the complex of thrombin with hirudin, the active site of thrombin in its complex with rscu-PA-40-kDa/Hir appears to be available for the synthetic peptide substrate S-2238, most likely because the N-terminal amino acids of hirudin, which normally bind in the catalytic site of thrombin in a direction opposite to that of the substrate-like inhibitor DPhe-Pro-ArgCH,CI [40], are absent in the chimera. It has been shown previously that hirugen (a synthetic peptide comprising residues 53-64 of hirudin) also interacts with thrombin without inhibiting its amidolytic activity

Our data furthermore indicate that the apparent thrombin concentration in plasma (thrombin time assay) is reduced to 50% at 20-fold lower molar concentration of hirudin as com- pared to the chimera. This is probably not due to unfavourable inhibition kinetics as a result of steric hindrance with the larger chimera, as suggested by our finding that also upon incubation for 10 min at 37°C a 30-fold molar excess of the chimera was required for complete inhibition of thrombin. It has been re- ported previously that the molar specific antithrombin activity of hirugen (measured in thrombin time assays in plasma) is about 50-fold lower than that of native hirudin 141 1. Thus, the intrinsic specific thrombin inhibitory activity of hirudin is not fully expressed in the chimera, but the activity is compatible with that expected for hirugen. This observation may, however, not represent a disadvantage. Indeed, our data indicate that higher concentrations of the chimera are required for inhibition of blood coagulation in the circulation than for lysis of a fibrin clot or for inhibition of thrombin-induced platelet aggregation. Therapeutic effective concentrations for clot lysis and preven- tion of platelet aggregation may thus be reached in plasma with- out compromising the systemic blood coagulation system and predisposing to bleeding complications. The relative effects of the chimera on the fibrinolytic and coagulation systems remain, however, to be investigated in more detail. It is an interesting possibility that the fibrin-specific mechanism of action of rscu- PA may direct and limit thrombin inhibition to the site of a thrombus, thus preventing or limiting more general systemic an- ticoagulant activity.

Chimeric molecules, combining the functional properties of a fibrin-specific plasminogen activator with the inhibitory poten- tial of a specific and potent thrombin inhibitor, may be useful for thrombolytic therapy. Their efficacy for the lysis of platelet- rich arterial blood clots and for the prevention of reocclusion after successful thrombolysis remains to be investigated in ap- propriate animal models.

~411.

We are grateful to J. Van Damme and P. Proost (Rega Institute, Uni- versity of Leuven, Belgium) for performing the amino acid sequence analysis. Skilful technical assistance by F. De Cock is gratefully ac- knowledged.

Lijnen et al. (Eur

REFERENCES I . Collen, D. & Lijnen, H. R. (1993) Am. J . Cardiol. 72, 1G-6G. 2. Collen, D. & Lijnen, H. R. (1991) Blood 78, 3114-3124. 3. Gold, H. K. (1990) N. Engl. J . Med. 323, 1483-1485. 4. Holmes, W. E., Pennica, D., Blaber, M., Rey, M. W., Guenzler, W.

A., Steffens, G. J. & Heynekel; H. L. (1985) Biotechnology 3, 923-929.

5. Ichinose, A,, Fujikawa, K. & Suyama, T. (1986) J . B id . Chem. 261, 3486-3489.

6. Stump, D. C., Lijnen, H. R. & Collen, D. (1986) J . Biol. Chem. 261, 17 120- 17 126.

7. Lijnen, H. R., Nelles, L., Holmes, W. E. & Collen, D. (1988) J . B id . Chem. 263, 5594-5598.

8. Lijnen, H. R., Stump, D. C. & Collen, D. (1995) in Cardiovu.scular drug therapy, 2nd edn (Messerli, F., ed.) Ch. 169, W. B. Saunders Company, Philadelphia, PA, in the press.

9. Liu, J., Pannell, R. & Gurewich, V. (1992) J . B id . Chem. 267, 15 289 - 15 292.

10. Liu, J . & Gurewich, V. (1992) Biochemistry 31, 6311-6317. 11. Fleury, V., Lijnen, H. R. & Anglts-Cano, E. [ 1993) J. B id . Chem.

12. Markwardt, F. (1970) Methods Enzymol. 19, 924-932. 13. Stone, S. R. & Hofsteenge, J. (1986) Biochemi.wy 25, 4622-4628. 14. Clore, G. M., Sukumaran, D. K., Nilges, M., Zarbock, J . & Gro-

nenborn, A. M. (1987) EMBO J. 6, 529-539. 15. Folkers, P. J. M., Clore, G. M., Driscoll, J., Dodt, J., Kohler, S. &

Gronenborn, A. M. (1989) Biochemistry 28, 2601 -2617. 16. Haruyama, H. & Wiithrich, K. (1989) Biochemistry 28,4301 -4312. 17. Griitter, M. G., Priestle, J . P., Rahuel, J., Grossenbacher, H., Bode,

W., Hofsteenge, J. & Stone, S. R. (1990) EMBO J. 9, 2361- 2365.

18. Rydel, T. J., Ravichandran, K. G., Tulinsky, A., Bode, W., Huber, R., Roitsch, C. & Fenton 11, J. W. (1990) Science 249, 277-280.

19. Lijnen, H. R., Van Hoef, B. & Collen, D. (1986) Biochirn. Biophys.

20. NeHes, L.. Lijnen, H. R., Collen, D. & Holrnes. W. E. (1987) J.

21. Lijnen, H. R., Van Hoef, B. & Collen, D. (1981) Eur J. Biochem.

268, 18554-18559.

A C ~ U 884, 402-408.

Biol. Chem. 262, 5682-5689.

120, 149-154.

J. Biochem. 234) 357

22. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl Acad. Sci.

23. Astrup, T. & Muellertz, S. (1952) Arch. Biochem. Biophys. 40, 346-

24. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 25. Clauss, A. (1957) Actu Haemat. 17, 237-246. 26. Teger-Nilsson, A. C. , Friberger, P. & Gyzander, E. (1977) Scand. J .

Clin. Lab. invest. 37, 403-409. 27. Brigelius-Flohe, R., Steffens, G., Strassburger, W. & FlohC, L.

(1 992) Appl. Microbiol. Biotechnol. 36, 640-649. 28. Steffens, G. J., Wnendt, S., Schneider, J., Saunders, D. & Heinzel-

Wieland, R. (1 994) Uffenlegungsschrift DE4323754CI. 29. Lijnen, H. R., Van Hoef, B., Vandenbossche, L. & Collen, D. (1992)

Fibrinolysis 6, 214-225. 30. PRIM1 Trial Study Group (1989). Lancet I , 863-868. 31. Ostermann, H., Schmitz-Huebner, U., Windeler, J., Bar, F., Meyer,

J. & van de Loo, J . (1992). Eur Heart J . 13, 122551232, 32. Weitz, J., Hudoba, M., Massel, D., Maganore, J. & Hirsh, J. (1990)

J . Clin. Invest. 86, 385-392. 33. Hogg, P. J . & Jackson, C. M. (1989) Proc. Natl Acad. Sci. USA 86,

34. Yao, S. K., McNatt, J., Anderson, H. V., Eidt, I., Cui. K. X., Golino, P., Glas-Greenwalt, P., Maraganore, J., Buja, L. M. & Willerson, J. T. (1992) Am. J. Physiol. 262, H374-H379.

35. Antman, E. M., for the TIMI-9A Investigators (1994) Circulation 90, 1624 - 1630.

36. The Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) IIa Investigators (1994) Circulation 90, 1631 - 1637.

37. Neuhaus, K.-L., v. Essen, R., Tebbe, U., Jessel, A., Heinrichs, H., Maurer, W., Doring, W., Harmjanz, D., Kotter, V., Kalhammer, E., Simon, H. & Horacek, T. (1994) Circulation 90, 1638-1642.

38. Phaneuf, M. D., Ozaki, C. K., Johnstone, M. T., Loza, J.-P., Quist, W. C. & LoGerfo, F. W. (1994) Thromh. Haemostas. 71, 481- 487.

USA 76, 4350-4354.

351.

361 9-3623.

39. Dewerchin, M. & Collen, D. (1991) Bioconj. Chem. 2, 293-300. 40. Bode, W., Mayr. I., Baumann, U., Huber, R., Stone, S. R. & Hof-

steenge, J . (1989) EMBO J . 8, 3467-3475. 41. Maraganore, J. M., Chao, B., Joseph, M. L, Jablonski, J. & Ra-

machandran, K. L. (1989) J. Biol. Chem. 264, 8692-8698.