THE JOURNAL OF Vol. Issue pp. for Printed in U.S.A. The ... · In addition, we have used rapid chemical quench and physical assessment of bond cleav- age to evaluate the effects of

THE JOURNAL OF BIOLM~~CAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular ' Biology, Inc.

Vol. 269, No. 44, Issue of November 4, pp. 2744-27450, 1994 Printed in U.S.A.

The Activation of Prothrombin by the Prothrombinase Complex THE CONTRIBUTION OF THE SUBSTRATE-MEMBRANE INTERACTION TO CATAZ;YSIS*

(Received for publication, June 24, 1994, and in revised form, September 8, 1994)

Randall K. Walker$ and Sriram KrishnaswamyP From the Department of Medicine, Division of HematologylOncology, Emory University, Atlanta, Georgia 30322

The conversion of prothrombin to thrombin requires the cleavage of two peptide bonds and is catalyzed by the prothrombinase complex composed of factors Xa and Va assembled on a membrane surface. Presteady- state kinetic studies of the effects of membranes on the proteolytic reaction were undertaken using model membranes composed of phosphatidylcholine and phosphatidylserine (PCPs). The concentration of PCPs was varied to alter the concentration of free phospholipid available for substrate binding without influencing the concentration of membrane-assembled prothrombinase. In fluorescence stopped-flow measurements, increasing concentrations of PCPs resulted in an increase in the rate of product formation. Assessment of bond cleavage by sodium dodecyl sulfate-polyacrylamide gel electrophoresis following rapid chemical quench using lzaI-prothrombin revealed that the activation reaction pro- ceeded through the ordered cleavage at ArgSrn-IleaZ4 followed by cleavage at Ar274-ThP at all concentrations of PCPs. Increasing the PCPs concentration resulted in a large increase in the Ar$2S-IleS24 cleavage reaction with a much smaller effect on the subsequent cleavage at A r ~ 7 4 - T h 9 7 5 , thereby leading to an increase in the extent of accumulation of the intermediate, meizothrombin. Fluorescence stopped-flow and rapid chemical quench measurements were also conducted using prethrombin 2 plus fragment 1.2 or meizothrombin as substrates to assess the influence of PCPs on the individual cleavage reactions. The rate of cleavage at ArgSrn-Ile3" by prothrombinase was increased "60-fold with increasing PCPs, whereas the cleavage at Ar2"- T h P 6 was increased by a factor of -5. These differential effects of PCPs on the two cleavage reactions adequately explain changes in the extent of accumulation of meizothrombin during prothrombin activation. Pro- teolytic removal of the membrane binding fragment 1 domain of the substrates, meizothrombin and prethrombin 2-fragment 1.2, had no effect on the cleavage at Arg274-ThP at saturating PCPs but completely eliminated the membrane-dependent rate enhancement for cleavage at ArgSz3-Ile324. Thus, membrane binding by the substrate is essential for the first cleavage reaction at

Grants HL-47465 and HL-52883 (to S. K.). Apreliminary account of this * This work was supported in part by National Institutes of Health

work was presented in poster form at the annual meeting of the ASBMB, San Diego, CA, May 30-June 3, 1993 (1). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ This work was completed in partial fulfillment for the Ph.D. degree from the Department of Biochemistry, University of Vermont, Given Health Sciences Center, Burlington, VT 05045. Present address: Sealy Center for Molecular Science, University of Texas Medical Branch, Galveston, TX 77555.

Oncology, Dept. of Medicine, Drawer A J , Emory University, Atlanta, GA 5 To whom correspondence should be addressed: Hematology/

30322. Tel.: 404-727-3806, Fax: 404-727-3404.

Ar$29-Ile324, which leads to the conversion of prothrombin to meizothrombin. In contrast, the substrate- membrane interaction mediated by the fragment 1 domain has no detectable effect on the second cleavage reaction at Ar&74-Thf176 which is required for the conversion of meizothrombin to thrombin.

The proteolytic activation of prothrombin to the serine protease a-thrombin is catalyzed by the prothrombinase complex (2-4). This enzyme complex forms through a calcium-dependent, reversible association of a protease, factor Xa, and a protein cofactor, factor Va, on a membrane surface.

Membranes containing anionic phospholipids are required for the optimal function of this enzyme complex (2, 5, 6). The phospholipid membrane facilitates the macromolecular assembly of the prothrombinase complex by providing a surface to which factors Xa and Va can reversibly bind ( 2 4 ) . Phospholipid membranes also reversibly bind prothrombin and thus localize substrate and enzyme on a surface (7).

The catalytic efficiency of prothrombin activation catalyzed by factor Xa alone is lower than the equivalent reaction catalyzed by the prothrombinase complex by a factor of -lo5 (8-10). The increased catalytic efficiency of prothrombinase results from a decrease in Km,app (-100-fold) and an increase in k,, (-3,000-fold) and reflects contributions of both factor Va and membranes (10, 11).

The increased kc,, of prothrombinase has been attributed to an effect of factor Va, evidently related to its ability to interact with both prothrombin and factor Xa (10, 11). The mechanistic basis for the - 100-fold decrease in Km,app for prothrombin activation following prothrombinase assembly is controversial (10- 13). In one kinetic model, the decrease in Km,app results from the ability of both enzyme and substrate to bind membranes, which raises the local prothrombin concentration in the vicinity of the enzyme complex (10, 11). In this case, membrane-bound prothrombinase would preferentially activate membrane-bound substrate and is consistent with the recent demonstration that diffusion of prothrombin to the membrane surface is rate- limiting in the action of prothrombinase assembled on planar lipid bilayers (14). In contrast, other initial velocity studies have shown that the Km,app is independent of the surface density of prothrombin and have concluded that the decreased Km,app is a result of a higher affinity of the assembled prothrombinase complex for prothrombin (13). The model derived from this study is consistent with the action of membrane-assembled prothrombinase on soluble prothrombin (12, 13). This explanation is consistent with the near equivalent kinetic constants for prothrombin activation obtained with membrane-assembled prothrombinase versus factor Xa saturated with high concentrations of factor Va in solution (15).

Prothrombin activation catalyzed by the prothrombinase complex proceeds via cleavage at Arg323-Ile324 followed by cleavage at A I - ~ ' ~ - T ~ ? ~ ~ (Scheme I), and meizothrombin transiently

27441

27442 Presteady-state Kinetics of Prothrombin Activation Meizothrombin

s-s

Prothrombin Thrombin

Fragment 1.2 . SCHEME I

accumulates as the sole or principal intermediate of this reaction (16-18). However, the extent of meizothrombin accumulation has been reported to vary with experimental conditions and is minimal when reactions are conducted on physiological membranes or in plasma (16, 19, 20). The mechanism(s) un- derlying these changes are unknown (20). Further, kinetic constants for the cleavage of the individual bonds using isolated intermediates and prothrombinase are unable to predict the observed kinetic constants for prothrombin activation or the ordered cleavage of the bonds in prothrombin (18).

These discrepancies in the literature are a t least partly related to a failure of rapid equilibrium type assumptions, the difficulties associated with using steady-state kinetic measurements to differentiate between the influence of membranes on complex assembly versus catalysis by the assembled complex, and the complexities associated with the nonproductive binding of substrate to membrane vesicles lacking assembled enzyme. The interpretation of steady-state measurements of prothrombin activation is further complicated by the fact that the conversion of prothrombin to thrombin requires two proteolytic cleavages (Scheme I). The intermediates and product differ in their abilities to cleave synthetic peptidyl substrates and in their interaction with active-site directed inhibitors such as dansyl-arginine-N-(3-ethyl-1,5-pentanediyl)amide (DAPA)' (21-23). Thus, measurements of prothrombin activation, which generally rely on the appearance of an active site, reflect differential contributions of the products of at least two enzymic reactions, and the significance of the measured initial velocity and derived kinetic constants is difficult to interpret.

In the present study we have investigated the effects of PCPs membranes on the rate of prothrombin activation by the prothrombinase complex using rapid kinetic techniques. This approach was used to permit measurements of reaction rate at appropriate reactant concentrations so that the free concentration of phospholipid could be varied without altering the concentration of assembled prothrombinase while assuring that all vesicles contained bound enzyme. In addition, we have used rapid chemical quench and physical assessment of bond cleavage to evaluate the effects of membranes on each of the two cleavages catalyzed by the prothrombinase complex.

' The abbreviations used are: DAPA, dansyl-arginine-N-(3-ethyl- 1,5-pentanediyl) amide; dansyl, 5-dimethylaminonaphthalene-1-sul- fonyl; DEGRck, dansyl-glutamyl-glycyl-arginine chloromethyl ketone; DEGR-Xa, Factor Xa modified with DEGRck; DEGR-mIIa, meizothrombin modified with DEGRck; Ecarin, prothrombin-activating protease isolated from E. carinatus venom; FPRck, o-phenylalanyl-prolyl-arginine chloromethyl ketone; 12581, N-dansyl-(p-guanidin0)-phenylalanine piperidide; PCPs, small unilamellar vesicles composed of 75% (w/w) L-a-phosphatidylcholine and 25% (w/w) L-a-phosphatidylserine; PAGE, polyacrylamide gel electrophoresis.

EXPERIMENTAL PROCEDURES

Materials Hepes, Tris base, hen egg L-a-phosphatidylcholine, bovine brain ~ - a -

phosphatidylserine, 1yophilizedEchis carinatus venom, soybean trypsin inhibitor, and Sephadex resins were purchased from Sigma. Dansyl- glutamyl-glycinyl-arginyl chloromethyl ketone (DEGRck) was from Calbiochem, and o-phenylalanylprolylarginyl chloromethyl ketone (FPRck) was from Bachem. N-Dansyl-(p-guanidin0)-phenylalanine piperidide (12581) was purchased from Chromogenix. Na**'I was obtained from Amersham Corp. Lactoperoxidase immobilized on Sepharose was obtained from Worthington. H-o-phenylalanyl-L-pipecoyl-L-arginine-p- nitroanilide (S2238) was from KabiVitrium. The fluorescent inhibitor DAPA was prepared as described (24). Small unilamellar vesicles composed of 75% (w/w) L-a-phosphatidylcholine and 25% (w/w) L-a-phosphatidylserine (PCPs) were prepared and assayed as described previously (25). Concentrations of PCPs are expressed in terms of monomeric phospholipid. The size distribution of the vesicles was determined by quasielastic light scattering using a NICOMP 370 particle sizer (NICOMP, Santa Barbara, CA). All kinetic measurements were conducted in 20 mM Hepes, 0.15 M NaCl, 2 mM CaCl,, pH 7.4 (assay buffer).

Proteins Factor X and prothrombin were prepared from bovine plasma as

described (17,26). Bovine factor Va was purified as described previously with additional purification using S-Sepharose (27, 28). Factor X was activated with the purified factor X activator from Russell's viper venom, and factor Xa was isolated using benzamidine-Sepharose as described (18, 29). Kinetic titration of factor Xa preparations with p- nitrophenol-p'-guanidinobenzoate (30) yielded 1.1 mol of active sited mol of Xa. Prothrombin used in this study was further purified by anion and cation exchange chromatography on S-Sepharose and DEAE-cellu- lose to remove all traces of contaminating prothrombin activation products. The prothrombin activator (Ecarin) from E. carinatus venom was partially purified as described previously (22) followed by chromatography using benzamidine-Sepharose to remove contaminants with hy- drolytic activity toward 52238. The preparation and purification of the prothrombin derivatives prethrombin 2 (Pre 2), fragment 1.2 (F1.2), fragment 2 (F2), and a-thrombin (IIa) have been described (31, 32). Meizothrombin was prepared by reacting prothrombin with Ecarin in the presence of DEGRck followed by purification as described to yield the fluorescent and inactivated derivatives of meizothrombin (DEGR- mIIa) or meizothrombin lacking the fragment 1 domain (DEGR-mIIa des F1) (15). Residual S2238 activity of the meizothrombin derivatives was below the detection limit and was estimated a t <0.01% that of thrombin. Purity of all proteins was assessed by SDS-PAGE followed by staining with Coomassie Brilliant Blue (33).

Protein concentrations were determined using the following molecular weights and extinction coefficients (E:$: prothrombin, 72,000, 1.44; mIIa, 72,000, 1.44; mIIa des F1, 50,200, 1.64; Pre 2, 37,400, 1.95; F1.2,34,800,1.12; F2,12,800,1.25 (31); IIa, 37,400,1.95 (32); factor Va, 168,000, 1.74 (27, 34); factor Xa, 45,300, 1.24 (35, 36).

Steady-state Anisotropy Measurements Fluorescence anisotropy was measured in T-format using an SLM

8000C fluorescence spectrophotometer (SLM Instruments, Urbana, IL). Reaction mixtures (2 ml) in 1 x 1-cm stirred cells maintained a t 25 "C contained 0.3 PM DEGR-Xa, 0.45 p~ Va with and without 0.3 PM prothrombin in assay buffer. Titrations were performed by incremental microliter additions of PCPs to the indicated concentrations. Fluores- cence anisotropy was measured as described previously using A,, = 330 nm, A, = 545 nm with long pass filters (KV-500, Schott, Duyrea, PA) in the emission beam (37). The incremental change in anisotropy (Ar) was determined by subtraction of the measured anisotropy in the absence of PCPS.

Fluorescence Stopped-flow Measurements Stopped-flow measurements were performed using a flow box (Kinet-

ic Instruments, Ann Arbor, MI) attached to the SLM 8000C fluorescence spectrophotometer as described (37, 38). Ratiometric fluorescence was monitored a t 25 "C using A,, = 280 nm and measuring broadband fluorescence (Aern > 520 nm) with a long pass KV-520 filter (Schott) in the emission path. mica l ly , eight replicate traces were averaged for each experimental condition and normalized using the t = 0 datum and a limiting signal collected 1-4 min following initiation of the reaction.

Measurements of prothrombin activation by prothrombinase were performed using DAPA and exploiting the large increase in fluorescence

Presteady-state Kinetics of Prothrombin Activation 27443 intensity which accompanies the binding of this fluorophore to the products (meizothrombin and thrombin) of the enzymatic reaction (22, 24). Reactions were initiated by rapidly mixing equal volumes of substrate solution (syringe A: 0.6 1.1~ prothrombin, 3 PM DAPA) with enzyme solution containing preassembled prothrombinase (syringe B: 0.6 p Xa, 0.9 p Va, 3 p DAPA, and the indicated Concentrations of PCPs). The final concentrations of reactants were therefore half the concentrations in each driving syringe.

The cleavage of DEGR-mIIa and DEGR-mIIa des F1 was studied using the same experimental design at the same concentrations of substrate and enzyme components except that these reactions were conducted in the absence of DAPA. Reaction progress was monitored by the decrease in fluorescence intensity of the dansyl dye which accompanies the conversion of these substrates to thrombin (15). Stopped-flow traces with DEGR-mIIa were characterized by a rapid initial increase in fluorescence over the first 50-80 ms. Control experiments established that this short lived fluorescence transient was unrelated to catalysis and probably reflected the binding of this substrate to PCPs andlor factor Va as noted previously (39). Therefore, the first 100 ms was eliminated from the stopped-flow traces prior to analysis to derive the rate constant for the cleavage reaction.

Analysis of Prothrombin Cleavage by SDS-PAGE The cleavage of prothrombin was directly assessed by SDS-PAGE

essentially using procedures developed previously for studies of bovine and human prothrombin activation (17, 18). Samples were prepared in 62.5 m Tris, 2% (w/v) SDS, 0.01% (w/v) bromphenol blue, 10% (v/v) glycerol, pH 6.8 (SDS sample buffer), in the absence (nonreduced) and presence (reduced) of 100 rm dithiothreitol followed by heating for 5 min at 80 "C. The samples were analyzed by electrophoresis in the buffer system of Laemmli (33), using gels (11 x 15 x 0.075 cm) composed of 9.5% acrylamide and 0.5% NJV"methylenebisacry1amide (10% T) at constant current. Protein bands were visualized by staining with Coomassie Brilliant Blue followed by destaining. Quantitation of stained protein was performed by laser densitometry (Molecular Dy- namics). Radioactivity was detected and quantitated using a Phosphor- Imager (Molecular Dynamics). Dried gels were exposed to an imaging plate followed by detection of luminescence by laser scanning. The densitometry images were quantitated by volume integration (density integrated over surface area) of bands. The integrated volumes were corrected for variability in the background using average pixel densities in the areas immediately adjacent to each band. For experiments with radiolabeled protein, control experiments established a linear relationship between the radioactivity loaded on the gel and the integrated volume. In addition, analysis by counting excised bands gave equivalent results.

Modification of Prothrombin with Iz5I

Prothrombin was labeled with lZ5I using immobilized lactoperoxidase (40). Lactoperoxidase-Sepharose (2 IU/ml packed bed) was washed with 0.2 M sodium acetate, pH 6.8, by repeated centrifugation prior to use. The iodination mixture (50 p1) contained lactoperoxidase beads (0.8 IU/ml reaction), 14 PM prothrombin, 10 rm KI, 1 mCi NalZ5I, and 28 1" H,O, in the same buffer. The reaction mixture was gently agitated at room temperature for 6 min, quenched by the addition of 15 mM NaN,, and chromatographed using Sephadex G-25-150 (0.4 x 24 cm) presatu- rated with unlabeled prothrombin and equilibrated in assay buffer. Radioactive protein appearing in the void volume was pooled, combined with an equal volume of glycerol, and stored at -20 "C. Typical yields were -2.2 x lo5 dpdpg protein. For each preparation, relative specific activities of the relevant cleavage products of prothrombin were determined following quantitative cleavage of reaction mixtures containing '251-prothrombin with Ecarin, factor Xa, thrombin, and prothrombinase followed by SDS-PAGE analysis and quantitation of radioactivity associated with the products. Typical relative specific activities (cpdmol) normalized to prothrombin were: prothrombin, 1.0; meizothrombin, 1.0; thrombin, 0.68; prethrombin 2, 0.68; fragment 1.2, 0.32; fragment 1, 0.22; fragment 2, 0.1; thrombin B chain, 0.58; thrombin A chain, 0.1; fragment 1.2-thrombin A chain, 0.42.

Comparison of lZ5I-Labeled and Unlabeled Prothrombin The validity of using lZ5I-labeled prothrombin to study the kinetics of

bond cleavage in the native substrate was established for each lZ5I- prothrombin preparation by comparing the fate of the labeled material with unmodified prothrombin during activation by prothrombinase. The conditions for this comparison were chosen on the basis of published results illustrating the fate of prothrombin, the intermediates, and thrombin under steady-state conditions (17, 18).

A reaction mixture (15 ml) containing 1.4 1.1~ prothrombin, lZ5I- labeled prothrombin (lo5 cpdml, 3.0 PM DAPA, 30 y PCPs, and 10 nM Va in assay buffer was initiated by the addition of 1 nM Xa. Aliquots (0.5 ml) were withdrawn at various times following initiation (18 aliquots, 0-20 min) and quenched by mixing with 0.5 ml of glacial acetic acid. The quenched samples were dialyzed against 0.2 M acetic acid, dried under vacuum in a centrifugal evaporator, and the residues were resuspended in 200 pl of SDS sample buffer. Aliquots (100 pl, 25 pg of total protein) of each sample were analyzed by SDS-PAGE with and without disulfide bond reduction followed by quantitation of unlabeled protein from stained gels or labeled protein by PhosphorImaging analysis (described above). The time courses for the disappearance of prothrombin, the transient appearance of meizothrombin, and accumulation of thrombin were equivalent to those described previously (17). The data obtained from analysis of stained protein were qualitatively and quantitatively indistinguishable from the results derived from the radioactive bands. It was concluded that under steady-state conditions, the kinetics of cleavage of lZ5I-labeled prothrombin is experimentally indistinguishable from that of unlabeled prothrombin.

Rapid Chemical Quench Measurements Rapid chemical quench experiments were conducted using a five-

syringe instrument (QFM-5, BioLogic, Echirolles, France) thermostat- ted at 25 "C. Reactions were conducted in the continuous flow mode so that reaction times could be controlled by varying syringe speeds without altering the delay lines.

Analysis by SDS-PAGE-The rate of bond cleavage in prothrombin was evaluated by SDS-PAGE of quenched samples containing 1251-prothrombin under reaction conditions designed to duplicate those used in the fluorescence stopped-flow measurements. Prothrombin activation was initiated by mixing enzyme solution (100 pl) containing preassembled prothrombinase (0.6 m Xa, 0.9 Va, 3 p~ DAPA, and variable concentrations of PCPs in assay buffer) with an equal volume of substrate solution (0.6 PM prothrombin, 2,500 cpdpl '251-prothrombin, 3 p~ DAPA in the same buffer). Following reaction for the indicated times, the mixture was rapidly mixed with 100 pl of quenching solution ( 10% (w/v) SDS, 50 mM EDTA, 25 PM FPRck, pH 7.4), collected, and stored at -20 "C prior to analysis. Quenched samples were prepared for SDS- PAGE by adjusting the concentration of buffer components to 83 mM Tris, pH 6.8,8.9% (v/v) glycerol, 0.013% (w/v) bromphenol blue with and without 58 mM dithiothreitol and analyzed by SDS-PAGE as described above.

Analysis by Fluorescence-The activation of prothrombin or prethrombin 2 was measured independently by rapid chemical quench followed by measurement of the increased fluorescence of DAPA complexed with the reaction products. For prothrombin activation, the substrate solution contained 0.6 p prothrombin. For studies with prethrombin 2, the substrate solution contained 0.6 p~ prethrombin 2 plus 1.2 p fragment 1.2 or 0.6 m prethrombin 2 plus 1.2 p~ fragment 2. Reactions were quenched at the indicated times by mixing with 100 pl of 0.5 M EDTA, 0.67 soybean trypsin inhibitor, pH 7.5. Immediately after collection quenched samples were diluted 4-fold into 20 mM Hepes, 0.15 M NaCl, 0.4 p~ soybean trypsin inhibitor, 0.13% (w/v) polyethylene glycol 8000, pH 7.4. Fluorescence intensity was measured in 1 x 1-cm quartz cuvettes using A,, = 280, A,, > 520 nm before ( F ) and after (F,) the addition of 20 FPRck to the sample to displace DAPAfrom active enzyme quantitatively (21). The fluorescence change was expressed as the ratio FIFO to correct for small variations in sample volume recovered from the quenched-flow instrument, normalized to the limiting value obtained at the 20-s time point and converted to concentration terms.

Data Analysis Calculation ofthe Concentration ofFree PCPS-The concentration of

free phospholipid (PCPs,,,) available for the substrate to bind was estimated according to Equation 1

PCPs,, = PCPs, - [X+.nE + (Va, - Xa,).n,I (Eq. 1)

where PCPs,, %, and Va, refer to total concentrations, nE is the empirically determined stoichiometry for the mol of phospholipid required to bind 1 mol of prothrombinase, and n,, is the stoichiometry for the interaction of factor Va with membranes. The published value of n,, = 42 determined from equilibrium binding measurements and stopped-flow kinetic studies under equivalent experimental conditions was used in the calculations (27, 38). Equation 1 requires the assumptions that 1 mol of factor Xa combines per mol of factor Va to form prothrombinase, Va is saturating relative to Xa, and the concentrations of both proteins are high relative to the Kd for their interaction on membranes.

27444 Presteady-state Kinetics of Prothrombin Activation Quantitation of '251-Prothrombin Activation-The concentrations of

prothrombin (11,) and the products, meizothrombin (mIIq), thrombin (IIa,), and fragment 1.2 (F1.2,) in the reaction mixture at time t following initiation were determined by quantitative densitometry using Equations 2-7.

11, = 11,. (f&)

mIIa, = F1.2AR. - (%ta?) IIa, = IIa,. - (kz)

(Eq. 3)

(Eq. 4)

(Eq. 5)

where fil, fdIa, fF1.=, fila, f,,,,, fFl, and fF2 represent the fractional specific activities of the various bands determined in separate experiments. The integrated volumes for bands corresponding to the indicated species (11, mIIa, F1.2A, IIa, F1.2, F1, and F2) were identified following SDS-PAGE with (subscript R) or without (subscript N R ) disulfide bond reduction. The volumes were normalized and converted to concentrations using the fractional specific activities, the total integrated volume for the lane (Total) and the initial substrate concentration (ST = 0.3 p). Mean values were calculated when the concentration of any species was determined in more than one way. The inclusion of F1 and F2 terms for the calculation of F1.2 provided a correction for the minor amounts of F1.2 degradation observed at long reaction times under the present experimental conditions. Only traces of prethrombin 2 were observed at very long reaction times even in heavily overexposed images in all studies with prothrombinase. This observation does not rule out significant flux via the initial formation of prethrombin 2 (Scheme I). Therefore, reactant flux through the initial formation of prethrombin 2 was estimated by comparing the initial velocity of the disappearance of prothrombin with the velocity for the appearance of meizothrombin. In addition, the concentration of thrombin formed via the initial formation of prethrombin 2 + fragment 1.2 (designated IIamt) was estimated by Equation 8.

1Iam,, = F1.2, - I14 (Eq. 8)

Cleavage of Meizothrombin or Prethrornbin 2Stopped-flow traces for the cleavage of DEGR-mIIa or DEGR-mIIa des F1 by prothrombinase were fit to a single exponential decay

Fobs = offset + Adoh*'' (Eq. 9)

where measured fluorescence (Fobs) as a function of time ( t ) yielded fitted values for offset, the amplitude of the fluorescence change ( A ) , and the observed rate constant (kobs). The fluorescence signal was converted to product concentration by assuming that the substrate was quantitatively converted to product at infinite time. When necessary, initial velocities for cleavage at Ar274-Th?75 were calculated using the relationship vi(Rz7I) = k,;A.

Progress curves for the disappearance of prothrombin or the cleavage of prethrombin 2 plus fragment 1.2 generated by rapid chemical quench experiments could be adequately described by a two-exponential decay (or rise)

pob = offset + Al~e-kl,obs't + & ~ - ~ Z . o b s " (Eq. 10)

where the concentration of prothrombin or product (Pobs) observed at time t was related to two amplitude terms (A, and A,) and two observed rate constants (kl,ob, and k2,0bJ. Initial velocities for cleavage at kgz3- Ile324 from these data using the relationship vi(M23) = kl,obB.Al + k2,0bs.A2.

RESULTS Measurement of the Membrane Requirement for Prothrombin-

ase Assembly-The concentration of monomeric PCPs needed to assemble 1 mol of the prothrombinase complex was determined by anisotropy measurements using DEGR-Xa as a re- porter group (Fig. 1). Increasing concentrations of PCPs led to

0.04 - 6.

U 0.02 - d /

~d

40 I 0.01 ~ P 0 0.0 g

0.00 1 30 60 90 120 150 180

[ P C P S l ~ ~ W4 , "

0 20 40 60 80 200 [PCPs] (PM)

FIG. 1. Assessment of the membrane requirements for prothrombinase assembly. Reaction mixtures containing 0.3 p DEGR-Xa and 0.45 p Va in 20 m~ Hepes, 0.15 M NaCl, 2 mM CaCl,, pH 7.4, were titrated with increasing concentrations of PCPs in the presence (0) or absence (0) of 0.3 prothrombin. Fluorescence anisotropy was measured at 25 "C using A, = 330 nm, A,, = 545 nm and Ar was calculated as described under "Experimental Procedures." The lines were drawn following linear regression analysis of the limits of the data to yield a stoichiometry of 89.3 mol of monomeric PCPS/mol of prothrombinase at saturation. Inset, predicted concentrations of PCPs,,, (dashed line) and assembled prothrombinase (solid line) as a function of the total PCPs concentration at 0.3 p~ factor Xa and 0.45 p factor Va.

a saturable increase in anisotropy reflecting the assembly of the prothrombinase complex (41). In agreement with previous observations, 0.3 p~ prothrombin had no effect on prothrombinase assembly (26). Since the experiments were conducted at fixed concentrations of DEGR-Xa and Va in excess (>300-fold) over the Kd for their interaction on the membrane surface, the stoichiometry for the membrane-prothrombinase interaction could be estimated graphically from the intersection point of straight lines drawn through the limits of the sharp saturation function obtained (Fig. 1). The validity of this approach has been established previously on the basis of the linked interactions leading to prothrombinase assembly (26). The calculated stoichiometry (nE, Equation 1) of 89.3 mol of PCPS/mol of prothrombinase at saturation is in agreement with the previously determined value of 88 under different experimental conditions (26). Thus, the concentrations of prothrombinase and free PCPs (PCPs,,,) can be calculated from the anisotropy data and from Equation 1 using the determined value of nE. Calcu- lations at 0.3 JIM Xa and 0.45 JIM Va (inset, Fig. 1) illustrate that increasing the concentration of total PCPs between 33 and 300 JIM results in a linear increase in PCPs,, with no change in the concentration of prothrombinase. Quasielastic light scattering measurements yielded a mean vesicle diameter of 21.1 * 9.0 nm. Calculations assuming a head group surface area of 14 k and a 60:40 distribution of phospholipids in the outer and inner leaflets, yield an average of 3,150 phospholipids/vesicle (42). Thus, with this experimental design, the concentration of prothrombinase is 300 nM with the concentration of vesicles varying between -11 and 95 I", ensuring that all vesicles contain assembled enzyme. Kinetic studies of prothrombin activation were therefore designed on the basis of these measurements to investigate the effects of membrane concentration on prothrombin activation by prothrombinase.

Stopped-flow Studies of Prothrombin Activation-Meas- urements of prothrombin activation by rapid kinetic techniques were necessitated by the high concentrations of factors Xa and Va required to ensure predictable changes in PCPs,,,, while maintaining a constant concentration of prothrombinase. Ex- periments were also conducted at 0.3 p~ prothrombin (So - E,) to reduce complexities due to the attainment of steady-state rates of product formation.

Presteady-state Kinetics of Prothrombin Activation 27445

1.2 I ’ I I

0.0 0.2 0.4 0.6 0.8 1 .o Time (s)

bin activation using DAPA. Prothrombin activation was initiated by FIG. 2. Stopped-flow fluorescence measurements of prothrom-

stopped-flow mixing of equal volumes of substrate solution (syringe A) containing 0.6 w prothrombin and 3 PM DAPA in 20 mM Hepes, 0.15 M NaCl, 2 mM CaCl,, pH 7.5, with preassembled enzyme solution (syringe B) containing 0.6 1.1~ Xa, 0.9 Va, and increasing concentrations of PCPs in the same buffer. Fluorescence intensity was monitored (Aex = 280 nm, A,, > 500 nm) at 25 “C as described under “Experimental Procedures.”All traces were normalized using the initial datum and the limiting fluorescence signal collected at the 20-s mark and were sys- tematically offset to facilitate presentation. The calculated free final concentrations of PCPs for the traces (bottom to top) were 0,17,42,117, and 267 PM.

Stopped-flow measurements of prothrombin activation detected using the fluorescent inhibitor, DAPA, are illustrated in Fig. 2. The traces were obtained by reacting prothrombin with preassembled prothrombinase using increasing concentrations of PCPs to vary PCPs,,,, between 0 and 267 m. The rate of product formation increased with increasing PCPs,,,, and reached a limiting rate at -117 m PCPs,,,. Since the concentration of prothrombinase was constant, the large change in the rate of product formation detected by DAPA with increasing PCPs,,,, indicates that the availability of free sites on the membrane surface directly influences prothrombin activation possibly through binding the substrate and mediating its delivery and/or presentation to prothrombinase. Prothrombin and the constituents of prothrombinase have been shown to exclude each other in binding to membranes (11). Thus, at 33 1.1~ PCPs (PCPs,,, - 0), the surface is saturated with prothrombinase with no sites available for prothrombin to bind. The low rate of product formation observed under these conditions suggests that direct encounters between solution-phase substrate and membrane-bound enzyme do not lead to efficient rates of product formation.

The progress curves also exhibited a lag phase that was dependent on the concentration of PCPs,,,, (Fig. 2). This suggests that a slow step precedes product formation as inferred from the increased fluorescence intensity of DAPA. Fluores- cence stopped-flow experiments (not shown) indicated that the binding of DAPA to thrombin was rapid (Kobs > 500 s-l, tw < 1.4 ms) under comparable pseudo first-order conditions. Thus, the lag phase is not a result of the slow binding of the indicator molecule to rapidly formed product. The evidence suggests that the effects of PCPs,, on prothrombin activation may be related to changes in the observed rate constant for steps that precede bond cleavage or product release by prothrombinase. However, the interpretation of the fluorescence traces is complicated by the unknown contributions of DAPA complexed to thrombin and the two possible intermediates (Scheme I) to the observed signal.

Kinetic Studies of Bond Cleavage during Prothrombin

Activation-The two possible pathways for prothrombin activation are illustrated in Scheme I. The substrate, intermediates, and products of prothrombin activation can all be resolved and quantitatively analyzed by SDS-PAGE (17). Un- ambiguous quantitation of the fate of prothrombin and its activation products in the stopped-flow measurements was therefore achieved by rapid chemical quench and SDS-PAGE using ‘251-prothrombin.

For rapid chemical quench, reactions were initiated by mixing prothrombin with preassembled prothrombinase using increasing concentrations of PCPs to vary PCPs,,,, between 0 and 267 PM without altering the concentration of prothrombinase. SDS-PAGE analysis of the quenched samples as described under “Experimental Procedures” yielded bands that could be ascribed to prothrombin, meizothrombin, and thrombin. Prog- ress curves describing the fate of these species following initiation of proteolysis are illustrated in Fig. 3.

Prothrombin disappeared monotonically, leading to the transient appearance of meizothrombin and the delayed accumulation of thrombin. These profiles are consistent with the sequential cleavage of the substrate by prothrombinase at A r $ 2 3

followed by cleavage at A r g 2 7 4 previously documented (17, 18). The disappearance of prothrombin showed a strong dependence on PCPs,,,, (Fig. 3A). The initial velocity increased by 60-fold and reached a limiting value at 117 PM PCPs,,,. Both the initial rate and the amplitude of the transient rise in meizothrombin also depended on PCPs,,,, (Fig. 3B). In contrast, progress curves for the accumulation of thrombin (Fig. 3C) showed a much smaller dependence on PCPs,,,,.

A band corresponding to prethrombin 2 was not detected over the time period illustrated a t any concentration of PCPs,,,, and the disappearance of prothrombin over the initial portion (0-50 ms) of the progress curves could quantitatively account for the appearance of meizothrombin over the same interval. The possible contribution of thrombin formation via the initial formation of prethrombin 2 was estimated as described under “Data Analysis.” Product formation via the prethrombin 2 pathway (Scheme I) represented less than 5% of the total thrombin formed a t any time. These observations suggest that product formation via the initial formation of prethrombin 2 is not significant at any phospholipid concentration under the present conditions.

Therefore, the data in Fig. 3A suggest that the rate of cleavage at Ar$23-Ile324 in prothrombin by prothrombinase is strongly dependent on the concentration of PCPs,,,,. However, the large increase in the extent of meizothrombin accumulation and the smaller effects of PCPs,,, on thrombin formation imply that the second proteolytic reaction (cleavage at ThP79 may exhibit a reduced dependence on PCPs,,,.

In contrast to fluorescence stopped-flow results (Fig. 2), progress curves describing the decrease in prothrombin as a function of time showed no obvious evidence of a lag. This observation suggested possible complexities arising from an extended dead time or delayed quenching reaction. The theoretical dead time of the instrument is -3 ms. The initial time points for the progress curves (Fig. 3 A ) are close to the zero points. In addition, eight different quenching reagents, including 10% (wh) trichloroacetic acid and 17.4 M acetic acid, all gave the same results. Although it is possible that the lag is obscured by a dead time that exactly matches its duration, the data also suggest that the lag phase in the fluorescence stopped-flow experiments may not reflect a slow step that precedes bond cleavage by prothrombinase.

The progress curves for prothrombin disappearance were also biphasic (Fig. 3 A ) with an initial rapid phase resulting in

27446 Presteady-state Kinetics of Prothrombin Activation

0.0 0.0 0.2 0.4 0.6 0.8 1.0 20.0

Time (s)

B ' ' h i E 3 0.10

a E

ON

- E

005 .* bl

E 0.00

0.0 0.2 0.4 0.6 0.8 1.0 20.0 Time (s)

0.0 0.0 0.2 0.4 0.6 0.8 1.0 20.0

Time (s)

equal volumes of substrate solution: 0.6 p prothrombin, lZ6I-prothrombin (2,500 cpdpl) and 3 p~ DAPAin 20 mM Hepes, 0.15 M NaCl, 2 mM CaCl,, FIG. 3. Physical assessment of bond cleavage during prothrombin activation. Rapid chemical quench studies were performed by mixing

pH 7.4, with enzyme solution: 0.6 PM Xa, 0.9 p Va preassembled with increasing concentrations of PCPs in the same buffer. Following incubation for the indicated times at 25 "C, the enzymatic reaction was rapidly quenched, subject to SDS-PAGE, and reaction progress was determined using a PhosphorImager. The concentration of PCPs was vaned to achieve PCPs,, concentrations of 0 (O), 17 p (O), 42 p~ (A), 67 p (A), 117 p (V), 167 p~ (V), and 267 p~ (13). Progress curves are illustrated for the substrate, prothrombin (panel A), the intermediate, meizothrombin (panel B), and the product, thrombin (panel C).

a burst of substrate depletion in the first 0.5 s followed by a much slower decrease to zero over the subsequent 20-30 s. The presence of a burst suggests that the rate of enzyme recycling is limited by a slow step in the reaction pathway following cleavage at Arg323-Ile324. The possible mechanisms that could account for this behavior include slow deacylation of an acylenzyme intermediate, slow dissociation of product from the enzyme, or a slow conformational change in the enzyme following each round of catalysis.

Comparison of Stopped-flow and Quenched-flow Measure- ments of Prothrombin Activation-A stopped-flow trace is compared with the predicted fluorescence calculated from the known concentrations of meizothrombin and thrombin determined by rapid chemical quench and SDS-PAGE analysis (Fig. 4). The expected fluorescence signal was calculated from concentration terms using relative fluorescence intensities of 1.0 and 1.5 determined previously for the thrombin-DAPA and meizothrombin-DAPA binary complexes (17). A progress curve obtained by rapid chemical quench followed by measurement of the fluorescence intensity of the products is also illustrated. Progress curves obtained by either rapid quench technique are experimentally indistinguishable from each other, further vali- dating the conclusions derived from the SDS-PAGE analysis of the kinetics of prothrombin activation using lZ5I-labeled protein. In contrast, the increase in DAPA fluorescence as measured by stopped-flow is significantly slower than the rapid quench measurements. The continuously measured fluorescence change, in fact, most closely approximates the appearance of thrombin (Fig. 3C) and indicates that meizothrombin is not detected by DAPA fluorescence at the reactant concentrations and time scale of the stopped-flow experiments. This is also suggested by the difference in the fluorescence signal obtained by the two techniques also illustrated in Fig. 4. Similar results were obtained in stopped-flow fluorescence measurements using I2581 as an indicator molecule. These observations suggests that DAPA cannot readily bind to andlor exhibit increased fluorescence with meizothrombin formed during the reaction, possibly because this product remains bound to the enzyme or because a slow conformational change is required following cleavage a t Arg323-Ile324 for an active site to be expressed. Taken together with the possible mechanisms pro- posed for the burst in the depletion of prothrombin (above), the data are consistent with our tentative conclusion that meizothrombin may remain tightly bound to prothrombinase or dis- sociate slowly from the enzyme after the cleavage of the Ile324 scissile bond.

Studies of the Cleavage Reactions Using Prothrombin Acti- vation Intermediates-Meizothrombin (Arg323-Ile324 cleaved,

Time (s) FIG. 4. Comparison of progress curves for prothrombin activa-

tion obtained by fluorescence stopped-flow and rapid chemical quench. Prothrombin activation was measured with final reactant concentrations of 0.3 p~ prothrombin, 3 p DAPA, 0.3 p Xa, 0.45 p Va, 150 PM PCPs (PCPs,,, = 117 PM) in 20 mM Hepes, 0.15 M NaCl, 2 m~ CaCI,, pH 7.4. A stopped-flow fluorescence trace with continuous moni- toring is compared with the results obtained by rapid chemical quench followed by direct fluorescence measurement of the quenched samples (0) or the fluorescence signal predicted from the concentrations of thrombin and meizothrombin determined under the same experimental conditions by rapid chemical quench and SDS-PAGE analysis (0). All three data sets were normalized to the limiting signal measured at 20 s to allow direct comparisons. The dotted line illustrates the difference between the signal from stopped-flow and rapid chemical quench measurements.

Scheme I) can be isolated in stable form following modification with DEGRck (39) . The dansyl moiety provides a convenient signal with which the conversion of this intermediate to thrombin can be monitored continuously (15). Therefore, the kinetics of cleavage a t Arg274-Thr275 by preassembled prothrombinase was assessed using DEGR-mIIa as a substrate with the same experimental design used for studies of prothrombin activation (Figs. 2 and 3). Product formation was detected either continuously by stopped-flow fluorescence measurements or by rapid chemical quench followed by measurements of fluorescence intensity. Within experimental error, both techniques yielded equivalent results. The fluorescence change was described by a single exponential with a rate constant comparable to the decay phase in the SDS-PAGE measurements of meizothrombin (Fig. 3B). Thus, the fluorescence measurements of DEGR-mIIa cleavage appear to correspond directly to cleavage at A r g 2 7 4 -

ThP5. The rate constant for DEGR-mIIa cleavage (not shown) increased by a modest 5-fold as PCPs,,, was increased from 0


0.0 1 0.3

I W

0.3 ‘ ‘ 0.0 0.0 0.2 0.4 0.6 0.8 1 .o

Time (s) FIG. 5. Rapid chemical quench analysis of cleavage at A&=-

IleSa4 in prothrombin or prethrombin 2 plus fragment 1.2. The cleavage of Pre 2-F1.2 was inferred from rapid chemical quench experiments followed by fluorescence detection of product as described under “Experimental Procedures.” Reactions were initiated by rapidly mixing substrate solution with preassembled prothrombinase to yield mixtures containing 0.3 p Pre 2,0.6 1.1~ F1.2,3 p DAF’A, 0.3 PM Xa, 0.45 p~ Va, and increasing concentrations of PCPs in 20 mM Hepes, 0.15 M NaCl, 2 m~ CaCI,, pH 7.4. The final concentrations of PCPs,, were 17 1.1~ (01, 42 PM (A), 117 1.1~ (V), and 267 p (0). The concentration of thrombin formed (left ordinate) was determined after normalizing the signal to the limiting fluorescence intensity observed at 20 s. The disappearance of prothrombin (closed symbols, right ordinate) from reaction mixtures containing identical concentrations of substrate, enzyme, and PCPs,,, is reproduced from Fig. 3 A .

to 267 p ~ . However, exactly the same results were also obtained using DEGR-mIIa des F1 (below), implying that the small membrane dependence for cleavage at Ar$74-Th?75 may reflect steric effects or “crowding“ when the surface is saturated with prothrombinase. This conclusion is also supported by a small change observed in the rate constant for the association of DEGRck with prothrombinase as PCPs,,,, was increased (43).

Prethrombin 2 binds tightly and reversibly (Kd 10”’ M) to the fragment 1.2 activation peptide through the fragment 2 domain (44). Thus, prethrombin 2 saturated with fragment 1.2 ( A r g 2 7 4 cleaved, Scheme I) was evaluated as a substrate analog for studies of the cleavage of the Arg323-Ile324 scissile bond by prothrombinase. Initial experiments indicated that stopped- flow fluorescence measurements of the cleavage of this substrate measured with DAPA deviated markedly from the time course of bond cleavage. Thus, further studies were performed by rapid chemical quench followed by fluorescence detection of the DAPA-thrombin complex formed in the quenched samples. Progress curves for the activation of prethrombin 2-fragment 1.2 by prothrombinase are illustrated in Fig. 5. The rate of cleavage of prethrombin 2-fragment 1.2 was strongly dependent on the concentration of PCPs,,,,. Progress curves obtained with this substrate also showed a distinct ‘%burst’’ phase that also increased and reached a limiting value at 117 p~ PCPs,,,. Also shown (Fig. 51, are the progress curves for the disappearance of prothrombin measured by rapid quench with analysis by SDS-PAGE at identical reactant concentrations. The data sets obtained for the cleavage at Arg323-Ile324 for either substrate are in agreement with each other. Within experimental error, the kinetics of cleavage at Arg323-Ile324 in prothrombin catalyzed by prothrombinase is adequately described by the kinetics of cleavage at the same bond when the substrate is the prethrombin 2-fragment 1.2 binary complex. For either substrate, the measured initial velocity for cleavage at Arg323-Ile324 increased 60-fold as PCPs,,, was increased from 0 to 267 p ~ .

Thus, the cleavage of the first bond (Arg323-Ile324) in prothrombin by prothrombinase exhibits a much larger dependence on

A 0.3 -1

0.3 0.0 0.2 0.4 0.6 0.8 1.0

Time ( s )

2 . 0 ’ ’ ’ ’ ’ I

0 I 2 3 4 Time (s)

FIG. 6. Assessment of the effect of the fragment 1 domain on the cleavage of the individual bonds in prothrombin. Panel A, cleavage at Ar223-Ile324. The cleavage of 0.3 p prethrombin 2 plus 0.6 p~ fragment 1.2 (0) was compared with cleavage of 0.3 1.1~ prethrombin 2 plus 0.6 1.1~ fragment 2 (0) in reaction mixtures containing 0.3 p~ Xa, 0.45 p Va, 3 p DAPA, and 300 p PCPs (PCPs,,, = 267 1.1~) in 20 mM Hepes, 0.15 M NaC1, 2 mM CaCl,, pH 7.4, a t 25 “C. Product formation was inferred from the increased fluorescence accompanying thrombin formation following rapid chemical quench. The lines were drawn following fitting to Equation 10 under “Data Analysis” to estimate initial velocities of 0.042 and 2.46 p thrombin formeds for the two traces, respectively. The disappearance of prothrombin (A, right ordinate) at PCPs,, = 0 from rapid chemical quench studies is illustrated for ref- erence. Panel B , cleavage a t Ar$74-ThP75. Stopped-flow measurements were conducted by mixing 0.6 p DEGR-mIIa (upper trace) or 0.6 p~ DEGR-mIIa des F1 in 20 mM Hepes, 0.15 M NaCl, 2 m~ CaCI,, pH 7.4, with an equal volume of enzyme solution: 0.6 1.1~ Xa, 0.9 PM Va, 600 1.1~ PCPs in the same buffer. Thrombin formation was measured by moni- toring the decrease in fluorescence intensity (A, = 280 nm, A,, > 500 nm) at 25 “C, and the traces were offset to facilitate presentation. The lines were drawn drawn according to Equation 9 using rate constants of 4.47 s-l and 4.28 s” for the upper and lower truces, respectively.

the concentration of PCPs,,, than the subsequent cleavage a t Ar$74-Thr275 which yields thrombin.

Assessment of the Contribution of the Substrate-Membrane Interaction To Bond Cleavage-The effects of PCPs,,, on enzyme crowding versus binding the substrate were further dis- tinguished by kinetic studies of the half-reactions using intermediates lacking the fragment 1 domain which mediates the high affinity interaction of prothrombin with membranes (45).

A progress curve for the cleavage of prethrombin 2-fragment 1.2 at saturating PCPs,,,, is compared with that obtained using prethrombin 2-fragment 2 in Fig. 6A. The rate of cleavage at Arg323-Ile324 in the absence of the fragment 1 domain is approximately 60-fold lower than the rate observed when fragment 1.2 is present. In addition, the progress curve for the activation of prethrombin 2-fragment 2 was essentially indistinguishable from the time course for cleavage of prothrombin or prethrombin 2-fragment 1.2 at PCPs,,, = 0. These data suggest that the large dependence of cleavage rate at A r g 3 2 3 -

Ile324 on PCPs,,,, is primarily related to the fragment l-mediated binding of prothrombin (or prethrombin 2-fragment 1.2) to membranes in steps that precede cleavage at Arg323-Ile324 by prothrombinase.

Stopped-flow experiments illustrating the cleavage of DEGR-mIIa or DEGR-mIIa des F1 by prothrombinase are illustrated in Fig. 6B. Essentially identical traces were obtained for either substrate. Thus, the presence of the fragment 1 do-


M + E f M-E kcah,, IIa-F1.2-E #' ' IIa-F1.2 + E KsM E K b a - F l 2 E

SCHEME I1

0'4 Y 0.01 , I I , I 1 I I I I , 10

0 50 100 150 200 250 300

FIG. 7. Prediction of changes in the accumulation of meizothrombin. The ratio of initial velocities for cleavage at Ar$23-Ile324 and Arg274-Th1.275 by prothrombinase (0) and the total amount of meizothrombin formed in studies of prothrombin activation calculated by numerical integration (A) are plotted as a function of PCPs,,,,.

main in the substrate has no detectable effect on the cleavage at Arg274-ThP75 by prothrombinase. The rate constants for the cleavage of DEGR-mIIa or DEGR-mIIa des F1 were equivalent to each other when the concentration of PCPs,,, or prothrombinase was varied over a wide range (not shown). Thus, neither the initial combination of meizothrombin with prothrombinase nor the turnover of the enzyme-substrate binary complex is influenced by the presence of the F1 domain. The data are consistent with the interpretation that the fragment 1 domain and the derived ability of the substrate to bind membranes have no detectable effect on cleavage a t Arg274-Th1.275 by prothrombinase.

Kinetic Accounting for Changes in the Accumulation of Meizothrombin-The large increase in the extent of meizothrombin accumulation during prothrombin activation with increasing concentrations of PCPS,,,, provided the initial sugges- tion that the cleavage of prothrombin and meizothrombin by prothrombinase may be differentially influenced by this reactant. The subsequent kinetic studies with the two intermediates indicate that membrane binding by the substrate is essential for efficient cleavage a t Arg323-Ile324 but not for cleavage at Arg274-ThP75. Since the conversion of prothrombin to thrombin via the formation of prethrombin 2 (Scheme I) was not experimentally significant in these studies with prothrombinase, conclusions derived from studies of the half-reactions should account for the observed changes in meizothrombin accumulation with increasing concentrations of PCPs,,,,.

The integrated areas of progress curves for the transient accumulation of meizothrombin determined by numerical integration of rapid quench studies of prothrombin activation (Fig. 3B) are plotted as a function of PCPs,,, in Fig. 7. Meizothrom- bin accumulation increased approximately 10-fold with increasing PCPs,,. Changes in the amplitude of meizothrombin formation were independently predicted by calculating the ratio of the initial velocities for the two cleavages from studies of the half-reactions as a function of PCPs,,,, which also increased about 10-fold over the same concentration range. The tolerable agreement between the two data sets implies that the differential membrane requirements for cleavage at Ar$',- Ile324 versus Arg274-Th?75 by prothrombinase provide an adequate explanation for the dependence of prothrombin activation on PCPs.

DISCUSSION

The results of this study imply an important role for the substrate-membrane interaction on prothrombinase function. The observed effects of PCPs on prothrombin activation cata-

lyzed by prothrombinase derive from a large effect on the first cleavage reaction (cleavage atArg323-Ile324) with a much smaller effect on the subsequent cleavage at Arg274-Th?75. In addition, proteolytic removal of the membrane binding fragment 1 domain leads to a 60-fold reduction in the rate of cleavage at Arg323-Ile324 with no detectable effect on the kinetics of cleavage at Ar$74-Th?75. This suggests that membrane binding by the substrate through the fragment 1 domain is necessary for efficient cleavage at Arg323-Ile324 but not for the second cleavage reaction at Arg274-Th1.275 catalyzed by prothrombinase. These conclusions derived from kinetic studies of the two half- reactions catalyzed by prothrombinase provide a reasonable explanation for the qualitative and quantitative effects of PCPs on prothrombin activation.

The conclusions of this study can be further summarized by the two sequential proteolytic reactions illustrated in Scheme 11. Both reaction sequences are based on the observations of the present work and the known mechanism of action of serine proteases on peptidyl bonds. The first cleavage reaction (cleavage at Arg323-Ile324) for the conversion of prothrombin (PI to meizothrombin (M) involves the initial binding of prothrombin to free sites (L) on the membrane surface before reversible binding and cleavage (determined by Ks(L.P,E) and kc,,,,,,,) by membrane-assembled prothrombinase (E). Following bond cleavage, the product (P') remains bound to the enzyme and requires a rate-limiting unimolecular step to yield meizothrombin (M) which can bind DAPA. The initial product P' could represent a cleaved form of prothrombin which requires a conformational change before it can either bind DAPA or develop increased fluorescence. An alternative possibility is that the L.P'.E ternary complex represents an acylenzyme intermediate that requires slow deacylation to yield meizothrombin but is disrupted by denaturing, basic conditions and heating required for SDS-PAGE sample preparation. In either case, the inclusion of such a rate-limiting unidirectional step is necessary to explain the burst kinetics for the cleavage at Ile324 and the large discrepancy between the stopped-flow and rapid chemical quench measurements for the cleavage at this bond by prothrombinase. Two subsequent reversible steps describe the dissociation of product from the ternary complex (K,,,.,,,,) and the subsequent dissociation of meizothrombin from the membrane surface.

A reaction pathway for the cleavage of meizothrombin a t Arg274-Thr275 to yield the terminal products thrombin (IIa) and fragment 1.2 (F1.2) is illustrated in the lower portion of Scheme 11. Kinetic measurements of this half-reaction indicate that efficient cleavage of meizothrombin by prothrombinase does not require the initial binding of this substrate to membranes. The reasonable agreement between stopped-flow and rapid chemical quench measurements also implies the lack of additional steps in the catalytic cycle which need to be accounted for. Thus, the second cleavage reaction catalyzed by prothrombinase involves the formation of the Michaelis complex through the direct interaction of solution phase M with membrane- assembled enzyme (E) determined by Ks(M,E), followed by catalytic turnover (kcat,R,74) and the reversible dissociation of the products (Kp~II,,,,,,,,) from the enzyme. Thrombin and fragment


1.2, the products of this cleavage reaction, interact nonco- valently (44).

The reactions illustrated in Scheme I1 imply that two distinct types of productive substrate-enzyme interactions are required for the activation of prothrombin by prothrombinase. Although this conclusion follows directly from the fact that two peptide bonds are specifically cleaved in the conversion of prothrombin to thrombin, it is further supported by the differential effects of factor Va on the two cleavages established previously (22) and by the differential effects of PCPs on the two reactions determined in the present work. The reactions illustrated in Scheme I1 also indicate that the intermediate, meizothrombin, must interact with membrane-assembled prothrombinase in a nonproductive or “product” mode as well as in a productive “substrate” mode. It is unknown whether these two binding modes are separated by dissociation and rebinding steps during prothrombin activation. Since the concentration of accumulating meizothrombin can be orders of magnitude greater than the concentration of prothrombinase under steady-state conditions (16-18), it seems reasonable that the two enzymic reactions are separated by the dissociation and rebinding of meizothrombin to prothrombinase as illustrated in Scheme 11. If this is the case, then the two points of substrate addition in the kinetic scheme are separated by product release, and the action of prothrombinase on prothrombin is best described in classical enzymology terms as an ordered ping-pong bi-bi reaction (46, 47).

Prothrombin activation by prothrombinase was found to pro- ceed via the initial formation of meizothrombin at all concentrations of PCPs in the present work. The contribution of thrombin formation via reactant flux through the alternate pathway (cleavage at Arg274-Thl.275 followed by cleavage at Arg323-Ile324, Scheme I) was estimated to be below the experimental error of quantitation following rapid chemical quench. These observations support previous conclusions of the ordered nature of bond cleavages by prothrombinase (17, 18).

Previous kinetic studies of the half-reactions have relied on the implicit assumption that measurements of prethrombin 2-fragment 1.2 activation by prothrombinase adequately describe the action of the enzyme complex on prothrombin which leads to cleavage at Arg323-Ile324 (18, 22). This assumption was verified in the present work. Thus, prior cleavage at A r g 2 7 4 -

T h P 5 has no detectable effect on substrate recognition and cleavage at Ar$23-Ile324 by prothrombinase. Such symmetry does not seem to apply for cleavage at Arg274-Thl.275. If substrate recognition and cleavage at Arg274-ThP by prothrombinase are equivalent in prothrombin and meizothrombin, our data would predict a large and systematic change in the contributions of the two pathways for prothrombin activation (Scheme I) as PCPs,,, is decreased. That this is not evident indicates that meizothrombin and prothrombin are not equivalent substrates for the cleavage at A r g 2 7 4 - T h l . 2 7 5 by prothrombinase. Therefore, this scissile bond or other structural elements required for the formation and turnover of a productive enzyme- substrate complex are in some way obscured in intact prothrombin and altered following initial cleavage at Ar$23-Ile324. This conclusion provides a potential explanation for the inabil- ity of the kinetic constants determined for the half-reactions to explain the ordered mechanism of bond cleavage during prothrombin activation by prothrombinase.

Thus far, steady-state kinetic studies of prothrombin activation have been interpreted in terms of two extreme possibilities describing the productive encounter between membrane-bound prothrombinase and prothrombin which can also bind reversibly to the surface. In the “bound substrate” model, membrane binding by the substrate is essential, and the rate enhance-

ment afforded by membranes is related to the steady-state concentration of membrane-bound prothrombin (10, 11) or the rate of the substrate-membrane interaction which regulates the rate of formation of productive E.S complexes (14). Either explanation would predict a dependence of reaction rate on the concentration of PCPs. Alternatively, in the “free substrate” model formulated by Pusey and Nelsestuen (131, the formation of the E.S complex results from encounters between solution- phase prothrombin and membrane-bound prothrombinase. In the limit, this model predicts that membrane binding by the substrate is inconsequential for catalysis, and the rate of the reaction would be independent of PCPs.

The results of the present study indicate that the first cleavage reaction catalyzed by prothrombinase (cleavage at A r $ 2 3 -

Ile324) has the features predicted by the bound substrate model. In contrast, the subsequent conversion of meizothrombin to thrombin (cleavage at Arg274-Th1375) has the features of the free substrate model. Thus, the reaction pathway for the conversion of prothrombin to thrombin by prothrombinase actually requires both models for adequate description. Although kinetic explanations have been presented to reconcile the two opposing models (14), it is possible that the discrepancies in the literature are also related to the use of different techniques to assess product formation during prothrombin activation. Initial velocity measurements that primarily reflect the conversion of prothrombin to meizothrombin would show a strong dependence on PCPs, whereas techniques that reflect only thrombin formation are likely to demonstrate a reduced dependence of initial velocity on the concentration of membrane-bound substrate.

Conclusions derived from kinetic studies with bovine proteins have been found to describe adequately the activation of human prothrombin by human prothrombinase (17, 18). It therefore seems reasonable to extrapolate the present findings to the human system. Several studies have investigated the possible functional significance of meizothrombin and the conditions that influence the accumulation of this intermediate (19,4842). Although the accumulation of substantial amounts of meizothrombin (to levels of -35% of added substrate) has been reported in systems using purified proteins and synthetic phospholipid vesicles (16-18), meizothrombin accumulation is reduced when prothrombin activation is supported by washed platelets and by synthetic vesicles containing either unsatur- ated lipids or low percentages of phosphatidylserine (51). Al- though little or no accumulated meizothrombin has been observed in prothrombin activation experiments conducted in plasma (191, meizothrombin has been reported to be the major intermediate observed during clotting of whole blood (52). These diverse findings may be explained by the differential membrane dependence of the two cleavage reactions identified in the present study. The extent of accumulation of meizothrombin would be expected to be influenced by the concentration of membrane sites available to support prothrombin binding. Thus, phospholipid membranes with altered prothrombin binding ability or the reduced site concentrations that might be expected to occur in the absence of blood cells at nonsaturating concentrations of platelets or in the presence of plasma proteins that may compete with the prothrombin-membrane interaction would yield much lower levels of meizothrombin.

In summary, the rapid kinetic techniques used to assess prothrombin activation indicate that the substrate-membrane interaction is essential for the efficient cleavage of prothrombin by membrane-assembled prothrombinase. Therefore, the membrane surface influences prothrombin activation by mediating prothrombinase assembly as well as the delivery of substrate to the enzyme. However, membrane-mediated delivery and/or ori-


ented presentation of the substrate to the enzyme is only es- 19. Tans, G., Janssen-Claessen, T., Hemker, H. C. , Zwaal, R. F. A,, and Rosing, J.

able effect On the subsequent conversion Of meizothrombin to 21. Hibbard, L. s., Nesheim, M. E., and Mann, K. G. (1982) Biochemistry 21, thrombin. Thus, the membrane binding ability of the substrate 2285-2292 and the availability of membrane sites can influence the path- 22. Nesheim, M. E., and Mann, K G. (1983) J. Biol. Chem. 258,5386-5391

way Of prothrombin activation, the accumulation O f the prOte0- 24. Nesheim, M. E., Prendergast, F. G., and Mann, K. G. (1979) Biochemistry 18,

for efficient at Arg323-11e324 and has no detect- 20, &,sing, J., and Tans, G, (1988) Thromb, Haemostasis 60, 355-360 (1991) J. Biol. Chem. 266,21864-21873

23. Doyle, M. F., and Haley, P. E. (1993) Methods Enzymol. 222,299-312

lytically active intermediate, meizothrombin, and also modu- late the rate of thrombin formation.

Acknowledgments-We are grateful to Dr. Jim Murtagh, for the use of the PhosphorImager and Dr. Ed Morgan, Dept. of Pharmacology, for the use of the laser scanning densitometer. We are also grateful to Drs. Pete Lollar and Michael Nesheim for reading the manuscript, helpful suggestions, and critical comments.

REFERENCES

2. Jackson, C. M., and Nemerson, Y. (1980) Annu. Reu. Biochem. 49, 765-811 1. Walker, R. K., and Krishnaswamy, S. (1993) FMEB J. 7, A1191 (abstr.)

3. Mann, K. G., Jenny, R. J., and Krishnaswamy, S. (1988)Annu. Reu. Biochem. 57,915-956

4. Mann, K. G., Nesheim, M. E., Church, W. R., Haley, P., and Krishnaswamy, S. (1990) Blood 76,l-16

5. Papahadjopolous, D., and Hanahan, D. J. (1964) Biochim. Biophys. Acta 90, 436-439

6. Tans, G., and Rosing, J. (1986) in Blood Coagulation: New Comprehensive Biochemistry (Zwaal, R. F. A,, and Hemker, H. C., eds) Vol. 13, pp. 59-85, Elsevier Science Publishers B. V., Amsterdam

7. Nelsestuen, G. L., and Lim, T. K (1977) Biochemistry 16,4164-4171 8. Esmon, C. T., Owen, W. G., and Jackson, C. M. (1974) J. Biol. Chem. 249,

9. Nesheim, M. E., Taswell, J. B., and Mann, K. G. (1979) J. Biol. Chem. 254,

10. Rosing, J., Tans, G., Govers Riemslag, J. W., Zwaal, R. F., and Hemker, H. C.

11. Nesheim, M. E., Tracy, R. P., and Mann, K. G. (1984) J. Biol. Chem. 259,

12. van Rijn, J. L., Govers Riemslag, J. W., Zwaal, R. F., and Rosing, J. (1984)

7798-7807

10952-10962

(1980) J. B i d . Chem. 255, 274-283

1447-1453

Biochemistry 23,455774564 13. Pusey, M. L., and Nelsestuen, G. L. (1983) Biochem. Biophys. Res. Commun.

14. Giesen, P. L. A,, Willems, G. M., and Hermens, W. T. (1991) J. Biol. Chem. 266,

15. Boskovic, D. S., Giles, A. R., and Nesheim, M. E. (1990) J. Biol. Chem. 265,

16. Rosing, J., Zwaal, R. F., and Tans, G. (1986) J. Biol. Chem. 261,4224-4228 17. Krishnaswamy, S., Mann, K. G., and Nesheim, M. E. (1986) J. Biol. Chem. 261,

18. Krishnaswamy, S., Church, W. R., Nesheim, M. E., and Mann, K G. (1987) J.

114,526-532

1379-1382

10497-10505

89774984

Biol. Chem. 262,3291-3299

25. 26. 27. 28.

99G1003 Higgins, D. L., and Mann, K. G. (1983) J. Biol. Chem. 268,65034508 Krishnaswamy, S. (1990) J. Biol. Chem. 266,3708-3718 Krishnaswamy, S., and Mann, K. G. (1988) J. Biol. Chem. 263,5714-5723 Kalafatis, M., Krishnaswamy, S., Rand, M. D., and Mann, K. G. (1993)

Methods Enzymol. 222, 224-236 29. Jesty, J., and Nemerson, Y. (1976) Methods Enzymol. 45,95-107 30. Chase, T., Jr., and Shaw, E. (1967) Methods Enzymol. 19, 20-27 31. Mann, K. G. (1976) Methods Enzymol. 45,123-156 32. Lundblad, R. L., Kingdon, H. S., and Mann, K. G. (1976) Methods Enzymol. 45,

33. Laemmli, U. K. (1970) Nature 227,680-685 34. Laue, T. M., Johnson, A. E., Esmon, C. T., and Yphantis, D. A. (1984) Biochem-

35. Jackson, C. M., Johnson, T. F., and Hanahan, D. J. (1968) Biochemistry 7,

36. Fujikawa, K., Coan, M. H., Legaz, M. E., and Davie, E. W. (1974)Biochemistry

156-176

istry 23,1339-1348

4492-4505

13,5290-5299

Methods Enzymol. 222,260-280

3823-3834

37. Krishnaswamy, S., Nesheim, M. E., Pryzdial, E. L. G., and Mann, K. G. (1993)

38. Krishnaswamy, S., Jones, K C., and Mann, K. G. (1988) J. Biol. Chem. 263,

39. Armstrong, S.A., Husten, E. J., Esmon, C. T., and Johnson,A. E. (1990) J. Biol.

40. David, G. S. (1972) Biochem. Biophys. Res. Commun. 48,464-471 41. Nesheim, M. E., Kettner, C., Shaw, E., and Mann, K. G. (1981) J. Biol. Chem.

42. Mason, J. T., and Huang, C . (1978) Ann. N. Y. Acad. Sci. 308, 29-49 43. Walker, R. K., and Krishnaswamy, S. (1993) J. Biol. Chem. 268, 13920-13929 44. Mvrmel. K. H.. Lundblad. R. L., and Mann. K. G. (1976) Biochemistry. 15,

Chem. 265,62104218

256,65374540

~. 1767-1773

45. Gitel, S. N., Owen, W. G., Esmon, C. T., and Jackson, C. M. (1973) PFOC. Natl.

46. Cleland, W. W. (1963) Biochim. Biophys. Acta 67, 104-137 47. Segel, I. H. (1975) Enzyme Kinetics: Behauiour and Analysis of Rapid Equi-

librium and Steady State Enzyme Systems, pp. 60M23. John Wilev & Sons, New York

Acad. Sci. U. S. A. 70,1344-1348

48. Schoen, P., and Lindhout, T. (1987) J. Biol. Chem. 262, 11268-11274 49. Doyle, M. F., and Mann, K. G. (1990) J. Biol. Chem. 266, 10693-10701 50. Tans, G., Nicolaes, G. A. F., Thomassen, M. C. L. G. D., van Zonnenveld, A.,

51. Govers Riemslag, J . W., Janssen, M. P., Zwaal, R. F., and Rosing, J. (1992)

52. Bovill, E. G., Hayes, T., Jenny, R., Bhushan, E, and Mann, K G. (1993)

Pannekoek, H., and Rosing, J. (1994) J. Biol. Chem. 269,15969-15972

Biochemistry 31, 10000-10008

Thromb. Haernostasis 69, 779 (abstr.)

Documents

THE JOURNAL OF Vol. Issue pp. for Printed in U.S.A. The ... · In addition, we have used rapid chemical quench and physical assessment of bond cleav- age to evaluate the effects of