Thermodynamic Linkage between the S1 Site, the Na+Site, and the Ca2+ Site in the Protease Domain of Human Coagulation Factor Xa

The serine protease domain of factor Xa (FXa) contains a sodium as well as a calcium-binding site. Here, we investigated the functional significance of these two cation-binding sites and their thermodynamic links to the S1 site. Kinetic data reveal that Na+ binds to the substrate bound FXa withK d ∼39 mm in the absence and ∼9.5 mm in the presence of Ca2+. Sodium-bound FXa (sodium-Xa) has ∼18-fold increased catalytic efficiency (∼4.5-fold decrease in K m and ∼4-fold increase ink cat) in hydrolyzing S-2222 (benzoyl-Ile-Glu-Gly-Arg-p-nitroanilide), and Ca2+ further increases this k cat∼1.4-fold. Ca2+ binds to the protease domain of substrate bound FXa with K d ∼705 μm in the absence and ∼175 μm in the presence of Na+. Ca2+ binding to the protease domain of FXa (Xa-calcium) has no effect on the K m but increases thek cat ∼4-fold in hydrolyzing S-2222, and Na+ further increases this k cat∼1.4-fold. In agreement with the K m data, sodium-Xa has ∼5-fold increased affinity in its interaction withp-aminobenzamidine (S1 site probe) and ∼4-fold increased rate in binding to the two-domain tissue factor pathway inhibitor; Ca2+ (±Na+) has no effect on these interactions. Antithrombin binds to Xa-calcium with a ∼4-fold faster rate, to sodium-Xa with a ∼24-fold faster rate and to sodium-Xa-calcium with a ∼28-fold faster rate. Thus, Ca2+and Na+ together increase the catalytic efficiency of FXa ∼28-fold. Na+ enhances Ca2+ binding, and Ca2+ enhances Na+ binding. Further, Na+ enhances S1 site occupancy, and S1 site occupancy enhances Na+ binding. Therefore, Na+ site is thermodynamically linked to the S1 site as well as to the protease domain Ca2+ site, whereas Ca2+ site is only linked to the Na+ site. The significance of these findings is that during physiologic coagulation, most of the FXa formed will exist as sodium-Xa-calcium, which has maximum biologic activity.

1 For comparison, the factor X amino acid numbering system is used. The numbers with a prefix c (e.g. c195) in parentheses refer to the chymotrypsin equivalents for the protease domain of FXa. Where insertions occur, the chymotrypsin numbering is followed by a capital letter, such as A. 2 The abbreviations used are: FXa, factor Xa; PABA, p-aminobenzamidine; FVa, factor Va; TFPI, tissue factor pathway inhibitor; EGF, epidermal growth factor; Gla, ␥-carboxyglutamic acid; IIa, thrombin; AT, antithrombin; S-2222, benzoyl-Ile-Glu-Gly-Arg-p-nitroanilide; Des-44-X or Xa, Gla domainless FX or FXa; Ch ϩ , choline; PEG, polyethylene glycol. therefore cannot participate in binding to Na ϩ (13)(14)(15). IIa performs two opposing functions in hemostasis. The procoagulant function involves conversion of fibrinogen to fibrin as well as activations of factors V and VIII (4). The anticoagulant function involves binding of IIa to thrombomodulin, which enables IIa to activate protein C to activated protein C. Activated protein C, the function of which is also regulated by Na ϩ (13, 16 -19), in turn, inactivates factors Va and VIIIa (4). Data exist indicating that the Na ϩ -free IIa preferentially performs the anticoagulant function, whereas the Na ϩ -bound IIa preferentially carries out the coagulant function (20). Thus, association or dissociation of Na ϩ creates a molecular switch in the functions of IIa.
In addition to IIa, factors VIIa, IXa, and Xa are each thought to possess a Na ϩ -binding site in their protease domains (13,(21)(22)(23). However, only the site in FXa is crystallographically defined (14). Importantly, in contrast to IIa, the Na ϩ site in FXa involves four carbonyl O atoms from two different loops (14). In this report, we have extensively characterized the function of Na ϩ -binding site in FXa and its relation to the protease domain Ca 2ϩ -binding site as well as to the S1 site. The present data indicate that choline chloride, which in recent studies (13,23) has been used as an inert ion to maintain constant ionic strength, is a potent inhibitor of FXa catalytic activity. We further establish that the monovalent Na ϩ site is thermodynamically linked to the protease domain divalent Ca 2ϩ site as well as to the S1 site. Thus, sodium enhances substrate binding, and substrate enhances sodium binding. Further, Ca 2ϩ binding to the protease domain increases the affinity for Na ϩ , and Na ϩ binding increases the affinity of Ca 2ϩ at this site in FXa. This results in maximal FXa enzymatic activity under physiologic conditions.

EXPERIMENTAL PROCEDURES
Proteins and Reagents-Benzoyl-Ile-Glu-Gly-Arg-p-nitroanilide (S-2222) was purchased from Helena Laboratories. Polyethylene glycol (PEG) 8000, p-aminobenzamidine (PABA), and N ␣ -p-tosyl-L-Phe chloromethyl ketone-treated bovine trypsin were obtained from Sigma. Lovenox, a low molecular weight heparin, was obtained from Rhone-Poulenc Rorer Pharmaceuticals. Human factor X and prothrombin were purified as described earlier (24). Two-domain TFPI (residues 1-161) was expressed and purified as described (25). It showed a single band (ϳ30,000 Da) on nonreduced and reduced SDS-PAGE (26). Des-44-X was prepared and purified by QAE-Sephadex A-50 column chromatography as described by Morita and Jackson (27). FXa and Des-44-Xa were prepared using insolublized Russell's viper venom as outlined previously (24,27). FXa was also purchased from Enzyme Research Laboratories, Inc. Human AT was obtained from Hematologic Technologies, Inc. The purity of both FXa and Des-44-Xa protein was examined by SDS gel electrophoresis (26) and was Ͼ98% as depicted earlier (8). The purity of AT was assured by the supplier (SDS gels were provided). Each protein sample was freed of Na ϩ by dilution or dialysis and/or by a desalting column exactly as described by Wells and Di Cera (28). The final concentration of Na ϩ after these steps was Ͻ1 mM as measured by a conductivity meter as outlined by Wells and Di Cera (28). The proteins were frozen at Ϫ80°C in 10 -15-l aliquots, thawed, and used immediately. This freezing and thawing step did not result in a measurable loss of activity.
Amino Acid Sequence Analysis-Automated Edman degradation of each protein component was performed using an Applied Biosystems model 477A gas phase sequencer. Approximately 0.5 nmol of protein was loaded on the filter cartridge. The proteins from SDS gels were transferred to polyvinylidene difluoride membranes as described by Rosenberg (29). The NH 2 -terminal sequence analysis of Des-44-X revealed two sequences of approximately equimolar amounts, one corresponding to the heavy chain (Ser-Val-Ala-Gln-Ala) and the other corresponding to the modified light chain (Lys-Asp-Gly-Asp-Gln). This indicates that Des-44-X does not contain the Gla domain.
SDS Gel Electrophoresis-SDS gel electrophoresis was performed using the Laemmli buffer system (26). The acrylamide concentration was 15%, and the gels were stained with Coomassie Brilliant Blue dye. All proteins used in the present study were ϳ98% pure.
Measurement of S-2222 Amidolytic Activity of FXa Proteins-The concentration of FXa and Des-44-Xa used was between 1 and 5 nM each. The S-2222 concentration ranged from 20 M to 3 mM. The buffer used was 50 mM Tris, pH 7.4, containing 0.1% PEG and various salt combinations. The p-nitroaniline release was measured continuously (⌬A 405 nm /min) for up to 30 min using a Beckman DU65 spectrophotometer equipped with a Soft-Pac kinetics module. An extinction coefficient of 9.9 mM Ϫ1 cm Ϫ1 at 405 nM was used in calculating the amount of p-nitroaniline released (30). All reactions were performed in triplicate. The K m and k cat values were obtained using the Enzyme Kinetics program from Erithacus Software.
PABA Binding-Binding of PABA was measured by increase in its intrinsic fluorescence upon binding to the active site of FXa using an SLM 8000C fluorescence spectrophotometer. The concentration of FXa used was 1.7 M, and the excitation wavelength was 336 nm (31). A titration of the protein solution (300 l) was performed by adding small increments (1-2 l) of 1 or 2 mM stock solution of PABA, and the resulting fluorescence at 376 nm was measured at each point after the attainment of equilibrium conditions (usually 1 min). Excitation bandwidth was set at 4 nm, and emission bandwidth was set at 2 nm. Because binding of PABA to FXa is expected to be relatively weak (32), high concentrations of PABA are needed to study this interaction. This leads to a significant absorption of the exciting light, and the data therefore require corrections for the inner filter effect (32,33). First, the nonlinear dependence of fluorescence on PABA concentration in the buffer (without protein) was evaluated according to Equation 1 (32).
where P is the concentration of PABA, m and i are, respectively, the slope and intercept of the expected linear dependence on PABA concentration in buffer, ⑀ is the extinction coefficient of PABA at 336 nm excitation wavelength (experimentally determined to be 1426 M Ϫ1 cm Ϫ1 ), and L is effective path length for the fluorescence cell. The value of L determined using Equation 1 was 0.4 Ϯ 0.2. The corrected fluorescence signal (F corr ) was then obtained using Equation 2 (32).
Binding data were then fitted to a single ligand binding site with a defined background using the following equation.
where K d app is the apparent dissociation constant, y is the inner filter corrected fluorescence at a given PABA concentration depicted by x, b is the corrected background fluorescence in the absence of protein, and c is the corrected fluorescence at saturating concentrations of PABA.   where Xa⅐TFPI is the steady state complex, which then isomerizes and results in the formation of Xa⅐TFPI*, k 1 and k 2 are the forward and reverse rates for the initial binding, and k 3 and k 4 are the forward and reverse rates for the isomerization step. To calculate the binding parameters, first the initial (v in ) and steady state (v st ) velocities were calculated using Equation 5, as described by Morrison and Walsh (34), where [P] represents the concentration of p-NA formed at time t, v in and v st are respectively the rates of substrate hydrolysis before and after the steady state is achieved, and k obs is the rate of conversion of v in to v st . For these reactions, TFPI concentrations were in large excess of FXa. Thus, the free TFPI concentration did not change significantly during the course of the reaction. Further, the rates in the control experiments performed in the absence of TFPI did not change over the course of measurement. Thus, substrate depletion did not contribute to v st . The values of k obs obtained above were fitted to Equation 3 to resolve the individual parameters. In this case in Equation 3, y represents k obs at a given concentration of two-domain TFPI depicted by x, whereas b and c, respectively, represent k 4 and (k 3 ϩ k 4 ), and K d app represents The value of K i (k 2 /k 1 ) was obtained from K i app using the following equation.
where [S] is the S-2222 concentration. The K m values obtained under different conditions (listed in Table I) were used to obtain K i . Binding of AT to FXa-Serpin AT binds FXa irreversibly and should displace PABA bound at the active site. Thus, the AT binding to FXa was monitored by measuring the decrease in fluorescence of FXa⅐PABA solution. The wavelengths for excitation (bandwidth, 4 nm) and emission (bandwidth, 2 nm) were set at 336 and 376 nm, respectively. For each salt condition, FXa at 1.7 M was mixed with PABA at the respective K d concentration (Table II)  , and 500 mM Na ϩ /0 mM Rb ϩ (crosses). K m app and V max values were calculated using the enzyme kinetics program from Erithacus Software; for clarity, higher concentrations of substrate used are not shown in the graph. The concentration of FXa was 2.5 nM. The value of K m app (M) at 0 mM Na ϩ was 514 Ϯ 21, at 15 mM Na ϩ it was 509 Ϯ 18, at 30 mM Na ϩ it was 454 Ϯ 41, at 50 mM Na ϩ it was 381 Ϯ 36, at 80 mM Na ϩ it was 306 Ϯ 32, at 100 mM Na ϩ it was 280 Ϯ 28, at 150 mM Na ϩ it was 276 Ϯ 29, at 200 mM Na ϩ it was 240 Ϯ 31, at 300 mM Na ϩ it was 146 Ϯ 24, at 400 mM Na ϩ it was 108 Ϯ 21, and at 500 mM Na ϩ it was 114 Ϯ 18 mM. For these experiments, two stock buffers were made. One buffer was 50 mM Trizma base, 0.1% PEG 8000, 1 mM EDTA (acid form), and 500 mM RbCl adjusted to pH 7.4 with HCl. In the second buffer 500 mM RbCl was replaced by 500 mM NaCl, and the pH was again adjusted to 7.4 with HCl. These two buffers were mixed in appropriate proportions to yield the desired concentration of Na ϩ and Rb ϩ . Thus, the monovalent ion concentration was always 500 mM, and the buffer containing 500 mM Rb ϩ had no added source of Na ϩ . B, V max as a function of Na ϩ . Data were fitted to Equation 3. Here, y represents maximum velocity at saturating concentrations of Na ϩ , b is the offset (maximum velocity at zero concentration of Na ϩ ), x is the concentration of Na ϩ , and K d app is the K d of Na ϩ for Xa(S) in the absence of Ca 2ϩ . The values of b and y (converted to k cat ) are given in Table I. saturating) low molecular weight heparin. AT was then added, and the decrease in the inner filter corrected fluorescence was determined at 1-5-s intervals for up to 30 min. The data were fitted to the equation given below.
where F t and F 0 are the inner filter corrected fluorescence values at time t and 0 seconds, respectively, and k obs is the observed pseudo first order rate constant. Under the conditions of experiment, the value of k obs is related to the kЈ obs as follows, where kЈ obs is pseudo first order rate constant extrapolated to 0 concentration of PABA. As the PABA concentration was fixed at the K d PABA in each case, the values of k obs were multiplied by a factor of 2 to obtain the values of kЈ obs . The values of kЈ obs were then plotted against the AT concentration to obtain k on values. Molecular Modeling-X-ray structures of FXa (7), IIa (14,15), and trypsin (36) were analyzed using the program O (37). Residue Tyr-c225 in FXa was mutated to Pro and manually adjusted based upon the coordinates of trypsin (36) and IIa Yc225P mutant (15). Local region was energy minimized using the program O (37).

Effects of Monovalent Cations on the Amidolytic Activity of
FXa-Hydrolysis of S-2222 by FXa was measured at various concentrations of Li ϩ , Na ϩ , K ϩ , Rb ϩ , and choline (Ch ϩ ). These data are presented in Fig. 1. In one set of experiments, the activity of FXa was measured at increasing cation concentrations from 0 to 0.2 M (Fig. 1A). FXa hydrolyzed S-2222 (100 M) in the absence of any ion (in 50 mM Tris, pH 7.4, 0.1% PEG 8000) at ϳ0.1 M/min/nM FXa. K ϩ increased this rate ϳ1.7fold, whereas Rb ϩ and Li ϩ had no significant effect. On the other hand, Ch ϩ decreased this rate ϳ3.5-fold. Further, Na ϩ was found to be the most potent ion, which potentiated FXa activity more than 10-fold (Fig. 1B). For these experiments chloride salt of each ion was used; therefore, these variations in rates are not due to changes in the anion concentration. Moreover, as the effects are specific for each ion, they cannot be due to changes in the ionic strength. Therefore, our data indicate that Na ϩ potentiates the activity of FXa and Ch ϩ is inhibitory toward this enzyme. Based upon this observation and available data in the literature (13,14), we conclude that the protease domain Na ϩ -binding site mediates these effects.
Although changes in ionic strength in the range of 0 -0.2 M Na ϩ did not affect the amidolytic activity of FXa (Fig. 1A), yet its effects on macromolecular recognition cannot be predicted. Therefore, we standardized a system to study the effects of Na ϩ at a constant monovalent ion concentration. In several previous studies with coagulation enzymes, including FXa, Ch ϩ has been employed as an inert ion (13,19,23). Based upon these reports and the data presented in Fig. 1A, we wished to further examine the effects of Ch ϩ and Rb ϩ on FXa activity. FXa activity was measured in reaction mixtures containing various ). However, use of Ch ϩ to compensate for ionic strength skewed the curve form a regular hyperbola toward being sigmoidal in nature. This can be explained by a combination of the potentiating effect of Na ϩ and the inhibitory effect of Ch ϩ operating simultaneously. On the other hand, the response curve obtained using Rb ϩ as an inert ion resembled the curve obtained in the absence of any compensating ion. These data establish that Ch ϩ is not an inert ion but inhibits FXa activity, whereas Rb ϩ is the ion of choice to compensate for ionic strength in studying FXa function. Data showing that Ch ϩ inhibits FXa activity are consistent with previous observations (38,39).
Because 200 mM Na ϩ does not appear to saturate the Na ϩ site in FXa, we opt to test whether or not we could use 500 mM NaCl as the highest salt concentration without affecting the catalytic activity of FXa. Thus when we performed experiments with buffers containing 500 mM RbCl, the rate of S-2222 hydrolysis (at 100 M or 200 M) was the same as with 200 mM RbCl. Similarly, when reaction buffer contained 200 mM Na ϩ and 300 mM Rb ϩ , the rate of hydrolysis was identical to that obtained with 200 mM Na ϩ without Rb ϩ . Thus, it appears that we could use up to 500 mM NaCl in our system to study the role of Na ϩ in FXa function.
Na ϩ Potentiation of S-2222 Hydrolysis by FXa in the Absence of Ca 2ϩ -To investigate the mechanism of Na ϩ -mediated potentiation of S-2222 hydrolysis, we determined K m app and V max at several concentrations of NaCl. As stated above, we used 500 mM NaCl as the highest salt concentration in these experiments. 1 mM EDTA was present in each buffer to eliminate the effect of divalent cations. These data are presented in Fig. 2A.
The values of K m app and V max were calculated for each salt concentration using the enzyme kinetics program from Erithacus Software. The results indicate that Na ϩ affects both the K m app and V max of this reaction. The K m app values obtained at each Na ϩ concentration are given in the legend to Fig. 2. When V max are plotted as a function of Na ϩ concentration (Fig. 2B), the midpoint of the curve should yield K d of interaction of Na ϩ with Xa(S) (Xa saturated with S-2222) in the absence of Ca 2ϩ ; this value was calculated to be 39 Ϯ 4 mM. Ca 2ϩ Potentiation of S-2222 Hydrolysis by FXa in the Absence of Na ϩ -These data are presented in Fig. 3. The kinetic data indicate that Ca 2ϩ does not change the K m app but increases the k cat ϳ4-fold. Importantly, the K d of Ca 2ϩ in its interactions with Xa(S) is ϳ700 M (Fig. 3B).
Na ϩ Potentiation of S-2222 Hydrolysis by FXa in the Presence of Ca 2ϩ -These data are presented in Fig. 4. The kinetic data indicate that Na ϩ decreases the K m app ϳ4.5-fold and increases the k cat ϳ1.4-fold. Importantly, Ca 2ϩ decreases the K d of Na ϩ interaction with Xa(S) to 9.5 mM as compared with 39 mM (Fig. 2) when Ca 2ϩ is absent. Thus Ca 2ϩ site is thermodynamically linked to the Na ϩ site.
Ca 2ϩ Potentiation of S-2222 Hydrolysis by FXa in the Presence of Na ϩ -These data are presented in Fig. 5. As noted previously (Fig. 3), Ca 2ϩ does not change the K m app but increases the k cat ϳ1.4-fold in the presence of Na ϩ as well. Further, Na ϩ decreases the K d of Ca 2ϩ interaction with Xa(S) to ϳ170 M as compared with ϳ700 M (Fig. 3) when Na ϩ is absent. Thus, as noted above, Na ϩ site is linked to the Ca 2ϩ site. The above experiments were repeated using Des-44-Xa, which has only one high affinity Ca 2ϩ site in the presence of Na ϩ . This Ca 2ϩ site is localized to the protease domain of FXa (8). The K d Na values for Xa(S) in the absence and presence of Ca 2ϩ were ϳ43 and ϳ11 mM, respectively. Similarly, the K d Ca values for Xa(S) in the absence and presence of Na ϩ were ϳ700 and ϳ 170 M, respectively. These data demonstrate that it is the protease domain Ca 2ϩ site that influences the K d for Na ϩ binding and vice versa. Moreover, when trypsin (400 pM) that lacks Na ϩ site was used to hydrolyze S-2222 (100 M), the rates in the presence and absence of Na ϩ were the same (1.1 M/ min), and Ca 2ϩ only minimally affected this rate (1.8 M/min).
Linkage between the Substrate Binding Site, the Na ϩ Site, and the Protease Domain Ca 2ϩ Site of FXa-The above data clearly establish the interdependence of the binding of Na ϩ and Ca 2ϩ in the protease domain of FXa. Fig. 6 depicts this as a thermodynamic cycle. Xa(S) can be converted to the state with both ions bound (sodium-Xa(S)-calcium) by acquiring either Na ϩ (via Xa(S)-calcium) or Ca 2ϩ (via sodium-Xa(S)). The ratio of the K d Na in the presence and absence of Ca 2ϩ is 4.1, and that of K d Ca in the presence and absence of Na ϩ is 4.2. Therefore, the net sum of binding energy over the cycle appears to be 0. This establishes the thermodynamic linkage between the Na ϩ site and the protease domain Ca 2ϩ -binding site of FXa.
Because Na ϩ affects the K m of S-2222 hydrolysis, there appears to be a link between the substrate binding and the Na ϩ -binding site. Both in the absence and presence of Ca 2ϩ , Na ϩ decreases the K m of S-2222 by ϳ4.5-fold. Using K m as an approximation of substrate affinity, one can complete the linkage using the thermodynamic principles (40). This is also depicted in Fig. 6. Therefore, the K d Na and K d Ca depicted on the

TABLE III Effects of calcium and sodium on the kinetic constants for binding of two-domain TFPI to FXa
The values are calculated using the data of Fig. 8 and Table I. The kinetic constants are defined under "Experimental Procedures." The concentration of Na ϩ in the absence of Ca 2ϩ was 500 mM, and in the presence of 5 mM Ca 2ϩ , it was 485 mM. Conditions a One should note that the equations for slow tight binding inhibition for multistep pathways are notorious for yielding poorly determined or ill conditioned parameters. Therefore, our rate constants must be interpreted with caution. Further, the possibility that Xa ⅐ TFPI complex may not be in a completely inhibited form cannot be ruled out. Nonetheless, our data indicate that Na ϩ affects the interaction of Xa ⅐ TFPI, whereas Ca 2ϩ does not. left side of the figure are not the experimentally obtained values but rather the calculated values.
Effects of Na ϩ and Ca 2ϩ on the Interaction of PABA with FXa-We next investigated whether Na ϩ and/or Ca 2ϩ affect the S1 part of the active site in FXa. We used the S1 site probe PABA for these studies. The interaction of PABA with FXa was determined under each of the four salt conditions. These data are presented in Fig. 7 and summarized in Table II. Under saturating concentration of Na ϩ , the affinity of PABA for FXa was increased ϳ5-fold as compared with that in its absence. Further, Ca 2ϩ had no effect on the K d PABA in the absence or presence of Na ϩ . These K d PABA data agree well with the K m data presented in the previous section where Na ϩ decreased the K m ϳ4-fold and Ca 2ϩ had no effect.
Effects of Na ϩ and Ca 2ϩ on the Binding of Two-domain TFPI and AT with FXa-In this section we investigated the role of Na ϩ and Ca 2ϩ in binding of FXa to its two physiologic inhibitors, namely, TFPI and AT. TFPI consists of three tandem Kunitz-type domains, and the second domain is responsible for binding to FXa with an overall K d in the nanomolar range (41). In this study we used two-domain TFPI, which contains the second FXa-binding domain. The data obtained were treated according to the tight binding inhibitory mechanism (34,35) and analyzed using Equations 5, 3, and 6, respectively. The kinetic parameters were obtained for all four forms of FXa depicted on the left side in Fig. 6. Representative data are presented in Fig. 8, and the kinetic parameters for all four conditions used are listed in Table III. The affinity for the first step of FXa⅐TFPI complex formation (Equation 4) was increased ϳ4-fold in the presence of saturating concentration of Na ϩ ; however, the slow isomerization step leading to the formation of FXa⅐TFPI* was not affected. Interestingly, Ca 2ϩ in the presence or absence of Na ϩ had no effect on any of these kinetic parameters.
The binding data for AT, which belongs to the serpin family of inhibitors are presented in Fig. 9 and the second order rate constants for all four forms of FXa (left side of Fig. 6) are listed in Table IV. Na ϩ increased the k on by ϳ24-fold in the absence of Ca 2ϩ , and addition of Ca 2ϩ had minimal further effect. Ca 2ϩ increased the k on by ϳ4-fold in the absence of Na ϩ , and addition of Na ϩ further increased this rate constant by ϳ7-fold. These AT binding data are consistent with the specificity constant data presented in Table I. DISCUSSION Recently, Dang and Di Cera (13) have reported that several serine proteases including those in coagulation possess a functional Na ϩ site. X-ray crystal structures of IIa (14,15) and FXa (7,14) are reported where the Na ϩ site in these molecules in defined. The Na ϩ site in IIa uses a single loop involving the carbonyl O atoms of residues c221A and c224 as well as four water molecules, whereas the Na ϩ site in FXa uses two loops involving the carbonyl O atoms of residues c185, c185A, c222, and c224 as well as two water molecules. The nature of the residue c225 plays an important role in orienting the carbonyl O atom of c224 toward the Na ϩ coordination shell (7,14,15). Further, FXa also possesses a Ca 2ϩ site in the protease domain involving the carboxyl groups of Asp-c70 and Glu-c80 as well as carbonyl O atoms of residues c72 and c75. The occupancy of Ca 2ϩ at this site prevents proteolysis in the autolysis loop, the integrity of which is essential for FXa biologic activity (8,42). The protease domain of FXa also contains a FVa-binding site that involves the c162-c169 helix (12), and residues Asp-c164 and Arg-c165 are important for this interaction (12,43). 3 Fig.  10A presents a schematic view of the spatial relationships of the Ca 2ϩ site, autolysis loop, Asp-c189 S1 site, Na ϩ site, FVabinding helix, and the catalytic triad. The objective of the present study was to examine the role of binding of Ca 2ϩ and Na ϩ in regulating the FXa catalytic efficiency.
In the absence Na ϩ , Ca 2ϩ increased the k cat of S-2222 hydrolysis ϳ4-fold (Table I). Similarly, it increased the AT binding rate ϳ4-fold (Table IV), which in all probability represents formation of the acyl intermediate. In the presence of Na ϩ , however, Ca 2ϩ had minimal effect (ϳ1.4-fold) on the hydrolysis of S-2222 or AT binding. Of interest, Ca 2ϩ in the presence or absence of Na ϩ had no effect on the K m of S-2222 hydrolysis as well as on the PABA binding and a tight binding reversible inhibitor, TFPI. Thus, Ca 2ϩ does not affect the ground state binding of substrates/inhibitors to FXa. The ϳ4-fold increase in k cat for S-2222 hydrolysis and in the formation of acyl covalent FXa⅐AT complex may be due to lowering of the transition state energy. In FXa, Ca 2ϩ loop is linked to the carbonyl O atoms of Gly-c193 and Asp-c194 through four water molecules (7). These two residues play an important role in positioning the Ser-c195 active site side chain for attacking the peptide bond in substrates and serpins (44). Further, the amide N atoms of c193 and c195 make H bonds with the oxyanion, which develops during the formation of the transition state tetrahedral intermediate (44,45). The Ca 2ϩ binding may stabilize the proper orientation of these two amide N atoms. In trypsin, the interaction of Ca 2ϩ loop with the carbonyl O atoms of c193 and c194 3 D. Zhong, K. Padmanabhan, and S. P. Bajaj, unpublished data. .  (Table  II). Experiments were initiated by the addition of AT to an equilibrated mixture of FXa⅐PABA, and the decrease in fluorescence was recorded up to 30 min. Data were fitted to Equation 7 to obtain k obs . Corrected k obs (kЈ obs ) were obtained as per Equation 8 and were then used to obtain the second order rate constant, k on . is absent (36), which may explain the lack of significant effect of Ca 2ϩ on its activity.
In the presence or absence of Ca 2ϩ , Na ϩ increased the affinity of S-2222 (Table I), PABA (Table II), or TFPI (Table III) binding to FXa ϳ4.5-fold. Because Ca 2ϩ does not influence the K m value of S-2222 and has no effect on the PABA or TFPI binding, it would appear that these effects are primarily mediated through the S1 site in FXa. These results can be readily rationalized, because the Na ϩ site in FXa is directly linked to the S1 specificity pocket residue Asp-c189 (Fig. 10B). Occupancy of the Na ϩ site could rigidify the c189 side chain for optimal interaction with the P1 residue (Arg) of substrates and inhibitors. Furthermore, the FVa binding c162-c169 helix in FXa is connected to the Tyr-c225 of the Na ϩ -binding loop via H bonds and van der Waals' contacts (Fig. 10B). These interactions may or may not be interrupted in the absence of Na ϩ . Coordinates (PDB code 1HCG) used are those of Padmanabhan et al. (7). ␤-Sheets are in red, ␣-helices are in blue, Ca 2ϩ -binding loop is in white, the two Na ϩ -binding loops are in yellow and white, respectively, and the autolysis loop is in magenta. The active site residues (Asp-c102, His-c57, and Ser-c195) are labeled D, H, and S, respectively. The amino-and carboxyl termini are marked N and C, respectively. Position of the FVa binding helix is indicated, and the Arg-c165 side chain, which is important for FXa-FVa interaction, is shown. Tyr-c225 side chain thought to be an important determinant for Na ϩ binding is also depicted. Sodium is shown as a white sphere, and the positions of four residues, namely, c185, c185A, c222b, and c224, which provide the carbonyl O atoms for coordinating to the Na ϩ , are indicated (14). Ca 2ϩ -binding site involves carbonyl groups of c72 and c75, and the carboxyl groups of c70 and c80. Binding of Ca 2ϩ prevents the cleavage of Arg-c150/Gln-c151 in the autolysis loop (8). Residue numbers are based upon chymotrypsin numbering. Ca, calcium; Na, sodium. B, relationships of the Na ϩ -binding site to the S1 site and FVa binding helix in FXa. The two Na ϩ -binding loops, c183-c189, colored red, and c221-c225, colored magenta, and the FVa binding helix colored by atom type are shown. Na ϩ is shown as a blue sphere, and water molecules are shown as red spheres. Dotted lines are H bonds, and dashed lines are van der Waals' interactions. The Na ϩ is coordinated by the carbonyl O atoms of c185, c185A, c222, and c224 and two water molecules. The benzene ring of Tyr-c225 makes van der Waals' contacts with Val-c163, and the OH group of Tyr-c225 is connected via two water molecules to Ser-c167 of the FVa binding helix. Tyr-c225 was mutated to Pro-c225 using the structure of trypsin (36) and thrombin mutant Yc225P (15). When residue c225 is proline, its carbonyl O atom points away from the Na ϩ site, and its side chain is unable to make contacts with the FVa binding helix. Residue numbers are based upon chymotrypsin numbering.
While this manuscript was in review, a study suggesting that FVa binding may not be affected by the Na ϩ site appeared in press (46). However, further critical data are needed to fully support this conclusion.
Of significance is the observation that Na ϩ site and the protease domain Ca 2ϩ site are thermodynamically linked. Thus Ca 2ϩ increases the affinity of Na ϩ by ϳ4.1-fold, and Na ϩ increases the affinity of protease domain Ca 2ϩ site by a similar fold (Fig. 6). The thermodynamic linkage between the Na ϩ site and the protease domain Ca 2ϩ site of FXa is schematically depicted in Fig. 6. This linkage is also supported by our extensive kinetic data. Thus, although the effects of Na ϩ and Ca 2ϩ at individual steps of the thermodynamic cycle depicted in Fig. 6 are different, the overall total change in K m and k cat is the same regardless of the pathway between FXa and sodium-FXa-calcium state. This further validates the thermodynamic linkage between the Na ϩ and Ca 2ϩ binding sites illustrated in Fig. 6. Furthermore, our data also indicate that the S1 site is linked to the Na ϩ site. Thus Na ϩ enhances substrate binding and substrate enhances Na ϩ binding. Thus during physiological clotting, FXa formed will mostly exist in sodium-Xa-calcium form that has maximum biological activity.
In an earlier study (23), Tyr-c225 was mutated to Pro to eliminate the Na ϩ -binding site in FXa. We also constructed and expressed this mutant. This mutant-Xa(S) bound Na ϩ in the presence of Ca 2ϩ with a dissociation constant of ϳ45 mM 4 ; the affinity of Na ϩ in the absence of Ca 2ϩ for this mutant has not yet been determined in our laboratory. Thus, this mutant is not totally impaired in binding to Na ϩ but has ϳ5-fold reduced affinity. This is in contrast to the results reported for Yc225P mutant of IIa, which does not bind Na ϩ (13,15). These observations could be explained on the basis of difference in the nature of Na ϩ site in IIa (15) and FXa (14). As stated earlier, the Na ϩ site in IIa involves only two carbonyl O atoms from the protein (14,15), whereas in FXa it involves four carbonyl O atoms (14). This is shown in Fig. 10B. A change of Tyr-c225 to Pro will result in the loss of one ligand out of two in IIa and one out of four in FXa (Fig. 10B). Thus, three ligands from the protein in Yc225P FXa may still be available to bind Na ϩ , and the results obtained with this mutant should be interpreted with caution. A Ec80K mutant of FXa, which is thought to mimic the Ca 2ϩ -binding site has also been used to study the role of Ca 2ϩ on Na ϩ binding (23). This mutant-Xa(S) in our hands binds Na ϩ with a dissociation constant of ϳ25 mM 4 in the presence or absence of Ca 2ϩ . This dissociation constant of ϳ25 mM is very close to the 40 mM K d Na obtained in the absence of Ca 2ϩ for wild-type FXa. Thus, Ec80K FXa mutant may not completely mimic the protease domain Ca 2ϩ occupied wild-type Xa(S). Our data with the Ec80K FXa mutant are consistent with our previous studies on Ec70K factor IX mutant, which did not completely mimic the properties of Ca 2ϩoccupied form of factor IXa (47). We are currently in the process of expressing several mutants of FX to further understand the nature of the Na ϩ site as well as the protease domain Ca 2ϩ site in this important coagulation protease.