Activation of human prothrombin by a procoagulant fraction from the venom of Echis carinatus. Identification of a high molecular weight intermediate with thrombin activity.

In the presence of a procoagulant fraction (Echis carinatus procoagulant) isolated from the venom of the saw-scaled viper Echis carinatus sochureki, purified human prothrombin (P1) is completely converted to thrombin. The first step is the removal of an NH2-terminal peptide (F1) representing approximately one-third of the prothrombin molecule. The remaining peptide (P2) is then cleaved by the action of E.c. procoagulant to yield a two-chain, disulfide-bridged protein (P'2) which has the same molecular weight as P2. P'2 has enzymic (thrombin) activity, as evidence by incorporation of radiolabeled diisopropylphosphate into its heavy chain (TB), hydrolysis of p-toluenesulfonylarginine methyl ester, and clotting of fibrinogen. Relative to thrombin, its esterolytic activity greatly exceeds its clot-promoting activity. Examination of the polypeptide chains obtained by reducing P'2 has shown that its larger chain (TB) is indistinguishable from the heavy chain of thrombin. Its other chain (F2TA) consists of the light chain (TA) of thrombin bound by peptide linkage to the protion of the prothrombin molecule which had been adjacent to F1. Removal of this portion (F2) is catalyzed by thrombin (and, evidently, by P'2), but not by the E.c. procoagulant. When F2 is removed from P'2, the remaining two-chian protein is indistinguishable from thrombin by any of the criteria applied--molecular weight, subunit chain composition, or enzymic activity. Polyacrylamide gel electrophoresis was carried out in sodium dodecyl sulfate before and after disulfide reduction of samples generated in the presence and in the absence of diisopropylphosphorofluoridate, which inhibits thrombin but not the E.c. procoagulant. Such experiments showed that thrombin (and probably P'2), as well as E.c. procoagulant, catalyzes the release of F1. Furthermore, thrombin brings about the cleavage of F1 to yield a two-chain, disulfidebridged protein (F'1). These observations, particularly those made in the course of characterizine P'2, have led to the conclusion that cleavage of the peptide bond connecting the TA and TB portions of the prothrombin molecule (or its derivatives) produces a serine active center and, hence, a molecule possessing thrombin activity. This cleavage is catalyzed by the E.c. procoagulant but not by thrombon itself.


SUMMARY
In the presence of a procoagulant fraction (Echis carinafus procoagulant) isolated from the venom of the saw-scaled viper Echis carinatus sochureki, purified human prothrombin (PI) is completely converted to thrombin. The first step is the removal of an NH2-terminal peptide (F1) representing approximately one-third of the prothrombin molecule. The remaining peptide (P2) is then cleaved by the action of E.c. procoagulant to yield a two-chain, disulfide-bridged protein (P'2) which has the same molecular weight as PB. P', has enzymic (thrombin) activity, as evidenced by incorporation of radiolabeled diisopropylphosphate into its heavy chain (Tn), hydrolysis of p-toluenesulfonylarginine methyl ester, and clotting of fibrinogen. Relative to thrombin, its esterolytic activity greatly exceeds its clot-promoting activity.
Examination of the polypeptide chains obtained by reducing P'Z has shown that its larger chain (TB) is indistinguishable from the heavy chain of thrombin. Its other chain (F2T,) consists of the light chain (TA) of thrombin bound by peptide linkage to the portion of the prothrombin molecule which had been adjacent to F1. Removal of this portion (F2) is catalyzed by thrombin (and, evidently, by P'2), but not by the E.c. procoagulant. When Fz is removed from P'Z, the remaining two-chain protein is indistinguishable from thrombin by any of the criteria applied-molecular weight, subunit chain composition, or enzymic activity.
Polyacrylamide gel electrophoresis was carried out in sodium dodecyl sulfate before and after disulfide reduction of samples generated in the presence and in the absence of diisopropylphosphorofluoridate, which inhibits thrombin but not the E.c. procoagulant. Such experiments showed that thrombin (and probably P'Z), as well as E.c. procoagulant, catalyzes the release of F1. Furthermore, thrombin brings about the cleavage of F1 to yield a two-chain, disulfidebridged protein (F'l).
These observations, particularly those made in the course . . of characterlzmg P12, have led to the conclusion that cleavage of the peptide bond connecting the TA and TB portions of the prothrombin molecule (or its derivatives) produces a serine active center and, hence, a molecule possessing thrombin activity. This cleavage is catalyzed by the E.c. procoagulant but not by thrombin itself. Considerable effort has been spent on elucidating the pathway by which the zymogen, prothrombin (Pi in the present terminology) , is converted to the active enzyme, thrombin.
Measurement of the appearance of new NHz-terminal amino acids has not only demonstrated the proteolytic nature of the activation but has shown that with the conversion methods used this is a multi-step reaction (1, 2). Analysis by other methods has allowed further delineation of the conversion so that a general scheme is now agreed upon (3-8). The reaction sequence seems to be the loss, first, of approximately one-half of the mass of the prothrombin molecule (in either one or two steps) followed by the splitting of the remaining peptide into a light chain (A chain) and a heavy chain (the B chain).
The currently accepted peptide structure and the nomenclature to be used in this communication are presented in Fig. 1. Table I  compares this nomenclature with that used by others (4, 6, 7). The studies cited above utilized conversion by plasma procoagulants, with the generation of thrombin being dependent on the presence of activated Factor X (X,). Taipan snake (Ozyurunus scutellatus) venom also converts prothrombin (9) to thrombin and apparently cleaves the zymogen at the same bonds as does X, (10).
Venom of the saw-scaled viper (Echis curinatus sochureki) has also been shown by Kornalik (11) (20 U.S. units/ml) and then incubating the resultant clot with 0.1 ml of the fraction being tested and recording the time of clot lysis. Those fractions which exhibited a clot formation time of 17 s or less and a clot lysis time of greater than 24 hours were pooled and used as E.c. procoagulant (Fig. 2).2Two different chromatographic pools have been used throughout the course of the following experiments. The Asa of these pools was in the range of 0.090. Two bands were seen when a sample of E.c. procoagulant was concentrated 50-fold (Minicon B15, Amicon Corporation) and analyzed by Na dodecyl-S04-electrophoresis. Prothrombin was measured by the two-stage prothrombin assay of Wagner et al. (20). Prothrombin two-stage reagent (Difco Laboratories) was used (21). This assay system, as well as those involving clot formation by E.c. procoagulant, was standardized with U.S. Standard thrombin (Lot B-3) so that 1 prothrombin unit gave 1 unit of thrombin.
Thrombin activity was measured by adding 0.1 ml of appropriately diluted sample (all dilutions in 0.15 M NaC1/0.02 M Tris-Cl, pH 7.4) to 0.2 ml of 0.25% human fibrinogen.
Tos-Arg-OMe was used to assay the esterase activity generated by the action of E.c. procoagulant on prothrombin. The sample to be tested was added to 3.0 ml of 0.01 M Tos-Arg-OMe in 0.15 M NaCl, and the pH was adjusted to 8.0 with NaOH. (acrylamide from Canalco) in the presence of 0.1% Na dodecyl-Sod. A current of 6 ma per gel was applied. Gels were stained with Coomassie blue (Brilliant blue R, Sigma Chemical Co.) in accordance with Weber and Osborn (22). Sample loads were within the range of 20 to 30 rg of protein per 5-mm diameter gel. Samples were prepared for electrophoresis by incubation for at least 2 hours at 37" after addition of 2 volumes of either a nonreducing medium (0.01 M Na phosphate, pH 7.0, 1.0% Na dodecyl-Sod) or a reducing medium (0.01 M Na phosphate, pH 7.0, 1.0% Na dodecyl-S04, 1.0% 2-mercaptoethanol) to 1 volume of sample so as to achieve a final protein concentration of 1 mg/ml.

EXPERIMENTS AND RESULTS
Activity Changes Produced by Treating Prothrombin with E.c. Procoagulant-Prior to each experiment the prothrombin was thawed, adjusted to pH 7.4 by the addition of 1 M Tris, and diluted with 0.15 M NaCl/O.OZ M Tris-Cl to a concentration of 3 mg/ml (2200 units/ml). Addition of 50 ~1 of E.c. procoagulant per ml of this prothrombin solution, followed by incubation, produced maximal thrombin activity (i.e. ability to clot fibrinogen) after approximately 150 min. However, the generation of this activity was nonlinear; an initial lag phase was followed by a rapid increase in thrombin activity over the 60-to 120-min interval (Fig. 3).
The Tos-Arg-OMe hydrolytic activity followed a markedly different pattern, with a rapid rise to a peak in 45 min followed by a gradual decrease (Fig. 3). A comparison of the thrombin activity with the Tos-Arg-OMe hydrolytic activity is presented in the inset of Fig. 3. During the interval of 15 to 30 min the ratio of Tos-Arg-OMe hydrolytic activity to thrombin activity (i.e. micromoles hydrolyzed per 100 min:U.S. thrombin units) was falling, but it averaged approximately 14; over 90 to 240 min it was stable at 0.90 f 0.05. For purified thrombin this ratio was usually 0.99. Two-stage prothrombin assays performed throughout the activation period showed that when the conversion was carried out in the presence of 2 mM iPrsP-F, E.c. procoagulant brought about a complete loss of prothrombin activity within 120 min (Fig. 4). In the absence of iPr,P-F, the prothrombin-E.c.
procoagulant mixture exhibited the "Cheshire Cat phenomenon,"a i.e. a rapid loss of two-stage activity reaching a minimum at approximately one hour, followed by a rise to the original level as thrombin (T) was formed.
Products Generated by Treating Prothrombin with E.c. Procoagulant or Factor X,-To investigate the chemical changes occurring as these activities were generated, samples taken at various stages during the activation of prothrombin by E.c. procoagulant were examined electrophoretically and chromatographically.
The elution patterns (from DEAE-cellulose) were indistinguishable from those observed (3) when Factor X, was used to catalyze the conversion. Furthermore, when the fragments formed during conversion with the E.c. procoagulant and those formed during activation with X, were compared by Na dodecyl-S04-electrophoresis, there was identity of all bands (Fig. 5A). When reduction was carried out prior to Na dodecyl-SO(-electrophoresis, the same bands were seen in both conversion systems (Fig. 5B), but their intensities and the temporal sequence of their appearance differed greatly.
It has recently been reported that the nature of the fragments 3 The term "Cheshire Cat phenomenon" was first applied by Dr. J. H. Milstone (circa 1939) to describe the apparent disappearance of two-stage activity and its subsequent reappearance. Though the term was never published by Dr. Milstone, its historical priority and its descriptive appropriateness justify its continued usage.
found after activation of bovine prothrombin by Factor X, is affected by the presence of iPrzP-F (4, 23). That is, when iPmP-F is present, activation peptides Fi and Ft occur as a single chain (Fn) rather than as the two separate peptides found in the presence of active thrombin. Furthermore, it had been reported that the E.c. procoagulant activity is not inhibited by iPrpP-F (12). This was confirmed by an experiment in which 1 M iPrzP-F was added to E.c. procoagulant (final concentration of iPrtP-F, 2 mM). After 30 min, 100 ~1 of this mixture was added to 0.2 ml of human plasma. The clotting time, 25 s, was identical with that of the control sample (i.e. involving E.c. procoagulant not treated with iPrzP-F). Therefore, iPreP-F was used to inhibit thrombin as it was formed by the action of the E.c. procoagulant.
A solution of prothrombin was made 2 mM in iPr#-F and treated with E.c. procoagulant (50 ~1 of procoagulant/ml of prothrombin solution). Under these conditions, none of the F1z (which would have appeared as a band migrating slightly slower than P3) was found. Fig. 6A shows that in the presence of 2 mM iPrzP-F there was minimal formation of thrombin despite the complete disappearance of the original prothrombin and the appearance of bands in the PZ and F1 positions.
Intermediate PI2 and Its Constituent Chain FtT,--Upon reduction, most of the PZ band disappeared, and two new bands became apparent (Fig. 6B). These corresponded to the B chain of thrombin (Ts) and component F2TA, respectively. The latter migrated somewhat faster than F1. Since component PZ is a single polypeptide chain, it was clear that the PZ band must have contained a two-chain (disulfide-bridged) protein which dissociated to yield Te and F2T,. This two-chain intermediate will be referred to as P'z; its structure appears to be F2TA---S-S-Ts. (Additional evidence regarding its formation and structure is presented later.) The small amount of bona$de PZ formed, disappeared with increasing incubation time. P'P was also formed in the absence of iPreP-F, as evidenced by the presence of the FtTA chain at 30 and 60 min (Fig. 6B). After 120 min the FzTA had essentially disappeared, concurrent with the appearance of FZ and TA.4 In the presence of iPrzP-F no F2 or T* was formed, even after 4 hours of incubation, and the level of FtT, remained essentially constant. ELUATE VOLUME (ml) only after reduction, it was clear that they arose from a two-chain that samples taken after appropriate incubation times (viz. 120 derivative of F1; the latter has been designated F'1 ( Fig. 6; cf. min with E.c. procoagulant, 480 min with X3 exhibited Sub- Fig. 1). fragment p bands of similar intensities (Fig. 5B, gels 1 and 2). The fact that Subfragments (Y and p were not formed in the presence of iPrzP-F (Fig. 6B) indicated that they were generated from F1 by the thrombin produced during the activation. To demonstrate this generation directly, F1 was prepared in purified form and allowed to react with thrombin.
Thrombin, devoid of X, activity, was added to prothrombin at a level of 50 units/ml. At the end of 1 hour the mixture was made 2 mM in iPrzP-F and chromatographed on DEAE-cellulose (Fig. 7). Material eluted near the end of the final peak was pooled and concentrated on a Minicon B15 filter to a concentration of 1.6 mg/ml. Na dodecyl-SOJ-electrophoresis before and after reduction showed that it consisted primarily of F1. A 1.4-ml portion of this solution was treated with 0.14 ml of a thrombin solution (final concentration 1000 units/ml), and samples were taken for Na dodecyl-S04electrophoresis after various periods of incubation. Electrophoresis of unreduced samples revealed no changes other than the disappearance of the small amount of Pp contaminating the F1 preparation and the presence of the added thrombin itself (Fig. 8). Upon reduction, however, the decrease in F1 during the 0-to 240-min interval was readily apparent, as was the concomitant production of Subfragments a! and @. The positive identification of these subfragments permitted a more valid comparison between the actions of E.c. procoagulant and Factor X, under the present experimental conditions. Whereas X, brought about the production of Pz, Pa, F1, and some FP within 120 min, no degradation of F1 to Subfragments cr and /3 was apparent until much later (Fig. 6C). By contrast, little F1 remained after 120 min of incubation of prothrombin with E.c. procoagulant (Fig. 6B) Moreover, the chromatographic system employed for isolating the F1 cleaved from prothrombin by thrombin was suitable for obtaining Pz (Fig. 7). Such PP was therefore prepared and purified, and subsequently treated with E.c. procoagulant in the presence and the absence of iPr*P-F. Thrombin, devoid of X, activity, was added to prothrombin (final thrombin concentration 50 units/ml) and the mixture was incubated for 1 hour. Following this, iPrsP-F was added (final concentration 2 mM) and chromatography on DEAE-cellulose was carried out. The material eluted just ahead of prothrombin was pooled and brought to a concentration of 1.8 mg/ml with a Minicon B15 filter (Fig. 7). Na dodecyl-S04-electrophoresis before and after reduction showed that it consisted entirely of Pz. E.c. procoagulant (final concentration 50 &ml) was added to this solution in the presence and in the absence of 2 mM iPmP-F. In the absence of iPrzP-F, thrombin activity was generated at a rate equal to that observed when intact prothrombin was treated with E.c. procoagulant.
Results of Na dodecyl-SOI-electrophoresis performed on samples taken at specified intervals are shown in Fig. 9.
In the presence of iPrgP-F no appreciable change was detected in the unreduced samples; in its absence new bands appeared in the positions of thrombin (T) and Fz, respectively (Fig. 9A). Na dodecyl-S04-electrophoresis of the reduced samples (Fig. 9B) revealed that even in the presence of iPrzP-F the E.c. procoagulant had cleaved Pz, hence reduction resulted in the appearance of a band indistinguishable from the heavy chain of thrombin Reduced samples are designated by BME (2-mercapto-material and must have arisen from the Pp band. The finding that ethanol) below the incubation time, which is given in minutes. it was converted to P, (cf. 0, 2-mercaptoethanol and 60, 2-mercap-In a subsequent experiment the same samples were subjected to toethanol) suggests that thrombin can cleave a peptide bond electrophoresis in Na dodecyl-SO., and then stained for glycopro-within a disulfide-bridged region of the Ff moiety. The lower arrow tein by the method of Zacharius et al. (24). The following (unre-indicates a contaminant which was present in the thrombin added. (i.e. Tn) and a second band which migrated slower than Fz. When iPrzP-F was absent from the incubation mixture, this second band disappeared, concurrent with the appearance of FP. Na dodecyl-S04-electrophoresis of combinations of reduced samples showed it to be identical with the FsTA chain formed by the action of E.c. procoagulant on prothrombin (Fig. 6B). Thus E.c. procoagulant per se was responsible for cleavage of the peptide bond between FzTA and Tn (yielding exclusively I"2 when the starting material was Pz), whereas the active thrombin produced in the absence of iPrzP-F catalyzed the removal of Fz (see Fig. 1).
Since early work (1,2) had shown that a single cleavage (yielding an isoleucine NH2 terminus) was associated with the generation of thrombin activity, it now became important to determine whether the two-chain protein, P'z, produced by E.c. procoagulant in the present study possessed such activity. Inasmuch as the experiments described above (Fig. 6B, inter &a) demonstrated that P's remains stable only under conditions in which thrombin is inhibited (e.g. in the presence of iPr,P-F), prothrombin was treated for 2 hours with E.c. procoagulant in the presence of iPrzP-F and immediately chromatographed on DEAE-cellulose (Fig. 10). To ensure complete removal of the E.c. procoagulant, the column was developed with 50 ml of the starting buffer (0.15 M NaCV0.02 M Tris-Cl, pH 7.4) prior to initiating the gradient (cf. Fig. 7). The elution pattern of this sample was similar to that resulting from DEAE-cellulose chromatography of prothrombin treated with thrombin (Fig. 7). The fractions containing P', (which appears at the same point in the elution pattern as Pz) were pooled and concentrated on a Minicon B15 filter. This pool5 (P'*,Pz in Fig. 10) and the F1 pool (Fig. 10) were compared with the analogous derivatives resulting from the action of thrombin on prothrombin (Fig. 7). Fragments Fi produced by E.c. procoagulant and by thrombin were found to be identical in their electrophoretic properties, as were P2 and P', , PZ (Fig. 11). When the chromatographically isolated P'2,Pe was incubated with fibrinogen, no thrombin activity (i.e. clotting) was detected. Electrophoretic examination before and after reduction (Fig. 12, 0 min) showed that this pool consisted primarily of the two-chain species, P'*. To P'2,Pa (concentration 3 mg/ml) was added either thrombin (final concentration 100 units/ml) or E.c. procoagulant (50 &ml). Incubation with the thrombin over a period of 240 min resulted in the appearance of F2, with only a slight decrease in the Pz band (Fig. 12B, With Thrombin) and no generation of thrombin activity.
In contrast, incubation with E.c. procoagulant brought about a rapid decrease in P:! (Fig. 12B in the presence of iPrsP-F and therefore initially possessed no thrombin activity. Since this generation accompanied a decrease in PZ (which was readily brought about by E.c. procoagulant but not by thrombin at the levels used in this experiment (see Fig. 12B)), the active species was evidently formed from Pp. In view of the fact that Fz appeared only after the production of thrombin activity, the early clotting activity must not have been due to thrombin (T) (which could have arisen only by removal of Fz from P'Z or P2) but rather to the PfZ formed from PZ by the action of E.c. procoagulant.
The Fz initially produced must then have been cleaved off by the enzymically active P'Z. In contrast to these results were the observations made when iPrzP-F was present in the incubation mixture. Under this latter condition neither thrombin activity nor Fs was generated regardless of whether the starting material was purified PS (Fig. 9A) or prothrombin itself (Fig. 6A). These findings indicated that the P'z generated by E.c. procoagulant must, like thrombin, have a serine active center which can react with iPrzP-F. If this reaction (which results in the incorporation of iPrzP) occurs immediately upon formation of P'z, no thrombin activity should be detectable, and peptides which are released by thrombin (but not by E.c. procoagulant) should not appear. Activity and Active Center of P'g-Although the results of the foregoing experiments could be adequately accounted for by this explanation, they did not provide direct evidence for the serine active center of P'z it presupposes. The following pair of experiments was therefore carried out. 11. Comparison of unreduced chromatographic fractions isolated after treatment of prothrombin with thrombin (see Fig. 7) or with E.c. procoagulant (see Fig. 10). Samples (1 mg of protein/ ml) were incubated in Na phosphate/Na dodecyl-SOa prior to electrophoresis.
Gel 9 shows a mixture of samples 1 + 2; Gel 6, a mixture of 4 + 6. Gel 7 shows the 240 min (unreduced) sample obtained by treating prothrombin with E.c. procoagulant in the presence of 2 mM iPrzP-F (see Fig. 6A). spectively, samples were removed and prepared for Na dodecyl-SOa-electrophoresis.
In the second experiment, the prothrombin and E.c. procoagulant were combined at these concentrations and incubated without iPrzP-F. At the intervals listed above, 0.2ml portions of the incubation mixture were mixed with l-p1 aliquots of the [a*P]iPr2P-F solution, held at 22" for 5 min, and then prepared for Na dodecyl-SO4-electrophoresis.
After electrophoresis the gels were cut into 1.2-mm segments, which were placed in vials containing 10 ml of Aquasol and counted. Fig. 13 shows the gels and the distribution of radioactivity found. to activate the abnormal prothrombin appears to be a function of its independence from the effects of ionic calcium (12). That is, Gitel et al. (26) have shown the importance of Fi in the interaction of normal prothrombin with Factor X,phospholipid-Ca2+; the portion of the prothrombin molecule represented by this fragment is altered by vitamin K antagonists (14, 27). On the basis of these findings one would predict that removal of Fi altogether should yield species which X,-phospholipid-Ca2f could convert to thrombin only slowly, but which an E.c. procoagulant system could activate readily. This prediction was verified by measuring two-stage activity during the incubation of prothrombin with E.c. procoagulant, though additional experiments were required to delineate individual steps in the activation. With the release of Fi (Fig. 6A) the two-stage activity dropped precipitously (Fig. 4), in agreement with the findings of Kisiel and Hanahan (6). When iPrlP-F was present in the E.c. procoagulant system, this decrease continued, owing to the formation of inactivated P'z. In the absence of iPrzP-F, however, two-stage activity reappeared (Fig. 4), coincident with a decrease in the amount of P'z, the disappearance of Pz, and the release of Fz (Fig. 6B). This paradoxical (Cheshire cat) phenomenon arises from the facts that (a) the two-stage assay used here measures both prothrombin and thrombin activity and (b) the thrombin activity (i.e. clotting activity) of P'z is less than that of thrombin (T) . As P'z, which is formed quite early (Fig. 6) and exhibits much more Tos-Arg-OMe hydrolytic activity than clotting activity (Fig. 3)) is converted to thrombin, considerable clotting activity (and hence two-stage activity) is developed (Figs. 3 and 4,60 to 120 min). As shown above (Fig.  12), conversion of P'z to thrombin can be brought about by thrombin and, evidently, by P'z itself; however, in the first stage of the two-stage assay (a 3-to 4-minute incubation in the method used here) catalysis of this conversion by the Factor X, complex must have been minimal.
Such differences in the abilities of intermediate activation products to serve as substrates for the X, complex or thrombin may underlie the observation made by Schieck et al. (13)  A, unreduced samples. B, reduced samples. Numbers under gels specify minutes of incubation.
The upper arrow indicates an unidentified component which does not appear to be Pt inas-much as it resolved from T before reduction and was not seen after reduction.
Its structure may be related to that of the component indicated by the upper arrow in Fig. 8