Localization of the Thrombin-binding Domain on Prothrombin Fragment 2*

Co-crystallographic studies have shown that the interaction of human prothrombin fragment 2 (F2) with thrombin involves the formation of salt bridges between the kringle inner loop of F2 and anion-binding exosite II of thrombin. When F2 binds to thrombin, it has been shown to evoke conformational changes at the active site and at exosite I of the enzyme. Using plasma, recombinant, and synthetic F2 peptides (F2, rF2, and sF2, respectively) we have further localized the thrombin-binding domain on F2. F2, rF2-(1–116), rF2-(55–116), and sF2-(63–116), all of which contain the kringle inner loop (residues 64–93) and the acidic COOH-terminal connecting peptide (residues 94–116), bind to thrombin-agarose. In contrast, analogues of the kringle inner loop, sF2-(63–90), or the COOH-terminal connecting peptide, sF2-(92–116), do not bind. Thus, contrary to predictions from the crystal structure, the COOH-terminal connecting peptide as well as the kringle inner loop are involved in the interaction of F2 with thrombin. F2 and sF2-(63–116) bind saturably to fluorescently labeled active site-blocked thrombin with K d values of 4.1 and 51.3 μm, respectively. The affinity of sF2-(63–116) for thrombin increases about 5-fold (K d = 10 μm) when Val at position 78 is substituted with Glu. F2 and sF2-(63–116) bind to exosite II on thrombin because both reduce the heparin-catalyzed rate of thrombin inhibition by antithrombin ∼4-fold. In contrast, only F2 slows the uncatalyzed rate of thrombin inactivation by antithrombin. Like F2, sF2-(63–116) induces allosteric changes in the active site and exosite I of thrombin because it alters the rates of thrombin-mediated hydrolysis of chromogenic substrates and displaces fluorescently labeled hirudin54–65 from active site-blocked thrombin, respectively. Both peptides also prolong the thrombin clotting time of fibrinogen in a concentration-dependent fashion, reflecting their effects on the active site and/or exosite I. These studies provide further insight into the regions of F2 that evoke functional changes in thrombin.

Prothrombin, a 581-amino acid plasma glycoprotein, is con-verted to the serine protease thrombin in the final stages of the blood coagulation cascade. The proteolytic conversion of prothrombin to thrombin is catalyzed by prothrombinase, an enzyme complex composed of the serine protease factor Xa, the cofactor Va, phospholipids, and calcium (1,2). During this reaction, prothrombin is cleaved into three fragments: fragment 1 (F1) 1 (consisting of a ␥-carboxyglutamic acid (Gla) domain and the kringle 1 domain), fragment 2 (F2) (consisting of the kringle 2 domain), and the catalytic domain (3). The Gla domain of F1 facilitates calcium-dependent binding of the proenzyme to phospholipid surfaces (4). Although the F2 domain has been shown to interact with factor Va (5), recent studies indicate that F2-factor Va interactions are not required for factor Va to enhance the catalytic efficiency of factor Xa within the prothrombinase complex (6). Rather, the function of F2 in prothrombin may be to alter the conformation of the proenzyme so that its scissile bond(s) is more complementary to the active site of factor Xa (6).
After release from prothrombin, F2 retains its ability to bind thrombin and influences thrombin function. Thus, F2 has been reported to slow the rate of thrombin inactivation by antithrombin (AT) (7), alter the environment of the catalytic site (8), enhance the esterolytic activity of thrombin (9), inhibit the clotting activity of thrombin (10) and modulate the calcium dependence of protein C activation (11). Crystallographic and chemical modification studies have shown that F2 interacts with a highly electropositive region in the COOH terminus of thrombin, also known as anion-binding exosite II or the heparin-binding region of thrombin (12)(13)(14). The crystallographic structure of F2 complexed with active site-blocked thrombin reveals numerous ionic interactions between the anionic inner loop of the F2 kringle and the COOH-terminal helix of thrombin, a region abundant in arginine and lysine residues (12)(13)(14)(15)(16).
In this study, we set out to further characterize the thrombin-binding domain within F2. Using plasma, recombinant, and synthetic F2 peptides (F2, rF2, and sF2, respectively), we report that, contrary to what would have been predicted from the crystallographic data, residues in the COOH-terminal connecting peptide as well as the inner loop of the F2 kringle are necessary for F2 interaction with thrombin. In addition, to identify the regions in F2 that evoke functional changes in thrombin, we also compared sF2-(63-116), 2 a thrombin-bind-ing analogue encompassing the kringle inner loop and the COOH-terminal connecting peptide, with F2 in terms of their ability to modulate thrombin function.
Synthetic Peptides-Cyclized synthetic peptides, purified by preparative reverse phase high performance liquid chromatography, were prepared by Chiron Mimotopes Peptide Systems (San Diego, CA) as described by Fredenburgh et al. (17). The cyclized peptides were resuspended in TBS (20 mM Tris-HCl, pH 7.4, 150 mM NaCl), and titrated to pH 7.0 with NaOH. The peptides yielded the expected composition on quantitative amino acid analysis, and their identities were confirmed by mass spectrometry. sF2-(63-116) corresponds to residues 63-116 of human F2 except that Cys 76 and Cys 93 were replaced with serine residues. A peptide variant of sF2-(63-116) with a scrambled sequence but retaining the original positions of the two Cys residues (Cys 64 and Cys 88 ) also was prepared. sF2-(63-94) and sF2-(63-90), truncated derivatives of sF2-(63-116), also were synthesized.
DNA Construction and Mutagenesis-Plasmid pBS-hFII containing the entire coding region of human prothrombin as a HindIII-SstI fragment in pBluescript (Stratagene Ltd., La Jolla, CA) was generously provided by Dr. Ross MacGillivray (University of British Columbia, Vancouver, BC). Full-length and truncated constructs of rF2 were prepared by PCR amplification (35 PCR cycles for 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C) using plasmid pBS-hFII as the template. The primers used in the PCR reactions contained terminal restriction sites and are shown in Table I. The PCR products were gel-purified (Qiagen Inc., Chatsworth, CA), cleaved with NcoI and XhoI, and cloned as NcoI-XhoI fragments into the bacterial expression vector pET22b(ϩ) (Novagen Inc., Madison, WI). The fidelity of each PCR-amplified construct was confirmed by double-stranded DNA sequencing using dideoxy chain termination and Sequenase 2.0 (U. S. Biochemical Corp., Cleveland, OH).
To ensure proper disulfide bond formation and to prevent the formation of intermolecular complexes, PCR mutagenesis was used to convert Cys 76 and Cys 93 to serine residues. Briefly, plasmid pBS-hFII was PCR amplified using primer B (Table I), which hybridizes downstream of the multiple cloning site of plasmid pBS-hFII, and the mutagenesis primer F (Table I), which contains a point mutation to convert Cys 76 and Cys 93 to serine residues (35 PCR cycles for 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C). The PCR product was gel purified and used as a primer on plasmid pBS-hFII for one PCR cycle (1 min at 94°C, 30 s at 60°C, and 40 s at 72°C). Primers C and D (Table I) were immediately added to the PCR reaction and 35 more cycles of PCR were carried out (1 min at 94°C, 1 min at 55°C, and 1 min at 72°C). Primer C hybridizes upstream of the multiple cloning site of pBS-hFII while primer D hybridizes downstream of primer B. To convert Cys 93 to serine, the mutagenesis procedure was repeated with primer G. The authenticity of the mutations, as well as the fidelity of the PCR-amplified constructs, were confirmed by double-stranded DNA sequencing.
Expression and Purification of Recombinant F2 Peptides-Recombinant F2 peptides were expressed in Escherichia coli BL21(DE3) by isopropyl-1-thio-␤-D-galactoside induction using the histidine-tag pET system (Novagen, Inc., Madison, WI). The expression vector used, pET22b(ϩ), directs the recombinant proteins to the periplasmic space thereby promoting proper folding and disulfide bond formation. The periplasmic fraction was obtained using the cold osmotic shock method described by Ausubel et al. (18). Histidine-binding Ni 2ϩ resin (Novagen, Inc.) was used to purify the recombinant proteins. The isolated proteins at a concentration of 1 mg/ml were subjected to electrophoresis in 15% SDS-polyacrylamide gels under reducing conditions (19), transferred to nitrocellulose, and visualized by Ponceau S staining and immunoblotting as described previously (20).
Purification of F2-F2 was obtained from a terminal prothrombin activation reaction as described by Fredenburgh et al. (17). F2 was dialyzed against TBS and the concentration was determined spectrophotometrically using ⑀ 280 0.1% of 1.1 (21). Thrombin-agarose Affinity Chromatography-Thrombin-agarose affinity chromatography was used to compare the affinities of plasmaand recombinant-F2 peptides for thrombin. 58 M thrombin was incubated for 30 min at 37°C with a 10-fold molar excess of biotin-FPR in TBS. After dialysis against TBS, the biotin-FPR-thrombin was passed over streptavidin-agarose in TBS and excess binding sites were blocked with biotin. 0.2 ml of purified F2 peptide at a concentration of 1 mg/ml was applied at 4°C to 0.5 ml of a streptavidin-agarose column containing 125 nmol of biotin-FPR-thrombin. The column had previously been equilibrated in buffer containing 20 mM Tris-HCl, pH 7.4, 10 mM NaCl. All subsequent steps were performed at 4°C. Elution of the protein was performed with 1 ml of 20 mM Tris-HCl, pH 7.4, containing increasing concentrations of NaCl (from 10 to 500 mM). Aliquots from the flowthrough and eluates were analyzed using SDS-PAGE and immunoblot- ting as described above. The protein elution profiles were obtained by laser densitometry scans of immunoblots using the UltroScanTM XL laser densitometer (Pharmacia LKB Biotechnology). The density of F2 in each fraction was expressed as a percentage of the total F2 density in the complete elution profile.
Reduction and Alkylation of F2 and rF2-(1-116)-The disulfide bonds in F2 and rF2-(1-116) were reduced and alkylated as described (22). The peptides were resuspended in 50 mM dithiothreitol at a concentration of 1 mg/ml and placed at room temperature for 30 min. Iodoacetamide (made up fresh in sterile distilled water) was then added to 50 mM and the reaction mixture was placed at room temperature for 45 min in the dark. The reaction mixture was dialyzed against 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.01% Tween 20 using Spectra/Por cellulose ester dialysis membrane (10-kDa cutoff) and the protein concentration determined as described above for intact F2.
Determination of the Affinity of F2 Peptides for Thrombin-The association between F2 peptides and thrombin was monitored by the F2-dependent fluorescence intensity change of anilinonaphthalene-6sulfonic acid-thrombin (ANS-FPR-thrombin). ANS-FPR-thrombin was prepared using ATA-FPR (N-((acetylthio)acetyl)-D-Phe-Pro-Arg-CH 2 Cl) as described by the supplier (Molecular Innovations Inc., Royal Oak, MI). Briefly, 10.6 M thrombin was incubated for 30 min at room temperature with 2.5-fold molar excess ATA-FPR in HBSE buffer (20 mM HEPES, pH 7.0, 100 mM NaCl, 1 mM EDTA). The reaction mixture was dialyzed in HBSE buffer, incubated with 10-fold molar excess 2-(4-(iodoacetamide)anilino)naphthalene-6-sulfonic acid (Molecular Probes Inc.) in the presence of hydroxylamine for 60 min at room temperature in the dark, and then redialyzed. 500 l of 100 nM ANS-FPR-thrombin was added to the same buffer in a semi-micro quartz cuvette. Using a Perkin-Elmer LS50B luminescence spectrometer with excitation and emission wavelengths set to 328 and 450 nm, respectively, and excitation and emission slit widths set to 12 nm, and an emission filter of 390 nm, readings were taken of ANS-FPR-thrombin alone (Io). Known quantities of F2 peptides (3.8 mM) were then added to the cuvette and, after mixing, the change in fluorescence was monitored (I). As a control, a scrambled variant of the 54-amino acid peptide was titrated to the same concentration, and any change in fluorescence due to nonspecific binding was used to correct the binding curve. The K d values were calculated by plotting I/Io versus F2 peptide concentration. The parameters K d and ␣ were calculated by nonlinear regression ("Tablecurve," Jandel Scientific, San Rafael, CA) using the equation where ␣ is the maximum fluorescence change and assuming a stoichiometry of 1 (17).
Displacement of Fluorescein-hirudin 54 -65 from FPR-thrombin by F2 and sF2-(63-116)-Sulfated hirudin 54 -65 was labeled with FITC as described by Liu et al. (23). 2 ml of TBS containing 10 nM FITChirudin 54 -65 was added to a 1 cm ϫ 1-cm quartz cuvette in a Perkin-Elmer LS50B luminescence spectrometer. Excitation and emission wavelengths were set at 492 and 522 nm, respectively, and excitation and emission slit widths were both set to 8 nm. The fluorescence intensity of 15 nM FPR-thrombin was measured before (Io) and after ( Effects of F2 on the Rate of Thrombin Inhibition by AT-The secondorder rate constant (k 2 ) for the inhibition of thrombin by AT was determined under pseudo first-order conditions in the absence or pres-ence of 1 unit/ml heparin. In a multiwell plate, aliquots (10 l) of thrombin (final concentration of 10 nM) were incubated for varying intervals with 10 l of 500 nM AT in TBSP (TBS containing 0.6% polyethylene glycol 8000) in the absence or presence of F2 or sF2-(63-116) in concentrations ranging from 10 to 500 M. All reactions were terminated by the addition of 200 M chromogenic substrate N-p-tosyl-Gly-Pro-Arg p-nitroanilide in 200 l of TBSP containing 10 mg/ml Polybrene. Thrombin inhibitory activity in the presence of 1 unit/ml heparin was determined using thrombin and AT concentrations of 5 and 50 nM, respectively. Residual thrombin activity was then calculated by measuring absorbance for 5 min at 405 nm using a Molecular Devices plate reader. The pseudo first-order rate constants (k 1 ) for thrombin inhibition were determined by fitting the data to the equation: o is initial thrombin activity and [P] t is thrombin activity at time t (24). The second-order rate constant, k 2 , was determined by dividing k 1 by the AT concentration.

RESULTS
Purification of F2 Peptides-F2 is a 116-amino acid prothrombin activation fragment that consists of a 14-residue interkringle peptide (region A), a 79-residue kringle (regions B, C, D, and E), and a 25-residue acidic COOH-terminal kringlecatalytic domain-connecting peptide (region F) (Fig. 1, panel A). For simplicity, the latter is referred to as the COOH-terminal connecting peptide. We mapped the thrombin-binding domain of F2 using deletion and point mutants prepared by recombinant and synthetic techniques. Schematic diagrams of the F2 peptides are shown in Fig. 1 (panel B) where F2, sF2, and rF2 denote plasma, synthetic, and recombinant F2 peptides, respectively. The deletion mutants encompass various portions of F2 including the kringle inner loop (residues 63-90), 2 the kringle outer loop (residues 1-55), the COOH-terminal connecting peptide (residues 92-116), and the COOH-terminal half of F2 (residues 55-116).
The predicted molecular masses of F2, rF2-(1-116), and the three deletion derivatives rF2-(1-55), rF2-(55-116), and rF2-(1-93) are 14, 15, 6, 7, and 11 kDa, respectively. These predicted values are consistent with the apparent molecular masses of the peptides as determined by SDS-PAGE and Western blot analysis (Fig. 2). No immunoreactive material was found in lysates of E. coli transformed with pET22b(ϩ), the parent vector lacking the F2 cDNA (data not shown). The presence of disulfide bonds in the F2 peptides was verified by comparing the electrophoretic mobilities of reduced versus nonreduced peptides. Disulfide bond-containing peptides migrated more slowly under reducing conditions than under non-reducing conditions (data not shown).
Thrombin-agarose Affinity Chromatography-The affinities of F2 peptides for thrombin were qualitatively assessed by subjecting the peptides to thrombin-agarose affinity chromatography. Biotin-FPR-treated thrombin was coupled to streptavidin-agarose as described under "Experimental Procedures." This directed coupling procedure ensures that all surface basic residues are unmodified and that all the thrombin molecules are coupled in the same fashion. F2 peptides were applied to the column in 10 mM NaCl, 20 mM Tris-HCl, pH 7.4, and the column was then washed with increasing concentrations of NaCl. Aliquots from the eluates were analyzed by SDS-PAGE and immunoblotting. The protein elution profiles were determined by laser densitometry scans of immunoblots. Intact F2 and rF2-(1-116) were retained on the thrombin-agarose column at concentrations up to 300 and 250 mM NaCl, respectively (Fig. 3, panel A). Upon reduction and alkylation of the disulfide bonds, both F2 and rF2-(1-116) eluted from the thrombinagarose column as sharp peaks at 80 mM NaCl (Fig. 3, panel A). Bovine serum albumin, used as a negative control for thrombin-agarose binding, eluted from the column at 60 mM NaCl (data not shown). Like rF2- (1-116), the deletion derivatives rF2-(55-116) and sF2-(63-116) were retained on the thrombin-agarose column at NaCl concentrations up to 250 and 300 mM NaCl, respectively (Fig. 3, panel B). Both of these peptides contain the kringle inner loop (residues 63-93) as well as the COOH-terminal connecting peptide (residues 94 -116). In contrast, deletion derivatives that contain only the kringle inner loop, sF2-(63-90), or the COOH-terminal connecting peptide, sF2-(92-116), did not bind to thrombin-agarose. Furthermore, neither the NH 2 -terminal interkringle peptide (residues 1-14) nor the outer loop of the kringle (residues 15-55) is required for thrombin-agarose binding since rF2-(1-55) and rF2-(1-93) bound weakly to thrombin-agarose (Fig. 3, panel B).
Determination of the Affinities of F2 Peptides for Thrombin-The affinities of F2 peptides for ANS-FPR-thrombin were determined quantitatively by monitoring the change in probe fluorescence during peptide titration as described by Bock (8). After titration, the I/Io values were plotted versus F2 peptide concentration and the data analyzed by nonlinear regression. Because of the low yields of rF2 peptides, K d values were only determined for F2 and sF2 peptides.  (Table II). A mixture of sF2-(63-90) and sF2-(92-116) also does not affect the fluorescence intensity of ANS-FPRthrombin, indicating that the COOH-terminal connecting peptide must be contiguous with the kringle inner loop for effective interaction with thrombin (Table II).
In an attempt to modify the number of salt bridges formed with exosite II of thrombin, five separate amino acid substitutions were introduced into sF2-(63-116), and the affinities of the mutant peptides for ANS-FPR-thrombin were measured (Table II) 4. Binding of F2 and sF2-(63-116) to ANS-thrombin. Increasing amounts of F2 (panel A) or sF2-(63-116) (panel B) were mixed with 100 nM ANS-thrombin and changes in fluorescent intensity were monitored. I/Io is plotted versus peptide concentration, where I is the fluorescent intensity at a given peptide concentration and Io is the initial fluorescence intensity. The K d values were determined by nonlinear regression analyses of the data (line).  (Fig. 5). In the absence of heparin, however, only F2 slows the rate of thrombin inhibition by AT (Fig. 5). Upon reduction and alkylation, neither F2 nor sF2-(63-116) has any inhibitory effect on the rate of thrombin inhibition by AT either in the absence or presence of heparin (data not shown). As shown in Fig. 6, both F2 and sF2-(63-116) increase the chromogenic activity of thrombin with substrates S2238, ChtPA, and S2222, whereas activity was reduced with S2444. With chromogenic substrates S2366 and S2288, opposing effects were induced by the two peptides. Both F2 and sF2-(63-116) also cause a concentration-dependent increase in the thrombin clotting time of fibrinogen (Fig. 7). Upon reduction and alkylation, sF2-(63-116) not only loses its ability to increase the thrombin clotting times, but appears to promote clotting to a small extent (Fig. 7). These data confirm previous reports that ligand binding to exosite II alters the conformation and function of the active site. However, the effect of the peptides on thrombin clotting times may also reflect ligand-induced changes at exosite I since both F2 and sF2-(63-116) displace fluorescein-labeled hirudin 54 -65 from FPR-thrombin. As shown in Table III 6. Influence of F2 and sF2-(63-116) (12,13). These positively charged residues on thrombin form salt bridges with the anionic motif DGDEE (residues 68 -72) in the kringle inner loop of F2. In this study, we experimentally defined the regions of F2 required for thrombin binding. Our data indicate that, in addition to the kringle inner loop (residues 64 -93), the COOH-terminal connecting peptide (residues 94 -116) also plays a role in the interaction of F2 with thrombin. The requirement of the COOHterminal connecting peptide of F2 for F2-thrombin interactions would not have been predicted from the F2-thrombin co-crystallized structure because the F2 interkringle connecting peptides (regions A and F, Fig. 1) were disordered in the crystal (12,13). However, recent molecular modelling suggests that the COOH-terminal connecting peptide is in contact with basic residues within exosite II (25), a concept supported by our data. The interaction of F2 with thrombin is conformation-dependent and ionic in nature because the binding of F2 and rF2-(1-116) to thrombin-agarose is abolished when the peptides are reduced and alkylated and binding of F2 deletion mutants to thrombin-agarose is dependent on the NaCl concentration. In support of the ionic nature of the F2-thrombin interaction, the reported K d values for the binding of F2 to thrombin in low ionic strength buffer (0.05 M Tris-HCl, pH 8.1) and in physiologic ionic strength buffer (0.05 M Tris-HCl, pH 7.5, 150 mM NaCl) are 770 pM (9) (7), fibrinogen clotting (10), and protein C activation (10). Thus, at physiological ionic strength, F2 binds thrombin with a K d of about 5 M. In contrast, F2 has been reported to bind prethrombin 2 with a K d of 33 nM (6). These differences could reflect structural diversity between prethrombin 2 and thrombin.
The K d values were determined by monitoring the changes in ANS-FPR-thrombin fluorescence intensity during peptide titration. Although both F2 and sF2-(63-116) elicit changes in the thrombin active site environment upon binding, the changes in fluorescence signal are diametric (Fig. 4). The diametric changes may reflect differences in the conformation of the COOH-terminal connecting peptide of sF2-(63-116) relative to full-length F2. However, three lines of evidence suggest that the structural integrity of the COOH-terminal connecting peptide is maintained in sF2-(63-116). First, the COOH-terminal connecting peptide alone does not bind to thrombin; only when this peptide is contiguous with the kringle inner loop of F2 does binding occur (Table II). Second, like F2, sF2-(63-116) loses its ability to bind thrombin-agarose upon reduction and alkylation of disulfide bonds (data not shown). Third, rF2-(1-93), which contains all but the COOH-terminal connecting peptide of F2, fails to bind to thrombin-agarose (Fig. 3).
The diametric changes in the thrombin active site environment elicited by the binding of F2 and sF2-(63-116) may instead arise from fewer sites of contact with thrombin for the truncated peptide relative to full-length F2. This is supported by the weaker affinity of sF2-(63-116) for thrombin compared with F2. The lower affinity of sF2-(63-116) may reflect the absence of residues in loop C (Leu 47 , His 50 , and Gln 51 ) that form interdomain contacts of less than 4 Å with thrombin (12,13). However, because rF2-(1-55) and rF2-(1-93), both of which contain loop C, fail to bind to thrombin-agarose, the NH 2terminal half of F2 likely plays only an indirect role in thrombin binding, either by contributing to the stability of F2 or by providing secondary contacts with thrombin. The distinct structural changes that occur in thrombin when it associates with F2 and sF2-(63-116) are similar to unique binding interactions of thrombin with different forms of thrombomodulin. Although thrombin binds to the fifth and sixth epidermal growth factor-like domains of thrombomodulin with high affinity (26), the fourth through sixth epidermal growth factor-like domains are needed for optimal protein C activation by thrombin (27,28) and for induction of structural changes in the active site of thrombin (29).
In an attempt to modify ionic interactions between the kringle inner loop of F2 and exosite II of thrombin, five separate amino acid substitutions were introduced into sF2-(63-116) and the affinities of the resultant peptides for thrombin were determined. Substitution of Val 78 with Glu increased the affinity of sF2-(63-116) for ANS-FPR-thrombin approximately 5-fold (K d decreased from 51.3 to 10.0 M) giving it an affinity similar to that of F2 (K d ϭ 4.1 M), whereas substitution of Tyr 87 with Glu produced only a modest increase in affinity (K d decreased from 51.3 to 40.7 M). Both of these mutations likely increase the number of ionic interactions formed between F2 and thrombin. In contrast, substitution of Lys 81 with Gln decreased the affinity of sF2-(63-116) for thrombin approximately 2-fold and substitution of Lys 81 with Glu abolished the binding of sF2-(63-116) to ANS-thrombin. Since Lys 81 was not observed to make ionic contact with thrombin (12, 13), one possible explanation for the decreases in thrombin affinity is that mutations in the second inner loop of the kringle distort the hairpin ␤-turn which has been shown to pivot at Val 78 and Asp 84 (12,13). The stability of F2 appears to be enhanced by aromatic stacking components since substitution of Phe 85 with Glu decreases the affinity of sF2-(63-116) for thrombin approximately 2-fold (K d increased from 51.3 to 99.8 M).
Functional comparison of F2 with sF2-(63-116), both of which contain the kringle inner loop (loops D and E) as well as the COOH-terminal connecting peptide (loop F), indicates that both peptides influence the chromogenic activity of thrombin (Fig. 6), producing similar effects on four substrates and opposing effects on two others. Although these results illustrate changes in the active site environment of thrombin due to ligand binding at exosite II, they also suggest that there may be additional or altered contacts of F2 compared with sF2-(63-116). Both F2 and sF2-(63-116) displace fluorescein-labeled hirugen from FPR-thrombin, indicating that binding of either ligand to exosite II affects the binding properties of exosite I, a phenomenon previously demonstrated by Fredenburgh et al. (17). F2 and sF2-(63-116) also prolong the thrombin clotting time of fibrinogen in a dose-dependent manner (Fig. 7). That F2 appears to be a more potent inhibitor of thrombin clotting activity likely reflects its higher affinity for thrombin. The effect of the peptides on thrombin clotting times may reflect conformational changes at the active site and/or allosteric changes at exosite I that result in decreased fibrinogen binding.
Walker and Esmon (7) reported that F2 slows the rate of thrombin inactivation by AT because it inhibits AT binding to thrombin. In this study, we demonstrated that, unlike F2, sF2-(63-116) has no inhibitory effect on the rate of thrombin inhibition by AT in the absence of heparin. One possible explanation for these findings is that the NH 2 -terminal connecting peptide or possibly the outer loop of F2, both of which are missing from sF2-(63-116), sterically block access of AT to the active site of thrombin. Alternatively, F2, but not sF2-(63-116), may evoke conformational changes in thrombin that limit its reactivity with AT. This is supported by the distinct changes in fluorescence of ANS-FPR-thrombin induced by F2 and sF2-(63-116) (Fig. 3) and by the work of others who proposed that individual residues within exosite II mediate various functions of the exosite (30). In contrast to their different effects in the absence of heparin, both F2 and sF2-(63-116) reduce the heparin-catalyzed rate of thrombin inactivation by AT about 4-fold likely by competing with heparin for binding to exosite II on thrombin.
In summary, our findings indicate that the functional changes produced by sF2-(63-116) binding to thrombin are quantitatively different from those evoked by F2. Some of these differences may simply reflect the 12-fold higher affinity of F2 for thrombin (K d values of 4.1 and 51.3 M, respectively). However, it is unlikely that thrombin binding affinity alone accounts for (a) the diametric changes in fluorescence signal elicited by the two ligands when they bind to thrombin, or (b) their differential effects on the uncatalyzed rate of thrombin inhibition by AT. Rather, these observations more likely reflect additional contacts of F2 relative to the smaller sF2-(63-116) peptide and suggest that ligand binding to various subsites within exosite II may have different effects on thrombin function.