Incorporation of factor Va into prothrombinase is required for coordinated cleavage of prothrombin by factor Xa.

Prothrombin is activated to thrombin by two sequential factor Xa-catalyzed cleavages, at Arg271 followed by cleavage at Arg320. Factor Va, along with phospholipid and Ca2+, enhances the rate of the process by 300,000-fold, reverses the order of cleavages, and directs the process through the meizothrombin pathway, characterized by initial cleavage at Arg320. Previous work indicated reduced rates of prothrombin activation with recombinant mutant factor Va defective in factor Xa binding (E323F/Y324F and E330M/V331I, designated factor VaFF/MI). The present studies were undertaken to determine whether loss of activity can be attributed to selective loss of efficiency at one or both of the two prothrombin-activating cleavage sites. Kinetic constants for the overall activation of prothrombin by prothrombinase assembled with saturating concentrations of recombinant mutant factor Va were calculated, prothrombin activation was assessed by SDS-PAGE, and rate constants for both cleavages were analyzed from the time course of the concentration of meizothrombin. Prothrombinase assembled with factor VaFF/MI had decreased k(cat) for prothrombin activation with Km remaining unaffected. Prothrombinase assembled with saturating concentrations of factor VaFF/MI showed significantly lower rate for cleavage of plasma-derived prothrombin at Arg320 than prothrombinase assembled with saturating concentrations of wild type factor Va. These results were corroborated by analysis of cleavage of recombinant prothrombin mutants rMz-II (R155A/R284A/R271A) and rP2-II (R155A/R284A/R320A), which can be cleaved only at Arg320 or Arg271, respectively. Time courses of these mutants indicated that mutations in the factor Xa binding site of factor Va reduce rates for both bonds. These data indicate that the interaction of factor Xa with the heavy chain of factor Va strongly influences the catalytic activity of the enzyme resulting in increased rates for both prothrombin-activating cleavages.

result of this process is the conversion of prothrombin to thrombin, which in turn cleaves fibrinogen to produce the insoluble fibrin mesh (1). Prothrombin is activated physiologically by the prothrombinase complex, which is composed of the non-enzymatic cofactor, factor Va, the enzyme, factor Xa, and the substrate, prothrombin, associated on a cell membrane surface in the presence of Ca 2ϩ (2,3). Activation of prothrombin to thrombin is a consequence of two cleavages at Arg 271 and Arg 320 by factor Xa. Depending on the order of peptide bond cleavage, different intermediates are formed (Fig. 1). Initial cleavage at Arg 271 will produce fragment 1⅐2 and prethrombin-2, whereas initial cleavage at Arg 320 results in the formation of meizothrombin, which has optimal amidolytic activity but diminished clotting activity (4 -13). The existence of these two pathways has long been established (4 -9), and the kinetics of the rate of cleavage determined (10 -13). Although factor Xa alone cleaves prothrombin sequentially at Arg 271 followed by Arg 320 to produce thrombin (Fig. 1, pathway I), the rate of the overall reaction for thrombin formation is slow and incompatible with survival. The prothrombinase complex catalyzes the activation of prothrombin following the second pathway (Arg 320 followed by Arg 271 , Fig. 1, pathway II), resulting in a 300,000-fold increase in the rate of the overall reaction compared with the rate of prothrombin activation observed with factor Xa alone and represents the physiological relevant pathway for prothrombin activation (3,12,13). Although both cleavages are phospholipiddependent, only initial cleavage of prothrombin at Arg 320 is strictly dependent on factor Va. The overall increase in the catalytic efficiency of prothrombinase when compared with the activation of prothrombin by factor Xa alone appears to arise from a 100-fold decrease in the K m and a 3,000-fold increase in the k cat of the enzyme (10, 14 -16).
Human prothrombin is composed of 579 amino acids and circulates in blood at a concentration of 1.4 M (17). Although the importance of the contribution of factor Va to the activity of factor Xa for rapid thrombin generation by prothrombinase at the place of vascular injury has been long established, the consequence of the interaction of the cofactor with the components of prothrombinase and the molecular mechanism by which factor Va reverses the order of cleavages and increases the rate of the overall catalysis remain enigmatic and are the subjects of intense investigation. Proteolytic processing of factor V by thrombin at Arg 709 , Arg 1018 , and Arg 1545 is required for the interaction of the cofactor with the members of prothrombinase and produces a factor Va molecule that consists of a heavy chain (M r 105,000) and a light chain (M r 74,000) (18,19). Although both chains of factor Va and Ca 2ϩ are required for the interaction with factor Xa, only the heavy chain of the cofactor binds prothrombin in a Ca 2ϩ -independent manner (20 -22). Cleavage of factor Va by APC at Arg 506 /Arg 679 results in de-creased affinity of the molecule for factor Xa and eliminates its interaction with prothrombin (21,23,24). Subsequent cleavage of the membrane-bound cofactor at Arg 306 , completely abolishes the ability of factor Va to interact with factor Xa (23,24).
The interaction of factor Va with factor Xa and prothrombin has been long established as a requirement for optimum rates of prothrombin activation (3,4,6,10,11,14,25,26). However, the consequences of these associations on the mechanism of catalysis and the discrete amino acid regions from the molecules that are involved in protein-protein interaction have remained puzzling. Furthermore, earlier findings as well as data using mutant prothrombin molecules have suggested that the rate of cleavage at Arg 320 largely benefits from incorporation of factor Va into prothrombinase while the rate of cleavage at Arg 271 appears to be less affected (12,13). We have recently identified binding sites for factor Xa and prothrombin in the middle portion and the COOH terminus of the heavy chain of factor Va, respectively (27)(28)(29)(30)(31). Our findings suggested that factor Va has a crucial effect on prothrombinase activity, orchestrating the spatial arrangement of enzyme and substrate, ultimately directing cleavage and activation of prothrombin by factor Xa at two spatially distinct proteolytic sites (28). The present study was undertaken to identify and understand the molecular mechanisms underlying the contribution factor Va to the two factor Xa-mediated prothrombin-activating cleavages following its incorporation into prothrombinase.

EXPERIMENTAL PROCEDURES
Materials, Reagents, and Proteins-L-␣-Phosphatidylserine (PS) 1 and L-␣-phosphatidylcholine (PC) were from Avanti Polar Lipids (Alabaster, AL). Normal reference plasma and the chromogenic substrate Spectrozyme-TH were from American Diagnostica, Inc. (Greenwich, CT). The thromboplastin reagent for the clotting assays was purchased from Organon Teknika Corp. (Durham, NC). The fluorescent thrombin inhibitor dansylarginine-N-(3-ethyl-1,5-pentanediyl)amide (DAPA), RVV-factor V activator, human factor Xa, human thrombin, a sheep anti-human factor V polyclonal antibody, and the monoclonal antibody ␣hFV#1 coupled to Sepharose were from Hematologic Technologies Inc. (Essex Junction, VT). Human prothrombin was purified according to the procedure detailed by Bajaj and Mann (32). A final step was included in the purification procedure that included the use of hydroxyapatite column chromatography (Bio-Rad) to eliminate a proteolytic product from prothrombin that was consistently present in all final prothrombin preparations. Recombinant prothrombin molecules rMZ-II (prothrombin with the substitutions Arg 155 3 Ala, Arg 284 3 Ala, and Arg 271 3 Ala) and rP2-II (prothrombin with the substitutions Arg 155 3 Ala, R 284 3 Ala, and R 320 3 Ala) were prepared and purified as described previously (13,33). The cDNA for factor V was purchased from American Type Tissue Collection (ATCC number 40515 pMT2-V, Manassas, VA). The sequence of this cDNA molecule is identical to the cDNA published by Jenny et al. (34). All restriction enzymes were from New England Biolabs (Beverly, MA), and all other molecular biology and tissue culture reagents and media were from Invitrogen or as indicated. Human plasma factor V and factor Va were purified and concentrated using methodologies previously described employing the monoclonal antibody ␣hFV#1 coupled to Sepharose and heparin-Sepharose (30,35). The cofactor activity of the plasma-derived and recombinant factor Va preparations was measured by a clotting assay using factor V-deficient plasma using an automated coagulation analyzer (START-4, Diagnostica Stago, Parsippany, NJ). Phospholipids vesicles composed of 75% PC and 25% PS (referred to as PCPS vesicles throughout the manuscript) were prepared as previously described (36). The concentration of phospholipid vesicles was determined by phosphorous assay as described earlier and is given as the concentration of inorganic phosphate (37).
Mutagenesis and Transient Expression of Recombinant Factor V Molecules-All mutant factor V molecules described herein were synthesized by PCR-based methods as described (28,31). The activity and the integrity of the molecules were verified before and after activation by thrombin by clotting assays using factor V-deficient plasma and by SDS-PAGE followed by Western blotting using well characterized monoclonal antibodies (provided by Dr. Ken Mann, Department of Biochemistry, University of Vermont).

Measurement of Rates of Thrombin Formation in a Prothrombinase
Assay-All factor V species (plasma and recombinant) were activated with thrombin for 15 min at 37°C followed by the addition of diisopropylfluorophosphate (2 mM, Sigma). The assay verifying the activity of the recombinant molecules was conducted as described by measuring thrombin formation by the change in the absorbance of a chromogenic substrate at 405 nm (Spectrozyme-TH) monitored with a Thermomax microplate reader (Molecular Devices, Sunnyvale, CA) (31). Absorbance at 405 nm was compared with a standard curve prepared daily using purified thrombin (0 -50 nM).
For the functional calculation of the apparent dissociation constants (K D app ) between the recombinant factor Va molecules and factor Xa, experiments were performed in the presence of a limited concentration of factor Xa (between 15 and 30 pM) and varying concentrations of factor Va (up to 20 nM). The initial rates of thrombin generation in the various assay mixtures under the conditions employed were linear, and in all experiments no more than 10% of prothrombin was consumed during the initial course of the assay. The initial rate of the formation of thrombin (initial velocity in nM⅐IIa⅐min Ϫ1 ) was calculated, and the data were analyzed and plotted assuming a 1:1 stoichiometry between factor Va and factor Xa (one binding site) using the software Prizm (Graph-Pad, San Diego, CA) as described (27,38). Once the value of the K D app for the bimolecular interaction of each species of factor Va with factor Xa was determined, the amount of recombinant factor Va required to completely saturate factor Xa and provide similar amount of enzyme (prothrombinase) when using various mutant recombinant factor Va  Prothrombin contains ten ␥-carboxyglutamic acid residues (hatched region), two kringle domains (checkered region), and a serine protease domain (black region) (17). Initial cleavage at Arg 271 by membrane-bound factor Xa in the absence of factor Va, results in the generation of fragment 1⅐2 and prethrombin-2 followed by subsequent cleavage at Arg 320 resulting in thrombin formation (pathway I). Initial cleavage at Arg 320 by prothrombinase (factor Va bound to factor Xa on a membrane surface in the presence of Ca 2ϩ ) results in the production of a catalytically active intermediate (meizothrombin), which has the same molecular weight as prothrombin under non-reducing conditions. Following reduction, the intermediate gives rise to fragment 1⅐2-A and the B chain of thrombin. Subsequent cleavage of this intermediate at Arg 271 produces thrombin and fragment 1⅐2 (pathway II). Once formed thrombin also cleaves prothrombin at Arg 155 to form prethrombin-1, and at Arg 284 to form the A chain of thrombin minus thirteen amino acids (dashed arrows).
For the calculation of the kinetic constants of prothrombinase assembled with the various recombinant mutant factor Va molecules as well as with plasma factor Va (K m and V max ), experiments were conducted in the presence of a limited amount of factor Xa (5 pM min, respectively, following the addition of factor Xa. A, factor Xa alone and no factor Va; B, prothrombinase assembled with plasma-derived factor Va; C, prothrombinase assembled with wild type factor Va; D, prothrombinase assembled with factor Va FF ; E, prothrombinase assembled with factor Va MI ; F, prothrombinase assembled with factor Va FF/MI . For the easy reading of the manuscript the factor Va species used for the reconstitution of prothrombinase are also shown under each panel. The identity of each fragment generated following cleavage of prothrombin by prothrombinase was verified in separate experiments by NH 2 -terminal sequencing and comparison of the data with the primary structure of prothrombin (17). The prothrombin-derived fragments are numbered consecutively from the NH 2 terminus of prothrombin and are shown as follows: II, prothrombin (amino acid residues 1-579); F1⅐2-A, fragment 1⅐2-A chain (amino acid residues 1-320); F1⅐2, fragment 1⅐2 (amino acid residues 1-271); B, B chain of thrombin (amino acid residues 321-579); F1, fragment 1 (amino acid residues 1-155); P1, prethrombin-1 (amino acid residues 156 -579); P2, prethrombin-2 (amino acid residues 272-579) (17). Gels were submitted to scanning densitometry as described under "Experimental Procedures," and the rates of prothrombin consumption are reported in Table I. a The rates of prothrombin consumption were obtained following scanning densitometry of the gels shown in Fig. 2. The final rate of prothrombin consumption in the presence of prothrombinase assembled with the various factor Va species, was extracted following plotting of the data according to the method described under "Experimental Procedures." action between the two molecules caused by a mutation may influence or modify the stability of the catalytic site of the enzyme and can be measured by the change in the transition-state stabilization of free energy for prothrombin activation as extensively described (41)(42)(43)(44)(45). Thus, the consequence of the mutations in factor Va affecting factor Xa catalytic efficiency and their additivity can be measured relative to the change in transition-state stabilization free energy (⌬⌬G ‡ ) of the enzyme as previously established (41)(42)(43)(44)(45).
To assess whether the two sets of mutations affecting the interaction of factor Va with factor Xa (E323F/Y324F; E330M/V331I) energetically interact, four separate free energies associated with the catalytic efficiency of prothrombinase must be measured: ⌬G wt , the functional free energy in prothrombinase assembled with wild-type factor Va; ⌬G FF , the functional free energy in prothrombinase assembled with factor Va FF ; ⌬G MI , the functional free energy in prothrombinase assembled with factor Va MI ; and ⌬G FF/MI , the functional free energy in prothrombinase assembled with factor Va FF/MI . In general, the perturbation to the function of prothrombinase assembled with wild type factor Va (state A), caused by a mutation in factor Va (state B) affecting the transition state, can be defined in general as follows, and because we are measuring prothrombinase activity assembled in the presence of various factor Va molecules against the same substrate (prothrombin), the changes in transition-state stabilization free energy (⌬⌬G ‡ ) during catalysis caused by a mutation in factor Va (state B) can be calculated from the following general equation, with R being the universal gas constant (2 cal⅐K Ϫ1 ⅐mol Ϫ1 ), T the absolute temperature in Kelvin (298 K in all our experiments), k cat is the turnover number, and K m is the Michaelis-Menten constant for prothrombinase assembled with either wild-type or mutant factor Va molecules. In addition, the free energy of the interaction between two mutations in factor Va (A 3 B and A 3 C; where A represents the wild type factor Va, B and C represent cofactor molecules with two different mutations, and B/C represent a double mutant factor Va molecule) can be calculated with the following general equation (42)(43)(44)(45).
The term ⌬⌬G int reflects an exchange in free energy between the side chains of the amino acids studied (i.e. 323 EY 324 and 330 EV 331 ). A value of ⌬⌬G int Ͼ0 or Ͻ0 implies that the side chains of the mutated amino acids interact with each other, reducing or enhancing, respectively, the catalytic efficiency of prothrombinase. Scanning Densitometry of SDS Gels and Calculation of the Rate of Prothrombin Consumption-The stained gels were scanned with a Lexmark printer/scanner, and the final images were captured in PowerPoint. They were subsequently imported as 8-bit grayscale images into Corel Photo Paint. They were analyzed in the National Institutes of Health Image module. Bands were isolated, and the total number of pixels and the average pixel intensities were determined on a scale from 0 to 255. A region of each lane was chosen as background. The average pixel intensity of the background was subtracted from that of each band. The total intensity in each band was then calculated by multiplying the result by the total number of pixels. To compensate for minor loading differences from lane to lane, all intensities were normalized to the sum of intensities of all the bands in each lane. The Coomassie Blue staining intensities of fragment 1⅐2-A and the B chain, relative to prothrombin, were determined based on the stoichiometric relationship that the sum of prothrombin plus the B chain is constant throughout the reaction in the presence of DAPA. Thus, for a given experiment, the intensity data were subjected to linear regression to the equation, intensity of B chain ϭ total intensity ϩ R⅐intensity of prothrombin, where R is the intensity per mole of B chain relative to that per mole prothrombin. The relationship was strictly linear for all extents of reaction, as expected. The value of R was typically ϳ0.65, indicating that, mole per mole, the B chain stained with 65% the intensity of prothrombin. Thus, fragment 1⅐2-A, which comprises the part of prothrombin not represented by the B chain, was assigned a relative staining intensity of 0.35. The molar concentration of prothrombin as observed on the gels was then calculated by normalizing its staining intensity to the initial prothrombin concentration (usually 1.4 M). The concentrations of fragments 1⅐2, and the B chain were calculated similarly, except that their staining intensities were divided by 0.35 and 0.65, respectively. The linearity of the relationship between amount of protein loaded and the intensity measured by densitometry was confirmed by loading known quantities of prothrombin and subjecting the stained gels to the above procedure.

Effect of the Interaction of Factor Va with Factor Xa on the
Rates of Prothrombin-activating Cleavages-To test the effect of the interaction of factor Va with factor Xa on the rates of cleavage of prothrombin by prothrombinase, we have used recombinant factor Va molecules mutated at the factor Xabinding sites and plasma-derived prothrombin. Membranebound factor Xa alone was a poor activator of prothrombin ( Fig. 2A and Table I). In the presence of plasma-derived factor Va, fragment 1⅐2-A, which is a transient fragment during prothrombin activation, and results from initial cleavage of prothrombin at Arg 320 (Fig. 1), is rapidly consumed following cleavage at Arg 271 to produce fragment 1⅐2 and thrombin (Fig. 2B, lanes 3-15). Similar results were found in the presence of wild type factor Va. These combined data demonstrate that prothrombin activation is significantly accelerated in the presence of the wild type and/or plasma cofactor molecule (Fig. 2, B and C, and Table I). Analysis of the products of prothrombin activation by gel electrophoresis demonstrated that the deficiency in cofactor activity of all mutant factor Va molecules appears to be a consequence of delayed thrombin formation because of delayed cleavage at both Arg 320 /Arg 271 (Fig. 2, D-F). Further, while the rates in prothrombin consumption were decreased by 3.1-and 2.3-fold when using factor Va FF and factor Va MI , respectively (Table  I), the rate of thrombin generation was diminished by ϳ17fold with factor Va FF/MI as compared with the rates of thrombin formation in the presence of the wild type cofactor molecule (Table I). The data demonstrate that mutating all four amino acids in the factor Xa binding site of factor Va heavy chain has a dramatic effect on factor Va cofactor activity than disabling either of the two sites individually.

Kinetic Analyses of the Effect of the Mutations in Factor Va on Prothrombinase Complex Assembly and Function-
The efficiency of prothrombinase in activating prothrombin is remarkable when compared with the efficiency of the catalytic subunit alone, factor Xa, on the same reaction (Refs. 10 and 11; also illustrated in Fig. 2, A-C). To evaluate the effect of the mutations in factor Va on prothrombinase assembly and function, kinetics experiments were designed to determine the dissociation constant for prothrombinase assembly as well as the K m and k cat values for prothrombinase function. The apparent dissociation constants (K D app ) for complex formation between membrane-bound factor Xa and the mutant factor Va molecules was inferred from kinetic experiments in which the steady-state rate of thrombin generation in the presence of a fixed concentration of factor Xa, prothrombin, and PCPS was determined as a function of factor Va (Fig. 3). The K D app values derived from these experiments are summarized in Table II. Wild type factor Va had a similar K D app to plasma-derived factor Va. The data show that, although factor Va FF and factor Va MI have ϳ4and 10-fold lower affinities for factor Xa than wild type factor Va, factor Va FF/MI has ϳ61-fold lower affinity than the wild-type cofactor molecule for the enzyme (K D app ϳ 10.9 nM). These data are consistent with our previous findings using synthetic peptides (30,31).
To elucidate which parameters of prothrombinase are affected by mutations in the regulatory subunit, we determined the kinetic parameters of the enzyme under conditions where all membrane-bound factor Xa was saturated with factor Va. Fig. 4 depicts the initial steady-state rate of prothrombin activation as a function of substrate concentration, and the calculated kinetic constants are provided in Table II. The data show that prothrombinase assembled with recombinant wild type factor Va as well as with recombinant factor Va FF , recombinant factor Va MI , and recombinant factor Va FF/MI were characterized by similar K m values (ϳ0.15 M). Furthermore, although prothrombinase assembled with the wild type cofactor molecule had similar k cat as prothrombinase assembled with plasma factor Va (Table II), prothrombinase assembled with factor Va FF/MI (30 nM) was characterized by a k cat value of ϳ406 nM IIa formed⅐min Ϫ1 ⅐nM prothrombinase Ϫ1 (Table II). Similar re-sults were found when using factor Va FF/MI at 50 nM (not shown). Prothrombinase assembled with the other two mutant recombinant factor Va molecules (5 nM) were characterized by normal k cat values (1806 and 2040 nM IIa formed⅐min Ϫ1 ⅐nM prothrombinase Ϫ1 , Table II). Comparable results were found with these two mutant molecules even when used at 30 nM (not shown). Altogether, these data demonstrate that substitution of either of the two sets of amino acid residues at the extremities of the factor Xa-binding site of factor Va has no significant effect for the overall catalysis at saturating cofactor concentrations and appear to be well tolerated by prothrombinase; however, substitution of both sets, all four amino acids, at the same time has a profound effect on the catalytic efficiency of the enzyme, decreasing the second order rate constant by ϳ5-fold (Table II). The inability of factor Xa to function optimally because of improper interaction with factor Va can be solely explained by the inability of factor Xa to efficiently convert prothrombin to thrombin because of diminished productive collisions (difference in k cat ). To quantify the interaction between the mutations in the factor Xa-binding site of factor Va heavy chain and their synergistic or additive effect on prothrombinase function, we have calculated the difference in free energy of the transition-state complex of the enzyme assembled with various mutant cofactor molecules. We have also constructed a thermodynamic cycle for cleavage of prothrombin by prothrombinase (provided in Scheme I). The positive value of ⌬⌬G int reveals that mutations at the extremities of the factor Xa-binding site on factor Va heavy chain are not additive, but their effect is synergistic and disruptive in nature, resulting in the destabilization of the transition-state complex and slower rate of catalysis (41)(42)(43)(44)(45). These data establish the fact that the interaction of factor Va with factor Xa through amino acids Glu 323 , Tyr 324 , Glu 330 , and Val 331 is crucial for efficient rates of cleavage of prothrombin by factor Xa.
Cleavage of Mutant Prothrombin Molecules-The kinetic data obtained with plasma-derived prothrombin were confirmed by using mutant recombinant prothrombin molecules (Fig. 5). Factor Xa alone was a poor catalyst for cleavage of both rMZ-II and rP2-II (Fig. 5A). In the presence of plasma and/or wild type factor Va the expected fragments were obtained following cleavage of rMZ-II (Fig. 5, B and C, lanes 1-9) or rP2-II (Fig. 5, B and C, lanes 10 -18) (13). Under similar experimental conditions when using prothrombinase assembled with wildtype factor Va, cleavage at Arg 320 in rMZ-II occurred 9-fold faster than cleavage at Arg 271 in rP2-II (Table III). Thus, incorporation of factor Va into prothrombinase is more favorable for cleavage of prothrombin at Arg 320 than for cleavage at Arg 271 (12,13). Cleavage of rMZ-II and rP2-II in the presence of prothrombinase assembled with factor Va FF , factor Va MI , and factor Va FF/MI was considerably delayed compared with the same cleavages in the presence of the wild type molecule (Fig. 5, panels D-F, lanes 1-18, and Table III). Interestingly, factor Va FF and factor Va MI had similar deficiency in cofactor activity for cleavage of both rMZ-II and rP2-II when compared with the cleavage of both recombinant prothrombin molecules by prothrombinase assembled with wild-type factor Va. From the data shown in Fig. 5F it is also evident that factor Va FF/MI has the largest deficiency in cofactor activity among the three molecules, and this is mostly apparent when using rMZ-II (Table III). These data are confirmatory of our kinetic studies and the findings with plasma-derived prothrombin.
It is well established that increased catalytic efficiency of prothrombinase as compared with the catalytic efficiency of factor Xa alone, is due to the fact that prothrombinase preferentially cleaves prothrombin through pathway II (Fig. 1) (12,  13). Hence, factor Va converts factor Xa from an enzyme that FIG. 3. Determination of the affinity of recombinant factor Va molecules for plasma-derived factor Xa. Initial rates of thrombin generation were determined as described under "Experimental Procedures." Prothrombinase assembled with wild-type factor Va is shown by filled squares, prothrombinase assembled with factor Va FF is depicted by filled inverse triangles, prothrombinase assembled with factor Va MI is depicted by filled triangles, and prothrombinase assembled with factor Va FF/MI is depicted by filled diamonds. Titrations were carried out to 20 nM factor Va; however, for graphical purposes the data show the titration for up to 10 nM cofactor. The solid lines represent a nonlinear regression fit of the data as detailed under "Experimental Procedures" using the software Prizm and the model for one binding site. The apparent dissociation constant (K D app ) for each species was derived from each titration performed at least in triplicate with four different preparations of recombinant molecules and is listed in Table II.
Factor Va-directed Catalysis slowly cleaves prothrombin initially at Arg 271 to an enzyme that efficiently cleaves prothrombin initially at Arg 320 . To ascertain the exact activation pathway of prothrombin by prothrombinase assembled with factor Va FF/MI , we followed pro-thrombin activation by gel electrophoresis under conditions where factor Xa was saturated with factor Va FF/MI (1 nM factor Xa with 100 nM factor Va FF/MI , corresponding to ϳ90% saturation of factor Xa). The data demonstrated that prothrombinase assembled with saturating amounts of mutant cofactor molecule activates prothrombin through pathway II with a slower rate than that observed for prothrombinase assembled with the wild type cofactor molecule (data not shown). Cleavage at Arg 320 and meizothrombin formation were considerably delayed. Further, meizothrombin did not accumulate, because the intermediate was cleaved immediately as it was formed to produce thrombin. These data verify our findings shown in Fig.  2F and strongly suggest that binding of factor Va heavy chain to factor Xa is required for efficient initial cleavage at Arg 320 and rapid meizothrombin formation. Overall, our findings suggest that the interaction of amino acids Glu 323 , Tyr 324 , Glu 330 , and Val 331 from the heavy chain of factor Va with factor Xa is required for the positioning of the active site of the enzyme in a position necessary for optimal productive collisions between the enzyme and prothrombin, efficient initial cleavage at Arg 320 , and competent thrombin formation. DISCUSSION Our data demonstrate that binding of the heavy chain of factor Va to factor Xa through amino acids Glu 323 , Tyr 324 , Glu 330 , and Val 331 is necessary for achieving optimal catalytic rates for cleavage of prothrombin by prothrombinase. Data obtained from the crystal structure of meizothrombin desF1 and prethrombin-2 demonstrate that 1) the two prothrombinactivating cleavages are 36 Å apart and 2) for the molecule to be initially cleaved at Arg 320 a rotation of ϳ150°around the Gly 319 and Gly 324 hinge points must occur (48 -51). As a consequence, and keeping in mind the results presented herein, it is logical to hypothesize that the cofactor, factor Va, rather that being an inert accessory protein, following its incorporation into prothrombinase and its interactions with factor Xa, directs catalysis of the substrate by the enzyme at two spatially distinct sites (i.e. Arg 320 and Arg 271 ).
Our data show that a mutation at either one of the two extremities of the factor Xa-binding site on factor Va heavy chain has no major effect on the kinetic parameters of prothrombinase assembled with saturating concentrations of factor Va; however, substitution of all four amino acids at both sites simultaneously results in a damaging effect on the ability of prothrombinase to efficiently convert prothrombin to thrombin through the meizothrombin pathway. These data imply that amino acid residues 330 and 331 together with any remaining factor Xa-binding site from the heavy chain and/or the factor Xa-binding site located on the light chain of the cofactor can compensate for substitutions at amino residues 323 and 324 and vice versa when factor Va is used in excess. However, substitution of all four amino acid residues at the same time SCHEME I. Thermodynamic cycle for prothrombin activation by prothrombinase following substitutions in factor Va that affect factor Xa binding. ⌬⌬G int is the free energy of interaction between the side chains of 323 EY 324 and 330 EV 331 of factor Va heavy chain and was calculated using Equations 2 and 3 described under "Experimental Procedures."

FIG. 4. Determination of kinetic parameters of prothrombinase assembled with various recombinant factor Va species.
Initial rates of thrombin generation were determined as described under "Experimental Procedures." Prothrombinase assembled with two different preparations of wild type factor Va is shown by filled squares (5 nM) and filled circles (15 nM), whereas prothrombinase assembled with plasma-derived factor Va (5 nM) is depicted by filled triangles. The open diamonds represent prothrombinase assembled with factor Va FF (5 nM), whereas the open circles show prothrombinase with factor Va MI (5 nM). Prothrombinase assembled with factor Va FF/MI is depicted by filled inversed triangles (30 nM). The values of the K m and V max extracted from these graphs are listed in Table II. a Apparent dissociation constants of recombinant factor Va for plasma-derived factor Xa (K D app ) were determined as described in the legend to Fig. 3 at limiting factor Xa concentrations (15 pM).
b Apparent V max , K m , and k cat were determined as described in the legend to Fig. 4 at limiting factor Xa concentrations (5 pM) and saturating concentrations of factor Va.
c The -fold decrease is the ratio of the second order rate constant (k cat /K m ) of prothrombinase assembled with wild type recombinant factor Va compared to the second order rate constant of prothrombinase assembled with the recombinant mutant factor Va molecules.
cannot be rescued by any other residue nor by the factor Xabinding site located on the light chain of factor Va resulting in diminished catalytic rate. Using prothrombinase assembled with factor Va FF/MI and plasma-derived prothrombin, we have shown a 16.7-fold decreased rate of prothrombin consumption compared with the rate of prothrombin consumption obtained by prothrombinase assembled with the wild type cofactor molecule. A similar decreased rate of prothrombin consumption by prothrombinase assembled with the same mutant cofactor mol-ecule under similar experimental conditions (ϳ17-fold) was observed with rMZ-II (which is only cleaved at Arg 320 ). Collectively, the data suggest that prothrombinase assembled with factor Va FF/MI has a lower k cat because of impaired initial cleavage at Arg 320 . These data were confirmed by experiments studying prothrombin cleavage by prothrombinase saturated with factor Va FF/MI showing that activation still proceeds through the intermediate meizothrombin, albeit less efficiently because of a considerable decrease in the rate of cleavage at FIG. 5. Analysis of the activation of mutant prothrombin molecules by prothrombinase assembled with mutant factor Va molecules. Mutant prothrombin molecules were incubated in different mixtures with PCPS vesicles, DAPA, and various factor Va molecules (10 nM). The reaction was started by the addition of factor Xa, and the samples were treated as detailed inder "Experimental Procedures." Lanes 1-9 represent samples of the reaction mixture following incubation of prothrombinase with rMZ-II, before (lane 1), or following 0.5 min, 1 min, 2.5 min, 4 min, 6 min, 10 min, 20 min, and 30 min incubation with factor Xa, respectively; lanes 10 -18 represent samples of the reaction mixture following incubation of prothrombinase with rP2-II before (lane 10), or following 0.5 min, 1 min, 2.5 min, 4 min, 6 min, 10 min, 20 min, and 30 min incubation with factor Xa, respectively. A, factor Xa alone no factor Va; B, prothrombinase assembled with plasma-derived factor Va; C, prothrombinase assembled with wild type factor Va; D, prothrombinase assembled with factor Va FF ; E, prothrombinase assembled with factor Va MI ; F, prothrombinase assembled with factor Va FF/MI . Positions of prothrombin-derived fragments are indicated at right as detailed in the legend to figure 2. For the easy reading of the manuscript, the factor Va species used for the reconstitution of prothrombinase are also shown under each panel. Gels were submitted to scanning densitometry as described under "Experimental Procedures," and the rates of prothrombin consumption are reported in Table III. a The -fold decrease is the ratio of the activation rates of recombinant mutant prothrombin molecules in the presence of wild-type factor Va and mutant factor Va species as measured by the disappearance of prothrombin following scanning densitometry of the gels shown in Fig.5. The final rate of prothrombin consumption in the presence of prothrombinase assembled with various factor Va species was extracted following plotting of the data according to the method described under "Experimental Procedures." Arg 320 . Because factor Va FF/MI is only impaired in its interaction with factor Xa, our findings demonstrate that one of the roles of factor Va in prothrombinase is to adjust (increase) the k cat of the enzyme by relocating the amino acids in and/or around the active site of factor Xa in an optimum position necessary to efficiently cleave at two sites distant by 36 Å (51). The thermodynamic data obtained demonstrate that mutation at residues Glu 323 , Tyr 324 , Glu 330 , and Val 331 of factor Va heavy chain have a cumulative effect on prothrombin catalysis by factor Xa and suggest that the side chains of these four amino acids act synergistically to stabilize allosterically the transition state complex of the enzymatic reaction (41)(42)(43)(44)(45). However, the change in free energy observed following substitution at the extremities of the factor Xa-binding site on factor Va could also be explained by subtle changes in the overall conformation of the 323-331 amino acid segment due to the two sets of double mutations. Taken as a whole, the results suggest that the interaction of factor Va heavy chain with factor Xa is required for the correct positioning of amino acids in/or around the active site of factor Xa with respect to the two scissile bonds in prothrombin, resulting in increased frequency of productive collisions between enzyme and substrate required for optimum rates of thrombin formation.
Conclusive evidence for subtle alterations of amino acid residues in and/or around the active site of factor Xa occurring upon its interaction with factor Va, was provided by studies using recombinant tick anticoagulant protein (TAP) and a recombinant mutant TAP molecule (Arg 3 3 Ala, R3A-TAP) (52)(53)(54)(55). TAP, purified from the soft tick Ornithodoros moubata, is a potent inhibitor of factor Xa (K I 180 pM), and its potency is considerably increased toward prothrombinase (K I 5.3 pM) (55). It has been established that the four NH 2 -terminal amino acid residues of TAP are responsible for its inhibitory activity and interact directly with amino acids from the active site of factor Xa (56). Further, although R3A-TAP has a 4,000-fold decreased K I toward factor Xa alone, its potency toward prothrombinase is only decreased by 30-fold. The fact that factor Va does not interact with TAP together with results demonstrating decreased inhibitory efficiency of recombinant TAP molecules mutated at residues Asp 47 and Asp 54 , which were thought to interact with a site remote from the active site of the enzyme (3-fold decrease in the K I ) (56), led to the suggestion that prothrombinase exposes cryptic exosites on factor Xa for substrate binding/recognition following incorporation of factor Va into the complex (54,55). Because these macromolecular recognition sites on factor Xa that are typically providing a prothrombin binding site in prothrombinase were assumed to be occupied by R3A-TAP, it was concluded that inhibition of factor Xa by R3A-TAP is favored kinetically when the enzyme is part of the prothrombinase complex (53)(54)(55)(56). These data are intriguing and of extreme physiological importance, because theoretically R3A-TAP would be able to discriminate between factor Xa and prothrombinase and could be used as a preventive therapeutic anticoagulant, albeit its slow on rate (53)(54)(55). Notwithstanding the potency of R3A-TAP and its potential therapeutic importance, the crystal structure of recombinant TAPfactor Xa clearly demonstrated that amino acid residues Asp 47 and Asp 54 from recombinant TAP, which are part of an amino acid sequence sharing a great deal of homology with prothrombin sequence 305-320 (preceding cleavage at Arg 320 ), interact with amino acids Lys 147 , Arg 222 , and Lys 224 from factor Xa, which in turn are adjacent to the active site of the enzyme (57). These latter amino acids play a preeminent role in optimum docking of the inhibitor and appear to be rearranged following the interaction of factor Xa with factor Va. These data support the hypothesis that one of the exosites provided by factor Xa upon its interaction with factor Va for prothrombin tethering appears to be composed of amino acids near and/or around the active site of the enzyme promoting optimal docking of the scissile bonds of the substrate in the catalytic groove of the enzyme. In conclusion, our findings, put in the context of the available literature, indicate that factor Va directs catalysis of prothrombin by factor Xa within prothrombinase at two spatially distinct sites.