Molecular mechanism of tissue factor-mediated acceleration of factor VIIa activity.

The mechanism of the acceleration of the catalytic activity of factor VIIa (VIIa) in the presence of tissue factor (TF) was investigated. To explore the VIIa's site(s) that correlates with TF-mediated acceleration, zymogen VII, VIIa, and active site-modified VIIa were prepared, and dissociation constants (Kd) for their bindings to TF or soluble TF in solution were determined. We found that conversion of zymogen VII to VIIa led to an increase in affinity (DeltaDeltaG = 4.3-4.4 kJ/mol) for TFs. Dansyl-Glu-Gly-Arg chloromethyl ketone (DNS-EGRck) treatment of VIIa led to a further increase in the affinity (DeltaDeltaG = 7.3-12 kJ/mol). Neither removal of the Gla domain from VIIa nor truncation of the COOH-terminal membrane and cytoplasmic regions of TF affected the affinity enhanced after DNS-EGRck treatment of VIIa. Treatment of VIIa with (p-amidinophenyl)methanesulfonyl fluoride also enhanced its affinity for soluble TF, whereas treatment with 4-(2-aminoethyl)benzenesulfonyl fluoride, phenylmethylsulfonyl fluoride, or diisopropyl fluorophosphate had a slight effect on the affinity. On the other hand, DNS-EGRck and (p-amidinophenyl)methanesulfonyl fluoride treatments, but not diisopropyl fluorophosphate treatment, of VIIa led to protection of its alpha-amino group of Ile-153 from carbamylation. Protection of the alpha-amino group was consistent with formation of a critical salt bridge between Ile-153 and Asp-343 in the protease domain of VIIa. Therefore, TF may preferentially bind to the active conformational state of VIIa. When one assumes that free VIIa exists in equilibrium between minor active and dominant zymogen-like inactive conformational states, preferential binding of TF to the active state leads to a shift in equilibrium. We speculate that TF traps the active conformational state of VIIa and converts its zymogen-like state into an active state, thereby accelerating the VIIa activity.

The mechanism of the acceleration of the catalytic activity of factor VIIa (VIIa) in the presence of tissue factor (TF) was investigated. To explore the VIIa's site(s) that correlates with TF-mediated acceleration, zymogen VII, VIIa, and active site-modified VIIa were prepared, and dissociation constants (K d ) for their bindings to TF or soluble TF in solution were determined. We found that conversion of zymogen VII to VIIa led to an increase in affinity (⌬⌬G ‫؍‬ 4.

3-4.4 kJ/mol) for TFs. Dansyl-Glu-Gly-Arg chloromethyl ketone (DNS-EGRck) treatment of VIIa led to a further increase in the affinity (⌬⌬G ‫؍‬ 7.3-12 kJ/mol). Neither removal of the Gla domain from
VIIa nor truncation of the COOH-terminal membrane and cytoplasmic regions of TF affected the affinity enhanced after DNS-EGRck treatment of VIIa. Treatment of VIIa with (p-amidinophenyl)methanesulfonyl fluoride also enhanced its affinity for soluble TF, whereas treatment with 4-(2-aminoethyl)benzenesulfonyl fluoride, phenylmethylsulfonyl fluoride, or diisopropyl fluorophosphate had a slight effect on the affinity. On the other hand, DNS-EGRck and (p-amidinophenyl)methanesulfonyl fluoride treatments, but not diisopropyl fluorophosphate treatment, of VIIa led to protection of its ␣-amino group of Ile-153 from carbamylation. Protection of the ␣-amino group was consistent with formation of a critical salt bridge between Ile-153 and Asp-343 in the protease domain of VIIa. Therefore, TF may preferentially bind to the active conformational state of VIIa. When one assumes that free VIIa exists in equilibrium between minor active and dominant zymogen-like inactive conformational states, preferential binding of TF to the active state leads to a shift in equilibrium. We speculate that TF traps the active conformational state of VIIa and converts its zymogen-like state into an active state, thereby accelerating the VIIa activity.
Factor VIIa (VIIa) 1 is a plasma serine protease that is essen-tial for the initiation of extrinsic blood coagulation. When VIIa forms a complex with cell-derived tissue factor (TF) in the presence of Ca 2ϩ and phospholipids, the protease activity of VIIa toward its natural substrates, factors X and IX, is enhanced by many orders of magnitude, and the coagulation cascade is triggered (1)(2)(3). In vitro, formation of the active complex can be evidenced by measuring the amidolytic activity of VIIa; this activity is also enhanced in the presence of TF and Ca 2ϩ without phospholipids (4 -8). To elucidate the molecular interaction between VIIa and TF, investigators explored the sites of VIIa interacting with TF. It was suggested that the ␥-carboxyglutamic acid (Gla) domain is important for high affinity binding of VIIa to TF (7-9). The first epidermal growth factor (EGF)-like domain also seems to be important for interactions of zymogen factor VII (VII) with TF (10,11). In our previous study, Gla-EGF1 peptide consisting of both Gla-domain and the first EGF-like domain of bovine VII showed a strong affinity for bovine TF, and this affinity was lost when cleaving the peptide at the hinge region between the two domains (12). Kazama et al. (13) also reported that the affinity of human TF for a fragment containing Gla domain and two EGF-like domains isolated from human VII was lost after removal of the Gla domain from the fragment. These findings suggest that the Gla domain-dependent site(s) interacting with TF exists in the Gla-EGF1 region of VII/VIIa. On the other hand, we find that VIIa has Gla domain-independent site(s) interacting with TF, whereas zymogen VII does not have corresponding site(s). An ␣-amino group of NH 2 -terminal Ile-153 of VIIa heavy chain is important for the Gla domain-independent interaction with TF (12). Furthermore, sites containing residues 195-206 (14), Glu-220 (15), Arg-304 (16,17), or other sites (18) located in the catalytic domain of human VIIa may interact with TF. The site of interaction correlating with TFmediated acceleration of VIIa activity has remained to be determined. Evidence that the ␣-amino group of Ile-153 in VIIa is protected from carbamylation but only after a complex formation with TF (12) has been reported, hence the TF-induced conformational change occurs in this region. Further evidence has now been obtained that supports the proposal that both the substrate-binding site and the ␣-amino group of Ile-153 in catalytic domain of VIIa are cooperatively involved in the TFinduced conformational change. Mechanisms related to the TF enhancement of the amidolytic activity of VIIa are discussed.
␣-Chymotrypsin was purchased from Worthington. The ammonium sulfate fraction (40 -65% saturation) of the barium citrate precipitate from bovine plasma was a gift from Mochida Pharmaceutical Co. Ltd. (Tokyo). All other chemicals were of analytical grade or the highest quality commercially available.
Proteins-Bovine VII was highly purified, as described (19). VIIa and Gla domainless VIIa (VIIa(GDϪ) were prepared from the bovine factor VII, as described (7). Bovine TF was purified from acetone powder of the placenta, using a modification of the method of Broze et al. (20) and was kept in the presence of 0.1% Triton X-100. The DNA encoding the extracellular domain of TF (amino acid residues 1-213, named recombinant bovine soluble TF, rsBTF) was constructed from the cDNA of bovine TF previously isolated (21). The protein was expressed in yeast and purified using the same procedures as used for human soluble TF, also expressed in yeast (22). The purified proteins gave single bands with a molecular mass of 37,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Protein concentrations of rsBTF, VII, and derivatives of VII were determined by amino acid analysis. The concentration of TF was determined by the dye binding method with bovine serum albumin as the standard (23), and the molecular mass used was 43,000 (24).
Treatments of VIIa and VIIa(GDϪ) with DNS-EGRck-As in previous work (7,12), 50 l of 5 mM DNS-EGRck dissolved in distilled water was added to 450 l of each of the protein solutions, which contained 1.6 nmol of VIIa or 1.3 nmol of VIIa(GDϪ) in 50 mM Tris-HCl, pH 8.0, containing 0.1 M NaCl and 0.01% NaN 3 (TBS), and each mixture was incubated at 25°C. After 4 h, excess DNS-EGRck was removed by extensive dialysis against TBS.
Treatments of VIIa with APMSF, ABSF, PMSF, or DFP-Fifty microliters each of 50 mM ABSF dissolved in distilled water, 40 mM PMSF, and 20 mM DFP dissolved in methanol were added to 450 l of the protein solution, which contained 1.6 nmol of VIIa in TBS, and the mixture was incubated at 25°C. After 4 h, excess reagents were removed by extensive dialysis against TBS. In the case of APMSF treatment of VIIa, 10 l of 100 mM APMSF newly dissolved in distilled water was added five times every 30 min to 450 l of the VIIa solution at 25°C. After the final addition of APMSF, incubation was continued for a further 2 h. The sample was then dialyzed against TBS.
Determination of Dissociation Constants for Binding of VIIa or Gla Domainless VIIa to TF or rsBTF-Various concentrations of rsBTF or TF were incubated with a fixed concentration of VIIa (2.0 or 0.2 nM) or 2.8 nM VIIa(GDϪ) in 230 l of TBS containing 5 mM CaCl 2 and 0.1% bovine serum albumin at 37°C for 15 min, and then 20 l of S-2288 (10 mM) was added. In the presence of TF, 0.02% Triton X-100 was also added to the reaction mixture. The rates of change in absorbance at 405 nm were measured to determine the velocity of hydrolysis of S-2288. The rate of hydrolysis of S-2288 in the absence of TF was also measured, as described above, and subtracted from the total rate of the substrate hydrolysis. The data were subjected to Woolf-Augustinsson-Hofstee plot analysis (25) to determine the apparent dissociation constants (K d where TF o and VIIa o are overall concentrations of added TF and those of VIIa, respectively. V obs and V satur are the observed velocity of hydrolysis of S-2288 by VIIa at each concentration of TF and that in the presence of a saturating concentration of TF, respectively. The data were fitted to linear regression curves. The ϪK d values were obtained from the x axis intercept. TF-partitioning Assay-VIIa and inactive derivatives of VII (VII*), such as zymogen VII or active site-masked VIIa, were mixed in fixed molar ratio, and various concentrations of the mixture were incubated with fixed concentrations of TF or rsBTF (see the legend for Fig. 2) in 230 l of TBS containing 5 mM CaCl 2 and 0.1% bovine serum albumin at 37°C for 15 min. In the presence of TF, 0.02% Triton X-100 was also added to the reaction mixture. After incubation, 20 l of S-2288 (10 mM) was added, and the rate of hydrolysis of S-2288 by VIIa complexed with TF or rsBTF was measured, as described above. We also measured the rate of hydrolysis of S-2288 by the VIIa⅐TF complex in the presence of a saturating concentration of VIIa instead of the VIIa/VII* mixture. The fraction of TF complexed with VIIa in equilibrium is given by the equation, where TF o represents the overall concentration of TF. Equation 3 can be derived from Equation 2, On the other hand, one can obtain the p values from the rate of hydrolysis of S-2288 by VIIa complexed with TF, using Equation 5, where V inh is an observed velocity of hydrolysis of S-2288 by VIIa complexed with TF in the presence of VII* and V satur is velocity of that in the presence of a saturating concentration of VIIa. As shown in Equations 3 and 4, the p value depends on the concentration of VIIa/ VII* mixture but is constant at a high concentrations of the mixture.
The p values versus concentrations of VIIa/VII* mixture were plotted to obtain the p value where the constant was reached (Fig. 2). The The mixture was incubated at 30°C for 0, 30, 60, 120, and 240 min. After incubation, 50 l of each sample taken from the reaction mixture was mixed with 20 l of 1.0 M hydroxylamine hydrochloride, pH 8.0, and incubated for 1 h at 25°C. Each of the samples was dialyzed against TBS at 4°C and incubated with 0.4 pmol of ␣-chymotrypsin at 25°C for 1 h. After incubation, the reaction was terminated by adding 70 l of 20% trichloroacetic acid, and the resulting precipitate was collected by centrifugation. The precipitate was sequentially rinsed with 10% trichloroacetic acid and with acetone and dissolved in 80% acetonitrile containing 0.1% trifluoroacetic acid. Each of the samples was then analyzed on an Applied Biosystems 477A gas-phase sequencer. Phenylthiohydantoin (PTH) derivatives were detected using an Applied Biosystems 120A PTH analyzer with an on-line system, as described previously (26), and the peak areas of PTH-Ile and PTH-Val were determined simultaneously in the first cycle of the analysis.

Determination of Dissociation Constants for Binding of VIIa or VIIa(GDϪ) to TF or rsBTF-Complex formation between
VIIa or VIIa(GDϪ) and TF or rsBTF was determined by measuring the rate of S-2288 hydrolysis by the each complex. The data were examined using Woolf-Augustinsson-Hofstee plot analysis. As shown in Fig. 1, all four linear regression curves showed much the same slope, indicating no significant difference in amidolytic activities among VIIa⅐TF, VIIa⅐rsBTF, VIIa(GDϪ)⅐TF, and VIIa(GDϪ)⅐rsBTF complexes. The four curves intercepted the x axis at different positions, indicating that K d values for the four interactions varied. The K d values are summarized in Table I.
Determination of Dissociation Constants for Binding of Inactive Derivatives of VII to TF or rsBTF-Affinities of zymogen VII, DNS-EGRck-treated VIIa (DNS-EGRck VIIa), and DNS-EGRck VIIa(GDϪ) for TF or rsBTF were estimated, using a TF-partitioning assay, as described under "Experimental Procedures." The K d values for binding of inactive derivatives of VII to TF or rsBTF were calculated from p values that reached the constant (Fig. 2), using the K d value for binding of VIIa to TF (0.24 nM) or that for binding of VIIa to rsBTF (14 nM), by Equation 4. The obtained K d values are also summarized in Table I.
Effect of Deletion of Gla Domain from Derivatives of VIIa on Their Affinities for TF or rsBTF-The K d values in Table I are used to calculate differences in free energy for binding of TF or rsBTF to various derivatives of VIIa before and after deletion of the Gla domain. As shown in Table II, deletion of the Gla domain from the derivatives of VIIa resulted in great decreases in affinities for TF (⌬⌬G ϭ 11-15 kJ/mol), whereas the deletion resulted in only modest decreases in affinities for rsBTF (⌬⌬G ϭ 4.4 -4.9 kJ/mol).
Effect of Truncation of Membrane and Cytoplasmic Regions from TF on Affinity for Derivatives of VIIa-Affinities of TF and rsBTF for derivatives of VIIa were then compared. As shown in Table III, truncation of membrane and cytoplasmic regions led to a decrease in affinity for VII, VIIa, or DNS-EGRck-treated VIIa (⌬⌬G ϭ 5.6 -10 kJ/mol). In contrast, truncation did not affect the affinity for Gla domainless derivatives of VIIa; thus, truncation of the COOH-terminal region of TF apparently results in loss of interaction with Gla domain-dependent site in VIIa.
Effect of Activation of Zymogen VII on Affinity for TF or rsBTF-As shown in Table IV, activation of zymogen VII resulted in an increase in the affinity for TF (⌬⌬G ϭ 4.4 kJ/mol).
A similar increase in the affinity for rsBTF (⌬⌬G ϭ 4.3 kJ/mol) was observed, which means that the enhanced affinities are probably not affected by truncation of the COOH-terminal region of TF.
Effect of DNS-EGRck Treatment of VIIa or VIIa(GDϪ) on Affinities for TF or rsBTF-After treatment of VIIa with DNS-EGRck, the binding affinity for TF or rsBTF was greatly enhanced (⌬⌬G ϭ 7.3-12 kJ/mol; Table V). DNS-EGRck treatment of VIIa(GDϪ) also resulted in a great increase in affinity for TF or rsBTF (⌬⌬G ϭ 12-13 kJ/mol), suggesting that neither deletion of the Gla domain from VIIa nor truncation of the COOH-terminal region of TF affects the affinity enhanced after treatment with DNS-EGRck.
Effect of Treatment of VIIa with APMSF, ABSF, PMSF, or DFP on the Affinity of VIIa for rsBTF-The results shown in Table V suggest that incorporation of the transition state analog of substrate as DNS-EGRck into VIIa induces a conformational change in the TF-binding site in VIIa. To examine the effects of other transition state analogs of substrate on the affinity of VIIa for rsBTF, VIIa was treated with APMSF, ABSF, PMSF, and DFP, and then affinities of the inactivated derivatives of VIIa for rsBTF were estimated, using the TFpartitioning assay. As shown in Table VI, DNS-EGRck-treatment proved to be the most effective enhancer of the affinity of VIIa for rsBTF. Treatment of VIIa with APMSF also greatly enhanced the affinity, and the K d value for binding of VIIa to rsBTF was reduced to 1 ⁄50 after the treatment. In contrast, treatment of VIIa with ABSF, PMSF, or DFP only slightly affected the affinity (Table VI).

Effect of DNS-EGRck Treatment of VIIa on Carbamylation Rate of ␣-Amino Group of Ile-153 in VIIa Heavy
Chain-Since the affinity of TF for DNS-EGRck VIIa or APMSF VIIa was much higher than that for native VIIa, the conformation of the active site-modified derivatives of VIIa may be similar to that of VIIa complexed with TF. To examine this possibility, carbamylation rates of the ␣-amino group of Ile-153 in various derivatives of VIIa were compared. After incubation with The values were obtained from results in Fig. 1. b The values were calculated from the p values obtained in Fig. 2  KNCO for various times, VIIa, DFP VIIa, APMSF VIIa, or DNS-EGRck VIIa was treated with ␣-chymotrypsin to remove the Gla domain, and PTH-Ile derived from NH 2 -terminal Ile-153 of the VIIa heavy chain and PTH-Val derived from the NH 2 -terminal Val-41 of VIIa(GDϪ) light chain were quantified, simultaneously, on a gas phase sequencer. As shown in Fig. 3, the ratio of PTH-Ile to PTH-Val decreased rapidly in the analysis on VIIa or DFP VIIa, suggesting that the ␣-amino group of Ile-153 in VIIa was rapidly carbamylated and was not converted to PTH-Ile. In contrast, the ratio of PTH-Ile to PTH-Val decreased slowly in the analysis on APMSF VIIa. The rate of carbamylation of DNS-EGRck VIIa was lowest among four derivatives of VIIa; thus, derivatives of VIIa that have gained high affinity for TF also afforded protection to the ␣-amino group of Ile-153. DISCUSSION We determined apparent K d values for binding of various derivatives of VIIa to TF or rsBTF and examined the interac-tion site(s) involved in the TF-mediated acceleration of VIIa activity. VIIa and VIIa(GDϪ) bound to TF or rsBTF with various K d values, and the resulting four complexes showed almost the same amidolytic activity toward the synthetic substrate S-2288 (Fig. 1). Hence, neither deletion of the Gla domain from VIIa nor truncation of the COOH-terminal membrane and cytoplasmic regions from TF affects the interaction(s) involved in the acceleration of VIIa activity. These data are consistent with documented studies (8,22,27). As deletion of the Gla domain from VIIa led to a great decrease in affinity for TF, the result was a modest decrease in the affinity for rsBTF (Table  II). Therefore, the Gla domain-dependent site(s) in VIIa may be partially involved in the interaction with the COOH-terminal region of TF. Truncation of the COOH-terminal region of TF led to a decrease in affinity for VIIa but did not affect the affinity for Gla domainless derivatives of VIIa (Table III); hence, the COOH-terminal region of TF is involved in interaction with the Gla domain-dependent site(s) in VIIa. On the a The values are the same as those in Table I. b The values were calculated from the relationship ⌬⌬G ϭ ϪRT ⅐ ln(K da /K db ), where K db and K da are the K d values for binding of TF to various VIIa before and after deletion of Gla domain, respectively. R ϭ 8.314 J ⅐ mol Ϫ1 ⅐ K Ϫ1 and T ϭ 310 K.  Table I. b The values were calculated from the relationship ⌬⌬G ϭ ϪRT ⅐ ln(K da /K db ), where K db and K da are the K d values for binding of various VIIa to TF before and after deletion of the membrane and cytoplasmic regions of TF, respectively. R ϭ 8.314 J ⅐ mol Ϫ1 ⅐ K Ϫ1 and T ϭ 310 K. a The values are the same as those in Table 1.
b The values were calculated from the relationship ⌬⌬G ϭ ϪRT ⅐ ln(K da /K db ), where K db and K da are the K d values for binding of TF to VII before and after activation, respectively. R ϭ 8.314 J ⅐ mol Ϫ1 ⅐ K Ϫ1 and T ϭ 310 K. a The values are the same as those in Table 1.
b The values were calculated from the relationship ⌬⌬G ϭ ϪRT ⅐ ln(K da /K db ), where K db and K da are the K d values for binding of TF to various VIIa before and after treatment with DNS-EGRck, respectively. R ϭ 8.314 J ⅐ mol Ϫ1 ⅐ K Ϫ1 and T ϭ 310 K. other hand, the activation of zymogen VII leads to an increase in affinity for TF or rsBTF. DNS-EGRck treatment of VIIa further enhances the affinity for TF or rsBTF, and this enhanced affinity is not affected by removal of the Gla domain (Table V). The interaction gained after activation of zymogen VII or that gained after DNS-EGRck treatment of VIIa seems to be independent of interactions that require the Gla domain of VIIa or the COOH-terminal region of TF. The relationship among effects of various modifications on the free energy of interaction between VIIa and TF is illustrated in Fig. 4.
We reported earlier that the ␣-amino group of Ile-153 in VIIa is important for the Gla domain-independent interaction with TF. Moreover, the TF-induced conformational change of VIIa results in protection from carbamylation of the ␣-amino group of Ile-153 (12). If the observed effect of the DNS-EGRck treatment is the result of a conformational change of VIIa, similar to that induced by TF, reactivity of the ␣-amino group of Ile-153 with cyanate ion would change after the treatment. The results in Fig. 3 show that the ␣-amino group is indeed protected from carbamylation. Only derivatives of VIIa that gained a high affinity for TF after modification of the active site have a protected ␣-amino group of Ile-153. The data are consistent with the view that TF preferentially binds to an active confor-mational state of VIIa which has a tight salt bridge between Ile-153 and Asp-343.
Trypsin-like serine proteases apparently have active and zymogen-like inactive conformational states, and both states are in equilibrium (28). A salt bridge formed between the ␣-amino group Ile-16 and the ␤-carboxyl group of Asp-194 in chymotrypsin stabilizes the active state; thus, the equilibrium shifts into an active state with conversion of chymotrypsinogen to chymotrypsin (28,29). Considering the analogy of the activation mechanism among trypsin-like serine proteases, we present in Fig. 5 a model for TF-mediated acceleration of VIIa activity. Equations pertaining to the cyclic equilibria of Fig. 5 are given below, and VIIaI, VIIaA, K f , K b , K dI , and K dA are defined in the figure legend.
K dI ϭ ͓VIIaI͔͓TF͔/͓VIIaI ⅐ TF͔ (Eq. 8) Using Equations 6 -10, the observed K d value (K dobs ) for binding of VIIa to TF can be represented as in Equation 11.
When one considers the interaction between VIIa and rsBTF (K d ϭ 14 nM), using the model shown in Fig. 5, one can assume that K dI is close to the K d value for binding of zymogen VII to rsBTF (74 nM). One can also assume that K dA is close to the K d After incubation, each sample was treated with ␣-chymotrypsin to remove the Gla domain, and the NH 2 -terminal Val and Ile of the two-chain Gla domainless derivatives of VIIa were analyzed simultaneously, using a gas phase sequencer as described under "Experimental Procedures." A ratio of PTH-Ile to PTH-Val obtained from analysis of each the derivatives of VIIa before incubation with KNCO is assumed to be 100%, and the ratios obtained from analyses of the carbamylated derivatives of  (Tables II-V) is schematically illustrated. G-site represents Gla domain-dependent interaction site, and A-site represents the active form-dependent interaction site.
FIG. 5. Scheme for the equilibrium between active and inactive conformational states of VIIa in the presence of TF. VIIaI and VIIaA, respectively, represent a zymogen-like inactive state and active state of VIIa. K f and K b represent a constant for the equilibrium between the two conformational states of VIIa in the absence of TF and that in the presence of a saturating concentration of TF, respectively. K dI and K dA , respectively, represent dissociation constants for binding of TF to VIIaI and that of TF to VIIaA. value for the binding of DNS-EGRck VIIa to rsBTF (0.13 nM), since DNS-EGRck VIIa is speculated to be in an active conformational state. If the values of K dA ϭ 0.13 nM, K dI ϭ 74 nM, and K dobs ϭ 14 nM are applied to Equations 10 and 11, the values for K f and K b are calculated to be 0.0076 and 4.3, respectively, indicating that the fraction of an active conformational state is 0.75% in free VIIa and that in VIIa⅐rsBTF complex it is 81%. Therefore, the activity of VIIa can be accelerated 110-fold in the presence of saturating rsBTF. This assumption is in good agreement with reports in which hydrolysis of peptidyl amide or ester substrate catalyzed by VIIa is accelerated 40 -150-fold in the presence of TF (5)(6)(7)22).
DNS-EGRck and APMSF, respectively, contain guanidino and amidino groups, which are designed to interact with Asp-189 (chymotrypsin numbering) in trypsin-like serine proteases. Some specific interaction between the subsite of VIIa and the transition state analogs of its specific substrate may provide the energy required for the conformational change in VIIa. Bode et al. (30) reported that binding of a pancreatic trypsin inhibitor to trypsinogen induces an active conformational state in zymogen and leads to uptake of Ile-Val dipeptide into zymogen. This seems analogous to the conformational change of VIIa induced by transition state analogs, since the inhibitor binding coupled with the salt bridge formation led to induction of an active conformational state in zymogen.
Trypsinogen, but not trypsin, has a flexible segment named the activation domain (31). Therefore, the putative activation domain in VIIa corresponding to that in trypsinogen/trypsin is one candidate for the effecting site of TF. Waxman et al. (32) detected a segmental motion in human VIIa; using a fluorescence anisotropy decay method, they found that the segmental motion is lost after a complex formation with TF. The putative activation domain of VIIa may be responsible for the segmental motion. During preparation of this manuscript, the crystal structure of human soluble TF⅐VIIa complex was reported (33). According to their data, TF does not have direct contacts with a region corresponding to the activation domain found in trypsinogen/trypsin. However, a neighboring loop corresponding to residues Cys-310 to Cys-329 of human VIIa is in contact with TF. This loop is longer than those of related serine proteases and may form an extended flexible activation domain in VIIa. Whether the putative activation domain corresponds to the region of conformational change in VIIa will need to be determined.