INTRODUCTION

The serine protease thrombin plays a central role in the blood coagulation cascade (1, 2). Thrombin catalyzes the conversion of fibrinogen to fibrin monomer and activates factor XIII to factor XIIIa, which in turn cross-links fibrin, thereby mechanically and chemically stabilizing the fibrin blood clot (3). Thrombin also activates factors XI, V, and VIII, important components of the thrombin-generating blood coagulation cascade. In addition, thrombin activates platelets (4). Activated platelets aggregate to form a platelet plug that, together with the fibrin clot, provides hemostasis. Activated platelets also provide the surface for assembly of components of the blood coagulation cascade (5). The ability of thrombin to catalyze the formation of several components involved in its generation is responsible for the explosive formation of thrombin when the blood coagulation cascade is triggered by vascular injury-induced exposure of tissue factor. Thrombin also down-regulates its own formation. Upon binding to endothelial cell membrane-associated thrombomodulin, the specificity of thrombin is altered. Essentially devoid of activity toward fibrinogen, thrombomodulin-bound thrombin efficiently catalyzes the activation of protein C, which in conjunction with activated protein S inactivates factors Va and VIIIa, thereby shutting down the thrombin-generating blood coagulation cascade (6, 7).

Early studies have shown that Na⁺ and certain other monovalent ions induce a conformational change in thrombin and alter the substrate specificity of the enzyme (8, 9). Wells and Di Cera (10) were the first to show that thrombin is an allosteric enzyme existing in two forms, designated as slow and fast form thrombin, and that these forms interconvert upon binding of Na⁺ to a specific site of the enzyme. The slow  fast transition is also accompanied by a significant increase in the intrinsic fluorescence of the enzyme (10-11). X-ray diffraction analysis of thrombin indicates that Na⁺ is coordinated octahedrally by 4 molecules of water and the amide oxygen atoms of Arg-221a and Lys-224 (12-14). Mutation of Arg-221a and Lys-224 to alanine diminishes the ability of thrombin to bind Na⁺ (15). The importance of the allosteric regulation mediated by Na⁺ is that the Na⁺-bound fast form cleaves fibrinogen with higher specificity, whereas the slow form shows a higher catalytic specificity in the activation of protein C both in the presence and absence of thrombomodulin (16). The existence of two interconvertible conformers provides a means for expanding the substrate specificity of thrombin and regulating the enzyme via interactions with allosteric effectors.

Although Na⁺-free thrombin has been designated as the slow form and Na⁺-bound thrombin has been designated as the fast form, the equilibrium distribution of the two interconvertible thrombin conformers in the Na⁺-free and Na⁺-bound states has not been evaluated. Additionally, the reaction pathway for the slow  fast transition has not been established. Elucidation of this pathway is crucial for understanding the mechanism of allosteric regulation of thrombin as a paradigm for serine protease specificity in general. The present study of the interaction of thrombin with Na⁺, the small active site-directed ligand L-371,912 (17), and the exosite I binding ligand hirugen (18) was initiated to address these issues and further characterize the effects of Na⁺ on the environment at the active site and exosite I.

EXPERIMENTAL PROCEDURES

Unless otherwise specified, all chemicals were purchased from Sigma. Z-Gly-Pro-Arg-7-amino-4-trifluoromethylcoumarin (Z-GPR-AFC)¹ was obtained from Enzyme Systems Products. L-371,912 (21) and hirugen (Ac-Asn-Gly-Asp-Phe-Glu-Glu-Ile-Pro-Glu-Glu-Tyr(SO₃)-Leu) were prepared by Dr. Terry A. Lyle and Dr. Victor M. Garsky, respectively, of the Medicinal Chemistry Department of Merck Research Laboratories. Human -thrombin (in 200 mM NaCl and 50 mM citrate buffer, pH 6.5) was obtained from Enzyme Research Laboratories, Inc. and used without further purification. R221aA and K224A thrombin variants were prepared as described previously (15).

Thrombin concentrations were determined from absorbance measurements (E₂₈₀ = 1.83 ml/(mg-cm) and a molecular weight of 36,500 (23). Active site titration indicated that the thrombin was >95% active (19). Solutions of the slow form were prepared by replacing sodium citrate buffer with Tris or Hepes buffer (50 mM, pH 7.4 at 5 °C) containing 300 mM choline chloride (ChCl) by centrifugation using a Centricon 10 (Amicon Inc.). Thrombin solutions 50-70 µM were stored in 20-µl aliquots at 70 °C. Unless otherwise specified, studies were performed at 5 °C, pH 7.4, in 50 mM Tris (or 50 mM Hepes) buffer containing 0.1% poly(ethylene glycol) 8000 and the indicated concentrations of NaCl and ChCl. Ionic strength was maintained by addition of ChCl. Several rate constants were determined in both Tris and Hepes buffers. Rate constants determined in the Tris buffer were indistinguishable from those determined in Hepes buffer. The kinetics for the binding of Na⁺ to thrombin were monitored using an Applied Photophysics SX17 stopped-flow fluorimeter (excitation 285 nm, emission 299 nm). Rate constants for the approach of the fluorescence to its final value or to a final linear decay due to photobleaching were evaluated by fitting the data to Equation 1 using a Kaleidagraph 3.0 least squares fitting routine.

(Eq. 1)

In Equation 1, i, f, and t denote values at the end of fast phase, at the end of slow phase, and at time t, respectively, and k is the apparent rate constant for photobleaching. Similar methods were used to study the kinetics of association of Rb⁺ and K⁺ with thrombin and the association of L-371,912 and hirugen with the slow form of wild-type and mutant thrombins.

The kinetics of dissociation of Na⁺ from the thrombin-Na⁺ complex were monitored by an Applied Photophysics SX17 stopped-flow fluorimeter (excitation 285 nm, emission 299 nm) using an asymmetric mixing method. Dissociation of Na⁺ was initiated by a 1:6 dilution of the solution containing thrombin (3 µM) and Na⁺ (100 mM or 30 mM) with 300 mM ChCl solution while maintaining the ionic strength at 0.3. Dissociation of Na⁺ (5 mM) from the thrombin-Na⁺ complex in the presence of the crown ether, 1,4,7,10,13-pentaoxacyclopentadecanether, 15-crown-5 (25 mM), in ChCl was conducted by the dilution procedure described above. Rate constants for the approach of the fluorescence to its final value or to a final linear decay due to photobleaching were evaluated from fits of the data to Equation 1, using a Kaleidagraph 3.0 least squares fitting routine.

The full time course of Na⁺ or L-371,912 binding to thrombin was simulated using Runge-Kutta digital integration methods, and rate constants for the approach of the fluorescence to its final value were evaluated from fits of the data to Equation 2 using a Kaleidagraph 3.0 least squares fitting routine.

(Eq. 2)

Inhibition constants for L-371,912 were determined from studies of the dependence of the initial velocity for the release of AFC from the substrate on the inhibitor concentration as described previously (20, 21) at 5 °C, /2 0.3, pH 7.4. The final concentrations were as follows: thrombin or thrombin mutants (10 nM), Z-GPR-AFC (12 µM or 100 µM), and inhibitor (0.1-6.0 µM). Since the inhibitor concentration is in large excess of the total enzyme concentration, the dependence of the initial rate of substrate hydrolysis (V_i) on the inhibitor concentration is described by Equation 3, where V₀ and K_i,app denote the initial rate of substrate hydrolysis in the absence of inhibitor and apparent inhibition constant, respectively.

(Eq. 3)

The inhibition constant (K_i) is related to K_i,app by Equation 4.

(Eq. 4)

The value of the observed pseudo first order rate constants (k_obs) for the binding of L-371,912 to thrombin was derived from nonlinear regression fits of progress curves to Equation 5 as described by Morrison and Walsh (22).

(Eq. 5)

where F₀, F_t, V₀, and V_s represent the initial fluorescence, the fluorescence at time t, and the initial and final rate of fluorescence change, respectively. The dependence of the first order rate constant (k_obs) on the inhibitor concentration is depicted in Equation 6.

(Eq. 6)

The association rate constant (k₁) was determined from the slope of a linear plot of k_obs versus [I] at constant substrate concentration.

The total change in the fluorescence of thrombin (0.5 µM) upon binding to Na⁺ was monitored as a function of the Na⁺ concentration at 5 °C, /2 0.3, pH 7.4. Static fluorescence measurements were performed on an SLM8000 fluorimeter (excitation 285 nm, emission 335 nm) using a thermostatted cell that was flushed with nitrogen and maintained at 5 °C with a circulating water bath. Each measurement was made after thermal equilibration. Thermal equilibration was confirmed by measurement of the temperature of the solution in the cuvette using a thermocouple.

RESULTS

Fig. 1 depicts the dependence of the difference in thrombin fluorescence (arbitrary units) in the presence of Na⁺ (F_f) and absence (F_o) of Na⁺ on the Na⁺ concentration at 5 °C. A value of 5.0 mM for the apparent dissociation constant K_d was obtained from fits of the dependence of the Na⁺-induced change in fluorescence (F) on the Na⁺ concentration ([Na⁺]) to the absorption isotherm defined by Equation 7 where F_l is the limiting value of the Na⁺-induced fluorescence change and K is equivalent to K_d (Fig. 1).

(Eq. 7)

A stopped-flow fluorescence study of the association of Na⁺ with thrombin at 5 °C with [Na⁺] = 5 mM is shown in Fig. 2. Inspection of Fig. 2 reveals that the binding of Na⁺ to thrombin is a biphasic process with a fast phase that occurs within the dead time of the stopped-flow fluorimeter and with a slow phase defined by a pseudo first order rate constant (k_obs) of 32 s¹. Such biphasic binding of a ligand to a protein has usually been interpreted in terms of one of two limiting kinetic pathways analogous to those depicted in Scheme 1, A and B (23). In Scheme 1A, Na⁺ binds to thrombin in a rapid reversible fashion to produce an initial thrombin-Na⁺ complex (E_S·Na⁺) with an increased fluorescence followed by a slow conformational change to produce a second thrombin-Na⁺ (E_F·Na⁺) complex. This scheme predicts that the rate constant for the slow phase should increase to a limiting value (k₂ + k₂) as the Na⁺ concentration is increased (Equation 8).

(Eq. 8)

In Scheme 1B, Na⁺ rapidly binds to the thrombin conformer E_s2 to produce an E_F·Na⁺ complex concomitant with a fluorescence increase. This process perturbs the equilibrium between free E_S2 and E_S1. In this pathway, the slow phase reflects conversion of E_S1 to E_S2. If Scheme 1B were operating, the pseudo first order rate constant for the slow phase should exhibit the Na⁺ dependence defined by Equation 9.

(Eq. 9)

Thus, the observed pseudo first order rate constant (k_obs) for the slow phase should decrease from a value equivalent to k₁ + k₁ to k₁ as the Na⁺ concentration is increased.

Fluorescence titration for the binding of Na⁺ to thrombin at 5 °C, pH 7.4, /2 0.3 (maintained by addition of ChCl), 500 nM thrombin.

[View Larger Version of this Image (12K GIF file)]

Biphasic binding of Na⁺ to thrombin at 5 °C, pH 7.4, /2 0.3, 5 mM Na⁺, 500 nM thrombin as monitored by stopped-flow fluorimetry.

[View Larger Version of this Image (18K GIF file)]

Scheme 1.

Since the Na⁺ concentration dependence of the pseudo first order rate constant for the slow phase should in theory distinguish between kinetic pathways A and B (Scheme 1, A and B), stopped-flow kinetic studies were performed at several Na⁺ concentrations. The data listed in Table I indicate that little change in the pseudo first order rate constant for the slow phase (k_obs) is seen upon increasing the Na⁺ concentration from 5 to 150 mM and that k_obs is changed little by increasing the ionic strength from 0.3 to 2. The relative constancy of k_obs observed in the range 5-150 mM Na⁺, together with the increasing k_obs with increasing Na⁺ concentration in the range 250-1000 mM, clearly deviates from the dependence of k_obs on the Na⁺ concentration predicted by either Equation 8 or 9. Thus, Na⁺ must bind to thrombin via a process more complex than that depicted in either Scheme 1A or Scheme 1B. Scheme 2, which might be considered as a modification of Scheme 1B wherein Na⁺ binds to two conformers of thrombin at different rates, is the simplest pathway that can account for the dependence of k_obs on the Na⁺ concentration.

Table I. Sodium ion concentration and ionic strength dependence of the pseudo first order rate constant (kobs) for the binding of Na+ to thrombin at 5 °C, pH 7.4  [Na+] kobs kobs a  2 = 0.3  2 = 2.0  mM s1 s1 5 32.2 34.3 32.1 17 32.5 34.6 30.3b 150 33.1 35.0 34.3 250 37.2 38.2 500 48.1 48.1 1000 65.4 68.3 a Rate constants for the approach of the simulated time dependence of fluorescence to its final value. Simulations were obtained by Runge-Kutta integration of Equations 22a-22d using values of k1 = 28 s1, k1 = 8.8 s1, k2 = 1.2 × 105 M1s1, k2 = 456 s1, k3 = 40 M1s1, and k3 = 0.053 s1 and EF·Na+ = 15.9 µM1, ES1 = 11.4 µM1, and ES2 = 14.5 µM1, together with Equations 15-18. b According to Scheme 2, the pseudo first order rate constant should first decrease from k1 + k1 to k1 as the Na+ concentration is increased, and then increase when the value of k3[Na+] becomes significant.

Scheme 2.

In Scheme 2, Na⁺ is envisaged as binding more rapidly to E_S2 than E_S1. At low Na⁺ concentrations ([Na⁺] < 150 mM), E_S1 is converted to E_S2 more rapidly than E_S1 binds to Na⁺ so that formation of E_F·Na⁺ from E_S1 proceeds via the pathway, E_S1  E_S2  E_F·Na⁺. An estimate of the equilibrium constant for the dissociation of E_F·Na⁺ to E_S2 and Na⁺ can be obtained by analyzing the dependence of the amplitude of the fast phase on the Na⁺ concentration (at [Na⁺]<150 mM) in terms of the absorption isotherm defined by Equation 7. A value of 3.8 mM was obtained from the fit of the data in Fig. 3 to Equation 7, where K is equivalent to k_-2 of Scheme 2. Analysis of the sum of the amplitudes of the fast phase (F_i  F_o) and slow phase (F_f  F_i) in terms of Equation 7 yields a value of K = 5.0 mM, which as expected is equivalent to the dissociation constant determined from the static fluorescence measurements. Equations 10-14 relate the equilibrium constant K_d to the equilibrium constants that define the processes depicted in Scheme 2.

(Eq. 10)

(Eq. 11)

(Eq. 12)

(Eq. 13)

(Eq. 14)

Thus, the experimentally determined values of K_d and K₂ of 5.0 mM and 3.8 mM yield values of 1.2 mM and 3.2 for K₃ and K₁. As the Na⁺ concentration is increased, and k₃[Na⁺] becomes comparable to k₁, formation of E_F·Na⁺ directly from E_S1 becomes important. Equations 15-18 relate fluorescence values observed in the stopped-flow fluorimeter to the concentration of the species in Scheme 2, where the subscripts o, i, f, and t denote values in the absence of Na⁺, at the end of the fast phase, at the end of the slow phase, and at time t after mixing, respectively, and  represents fluorescence response factors for the subscripted species.

(Eq. 15)

(Eq. 16)

(Eq. 17)

(Eq. 18)

The observed values of F_o, F_i, and F_f in experiments with 150 mM [Na⁺] and 0.5 µM thrombin, and Equations 19-21 were used to calculate values for the fluorescence response factors _ES1, _ES2, and _EF·Na⁺ of 11.4 µM¹, 14.5 µM¹, and 15.9 µM¹ respectively.²

(Eq. 19)

since [E_S1]_i = [E_S1]_o.

(Eq. 20)

(Eq. 21)

Runge-Kutta digital integration of the set of differential equations (Equations 22a-22d) for Scheme 2, with k₁ = 28 s¹, k₁ = 8.8 s¹, k₂ = 1.2 × 10⁵ M¹ s¹, k₂ = 456 s¹, k₃ = 40 M¹ s¹, and k₃ = 0.053 s¹, yielded the time dependence of [E_S1], [E_S2], and [E_F·Na⁺], which, when substituted into Equation 18, provided the best fit for the time-dependent changes in fluorescence observed upon mixing Na⁺ with thrombin in the stopped-flow fluorimeter (Fig. 4). Additionally, the apparent first order rate constants for the approach of the calculated fluorescence to its final value agreed with the corresponding apparent first order rate constants obtained from the observed time dependence of fluorescence over the entire range of Na⁺ concentrations studied (Table I). Primes in Equations 22a-22d denote the derivative of the designated concentrations with respect to time.

(Eq. 22a)

(Eq. 22b)

(Eq. 22c)

(Eq. 22d)

Consistent with Scheme 2, digital integration of Equations 22a-22d using the rate constants listed above accounts for the observed time dependence of fluorescence in experiments where the thrombin-Na⁺ complex is dissociated by dilution with Na⁺ free buffer (Fig. 5). Fig. 6 depicts the effect of diluting the thrombin-Na⁺ complex with a Na⁺ free buffer containing a crown ether (15-crown-5) that sequesters free Na⁺ and thereby promotes further dissociation of Na⁺ from thrombin.

Fluorescence titration for the binding of Na⁺ to E_S2 as determined from the Na⁺ concentration dependence of the amplitude of the fast phase in the stopped-flow fluorimeter at 5 °C, pH 7.4, /2 0.3, 500 nM thrombin.

[View Larger Version of this Image (12K GIF file)]

Fit of the time dependence of the fluorescence upon mixing thrombin (500 nM) with Na⁺ at 5 °C, pH 7.4, /2 0.3. 

1

1

1

1

1

2

1

1

1

1

[View Larger Version of this Image (28K GIF file)]

Fit of the time dependence of the fluorescence upon a 6-fold dilution of thrombin (3 µM) in a Na⁺ (100 mM or 30 mM) containing buffer with Na⁺-free buffer at 5 °C, pH 7.4, /2 0.3; Final Na⁺ concentration.

1

1

1

1

1

2

1

1

1

1

[View Larger Version of this Image (27K GIF file)]

Time dependence of fluorescence after a 6-fold dilution of buffered solution (/2 0.3) containing 30 mM Na⁺ and 3.0 µM thrombin with a buffer (/2 0.3) containing no thrombin or Na⁺ at 5 °C, pH 7.4, and either no other additions (a) or 25 mM of crown ether (15-crown-5) (b).

1

1

[View Larger Version of this Image (24K GIF file)]

Additionally, the alkali metal ions Rb⁺ and K⁺ that bind to thrombin and result in the slow  fast conversion do so with a limiting rate constant of 33 s¹ (data not shown). This rate constant is close to the value for the rate constant for conversion of E_S1 to E_S2 observed in studies of Na⁺ binding to thrombin, as expected for the pathway depicted in Scheme 2.

Although Scheme 2 depicts the simplest kinetic pathway for the binding of Na⁺ to thrombin that accounts for our experimental observations, Na⁺ probably binds to thrombin via a more complex pathway involving formation of Na⁺ complexes of E_S1 and E_S2 that undergo conformational rearrangements to E_F·Na⁺ (Scheme 3). Rapid rates of conversion of E_S1·Na⁺ and E_S2·Na⁺ to E_F·Na⁺, or the failure to saturate a weak E_S1-Na⁺ or E_S2-Na⁺ interaction may account for our inability to detect to E_S1·Na⁺ and E_S2·Na⁺.

Scheme 3.

At 150 mM Na⁺ (37 °C), ~50% of the thrombin contains no bound Na⁺ (9). The Na⁺-free thrombin (slow form thrombin) appears predominantly E_s2 as judged by the ratio of the amplitudes between fast and slow phases (~7:1), assuming that the relative fluorescence response factors for the different forms of thrombin are independent of temperature. Pseudo first order rate constants of 184 ± 6 s¹ and 463 ± 20 s¹ were determined for the binding of 300 and 700 mM Na⁺ to thrombin at 25 °C and 37 °C, respectively. The observation of biphasic Na⁺ binding at 25 °C and 37 °C similar to that observed at 5 °C, suggests that the Na⁺-induced slow to fast transition is not qualitatively altered in the temperature range 5-37 °C.³ Consistent with this view, we found no evidence for temperature-induced changes in thrombin structure in the temperature range (5-37 °C), as judged by the temperature dependence of affinity for the active site-directed inhibitor and catalytic activity (data not shown).

Recent studies have shown that the mutants R221aA and K224A bind Na⁺ with reduced affinity (19). Values of 48 mM and 65 mM for the equilibrium constants for dissociation of Na⁺ from R221aA and K224A at 5 °C were determined by fluorescence titration (data not shown). Interestingly, at 150 mM NaCl, 5 °C, the binding of Na⁺ to R221aA follows a biphasic time course with a fast phase that occurs within the dead time of the stopped-flow fluorimeter, and a slow phase characterized by a rate constant of 31 s¹, which was independent of the Na⁺ concentration (Fig. 7). The K224A variant, however, binds Na⁺ in a monophasic reaction (Fig. 7). The observation of a rate constant of 30 s¹ for the binding of Na⁺ to K224A, which was independent of the Na⁺ concentration in the range 40-300 mM, suggests that Scheme 3 is operative with K224A, but no fluorescence change is associated with the transition of E_S2 to E_F·Na⁺ with the K224A mutant. This conclusion implies that the integrity of the Lys-224-Glu-217 ion pair is important for the fluorescence change induced by Na⁺ binding (15).

Binding of Na⁺ to thrombin mutants at 5 °C, pH 7.4, /2 0.3, 150 mM Na⁺, 500 nM K224A (panel A) and R221aA (panel B) as monitored by stopped-flow fluorimetry.

[View Larger Version of this Image (29K GIF file)]

To determine how the binding of Na⁺ to thrombin might allosterically affect the environment at the active site, we studied the effect of Na⁺ on affinity and rate of binding of the active site-directed inhibitor L-371,912 (Structure 1). Equilibrium and rate constants for the interaction of L-371,912 with thrombin are listed in Table II. These thermodynamic and kinetic parameters were determined from studies of the time dependent inhibition of thrombin catalyzed hydrolysis of Z-GPR-AFC by L-371,912 (Figs. 8, 9, 10). Inhibition constants were deduced from the dependence of the limiting velocity on inhibitor concentration (obtained from plots such as those illustrated in Fig. 8). Pseudo first order rate constants for the binding of the active site-directed inhibitor to thrombin were obtained from an analysis of the time dependent approach of the velocity to its final value using previously described methods (20, 21). The slopes of linear plots of pseudo first order rate constants versus inhibitor concentration (Fig. 10) yielded the second order rate constants listed in Table II for the binding of L-371,912 to thrombin. In each case, the inhibitor appears to bind the fast form tighter and more rapidly than the slow form. Interestingly, although the inhibitor binds the fast form of wild-type and mutant thrombins with similar affinities and rate constants, the inhibitor shows a >10-fold reduction in affinity for the slow form of the K224A and R221aA thrombin mutants.

structure 1.

Table II. Values of Ki and kon for the binding of L-371,912 to slow and fast form thrombin at 5 °C, pH 7.4  Ki KiS/KiF kon konS/konF Fast forma (KiF) Slow form (KiS) Fast formb (konF) Slow form konS nM µM1s1 Wild-type 5.1 23.9 4.7 0.81 0.10 0.12 K224A 2.7 319 118.1 1.76 0.052 0.029 R221aA 9.3 380 40.9 1.76 0.12 0.068 a KiF, the equilibrium constant for dissociation of L-371,912 from the thrombin-Na+-L-371,912 ternary complex, was calculated using the relationship KiF = Ki,obs × KiS × [Na+]/(KiS(Kd + [Na+])  Ki,obs × Kd), where Ki,obs is the apparent equilibrium constant for dissociation of L-371,912 from thrombin at the Na+ concentration (300 mM) used for the measurement, Kd is the equilibrium constant for dissociation of Na+ from thrombin and KiS is the equilibrium constant for dissociation of L-371,912 from thrombin in the absence of Na+. b Values for konF, the second order rate constant for the binding of L-371,912 to the thrombin-Na+ binary complex, were evaluated using the relationship konF = (kon,obs × (Kd + [Na+])  Kd × konS)/[Na+], where kon,obs is the apparent second order rate constant for binding of the inhibitor to thrombin at the Na+ concentration (300 mM) used for the measurement, and konS is the second order rate constant for binding of L-371,912 to thrombin in the absence of Na+.

Progress curves depicting the inhibition of wild-type thrombin by L-371,912 (I) in the presence of a fluorogenic substrate (10 µM Z-GPR-AFC) at 5 °C, pH 7.4, /2 0.3, 150 mM Na⁺, 10 nM thrombin as monitored in the stopped-flow fluorimeter.

[View Larger Version of this Image (19K GIF file)]

Dependence of the apparent first order rate constant (k_obs) for inhibition of wild-type and the mutant thrombins by L-371,912 in the presence of 150 mM Na⁺ on the concentration of L-371,912 at 5 °C, pH 7.4, /2 0.3, 10 nM wild-type thrombin (), R221aA (), or K224A ().

[View Larger Version of this Image (12K GIF file)]

Dependence of the apparent first order rate constant (k_obs) for inhibition of wild-type and mutant thrombins by L-371,912 in the absence of Na⁺ on the concentration of L-371912 at 5 °C, pH 7.4, /2 0.3, 10 nM wild-type thrombin (inset), R221aA (), or K224A ().

[View Larger Version of this Image (15K GIF file)]

Like the binding of Na⁺ to the slow form, the binding of inhibitor to the slow form is associated with a significant increase in fluorescence. Upon binding of L-371,912 to the fast form, however, no significant change in fluorescence is observed (data not shown). This observation suggests that, upon binding to thrombin, L-371,912 induces a conformational change similar to that induced by Na⁺. The change in intrinsic protein fluorescence accompanying the binding of L-371,912 to thrombin allowed us to study the binding of this inhibitor at higher inhibitor concentrations and thereby enabled us to determine whether L-371,912, like Na⁺, binds selectively to E_S2. As shown in Fig. 11, the apparent pseudo first order rate constant for the binding of L-371,912 to thrombin as measured in the stopped-flow fluorimeter increases with increasing inhibitor concentration and approaches a limiting value of 29 s¹ consistent with the pathway illustrated in Scheme 4. At high concentrations of L-371,912 (1 mM), binding of this ligand to thrombin was a biphasic process, as in the case of the binding of Na⁺ to thrombin. The failure to observe a biphasic reaction at low concentrations of L-371,912 probably reflects the fact that at low concentrations of L-371,912 formation of the E_S2·L-371,912 complex is slower than conversion of E_S1 to E_S2. The second order rate constant for binding of L-371,912 to E_S2 of 1.3 × 10⁵ M¹ s¹ indicates that a concentration of >200 µM L-371,912 would be required for the rate of binding L-371,912 to E_S2 to be equivalent to the rate of conversion of E_S1 to E_S2. The inset in Fig. 11 shows the linear dependence of k_obs on the concentration of L-371,912 obtained at [L-371,912] < 32.0 µM in studies where changes in thrombin fluorescence were used to monitor inhibitor binding. The value of the overall second order rate constant for the formation of thrombin-L-371,912 complex obtained from the slope of the plot (Fig. 11, inset) yields a value of 0.09 µM¹ s¹, which was as expected similar to the value obtained for this rate constant (0.1 µM¹ s¹) by monitoring the time dependent effect of L-371,912 on substrate hydrolysis (Fig. 10, Table II). Fig. 11 illustrates the fit of the observed data to Scheme 4. The open circles plot rate constants for the first order approach of the observed fluorescence to its final value. The filled circles plot rate constants obtained by simulating the time dependence of fluorescence accompanying the binding of L-371,912 to thrombin using the kinetic parameters stipulated in Scheme 4 and determining the apparent first order rate constant for the approach of the simulated time dependence of fluorescence to its final value.

Dependence of the pseudo first order rate constants (k_obs) for binding of L-371,912 to slow form thrombin on the concentration of the inhibitor at 5 °C, pH 7.4, /2 0.3, 500 nM thrombin.

1

1

1

1

1

2

1

1

1

1

[View Larger Version of this Image (16K GIF file)]

Scheme 4.

To assess the effect of the slow  fast transition of thrombin on the environment of exosite I, we investigated the effect of Na⁺ on the binding of hirugen to exosite I of thrombin. Fluorescence titration indicated equilibrium constants of 0.5 µM and 2.7 µM for the dissociation of thrombin-hirugen complexes of fast and slow form thrombin at 5 °C, /2 0.3. Rate constants for the binding of hirugen to thrombin in the absence of Na⁺ were determined by the increase in fluorescence associated with this process. As in the case of the binding of Na⁺ and L-371,912 to thrombin, the pseudo first order rate constant for this process approached a limiting value as the concentration of hirugen increased. The limiting value of 26 s¹ (at 17 µM hirugen) observed for the binding of hirugen to thrombin is somewhat lower than that observed for the binding of L-371,912 to thrombin. The limiting rate constant obtained with hirugen may have been depressed by secondary effects of hirugen. At high concentrations of hirugen both the fluorescence change and the pseudo first order rate constant for the binding of hirugen to thrombin decrease with increasing hirugen concentration (data not shown). This behavior may reflect binding of a second molecule of hirugen to the thrombin hirugen complex at high hirugen concentrations. At low concentrations of hirugen (<17 µM), however, our observations are consistent with one molecule of hirugen binding selectively to the E_S2 conformer of thrombin and subsequently mediating the slow  fast form transition. Thus, binding of L-371,912 to the active site, binding of hirugen to exosite I, and binding of Na⁺ to the alkali metal binding site of thrombin are thermodynamically linked processes that appear to occur via a common kinetic pathway.

DISCUSSION

The present study indicates that Na⁺-free thrombin consists of two conformers in a 3:1 ratio (E_S2:E_S1) at 5 °C with E_S2 binding Na⁺ 3 orders of magnitude faster than does E_S1. Since E_S1 and E_S2 exhibit similar affinities for Na⁺, the rate constant for dissociation of Na⁺ from E_S2 must be much larger than that for dissociation of Na⁺ from E_S1. These observations suggest that steric constraints may block entry and egress of Na⁺ to and from the Na⁺ binding site of E_S1. The fact that the active site-directed inhibitor, L-371,912, and the exosite I ligand hirugen, like Na⁺, bind more rapidly to E_S2 suggests that concerted conformational changes occur at the Na⁺ binding site, the active site and exosite I of thrombin concomitant with the transition from E_S1 to E_S2.

Upon binding Na⁺, thrombin exhibits a 4.7-5.4-fold increase in its affinity for L-371,912 and hirugen. The fact that the bound Na⁺ is spatially removed from the active site of thrombin and exosite I makes it unlikely that a direct ionic effect of Na⁺ at the active site or exosite I is responsible for the increased affinity of thrombin for L-371,912 and hirugen in the presence of saturating concentrations of Na⁺. The simplest explanation for the increase in affinity of thrombin for L-371,912 and hirugen observed in the presence of saturating concentration of Na⁺ is that the thrombin conformer stabilized in the E_F·Na⁺ complex has a greater affinity for these ligands. A detailed pathway for the slow to fast transition is depicted in Scheme 3. Thus, as depicted in Scheme 3, upon binding Na⁺, the E_S2 conformer of thrombin, which binds Na⁺ faster than the E_S1 conformer, undergoes a further conformational transition to E_F·Na⁺, which exhibits an increased affinity for the active site and exosite ligands L-371,912, and hirugen.

Recognition that thrombin undergoes a conformational change upon binding Na⁺ resulted in the designation of Na⁺-free thrombin as slow form thrombin and Na⁺-bound thrombin as fast form thrombin. The kinetic pathway (Scheme 3) for the binding of Na⁺ to thrombin delineated in the present study indicates that slow form thrombin is an equilibrium mixture of two conformers with the more abundant conformer E_S2 binding Na⁺ faster than the less abundant one. Fast form thrombin appears to be an equilibrium mixture of three thrombin conformers. All observations regarding the alterations in the properties of thrombin induced by the binding of Na⁺ are consistent with E_F·Na⁺ being the predominant species and conformationally distinct from E_S1·Na⁺ and E_S2·Na⁺.

It is important to note that the Ala substitution at Lys-221a and Arg-224 had little effect on the affinity of thrombin for L-371,912 in the presence of saturating Na⁺, whereas in the Na⁺-free state these substitutions reduced the affinity of thrombin for L-371,912 by ~15-fold. This observation is consistent with the view that the Ala substitutions selectively perturb the slow form. The observation that binding of hirugen to thrombin exosite I, like the binding of Na⁺ to the alkali metal binding site and the binding of L-371,912 to the active site, appears to induce the slow  fast transition, indicates that the equilibrium between slow and fast forms can be altered by interactions at three structurally distinct domains of thrombin. The occurrence of concerted alterations in the allosteric sites of the enzyme reported here suggests the possible existence of physiologically important allosteric effectors of thrombin that selectively stabilize the slow or fast form, and thereby regulate the anticoagulant and procoagulant activities of the enzyme.