The Use of Prothrombin(S525C) Labeled with Fluorescein to Directly Study the Inhibition of Prothrombinase by Antithrombin during Prothrombin Activation*

Serine 525 of human prothrombin was mutated to cysteine and covalently labeled with fluorescein to make II(S525C)-fluorescein. Kinetics of cleavage of this derivative by prothrombinase are identical to those of wild-type prothrombin. Cleavage is coincident with a 50% increase in fluorescence intensity and the product is catalytically inactive. Thus, it allows convenient monitoring of prothrombin activation without generating active thrombin. The kinetics of inhibition of factor Xa (FXa) by antithrombin (AT) and AT-heparin were measured by monitoring activation of II(S525C)-fluorescein and the hydrolysis of the chromogenic substrate S2222 in the presence of AT. With S2222 as the substrate the rate constant for inhibition of FXa, Ca2+, and unilamellar vesicles of phosphatidylcholine and phosphatidylserine (75:25) (PCPS) vesicles by AT was 3.51 × 103 m −1 s−1; when factor Va (FVa) was included the rate constant was 1.55 × 103 m −1 s−1. In the absence of FVa, II(S525C)-fluorescein had no effect on inhibition. When II(S525C)-fluorescein was the substrate, however, FVa at saturating concentrations profoundly protected FXa from inhibition by AT, increasing the half-life from 3 min with FXa, Ca2+, PCPS, and II(S525C)-fluorescein, to greater than 69 min when FVa was included. Thus, both FVa and prothrombin are necessary for this level of protection. In the absence of prothrombin, FVa decreased the second order rate constant for inhibition by the AT-heparin complex from 1.58 × 107 m −1s−1, for FXa, Ca2+, and PCPS, to 7.72 × 106 m −1 s−1. II(S525C)-fluorescein and factor Va together reduced the rate constant to less than 1% of that for FXa, Ca2+, and PCPS. At a heparin concentration of 0.2 unit/ml, this corresponds to a half-life increase from 1 s to 136 s.

One of the key steps of coagulation involves activation of prothrombin to its active form thrombin by the multicomponent complex, prothrombinase. Prothrombinase comprises the serine protease factor Xa, the activated protein cofactor factor Va, calcium ions, and an appropriate cell membrane or phospholipid surface (1)(2)(3)(4)(5). Although factor Xa alone can slowly generate thrombin by an extremely inefficient reaction, the rate of thrombin generation is enhanced by several orders of magnitude by incorporation of the cofactor protein factor Va and the procoagulant surface (5).
The activities of the clotting serine proteases, such as thrombin, are regulated, in part, by plasma protease inhibitors of the serpin superfamily. Of these, antithrombin appears to be the most important. Antithrombin targets both the product of prothrombin activation, thrombin, and the enzyme responsible for the reaction, factor Xa. Inhibition of serine proteases by serpins occurs through the formation of a stable complex between the serpin and the active site of the serine protease. Inhibition by antithrombin is markedly enhanced by the glycosaminoglycan cofactor, heparin. Heparin accelerates rates of inhibition of both thrombin and factor Xa several thousand fold (6). Upon binding, heparin induces a conformational change in antithrombin (7)(8)(9), which is sufficient to accelerate the inhibition of factor Xa through an allosteric mechanism. The acceleration of thrombin inhibition, however, is dependent on the formation of a ternary complex between heparin, antithrombin, and thrombin (10 -12).
Several studies demonstrated that incorporation of factor Xa into the prothrombinase complex changes the ability of the antithrombin-heparin complex to inhibit it. Calcium ion accelerates the heparin catalyzed inactivation of factor Xa by antithrombin through a template mechanism (13). Phospholipids and factor Va play protective roles and decrease the ability of the heparin-antithrombin complex to inhibit factor Xa both on synthetic surfaces (14 -19) and in whole blood clots (20). Herault et al. (21), however, indicated that the antithrombin-heparin complex is an effective inhibitor for phospholipid/factor Va-bound factor Xa in solution and on blood cells. The substrate prothrombin also plays an important role in the protection observed when factor Xa is incorporated into the prothrombinase complex (14).
Prothrombinase activity is usually inferred through measurements of thrombin activity over time. Therefore, studying heparin-dependent antithrombin inhibition of prothrombinase is challenging because the inhibitor recognizes both the product of the reaction, thrombin, and the enzyme component of prothrombinase, factor Xa. To resolve this problem in this study, a mutant form of prothrombin (II(S525C)), in which serine 525 (the active site serine of thrombin) is replaced with cysteine, was expressed in a mammalian expression system, isolated, and covalently labeled at Cys 525 with fluorescein. II(S525C)-fluorescein exhibits an increase in fluorescence intensity upon activation to thrombin but is catalytically inactive to both small and macromolecular substrates. Thus, it allows convenient monitoring of prothrombin activation in the presence of antithrombin without generating active thrombin. In this study inhibition of free factor Xa or factor Xa bound to vesicles and factor Va, by antithrombin or the antithrombin heparin complex was determined by monitoring over time the hydrolysis of the small factor Xa substrate S2222 or the activation of II(S525C)-fluorescein.

Methods
DNA Construction and Mutagenesis-pBluescript(SKϩ) containing full-length human prothrombin cDNA (26) was digested with PstI and XbaI restriction endonucleases. The resulting 393-bp fragment, encompassing nucleotides 1580 -1973, was subcloned into pBluescript(SKϩ). Oligonucleotides 1 and 2 (Table I) were constructed to facilitate mutation of A 1714 3 T resulting in a serine to cysteine substitution at amino acid Ser 525 in the expressed protein product. PCR using Pfu DNA Polymerase with oligonucleotide 1 and the T3 promoter primer (Stratagene) produced a 450-bp product (PCRI), whereas that with oligonucleotides 2 and the T7 promoter primer (Stratagene) produced a 217-bp product (PCRII). Each PCR product was ligated into the EcoRV site of pBluescript(SKϩ). PCRI and PCRII in pBluescript were digested with BspHI and XbaI, and BspHI and PstI, respectively. The resulting fragments were ligated at their corresponding BspHI sites and subcloned into pBluescript(SKϩ) digested with PstI and XbaI. The presence of mutations and correct PCR amplifications were verified by DNA sequence analysis using the T3 and T7 promoter primers (Stratagene). The mutated fragment was ligated back into full-length prothrombin cDNA to make II(S525C)-cDNA. For expression in mammalian cells, the II(S525C)-cDNA was excised from pBluescript(SKϩ) with XbaI, followed by incubation with T4 DNA Polymerase, and ligated into the pNUT vector (27) at the SmaI site. This site is downstream of the zinc-inducible mouse metallothionein I promoter and upstream of the human growth hormone polyadenylation signal. Proper orientation of the construct was determined using restriction digest analysis and DNA sequence analysis using oligonucleotides 3 and 4. The pNUT vector encodes a modified dihydrofolate reductase gene, which allows for selection under high methotrexate concentrations.
Cell Culture, Transfection, and Selection-BHK cells were cultured in Dulbecco's modified Eagle's medium/F-12 nutrient mixture (1:1) supplemented with 5% newborn calf serum during transfection and selection. Transfection of II(S525C)-cDNA was carried out using the calcium phosphate co-precipitation technique (28). Sixteen hours after transfection, the medium was supplemented with 0.44 mM methotrexate to select for pNUT. Fifteen days after transfection, individual colonies were screened for II(S525C) production by enzyme-linked immunosorbent assay with horseradish peroxidase-conjugated sheep anti-human prothrombin antibody. High expressing clones were seeded into 500cm 2 triple flasks for large scale production. At confluence, the growth medium was replaced by serum-free Opti-MEM I, supplemented with 50 M ZnCl 2 , 10 g/ml vitamin K 1 , and penicillin/streptomycin/Fungizone mixture. The medium was collected at 24-h intervals, supplemented with glutathione (1 mM), and stored at Ϫ20°C.
Recombinant Protein Purification-Stored medium was thawed at 4°C and loaded onto XAD-2 (2.5 ϫ 15 cm) and Q-Sepharose (1.4 ϫ 8 cm) columns in tandem at either 4°C or 21°C. The Q-Sepharose column was then washed with 5 column volumes of 0.02 M Tris-HCl, 0.15 M NaCl, pH 7.4 (TBS), followed by elution of II(S525C) with 0.02 M Tris-HCl, 0.5 M NaCl. Protein containing fractions were identified using a Bio-Rad Assay and pooled. Fractions were collected until all protein was eluted from the column. Fluorescent labeling was performed directly on this pooled fraction by adding a 30 M excess of 5-IAF from a 20 mM stock solution of 5-IAF in N,N-dimethyl formamide and incubating the sample for 2 h at room temperature in the dark. The resulting II(S525C)-fluorescein was separated from the excess label by addition of sodium citrate to a final concentration of 0.025 M, followed by addition of a 1.0 M BaCl 2 solution to a final concentration of 0.08 M. The solution was stirred at 4°C for 1 h and centrifuged. The resulting pellet was washed with the supernatant from a parallel precipitation carried out in 0.02 M Tris-HCl, 0.5 M NaCl. The adsorbed protein was eluted by dissolving the barium citrate pellet in 0.2 M EDTA, pH 8.0 (1/6 of the original volume). The sample was then dialyzed against TBS, at 4°C, and subjected to anion-exchange chromatography on a Amersham Pharmacia Biotech fast-protein liquid chromatography Mono-Q HR 5/5 column at 4°C. This step was carried out to resolve fully ␥-carboxylated species from partially ␥-carboxylated species of II(S525C)-fluorescein (26,29). The protein was eluted with a 0 -30 mM CaCl 2 gradient in TBS (30-ml total volume, flow rate 0.5 ml/min). The first peak containing the fully ␥-carboxylated II(S525C)-fluorescein eluted at 15.0 mM CaCl 2 , whereas the second eluted between 18 and 25 mM CaCl 2 . Fractions from the first peak were pooled, and protein concentrations and labeling efficiency values were determined by absorbance readings at 280 and 495 nm, respectively.
SDS-PAGE Analysis of Cleavage of II(S525C)-fluorescein-II(S525C)-fluorescein (800 nM) was activated with 50 M PCPS, 5 nM FVa, 0.1 nM FXa, and 5 mM CaCl 2 in TBS. Fluorescence intensity was monitored at room temperature with an excitation wavelength of 495 nm and an emission wavelength of 535 nm, slit widths of 5 nm and 10 nm, respectively, and a 515-nm cut-off emission filter in place. Aliquots were removed from the reaction and added to an equal volume of 0.2 M acetic acid. Samples were concentrated and reduced and subjected to SDS-PAGE on a 5-15% minigel, which was subsequently photographed over UV light.   Reactions were started by the addition of FXa, and 10-l aliquots were removed and diluted in a 96-well plate, which had been pretreated with TBS, 1% Tween 80. The wells contained TBS, 0.01% Tween 80, with S2222, and CaCl 2 in sufficient amounts to reach final diluted concentrations of 400 M and 5 mM, respectively. Absorbance at 405 nm was read over time in a Spectromax Plus spectrophotometer (Molecular Devices, Sunnyvale, CA) at room temperature. Initial rates were calculated. The natural logarithms of these were then plotted against time and the corresponding slope was calculated to determine the second order rate constants of inhibition. Inhibition by S2222 Product-Reactions containing S2222 at various concentrations (0 -1000 M), 2 nM FXa, and 5 mM CaCl 2 , in TBS, 0.01% Tween 80, were monitored at 405 nm to completion in a 96-well plate. Additional S2222 (250 M) was then added, and the resulting reactions were monitored at 450 nm. Initial rates were determined and plotted versus [S2222 product], and a K P value for the inhibition of FXa by the product of S2222 hydrolysis was determined according to the following equation, where [P] is the concentration of S2222 product:

SDS-PAGE Analysis of Activation of II(S525C)-fluorescein Versus II (Wild-type) in the Presence and
In  Tween 80. Reactions were started by addition of FXa and monitored at 405 nm to completion. End point data were analyzed to determine the pseudo first order rate constant. The following rationale was used, taking into account that the product of S2222 hydrolysis as well as AT inhibit the enzyme. In this analysis, E is the enzyme, S is the substrate, P is the product, I is the inhibitor (AT), and ␣ is the pseudo first order rate constant for inhibition of FXa by AT. ␣ is equal to the second order rate constant (k 2 ) multiplied by the inhibitor concentration.
For product formation, Equating Equations 4 and 7 yields Equation 8,

Integration of Equation 8 yields Equation 9
, When the reaction with AT stops, Consequently Integration of Equation 11 over time gives Equation 12,

Equation 12 is rewritten as Equation 13
, Raw data were subjected to nonlinear regression to Eq. 13 to find ␤. Elimination of k[E] T between Equations 10 and 13, yields Equation 14, from which the value of ␣ was calculated.
The second order rate constant (k 2 ) was then calculated using

RESULTS
Production and Characterization of II(S525C)-fluorescein-To obtain II(S525C), stable BHK cell lines expressing II(S525C) were grown in serum-free medium (Opti-MEM) in triple flasks. Typically 4 liters of pooled conditioned media were used as the starting material for purification as outlined under "Methods." II(S525C) behaved identically to wild-type recombinant prothrombin in all of the purification steps. II(S525C) was found to be present in conditioned media at a concentration of 5-10 mg/liter. Only 10% of it, however, was fully ␥-carboxylated. II(S525C)-fluorescein was produced with a typical labeling efficiency of ϳ50%. Purified II(S525C)-fluorescein comigrated with plasma prothrombin when subjected to SDS-PAGE (data not shown). Upon activation of II(S525C)-fluorescein, an increase in fluorescence intensity of ϳ50% was observed, which was found to correlate well with cleavage (Fig. 1). The fluorescence intensity profile exhibited a transient maximum in intensity, which is indicative of the expected meizothrombin intermediate (26,30). Additionally, upon cleavage, the band that retains the fluorescent properties corresponds to the B-chain of thrombin, which contains the active site serine. This confirms that the fluorescein moiety attaches specifically to Cys 525 (Fig. 1). When compared with plasma prothrombin, II(S525C)-fluorescein activated at similar rates in both the presence and absence of FVa (Fig. 2). The thrombin generated from activation of II(S525C)-fluorescein was catalytically inactive toward small substrates, as determined by S2238 hydrolysis, in which II(S525C)-fluorescein was observed to have at least 130,000-fold less activity (data not shown). II(S525C)-fluorescein was also catalytically inactive toward macromolecular substrates such as Factor V, fibrinogen, and itself (data not shown).

Determination of the Second Order Rate Constant Values (k 2 ) by Both Subsampling and End Point Analysis Using S2222
Hydrolysis-The second order rate constants for inhibition of FXa by antithrombin in the presence and absence of Ca 2ϩ , PCPS vesicles, and FVa were determined from the time courses of hydrolysis of the FXa substrate S2222. Two methods were employed. One involved subsampling at various intervals and measuring residual FXa activity with the substrate. The other involved monitoring the reactions containing the substrate to completion and calculating the value of the inhibition constants from the level of unconsumed substrate at the end of the reaction carried out in the presence of inhibitor. In the first method the time course of decay of enzymatic activity is directly measured; thus the results are unambiguous. In the second, the validity of the inferred value depends on the validity of the model used to interpret the data. The second method was used with this substrate in anticipation of the experiments with II(S525C)-fluorescein, in which the direct method would not be feasible because of the low levels of FXa used. Thus, the two types of measurements with S2222 were done to test the validity of the end point method. The analyses involving sub-  sampling are shown in Fig. 3. Two time courses of the reactions, along with their controls without AT, carried out with the substrate included (end point analysis) are shown in Fig. 4. In the subsampling experiments, plots of the natural logarithm of the residual activity versus time were linear and pseudo first order rate constants were measured from the slopes. In the end point analyses, reactions containing AT did not go to completion because the enzyme was consumed, which allowed determination of the pseudo first order rate constant for inhibition from the residual substrate concentration and the v max /K m ratio for the control. Initial efforts to find the v max /K m ratio by application of the integrated Michaelis-Menten equation were unsuccessful, because the time courses of product formation were first order, regardless of the initial substrate concentration, which is indicative of equal affinity binding of both the substrate and product to the enzyme. Thus, the K m was found by standard initial rate measurements and the K P for inhibition by product was found by measuring initial rates of substrate hydrolysis in the presence of the product at various concentrations (Fig. 5). K m was determined to be the same regardless of the presence or absence of Ca 2ϩ , PCPS, and FVa. The value inferred by non-linear regression of the data from Fig. 5 to the Michaelis-Menten equation was 234 Ϯ 40 M. The K P value for product inhibition obtained for the data inset in Fig. 5 was 338 Ϯ 14 M.
The k 2 values for the inhibition of FXa by antithrombin were determined for FXa alone, FXa ϩ Ca 2ϩ , FXa ϩ Ca 2ϩ ϩ PCPS, FXa ϩ Ca 2ϩ ϩ PCPS ϩ FVa, and FXa ϩ Ca 2ϩ ϩ PCPS ϩ II(S525C)-fluorescein. Table II summarizes the k 2 values determined using both methods. The values determined were in good agreement with each other. Thus, end point analysis was determined to be a valid method for measuring k 2 and was used in subsequent experiments with II(S525C)-fluorescein. Upon addition of Ca 2ϩ , FXa was more susceptible to inhibition by AT, which was expected (13). Addition of PCPS or PCPS ϩ II(S525C)-fluorescein to the system had little to no effect when compared with FXa ϩ Ca 2ϩ . Finally, addition of FVa decreased the second order rate constant by ϳ50% when compared with that for FXa ϩ Ca 2ϩ ϩ PCPS. In the absence of FVa, II(S525C)fluorescein had no effect on inhibition of FXa by AT. ried out, and the raw data were fit to the integrated Michaelis-Menten equation to determine K m and v max values. Unlike S2222 cleavage, prothrombin activation did not appear to exhibit strong product inhibition. Fig. 6 illustrates control and AT-containing experiments for reactions containing FXa, Ca 2ϩ , PCPS, FVa at various concentrations, and II(S525C)-fluorescein. The end points for the experiments containing AT increase with increasing FVa, suggesting increasing protection. Fig. 7A illustrates that, at saturating levels of FVa, prothrombinase becomes completely protected and the k 2 values for inhibition by AT fall to undetectable levels. The half-life at a FVa concentration of 0.75 nM is 69 min, which can be compared with the half-life in the absence of FVa of 2.9 min. Because this latter value is very similar to the half-life predicted from the FXa ϩ Ca 2ϩ ϩ PCPS ϩ II(S525C)-fluorescein subsampling experiments, prothrombin alone does not mediate protection of FXa from inhibition by AT. When a similar titration of FVa was carried out using S2222 as the substrate, the k 2 values were observed to fall by only ϳ50% at saturating levels of FVa (Fig.  7B). Therefore, in the presence of Ca 2ϩ and PCPS, FXa was observed to have a half-life of 2.0 min, which then increased to 3.7 min at a FVa concentration of 50 nM, suggesting that FVa alone does not mediate complete protection of FXa from inhibition by antithrombin. Thus, FVa and prothrombin must both be present to observe complete protection.

Effects of FVa on AT Inhibition of Prothrombinase in the
In the FVa titration in which II(S525C)-fluorescein was the substrate, a half-maximal effect was obtained at a FVa concentration of ϳ20 pM. Saturation of the effect was also observed when S2222 was the substrate, but the concentration of FVa at half-maximal effect was 3.0 nM.
Effects of Heparin on AT Inhibition of Prothrombinase-Heparin accelerates inhibition of FXa by AT by several orders of magnitude. Fig. 8 shows the calculated pseudo first order rate constants for a variety of heparin concentrations deter-mined by monitoring II(S525C)-fluorescein activation by prothrombinase. The corresponding k 2 value was determined from the slope of the line to be 1.27 ϫ 10 5 M Ϫ1 s Ϫ1 . The inset in Fig.  8 illustrates the pseudo first order rate constants obtained for analysis of the heparin-dependent AT inhibition of FXa ϩ FVa ϩ PCPS ϩ Ca 2ϩ determined by S2222 hydrolysis. The corresponding k 2 value was determined to be 7.72 ϫ 10 6 M Ϫ1 s Ϫ1 . The second order rate constant for FXa ϩ Ca 2ϩ ϩ PCPS was also determined in a similar manner and found to be 1.58 ϫ 10 7 M Ϫ1 s Ϫ1 . Table III summarizes the k 2 values determined in both the absence and presence of heparin. Heparin increased the k 2 for inhibition of FXa in the presence of Ca 2ϩ and PCPS by 5300-fold. This increase was also observed for FXa ϩ Ca 2ϩ ϩ PCPS ϩ FVa as determined by S2222 hydrolysis. It was impossible to calculate a -fold difference for prothrombinase inhibition in the presence of II(S525C)-fluorescein, because prothrombinase appears to be completely protected at saturating levels of FVa in the absence of heparin under these circumstances but is inhibited with a second order rate constant of 1.27 ϫ 10 5 M Ϫ1 s Ϫ1 upon heparin addition. Table III also summarizes the k 2 values in the presence of heparin relative to that for FXa ϩ Ca 2ϩ ϩ PCPS. The k 2 for prothrombinase in the presence of II(S525C)-fluorescein was found to be 0.8% of that for FXa ϩ Ca 2ϩ ϩ PCPS, implying profound protection of FXa, when it is incorporated into the prothrombinase complex and in the presence of prothrombin, even in the presence of heparin. The k 2 for prothrombinase as determined by S2222 hydrolysis, was one-half that obtained with FXa ϩ Ca 2ϩ ϩ PCPS. This value was 60-fold greater than that obtained with prothrombinase plus II(S525C)-fluorescein. The profoundly smaller value observed when II(S525C)-fluorescein is the substrate again suggests that the presence of both FVa and prothrombin causes increased protection of FXa from AT, even in the presence of heparin. Effects of the Concentration of Prothrombin on AT Inhibition of Prothrombinase-Reactions containing FXa ϩ FVa ϩ PCPS ϩ Ca 2ϩ and II(S525C)-fluorescein at various concentrations, in the presence and absence of 2 M AT and 60 pM heparin, were monitored until completion. End point analysis was carried out and the k 2 values were determined. Fig. 9 illustrates that reactions containing II(S525C)-fluorescein at concentrations from 50 to 500 nM exhibited similar k 2 values. Results in Table III indicate that the k 2 value in the absence of prothrombin should be 7.72 ϫ 10 6 M Ϫ1 s Ϫ1 , which is 50-fold higher than the average k 2 value in Fig. 9. Thus, prothrombin does appear to play a protective role, but, perhaps similar to FVa, the protective effects of prothrombin saturate at very low concentrations. Unfortunately, it is impossible to accurately monitor activation of II(S525C)-fluorescein at concentrations below 50 nM, because the change in fluorescence is too small. DISCUSSION We have expressed a variant of prothrombin in which the serine of the thrombin catalytic triad has been replaced with a cysteine that was subsequently labeled with fluorescein. Characterization of the change in fluorescence upon cleavage, the rate of activation in the presence and absence of FVa and the absence of activity upon activation of this variant showed it to be a good model substrate to investigate its activation by the enzyme FXa in the presence of the physiological inhibitor antithrombin. This variant allows easy monitoring of the conversion of prothrombin to an inactive thrombin and would be a useful tool for monitoring prothrombin activation in the absence of all thrombin feedback reactions, such as the activation of FV, and cleavage of itself.
It has been well established that FXa is less susceptible to inhibition by antithrombin and the antithrombin-heparin complex when it is incorporated into the prothrombinase complex (14 -20). The extent of this protection and the mechanism behind it, however, have not been clearly established. Inhibition of FXa by antithrombin and the antithrombin-heparin complex under various conditions was quantitatively determined in this study. In the absence of FVa, the combination of phospholipid vesicles and prothrombin had no effect on the second order rate constant of inhibition of FXa ϩ Ca 2ϩ by antithrombin; FVa in the absence of prothrombin showed only a modest effect, decreasing the rate constant approximately 2-fold. Upon addition of both FVa and prothrombin, however, the second order rate constant decreased to an immeasurable level, suggesting that, when both FVa and prothrombin are present, prothrombinase is extremely protected from antithrombin, with its half-life increasing from 3 min in the absence of FVa to greater than 69 min with prothrombin and FVa. Upon addition of heparin, FVa in the absence of prothrombin was again observed to decrease the second order rate constant of inhibition by ϳ50%. In the presence of both FVa and prothrombin, the second order rate constant decreased to 0.8% of that for FXa ϩ Ca 2ϩ ϩ PCPS, implying profound protection of prothrombinase even in the presence of heparin. In the presence of heparin at 0.2 unit/ml (40 nM), the half-life increased from 1 s without prothrombin to 136 s with it.
Prothrombin clearly effects the protection of prothrombinase. Previous studies have suggested that the thrombin produced at the catalytic surface consumes antithrombin and therefore lowers its concentration in the vicinity of the prothrombinase complex (15). In the present study however, active thrombin was not produced and no change in the fluorescence of IIa(S525C)-fluorescein was observed upon antithrombin addition, implying that the variant used in this study does not associate with and therefore does not consume antithrombin.
However, the remarkable extent to which FXa is protected from inhibition by both antithrombin and the antithrombinheparin complex also cannot be rationalized based on straight forward competition between prothrombin and antithrombin according to a simple Michaelis-Menten model. Thus, some other aspect of the conversion of prothrombin to thrombin by prothrombinase is responsible for the protection. One possibility is that the interactions of FXa with both prothrombin and antithrombin require an exosite on FXa distinct from the S1 subsite of the enzyme. That such an exosite exists on FXa is supported by the data of Krishnaswamy and Betz (31). If this rationalizes the protection, the exosite is not recognized by FIG. 8. Effects of heparin on the pseudo first order rate constant. Reactions containing heparin at various concentrations, 2 M AT, 60 pM FXa, 50 M PCPS, 5 nM FVa, 1 M II(S525C)-fluorescein, and 5 mM CaCl 2 in TBS, 0.01% Tween 80, were monitored until completion at excitation and emission wavelengths of 490 nm and 538 nm, respectively, with a 515-nm emission filter in place. Reactions were started by addition of prothrombinase. End point analysis was carried out as outlined under "Experimental Procedures" to determine the pseudo first order rate constants, which were plotted versus [heparin]. The corresponding slope was used to calculate k 2 . Inset, reactions containing heparin at various concentrations, 2 M AT, 15 nM FXa, 50 M PCPS, 50 nM FVa, 400 M S2222, and 5 mM CaCl 2 in TBS, 0.01% Tween 80, were monitored until completion and treated as above.  9. Effects of prothrombin concentration on k 2 . Reactions containing 60 pM FXa, 5 nM FVa, 50 M PCPS, and II(S525C)-fluorescein at various concentrations, in the presence and absence of 2 M AT and 60 nM heparin were monitored until completion at excitation and emission wavelengths of 490 nm and 538 nm, respectively, with a 515-nm emission filter in place. End point analysis was carried out as outlined under "Experimental Procedures," and k 2 values were determined. prothrombin in the absence of FVa, because prothrombin alone did not protect FXa. Alternatively, the protection is inherent in the dynamics of the catalytic cycle. Conceivably, for most of the cycle, FXa is in a conformation or state in which it does not recognize antithrombin. If this is so, then FXa would be accessible for only about 1% of the cycle, because the rate constant for inhibition decreases by a factor of ϳ100 for inhibition of FXa by the antithrombin-heparin complex. Regardless of the mechanistic reason for the protection, the fact that it exists suggests that prothrombin activation cannot be accounted for by simple substrate competition and the Michaelis-Menten model.
The serine proteases of other multicomponent enzymatic complexes of the coagulation cascade (factor VIIa and factor IXa) are also inhibited by antithrombin. Perhaps they are similarly protected from inhibition in the presence of the substrate (factor X), and thus the phenomenon reported here for prothrombinase may apply generally to all of the reactions of the cascade. If this is the case, the reactions of the cascade localized at the site of injury would be relatively protected from antithrombin, whereas enzymes that might escape would be susceptible to inhibition, thereby both allowing efficient thrombin formation locally, while preventing or attenuating it systemically.