Role of the α-Helix 163-170 in Factor Xa Catalytic Activity*

Factor Xa (FXa) is a key protease of the coagulation pathway whose activity is known to be in part modulated by binding to factor Va (FVa) and sodium ions. Previous investigations have established that solvent-exposed, charged residues of the FXa α-helix 163-170 (h163-170), Arg165 and Lys169, participate in its binding to FVa. In the present study we aimed to investigate the role of the other residues of h163-170 in the catalytic functions of the enzyme. FX derivatives were constructed in which point mutations were made or parts of h163-170 were substituted with the corresponding region of either FVIIa or FIXa. Purified FXa derivatives were compared with wild-type FXa. Kinetic studies in the absence of FVa revealed that, compared with wild-type FXa, key functional parameters (catalytic activity toward prothrombin and tripeptidyl substrates and non-enzymatic interaction of a probe with the S1 site) were diminished by mutations in the NH2-terminal portion of h163-170. The defective amidolytic activity of these FXa derivatives appears to result from their impaired interaction with Na+ because using a higher Na+ concentration partially restored normal catalytic parameters. Furthermore, kinetic measurements with tripeptidyl substrates or prothrombin indicated that assembly of these FXa derivatives with an excess of FVa in the prothrombinase complex improves their low catalytic efficiency. These data indicate that residues in the NH2-terminal portion of the FVa-binding h163-170 are energetically linked to the S1 site and Na+-binding site of the protease and that residues Val163 and Ser167 play a key role in this interaction.

this process, FX is activated to FXa and forms a high affinity macromolecular complex with other components of the prothrombinase complex, factor Va (FVa), negatively charged phospholipid surfaces, and calcium to activate prothrombin to thrombin (1)(2)(3)(4)(5)(6). These macromolecular interactions lead to an increase of 5 orders of magnitude in the catalytic efficiency of FXa toward prothrombin (2,7). Enhancement of the k cat of the reaction is mainly due to the cofactor function of FVa. Two basic residues of h163-170 5 of the protease domain of FXa, namely Arg 165 and Lys 169 , directly interact with FVa (8,9). All known sequences from different species in this surface-exposed helix of FXa are similar. Interestingly despite being stimulated by different cofactors, the catalytic domains of other blood coagulation proteins, such as factor IXa (FIXa) and factor VIIa (FVIIa), share the same cofactor-dependent activity binding site based on the structural equivalences with chymotrypsin (10 -12).
Like other serine proteases of blood coagulation, small ligands such as calcium and sodium can allosterically modulate the activity and the specificity of FXa (13)(14)(15)(16)(17)(18)(19)(20) by binding to several exposed surface loops near or remote from the catalytic pocket of the enzyme (21). According to the three-dimensional structure of FXa (22), the FXa Na ϩ -binding site is close to the catalytic pocket of the enzyme and to the FVa-binding h163-170 ( Fig. 1). Furthermore there is some evidence that both the FVa-and Na ϩ -binding sites of FXa are energetically linked (20,23). Altogether these observations suggest that there is an allosteric linkage between the Na ϩ -binding site and FVa-binding h163-170.
In the current study, the relationship between h163-170, which is a crucial FVa-binding site, and the Na ϩ -binding site were investigated. To this end, FXa derivatives were designed in which h163-170 was either substituted by the corresponding region of FVIIa and FIXa or mutated (Fig. 2) at residues known, based on the three-dimensional structure of FXa, to be solventexposed (Arg 165 ) or to be proximate to the Na ϩ -binding site (Val 163 and Ser 167 ). The FXa derivatives were expressed in mammalian cells, and the purified activated forms were functionally characterized.
Enzyme-linked immunosorbent assay using anti-FX polyclonal antibodies was used to assay pd-FX and FX derivative proteins. FX was expressed in units where 1 unit is the amount present in 1 ml of normal human plasma. Proteins were quantified by the method of Bradford (24) using BSA as a standard. Molar concentrations of FXa (recombinant or plasmatic) and ␣-thrombin were determined by active site titration (see below).
Recombinant FX Derivatives-The mammalian expression plasmid pKG5 containing human FX cDNA and encoding wildtype (wt) FX (25) was used as a template for standard PCR mutagenesis to generate cDNAs encoding FX mutants carrying specific residues of FIX or FVII named FX/FVII 163-170 , FX/FVII 163-167 , FX/FVII 168 -170 , FX/FIX 163-170 , and FX R165A (Fig. 2). The mutated full-length cDNAs cloned into the expression vector pKG5 were checked by DNA sequence analysis using the ABI PRISM Dye Terminator Cycle Sequencing Reaction kit version 3.1 (Applied Biosystems Applera, Courtaboeuf, France) on an ABI PRISM 310 DNA sequencer according to the manufacturer's specifications. All constructs were transfected into Madin-Darby canine kidney cells using calcium phosphate precipitation, and cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Stable cell lines producing recombinant FX derivatives and wt-FX were prepared as already described (25) and maintained in 300-cm 2 flasks for protein production in serum-free Dulbecco's modified Eagle's medium/F-12 supplemented with 800 g/ml Geneticin G-418, 100 units/ml penicillin, 100 g/ml streptomycin, 5 g/ml vitamin K 1 , and 1% insulin-transferrin-selenium-X. FXcontaining medium was harvested every 48 h. Benzamidine and phenylmethylsulfonyl fluoride were added to a final concentration of 10 and 2 mM, respectively, and the medium was centrifuged (6000 ϫ g), passed over cellulose acetate membranes (0.45 m) to eliminate cell debris, and stored at Ϫ80°C. Conditioned medium was thawed at 37°C. EDTA was added to a final concentration of 5 mM. The medium was diluted to bring the final NaCl concentration to 60 mM. The mixture was then  (22). The backbone structure is represented by a ribbon. The catalytic triad residues, located in the center of this representation, are (from left to right) Asp 102 , His 57 , and Ser 195 (chymotrypsin numbering) and are shown as black ball-and-stick; the S1 site (Asp 189 ) is also shown. The h163-170, which is an FVa-binding site (8,9), and the two loop segments 183-189 and 221-225, which constitute a Na ϩ -binding site (35), are shaded black. stirred at room temperature for 30 min with QAE-Sephadex A-50 beads to achieve a final concentration of 0.25% (w/v). Beads were washed before elution with 50 mM Tris (pH 7.4), 500 mM NaCl, and 10 mM benzamidine. Recombinant FX contained in the eluted fractions (enzyme-linked immunosorbent assay) was immediately dialyzed against 15 mM Tris (pH 7.4), 15 mM NaCl containing 10 mM benzamidine, and 4 mM EDTA. Proteins were incubated at 4°C overnight with Q-Sepharose Fast Flow resin (1 ml of resin/50 ml of concentrated proteins). The resin was pre-equilibrated in 15 mM Tris (pH 7.4), 15 mM NaCl containing 10 mM benzamidine, and 4 mM EDTA. The resin was loaded on a column, washed with 200 ml of equilibration buffer containing 2 mM EDTA, and then washed with 100 ml of equilibration buffer without EDTA. Recombinant FX was eluted from the column with a 0 -60 mM CaCl 2 gradient at a flow rate of 1 ml/min. Recombinant FX fractions were pooled, concentrated on Amicon Ultra 30,000 molecular weight cutoff filters, dialyzed extensively against 50 mM Tris (pH 7.4) and 50 mM NaCl, and stored at Ϫ80°C. Before analysis, a final pass over a benzamidine-Sepharose column equilibrated with 50 mM Tris (pH 7.4) and 50 mM NaCl was used to eliminate trace contaminants of FXa that may have been generated during production or purification of the recombinant protein. Furthermore 1 h prior to use as a zymogen, FX derivatives were incubated with 1 mM phenylmethylsulfonyl fluoride to neutralize any trace of FXa. Control experiments indicated that after 30 min in Tris-HCl buffer phenylmethylsulfonyl fluoride was fully hydrolyzed and would not interfere with other reactions. Protein purity was assessed using 15% SDS-polyacrylamide gel electrophoresis analysis of the FX derivatives under reducing (50 mM dithiothreitol, final concentration) and non-reducing conditions followed by staining with Coomassie Brilliant Blue R-250. NH 2 -terminal sequence analysis was carried out after the purified recombinant FX derivatives were reduced and loaded onto a 15% SDS-polyacrylamide gel. The resolved proteins were transferred to an Immobilon membrane and stained with Ponceau S. The light chains were excised and sequenced using an Applied Biosystems Procise model 494 sequencer in the sequencing facility of the Institut de Biologie et Chimie des Protéines (Lyon, France). This analysis indicated that for all FX derivatives the signal sequence and the propeptide were accurately and efficiently removed before secretion. In addition, sequence analysis disclosed that the average yield for the two glutamic acid residues at the NH 2 terminus was less than 5% of the average yield of the two subsequent residues. Because ␥-carboxylation reduces the yield of Glu residues, these data indicate that the glutamic residues were properly modified.
Activation of FX by RVV-X-Recombinant FX derivatives and wt-FX were activated by RVV-X (26); 1 mg of it was coupled to 1 ml of CNBr-activated Sepharose 4B according to the manufacturer's instructions. Recombinant FX derivatives (1 M) were incubated with coupled RVV-X (30 nM) in 50 mM Tris (pH 7.4) and 100 mM NaCl containing 10 mM CaCl 2 . Time course analysis of the activation reactions by SDS-PAGE indicated that all FX zymogens were fully converted to their active forms within 1 h under these experimental conditions. After 2 h, the reaction was stopped by addition of 15 mM EDTA. Activated recombinant FX derivatives were loaded onto a 1-ml HiTrap heparin-Sepharose column and eluted with 50 mM Tris-HCl (pH 7.4) containing 0.5 M NaCl and 5 mM CaCl 2 . The eluate was immediately dialyzed against 50 mM Tris (pH 7.4) and then against 50 mM Tris (pH 7.4) and 75 mM NaCl and precipitated by addition of solid ammonium sulfate to 80% saturation. The precipitated proteins were dissolved in 50 mM Tris (pH 7.4) and 10 mM NaCl, and the protein solution was immediately dialyzed against the same buffer containing 50% glycerol (v/v) and stored at Ϫ20°C until use. Recombinant FXa appeared pure by SDS-PAGE and was indistinguishable from commercially available pd-FXa (Kordia). Furthermore wt-FXa was identical to pd-FXa with respect to its K m and k cat values for the hydrolysis of the chromogenic substrates S-2765 and SpeFXa as well as the rate of prothrombin activation within prothrombinase. Recombinant wt-FXa and commercial pd-FXa were nevertheless systematically compared in all experiments reported in this study and found to be virtually identical.
Interaction of FXa with D-FFR-CK-The rate constant k on for the interaction between D-FFR-CK and pd-FXa, wt-FXa, or recombinant FXa derivatives was evaluated as follow. First titrated ␣-thrombin, with p-nitrophenyl-pЈ-guanidinobenzoate hydrochloride, was used to determine the precise concentration of D-FFR-CK aliquots in 1 mM HCl. Briefly 250 nM ␣-thrombin in kinetic buffer (50 mM Tris-HCl (pH 7.5) containing 75 mM NaCl, 5 mM CaCl 2 , 0.1% (w/v) BSA, and 0.1% polyethylene glycol 8000) was incubated for 3 h at room temperature with various amount of D-FFR-CK (0.03-6 mM). The reaction mixture was diluted 1:10 in kinetic buffer containing 100 mM S-2238, and the remaining enzyme concentration was estimated from the rate of A 405 increase. The release of paranitroanilide (pNA) was recorded at 405 nm as a function of time (i.e. the initial rate of S-2238 hydrolysis) in a kinetics microplate reader (Bio-Tek Instruments, Winooski, VT). The initial concentrations of D-FFR-CK aliquots were deduced from the intercept to the x axis of a linear plot of the remaining activity versus the amount of inhibitor added.
Subsequently a sufficient amount of activated enzyme (30 -500 nM, estimated from the published extinction coefficient of FXa of 1.25 ml mg Ϫ1 cm Ϫ1 at 280 nm) (27) to obtain a readily detectable amidolytic activity (10% hydrolysis of S-2765 in 30 min) was incubated for a given period of time (10 s to 5 h) in the presence of a fixed concentration of D-FFR-CK in kinetic buffer at 25°C. Typically three concentrations of D-FFR-CK were used that correspond to 10, 20, and 40 times the target concentration to ensure that the reaction occurs under pseudo-first order conditions. At the end of each incubation, 100 or 500 M S-2765 was added, and the residual amidolytic activity was monitored at 405 nm in a kinetics microplate reader. By plotting the rate of S-2765 hydrolysis as a function of the incubation time of D-FFR-CK and its target enzyme, a curve was obtained. By non-linear regression, using Equation 1, the rate constant for inactivation of the target enzyme can be estimated.
The parameters A t , A 0 , and A min represent the residual activity at time t, the initial activity at time 0, and the activity at infinite time, respectively. Under pseudo-first order condition, the OCTOBER 26, 2007 • VOLUME 282 • NUMBER 43 JOURNAL OF BIOLOGICAL CHEMISTRY 31571 value obtained for k is equal to the concentration of inhibitor multiplied by its k on for the enzyme.

Characterization of a Factor Xa Cofactor-binding Site
Titration of FXa-The active site concentrations of FXa (plasmatic and activated recombinant FX derivatives) were determined by titration with known concentrations of D-FFR-CK (see above). For FXa (plasmatic and recombinant derivatives) titration, 1 mM enzyme (determined from the published extinction coefficient of FXa at 280 nm (27)) was incubated with 50 nM to 10 mM D-FFR-CK in kinetic buffer at room temperature. The incubation was sustained until the reaction was complete, i.e. a minimum of 10 half-lives were covered. The half-life of the reaction is equal to the natural logarithm of 2 divided by the product of the k on value of D-FFR-CK for each FXa derivatives (see above) and the concentration of the inhibitor. At the end of this incubation, remaining free enzyme was measured with 100 M S-2765 as substrate. The residual amidolytic activity was monitored at 405 nm in a kinetics microplate reader. By plotting the rate of S-2765 hydrolysis as a function of D-FFR-CK concentration, a straight line was obtained for which the abscissa at the origin corresponds to the initial concentration of the active enzyme.
Activation of Prothrombin by FXa-The initial rate of prothrombin activation by FXa (plasmatic and activated recombinant FX derivatives) was measured in both the absence and presence of FVa as described previously (28). Briefly in the absence of the cofactor, the concentration dependence of prothrombin activation was studied by incubating each FXa derivative (5-10 nM) with increasing concentrations of the zymogen (20 -3000 nM) in the presence of a saturating concentration (30 M) of phospholipid vesicle preparation (PC:PS, 3:1) in 50 mM Tris (pH 7.5) and 100 mM NaCl with 5 mM CaCl 2 and 0.2% (w/v) BSA at 25°C. Phospholipid vesicles (PC:PS, 3:1) of nominal 100-nm diameter were synthesized by the method of membrane extrusion (29). Phospholipid concentration was determined by phosphate analysis. Aliquots of the reaction mixture were taken at specified times, and the reaction was stopped in EDTA (10 mM final concentration). Thrombin formation in each aliquot was determined by measuring the amidolytic activity of the sample toward the synthetic substrate S-2238 (250 M) in the presence of 15 g/ml soybean trypsin inhibitor to inhibit amidase activity of FXa. During the assay, it was ensured that less than 5% of prothrombin was converted to thrombin, and thrombin formation was linear. Conversion of substrate was monitored at 405 nm. Concentrations of thrombin generated in the activation reactions were determined from a standard curve derived from the cleavage rate of S-2238 by known concentrations of thrombin under the same conditions.
In experiments using various concentrations of FVa, the rates at which activated recombinant FX derivatives, wt-FXa, and pd-FXa can activate prothrombin to thrombin in the presence of a saturating concentration of phospholipid vesicles were compared as described previously with slight modifications (28). Briefly 50 M phospholipid vesicles (PC:PS, 3:1) and 20 pM FXa were incubated at 37°C for 5 min with various concentrations of FVa (0 -25 nM). The reaction was started by the addition of 2 M prothrombin, and the assay was performed at 37°C in 50 mM Tris (pH 7.4), 100 mM NaCl containing 5 mM CaCl 2 , and 0.2% (w/v) BSA. Aliquots of the reaction mixture were taken and stopped in EDTA (10 mM final concentration) at the specified times. In experiments using various concentrations of prothrombin (0 -4000 nM) and a saturating concentration of FVa (25 nM), the same protocol as described above was followed. During all assays, it was ensured that less than 10% of prothrombin was converted to thrombin, and thrombin formation was linear. Thrombin concentration was calculated as described above.
Initial velocity measurements of prothrombin hydrolysis by FXa derivatives alone or in the prothrombinase complex were analyzed by fitting the data to the Henri-Michaelis-Menten equation to yield fitted values for K m and k cat . Non-linear regression was used to derive apparent dissociation constants K d(app) for the interaction between FXa derivatives and phospholipid vesicle-bound FVa using the initial rate of thrombin formation in the presence of different concentrations of FVa and a single ligand-binding site equation.
Chromogenic Substrate Cleavage by FXa-The steady-state kinetics of hydrolysis of S-2765 and SpeFXa by FXa derivatives were assayed at 37°C in kinetic buffer with various salt combinations. Kinetics parameters of substrate hydrolysis were determined using enzyme concentrations of 0.5, 2.0, or 6.0 nM and various substrate concentrations ranging from 20 to 2000 M. The release of pNA was monitored at 405 nm at 37°C in a kinetics microplate reader.
Apparent dissociation constants K d(app) for the interaction between FXa derivatives and Na ϩ were obtained from the dependence of the initial rate of pNA formation on the concentration of Na ϩ (10 -800 mM) in the presence of various concentrations of S-2765 (25-2500 M). It has been demonstrated previously that the amidolytic activity of FXa is not dependent on the ionic strength of the reaction buffer in the range of NaCl used (19,23,30). Thus, no compensating chloride salt was added to the reactions. The data were fitted to a single ligandbinding site with a defined background amidolytic activity in the presence of 10 mM Na ϩ using Equation 2, where K d(app) is the apparent dissociation constant of Na ϩ for an FXa derivative in the kinetic buffer used, y is the k cat of the substrate hydrolysis at a given Na ϩ concentration depicted by x, k cat(min) is the corrected background k cat in the absence (in fact 10 mM) of Na ϩ , and k cat is the corrected k cat at saturating concentration of Na ϩ .
In the presence of a saturating concentration of FVa (25 nM), the concentration dependence of SpeFXa cleavage was studied by incubating each FXa derivative (0.5 and 2 nM) with increasing concentrations of the substrate (10 -2000 M) on phospholipid vesicles prepared at a saturating concentration (50 M). Initial velocity measurements of pNA generated by FXa derivatives alone or in the prothrombinase complex were analyzed by fitting the data to the Henri-Michaelis-Menten equation to yield fitted values for K m and k cat .
Inhibition of FXa Derivatives Alone or in the Prothrombinase Complex by PAB-The inhibitory constant (K i ) of PAB for FXa derivatives alone or in the prothrombinase complex was assessed assuming classical competitive inhibition by initial velocity measurements of S-2765 hydrolysis by each enzyme using increasing concentrations of substrate (0 -1000 M) at different fixed concentrations of PAB (0 -500 M) as described previously (31). Initial velocities were measured at 37°C in kinetic buffer using enzyme concentrations of 0.5, 2.0, or 6.0 nM with or without 25 nM FVa and 50 M phospholipid vesicles (PS/PC). The concentrations were chosen to saturate the prothrombinase complexes with the FXa derivatives based on the equilibrium constants determined above. The linear dependence of rate on the concentration of FXa at saturating concentrations of FVa and PS/PC was established in separate experiments. Initial velocity measurements of substrate hydrolysis by FXa derivatives alone or in the prothrombinase complex were analyzed according to the rate expression for linear competitive inhibitor to yield the fitted values for K m , V max , and K i .
Inhibition of FXa Derivatives Alone or in the Prothrombinase Complex by Antithrombin III-The rate of inactivation of FXa derivatives alone or in the prothrombinase complex by ATIII was measured under pseudo-first order rate conditions by a discontinuous assay. For the inhibition of FXa derivatives, ATIII (0 -0.7 M) was incubated with FXa (5 or 10 nM) in kinetic buffer up to 120 min (12 time points). At the end of the time course, S-2765 was added (100 M final concentration) to monitor residual enzyme activity. For the inhibition of FXa derivatives in the prothrombinase complex, ATIII (0 -7 M) was incubated with FXa (5 nM), FVa (30 nM), and phospholipids vesicles (50 M) in kinetic buffer for up to 60 min (12 time points). Residual enzyme activity was monitored as described above. By plotting the rate of S-2765 hydrolysis as a function of the incubation time of ATIII and its target enzyme, a curve was obtained. By non-linear regression, using Equation 1 (see above), the rate constant for inactivation of the target enzyme was estimated. In both assays, each measurement was made in three separate experiments.
Molecular Modeling-To explore the possible structural consequences of the FXa V163A and FXa S167A mutations near the interface with Tyr 225 , energy minimization and molecular dynamics simulations were performed with GROMACS 3.3.1 (32) based on the 2.2-Å structure of wt-FXa (Protein Data Bank accession code 1HCG (22)). The wild-type structure was used as a control and treated in an identical fashion to the mutants. A strategy using positional restraints was chosen to broadly constrain the conformation of the protein to that observed in the crystal structures of FXa while permitting deviation particularly in the vicinity of Tyr 225 , Val 163 , and Ser 167 . Side-chain C␥-1 and C␥-2 and O␥ atoms were deleted to generate the FXa V163A and FXa S167A mutants. The two mutants and wildtype protein were modeled in an 81 ϫ 81 ϫ 81-Å cubic box with boundaries at least 9.5 Å from the protein, solvated using the Simple Point Charge model for solvent with ϳ18,800 water molecules, and subjected to 1000 steps of steepest descent energy minimization using a grid search and default parameters. Subsequent molecular dynamics simulations used periodic boundary conditions, 2-fs steps, weak coupling (0.1 s Ϫ1 ) to a 300 K temperature bath separately for protein and solvent, coupling (0.5 s Ϫ1 ) to a pressure sink (1 bar), Linear Constraint Solver restraints on bond lengths, the Particle Mesh Ewald model for electrostatic interactions with a 10-Å cutoff, and the ffG43a1 (GROMOS96) force field (33). Initially a 100-ps molecular dynamics simulation applying medium side-chain/high main-chain (1000 kJ mol Ϫ1 nm Ϫ2 /10,000 kJ mol Ϫ1 nm Ϫ2 ) positional restraints was undertaken. Following this, these restraints were relaxed to 10/100 for residues 145-151 and 224 -227 and 100/1000 for residues within 10 Å with a smooth transition to 1000/10,000 at a distance of 18 Å or more. A 1-ns molecular dynamics simulation was then performed. For each variant, the average structure was determined from the simula-   Table 1. OCTOBER 26, 2007 • VOLUME 282 • NUMBER 43 JOURNAL OF BIOLOGICAL CHEMISTRY 31573 tion and subjected to energy minimization as detailed above. The progress of the simulation and the resulting structures were visualized using PyMOL (36).

RESULTS
Recombinant Proteins-All recombinant FX proteins were subjected to digestion by RVV-X under conditions similar to those of pd-FX. All FX derivatives could be converted into their active form, and the final activated preparations were more than 90% active as determined by active site titration. Mutations introduced within h163-170 of the protease domain are shown in Fig. 2.
Prothrombin Activation with Increasing FVa Concentrations-Previous studies have indicated that substitution of two basic residues exposed at the surface of h163-170 interferes with the formation of the FXa-FVa complex (8,9,34). We decided to explore the effect of substituting other residues within h163-170 on prothrombin activation by the FXa-FVa complex. For this purpose, the rate of thrombin generation was studied as a function of increasing FVa concentration in the presence of an excess of phospholipid (50 M) and prothrombin (2 M). Under these conditions, prothrombin represents a saturating concentration of substrate for all constructs (see results below) and twice the physiological concentration found in plasma. For all the recombinant FXa constructs tested, the rate of thrombin formation was saturable and dependent on the concentration of FVa (Fig. 3).
The apparent affinity (K d ) of wt-FXa for FVa was determined to be 0.11 nM, and a similar value (0.09 nM) was obtained for the FX/FIX 163-170 chimera (Table 1). In contrast, substitution of h163-170 by the corresponding residues of FVII had a signifi-  Table 1.

TABLE 1 Kinetic constants for the cleavage of prothrombin by FXa derivatives in the absence and presence of FVa and the apparent dissociation constants for their interaction with FVa
The kinetic parameter k cat and K m(app) values were determined from the concentration dependence of prothrombin activation by FXa derivatives on phospholipid vesicles in the absence and presence of saturating concentration of FVa ( (8,9). This defect was observed to a lesser extent for FXa V163A and FXa S167A . Prothrombin Activation with Increasing Prothrombin Concentration-Next the rate of thrombin generation was studied as a function of increasing concentrations of prothrombin in the presence of an excess of phospholipids (50 M) and a concentration of FVa representing the physiological concentration of FV found in plasma (25 nM). Under these conditions, the assembly of the FXa derivatives in the prothrombinase complex partially restored, when it was defective, the catalytic efficiency of the enzyme toward prothrombin to that observed with wtand pd-FXa (Table 1).
In contrast, the absence of FVa yielded a different profile of prothrombinase activity among the FXa variants. The rate of thrombin formation was assessed in the presence of saturating concentrations of phospholipids (50 M) and various concentrations of prothrombin but in the absence of FVa. Under these conditions, the FXa/FIX 163-170 chimera displayed the same catalytic efficiency as wt-FXa toward prothrombin (Fig. 4A) as observed in the presence of an excess of FVa (Fig. 4B). Similarly and as previously reported (9), FXa R165A had the same catalytic efficiency as wt-FXa (Table 1 and Fig. 4A). In contrast, substitution of h163-170 by the corresponding residues of FVII and the V163A and S167A mutations had a significant effect on the catalytic efficiency of FXa toward prothrombin in the absence of FVa (Table 1 and Fig. 4A). These results suggest that h163-170, more specifically the residues at the amino-terminal part of h163-170 but not Arg 165 , had a deleterious effect on the catalytic activity of FXa toward prothrombin. The assembly of the defective FXa derivatives in the prothrombinase complex, containing 25 nM FVa, was found to correct their catalytic efficiency toward prothrombin.
Amidolytic Activity of Uncomplexed Protease-To further explore whether substitutions in h163-170 affect the catalytic groove of FXa, the amidolytic activity of activated forms of pd-FX, wt-FX, and all FXa derivatives were compared. To this purpose, hydrolysis of various concentrations of two synthetic substrates, namely S-2765 and SpeFXa, was monitored as described under "Experimental Procedures." FXa/FIX [163][164][165][166][167][168][169][170] and FXa R165A displayed rates of catalytic efficiency similar to those of wt-and pd-FXa (Table 2). These results indicate that the catalytic groove of these two constructs is identical to that of pd-and wt-FXa. In contrast, an increase in K m and a decrease k cat were observed for FXa/FVII 163-167 , FXa/FVII 163-170 , FXa V163A , FXa S167A , and to a lesser extent FXa/FVII 168 -170 compared with those of wt-and pd-FXa (Table 2). These data indicate that these mutations altered the amidolytic activity of FXa, and hence suggest that the substrate binding cleft (S1-S3 site) is changed in some fashion.
Amidolytic Activity of Complex-bound Protease-In contrast, the assembly of the FXa derivatives in the prothrombinase complex restored or partially restored their catalytic efficiency TABLE 2 Kinetic constants for the cleavage of peptidyl substrates by FXa derivatives alone or in the prothrombinase complex against the peptidyl substrate tested to that observed with wtand pd-FXa with the exception of FXa/FVII 163-167 and FXa/ FVII 163-170 . This was mediated through changes in K m values, whereas the k cat values were not significantly changed ( Table 2). The minimal restoration of the catalytic efficiency observed for FXa/FVII 163-170 and FXa/FVII 163-167 will be discussed later. In conclusion, the data obtained are consistent with the hypothesis that prothrombinase complex assembly corrects defective binding of peptidyl substrates introduced by mutations in h163-170 of FXa.
Interaction of a Probe with the S1 Site-The binding of a well defined S1 probe, PAB, to FXa was also investigated. FXa/ FIX 163-170 and FXa R165A displayed affinity for PAB similar to that of wt-FXa or pd-FXa, whereas FXa/FVII 163-167 , FXa/ FVII 168 -170 , FXa/FVII 163-170 , FXa V163A , and FXa S167A displayed lower rates of affinity, indicating an alteration of their S1 specificity pocket (Table 3). Similar results were also observed for the interaction with both D-FFR-CK and ATIII (Table 3). In contrast, assembly of the FXa derivatives in the prothrombinase complex, containing an excess of FVa, restored ATIII and PAB binding to that observed with wt-and pd-FXa with the exception of FXa/FVII 163-167 and FXa/FVII 163-170 (Table 3). In conclusion, the data obtained show that prothrombinase complex assembly corrects defective binding of peptidyl substrates into the S1 pocket by molecules containing mutations in h163-170.
The Effect of Sodium Ions on Amidolytic Activity-Previous results (20,23) have indicated that there is an allosteric linkage between the FVa-binding site(s) and the Na ϩ -binding site of FXa. To test the hypothesis that h163-170, which is an FVabinding site, is linked to the Na ϩ -binding site, Na ϩ dependence of the amidolytic activity toward S-2765 of pd-FXa, wt-FXa, and recombinant FXa derivatives was assayed. Na ϩ improved the amidolytic activity of wt-FXa and all FXa derivatives (Fig. 5). At low (10 mM) Na ϩ concentration, the k cat values for FXa/ FVII 163-170 and FXa/FVII 163-167 were reduced compared with wt-FXa, whereas all the other FXa derivatives had k cat values similar to that of wt-FXa (Fig. 5A). At 150 mM Na ϩ , which is the physiological concentration tightly regulated in blood, only FXa/ FIX 163-170 and FXa R165A displayed catalytic efficiency similar to that of wt-FXa, whereas the other FXa derivatives displayed reduced catalytic efficiencies (Fig. 5B). Comparison of the catalytic parameters as a function of Na ϩ concentration showed that Na ϩ diminished the K m values and to a lesser extent increased the k cat values (Fig. 6).
Notably it was observed that Na ϩ reduced the differences between the catalytic parameters of the variants toward the peptidyl substrate tested (S-2765) compared with wt-and pd-FXa. This was affected predominantly through changes in K m values. When k cat and K m values are  plotted as a function of Na ϩ concentration (Figs. 6, A and B, respectively) the midpoint of the curves yields the apparent K d for Na ϩ interaction with wt-FXa and FXa derivatives in the amidolytic assay using S-2765. Calculated K d values show a significant decrease of the affinity of FXa V163A , FXa S167A , FXa/ FVII 163-170 , and FXa/FVII 163-167 for Na ϩ compared with wt-FXa (Table 4). Altogether these results support the notion that in the presence of 150 mM Na ϩ (close to the physiological concentration found in plasma) the impairment of the amidolytic activity of FXa derivatives, when observed, is caused by their impaired interaction with Na ϩ .

DISCUSSION
In the present investigation, we extended the finding of previous studies that have implicated the basic residues of h163-170 (in particular Arg 165 ) in the regulation of FXa by cofactors (8,9). Indeed our data show (Fig. 3) that FXa derivatives with a mutation at position Arg 165 exhibited a severe defect in thrombin formation in the presence of small amount of FVa. The same defect was observed, albeit to a lesser extent, in FXa/ FVII 168 -170 bearing the substitutions K169L and L170Q. Surprisingly replacement of h163-170 with the FIXa h163-170 was without effect on FVa binding (Fig. 2). This is consistent with the structure of FXa and the data presented in this study, however. Both Lys 169 and Leu 170 are solvent-exposed and protrude away from the enzyme; hence neither mutation would be expected to alter the local conformation of h163-170. It has also been reported that basic residues are important for the interaction of FXa with its cofactor (8), and FIXa possesses an Arg at position 170; essentially the hydrophobic and charged side chains have been transposed. Furthermore consideration of the three-dimensional structures of FXa and FIXa (21,22) suggests that the other residues of h163-170 are equivalent in the two enzymes. In particular, residue Ser 167 is replaced in FIXa by threonine, which is able to adopt a similar rotamer and form the hydrogen bond with Asp 164 identified in this work as important to the function of h163-170. Thus, the absence of effect on prothrombinase activity observed in FXa/FIX 163-170 can be explained by the presence of a surface-exposed basic residue at position 169 or 170 (as in FXa or FXa/FIX 163-170 , respectively), the conserved hydrogen bond between residues 164 and 167, and the conserved Arg 165 residue.
The present study demonstrates that FXa positions other than the surface-exposed arginine and lysine residues of h163-170 are important for the interaction with FVa. When Val 163 or Ser 167 were changed to alanine, FXa displayed a defect in thrombin formation in the presence of small amounts of FVa ( Fig. 3 and Table 1). The effect of the mutations could occur because Val 163 or Ser 167 is directly important for FVa binding. This notion is supported by the reduced affinity of FXa V163A and FXa S167A for FVa (Table 1 and Fig. 3). Furthermore the same observation has been reported previously at these two positions in FIXa-FVIIIa interaction (12). Residues Val 163 and Ser 167 are on the same side of h163-170, making contacts with  Table 4. In panel B, data were fitted to a single ligand-binding site with a defined background K m in the presence of 10 mM Na ϩ using Equation 2 defined under "Experimental Procedures" where k cat was replaced by K m . residues inside the enzyme (22). There is a hydrogen bond between Ser 167 and Asp 164 that appears to orient the acidic side chain toward h163-170 (Fig. 7B). Val 163 makes symmetric van der Waals contacts with the aromatic ring of Tyr 225 , restricting the conformational mobility of the side chain (Fig. 7B). A molecular dynamics simulation, undertaken to explore a potential change in conformational freedom of Tyr 225 , revealed that in FXa V163A the tyrosine side chain would gain additional mobility (Fig. 7D, red circles) with a corresponding predicted shift in position (Fig. 7C) with respect to the wild-type enzyme (Fig. 7D, blue circles). In the case of FXa S167A under the same conditions, this increase in conformational freedom is not evident for Tyr 225 (Fig. 7E). However, loss of the hydrogen bond with Asp 164 and the concomitant increase in mobility of this side chain may still elicit structural consequences. For example FVa binding may cause a slight shift in the position of h163-170; loss of the Ser 167 -Asp 164 hydrogen bond would reduce the degree to which this shift is transmitted to the adjacent Val 163 residue. The observation that a saturating concentration of FVa restored or partially restored the catalytic defect of FXa derivatives toward prothrombin (Table 1 and Fig. 4), chromogenic substrate ( Table 2), and interaction with PAB, D-FFR-CK, and ATIII (Table 3) supports this notion.
The effect of these mutations is likely mediated indirectly through allosteric changes in the conformation of the S1 site of the protease. These defects were prominently observed for FXa derivatives with substitutions or mutations of residues at the amino-terminal end of h163-170 except Arg 165 . Therefore, the data presented demonstrate that FVa, through interaction with h163-170, optimizes the conformation of the S1 site for FXa interaction with its substrates in the activation complex.
Thermodynamic linkage analyses have shown an allosteric coupling between FVa and Na ϩ sites in FVa (20,23). This study shows that the alteration of the interaction of FXa V163A and FXa S167A with their substrates or inhibitors paralleled their impaired Na ϩ binding (Fig. 6). The same phenomenon was observed for FXa/ FVII 163-167 and FXa/FVII 163-170 . Although it cannot be excluded that the more extensive changes in FXa/FVII 163-167 and FXa/FVII 163-170 might have resulted in perturbations affecting the intrinsic catalytic activity of FXa, the substantial increase in K d(app) for Na ϩ is consistent with overpacking introduced by the V163L mutation. Thus, the results of the current study indicate that changing residues Val 163 and Ser 167 to Ala has a deleterious effect for Na ϩ interaction with FXa (Figs. 5 and 6 and Table 4). This reduced apparent affinity for Na ϩ of FXa V163A and FXa S167A is responsible for the defective interaction, observed at 150 mM NaCl, of the enzymes with substrates that bind into their S1 site (Tables 2 and 3 and Figs. 5 and 6). These defects can be overcome in the prothrombinase complex with an excess of FVa (Tables 2 and 3). In conclusion, our study reveals the existence of a functional link in FXa among h163-170, an  (22) oriented to highlight the position of h163-170 (yellow) and loops that coordinate the sodium ion (dark blue tubes). Other helices are pale blue, and strands are pink. The positions of the calcium ion (green sphere) and sodium ion (red sphere), inferred from crystal structures (Protein Data Bank codes 2J4I and 2BOK) are also shown along with their coordinating bonds. The S1 subsite (gray transparent surface) and the catalytic residues (orange stick), on the opposite face of the enzyme, are partially visible. Tyr 225 , Val 163 , Asp 164 , and Ser 167 are shown in black-and-red stick, and Arg 165 is shown in blue stick. In panel B, a close-up view of the FVa-binding region of wt-FXa is shown, illustrating the close contacts between Val 163 and Tyr 225 (gray sticks inside transparent van der Waals spheres). The sodium ion and S1 subsite are shown for reference. In panel C, the equivalent region in the FXa V163A mutant shows both a difference in packing and a change in the orientation of Tyr 225 . In panel D, a graphical representation of the relative mobility of the Tyr 225 side chain during the molecular dynamics simulation for wt-FXa (blue) and the V163A mutant (red) is shown. The abscissa reflects the relative dihedral angle around the C␣-C␤ bond with respect to that of the initial, energy-minimized structure of wt-FXa; the ordinate reflects the relative angle around the C␤-C␥ bond. It can be seen that the FXa V163A mutant is able to explore a much greater angular range than the wild-type protein.
In panel E, the FVa-binding region of the FXa S167A mutant highlights the loss of the hydrogen bond between Ser 167 and Asp 164 visible for wt-FXa (A) with the consequence that the side chain of the aspartic acid residue is no longer held in place. Structural figures were prepared using PyMOL (36).
FVa-binding site, Na ϩ , and S1 sites that is modulated by residues Val 163 and Ser 167 .