The Conformation of the Activation Peptide of Protein C Is Influenced by Ca 2 (cid:1) and Na (cid:1) Binding*

Previous studies have suggested that the conformation of the activation peptide of protein C is influenced by the binding of Ca 2 (cid:1) . To provide direct evidence for the linkage between Ca 2 (cid:1) binding and the conformation of the activation peptide, we have constructed a protein C mutant in the (cid:2) -carboxyglutamic acid-domainless form in which the P1 Arg 169 of the activation peptide is replaced with the fluorescence reporter Trp. Upon binding of Ca 2 (cid:1) , the intrinsic fluorescence of the mutant decreases (cid:1) 30%, as opposed to only 5% for the wild-type, indicating that Trp 169 is directly influenced by the divalent cation. The K d of Ca 2 (cid:1) binding for the mutant protein C was impaired (cid:1) 4-fold compared with wild-type. Interestingly, the conformation of the activation peptide was also found to be sensitive to the binding of Na (cid:1) , and the affinity for Na (cid:1) binding increased (cid:1) 5-fold in the presence of Ca 2 (cid:1) . These findings suggest that Ca 2 (cid:1) changes the conformation of the activation peptide of protein C and that protein C is also capable of binding Na (cid:1) , although with a weaker affinity compared with the mature protease. The mutant protein C can no longer be activated by thrombin but remarkably it can be h atroom temperature in TBS containing either 5 m M EDTA or 5 m M Ca 2 (cid:2) . Fluorescence Measurements— Direct binding of Na (cid:2) and Ca 2 (cid:2) to pro- tein C derivatives was evaluated from the changes in the intrinsic protein fluorescence associated with the binding of the cations to the zymogens. Equilibrium dissociation constants were determined by fluorescence titration with a Spex FluoroMax-3 spectrophotometer (Jobin- Yvon, Edison, NJ). The excitation and emission wavelengths were 295 and 345 nm, respectively. The titration experiments with monovalent cations were carried out in 50 m M Tris-HCl (pH 8.0), 1% PEG 8000 at an ionic strength of 800 m M at 25 °C as described (19). Titrations were carried out by adding aliquots of wild-type or R169W GD-PC (1 (cid:3) M ) in solution containing [M (cid:2) ] (cid:4) 800 m M Cl (cid:3) salt to a solution containing the same concentration of each zymogen in 800 m M choline chloride. Monovalent cation binding profiles were obtained in both the absence (1 m M EDTA) and presence of 2.5 m M CaCl 2 . In all titrations, the ionic strength, protein concentration and [Cl (cid:3) ] of the samples were held constant, whereas the monovalent cations [M (cid:2) ] were varied. The Ca 2 (cid:2) binding titrations were carried out in 0.1 M NaCl, 20 m M Tris-HCl, pH 7.5, at 25 °C as described (15). The values of the intrinsic fluorescence ( F ) for the protein C derivatives as a function of each metal ion concen- tration [M] were fit according to the equation (19),

Previous studies have suggested that the conformation of the activation peptide of protein C is influenced by the binding of Ca 2؉ . To provide direct evidence for the linkage between Ca 2؉ binding and the conformation of the activation peptide, we have constructed a protein C mutant in the ␥-carboxyglutamic acid-domainless form in which the P1 Arg 169 of the activation peptide is replaced with the fluorescence reporter Trp. Upon binding of Ca 2؉ , the intrinsic fluorescence of the mutant decreases ϳ30%, as opposed to only 5% for the wild-type, indicating that Trp 169 is directly influenced by the divalent cation. The K d of Ca 2؉ binding for the mutant protein C was impaired ϳ4-fold compared with wild-type. Interestingly, the conformation of the activation peptide was also found to be sensitive to the binding of Na ؉ , and the affinity for Na ؉ binding increased ϳ5-fold in the presence of Ca 2؉ . These findings suggest that Ca 2؉ changes the conformation of the activation peptide of protein C and that protein C is also capable of binding Na ؉ , although with a weaker affinity compared with the mature protease. The mutant protein C can no longer be activated by thrombin but remarkably it can be activated efficiently by chymotrypsin and by the thrombin mutant D189S. Activation of the mutant protein C by chymotrypsin proceeds at a rate comparable to the activation of wild-type protein C by the thrombin-thrombomodulin complex.
Protein C is a multidomain vitamin K-dependent serine protease zymogen in plasma, which, upon activation by the complex of thrombin and thrombomodulin (TM), 1 down-regulates the coagulation cascade by degrading cofactors Va and VIIIa by limited proteolysis (1)(2)(3). Activated protein C has a light and a heavy chain held together by a single disulfide bond, with the N-terminal light chain containing the non-catalytic ␥-carboxyglutamic acid domain and two epidermal growth factor-like domains (4,5). The catalytic domain of activated protein C is located in the C-terminal heavy chain of the molecule (6) and contains functionally critical and conformationally linked binding sites for both Na ϩ and Ca 2ϩ (7,8). The Ca 2ϩ binding site is located in the 70-loop (chymotrypsin numbering (9)) by analogy with trypsin (10) and optimizes the catalytic activity of the protease (11,12). Ca 2ϩ binding to the 70-loop is also required for the rapid activation of protein C by the thrombin-TM complex, whereas such binding inhibits the conversion to activated protein C in the absence of TM (13,14). Based on these and other observations, it has been postulated that binding of Ca 2ϩ to the 70-loop of protein C changes the conformation of the activation peptide of the zymogen in a way that promotes cleavage by thrombin in the presence but not the absence of TM (1,15). However, no direct evidence for the Ca 2ϩ -induced conformational change in the activation peptide of protein C has been presented so far.
In addition to a Ca 2ϩ binding site in the 70-loop, the catalytic domain of activated protein C possesses a functionally important Na ϩ binding site between the 220-and 186-loops as in other clotting proteases carrying Tyr or Phe at position 225 (8,16,17). We have recently demonstrated that the Ca 2ϩ and Na ϩ binding sites of activated protein C are allosterically linked (8). It is not known, however, whether Na ϩ binds to the zymogen protein C and is capable of eliciting effects of functional significance, including a linkage with the Ca 2ϩ binding site in the 70-loop that is so critical for the conversion to activated protein C by the thrombin-TM complex. To address these questions, we have constructed a mutant of protein C with the P1 Arg 169 of the activation peptide replaced by a Trp. The introduction of the fluorophore at position 169 eliminates the site of cleavage by thrombin and provides a direct probe for monitoring the effect of cation binding on the conformation of the activation peptide.

Construction and Expression of Recombinant
Proteins-Construction, expression, and purification of wild-type protein C in the ␥-carboxyglutamic acid-domainless form (GD-protein C) in the RSV-PL4 expression purification vector system has been described previously (18). The P1 Arg 169 to Trp (R169W) mutant of protein C was constructed by PCR mutagenesis methods and expressed in the same vector system in HEK293 cells as described (18). Both wild-type and mutant protein C were purified by immunoaffinity chromatography using the Ca 2ϩ -dependent monoclonal antibody, HPC4, linked to Affi-Gel 10 (Bio-Rad) as described (18). Following purification to homogeneity, both wild-type and mutant protein C were desalted by passing through a PD-10 gel filtration column (Amersham Biosciences) equilibrated with Chelextreated (Bio-Rad) 5 mM Tris-HCl (pH 8.0) buffer containing 0.1% polyethylene glycol (PEG) 8000 and stored at Ϫ80°C. Following activation by chymotrypsin or thrombin D189S, the activated GD-PC mutant exhibited identical amidolytic activity toward Spectrozyme PCa suggesting that a Trp at the C-terminal end of the light chain does not alter the catalytic property of the mutant protein. The expression, purifica-tion, and characterization of the GD-PC mutant in which the Na ϩbinding loop of the zymogen from residues Gly 221 to Tyr 225 was replaced with the corresponding sequence of trypsin (GD-PC trypsin loop) has been described previously (8). The expression, purification, and characterization of the Asp 189 to Ser substitution mutant of thrombin (D189S) has been described previously (19). Recombinant human thrombin (20), thrombomodulin fragment 456 (TM456) (18), and antithrombin (21) were expressed in mammalian cells and purified to homogeneity as described in the cited methods. Bovine ␣-chymotrypsin and hirudin were purchased from Sigma, and Spectrozyme PCa was purchased from American Diagnostica (Greenwich, CT).
Zymogen Activation-The initial rates of concentration dependence of wild-type and R169W GD-PC activation by thrombin or chymotrypsin were monitored in both the absence and presence of Na ϩ and Ca 2ϩ as described (22). The activation reactions were carried out in 20 mM Tris-HCl (pH 7.5), 0.1 M NaCl (TBS) or 0.1 M choline chloride containing 1 mg/ml bovine serum albumin and 0.1% PEG 8000. In both cases, the activation of wild-type GD-PC (0.3-10 M) by thrombin (5-50 nM) and R169W (0.3-10 M) by chymotrypsin (1 nM) was monitored in the absence or presence of 2.5 mM Ca 2ϩ for 5-30 min at room temperature. After inhibition of thrombin activity by antithrombin, the initial rates of activations were measured from the rate of activated protein C generation in an amidolytic activity assay using 400 M Spectrozyme PCa in TBS as described (22). Under the experimental conditions used for the activation of GD-PC R169W, chymotrypsin did not exhibit a measurable amidolytic activity toward Sectrozyme PCa. The rate of hydrolysis was measured at 405 nm at room temperature in a V max kinetic plate reader (Molecular Devices, Menlo Park, CA). The concentration of active protein C in each reaction was determined by reference to a standard curve, which was prepared by total activation of each protein C derivative at the time of experiments. This was accomplished by activation of 1-2.5 M wild-type GD-PC with 10 nM wild-type thrombin in complex with 500 nM TM456 and R169W GD-PC with 100 nM thrombin D189S in complex with 500 nM TM456 in TBS containing 2.5 mM Ca 2ϩ for 2-4 h at 37°C. Under these conditions both zymogens were completely activated to their enzymatic forms. The K m and k cat values were calculated from the Michaelis-Menten equation. The initial rate of concentration dependence of GD-PC R169W activation was also evaluated by the thrombin D189S mutant. In this case, GD-PC R169W (0.8 -25 M) was activated by the thrombin mutant (50 nM) in complex with TM456 (500 nM) in TBS containing 2.5 mM Ca 2ϩ . After 10 -20 min of incubation at room temperature, the activity of thrombin D189S was inhibited by hirudin (200 nM), and the rate of activated protein C generation was measured by an amidolytic activity assay as described above. The TM456 dependence of GD-PC activation suggested that the TM456 concentration was saturating under these conditions. The initial rate of the activation of the protein C mutant by the thrombin mutant was also measured in the absence of TM in both the absence and presence of Ca 2ϩ . In this case, the activation conditions were the same as above except that the concentration dependence of the zymogen activation was monitored for 3 h at room temperature in TBS containing either 5 mM EDTA or 5 mM Ca 2ϩ .
Fluorescence Measurements-Direct binding of Na ϩ and Ca 2ϩ to protein C derivatives was evaluated from the changes in the intrinsic protein fluorescence associated with the binding of the cations to the zymogens. Equilibrium dissociation constants were determined by fluorescence titration with a Spex FluoroMax-3 spectrophotometer (Jobin-Yvon, Edison, NJ). The excitation and emission wavelengths were 295 and 345 nm, respectively. The titration experiments with monovalent cations were carried out in 50 mM Tris-HCl (pH 8.0), 1% PEG 8000 at an ionic strength of 800 mM at 25°C as described (19). Titrations were carried out by adding aliquots of wild-type or R169W GD-PC (1 M) in solution containing [M ϩ ] ϭ 800 mM Cl Ϫ salt to a solution containing the same concentration of each zymogen in 800 mM choline chloride. Monovalent cation binding profiles were obtained in both the absence (1 mM EDTA) and presence of 2.5 mM CaCl 2 . In all titrations, the ionic strength, protein concentration and [Cl Ϫ ] of the samples were held constant, whereas the monovalent cations [M ϩ ] were varied. The Ca 2ϩ binding titrations were carried out in 0.1 M NaCl, 20 mM Tris-HCl, pH 7.5, at 25°C as described (15). The values of the intrinsic fluorescence (F) for the protein C derivatives as a function of each metal ion concentration [M] were fit according to the equation (19), where F o and F 1 are the values of fluorescence in the absence and presence of saturating concentrations of the monovalent or divalent cations and K d,M is the equilibrium dissociation constant for the metal ion binding.

RESULTS
Protein C Activation-GD-PC activation by thrombin in both the absence and presence of TM and Ca 2ϩ has been extensively studied in the past (13,18). It is known that Ca 2ϩ , in the presence of TM, accelerates protein C or GD-PC activation by thrombin but functions as a potent inhibitor of protein C activation in the absence of TM (1). Consistent with these findings, Ca 2ϩ inhibited GD-PC activation by thrombin in the absence of TM in the presence or absence of Na ϩ (Fig. 1). Individual Michaelis-Menten parameters could not be resolved in the presence of Ca 2ϩ up to 10 M GD-PC. However, K m and k cat values of 10 Ϯ 3 M and 3.0 Ϯ 0.4 nM/min/nM for GD-PC activation by thrombin in the absence of Ca 2ϩ were obtained in NaCl-and choline chloride-containing buffers (Table I). GD-PC activation in the presence of a saturating concentration of TM456 showed similar catalytic rates in NaCl-and choline chloride-containing buffers, confirming that the activation of protein C by thrombin does not require Na ϩ (23,24).
The R169W mutant of GD-PC was not activated by thrombin in either the absence or presence of Ca 2ϩ or TM but was activated by chymotrypsin with comparable rates in the absence or presence of Ca 2ϩ and/or Na ϩ (Fig. 2, Table I). Remarkably, the K m and k cat values for the activation of mutant GD-PC by chymotrypsin in the presence of Ca 2ϩ (5.3 M and 13.1 min Ϫ1 , Table I) were the same as those obtained for wild type GD-PC activation by the thrombin-TM complex in the presence of Ca 2ϩ (6.8 M and 13.8 min Ϫ1 (22)). Both the K m and k cat values for the mutant GD-PC activation by chymotrypsin in the absence of Ca 2ϩ and Na ϩ were reduced ϳ3-fold (Table I) cause chymotrypsin does not bind Ca 2ϩ or Na ϩ , the effects on the kinetic parameters must be due to cations binding to the substrate. The effect of Ca 2ϩ on the mutant protein C activation by chymotrypsin was more pronounced in the absence of Na ϩ , suggesting that the binding of Ca 2ϩ and Na ϩ elicits distinct conformational changes around the P1 residue of the activation peptide that are detected by chymotrypsin.
The activation of the mutant GD-PC was also studied in the presence of the thrombin mutant D189S, where the residue in the primary specificity site was converted to that of chymotrypsin. Under saturating conditions of TM456, activation required Ca 2ϩ with a half-maximal value of 90 Ϯ 12 M (Fig. 3A) as opposed to 15 M seen for wild-type thrombin (25). This observation suggests that the R169W mutation has affected the ability of protein C to interact with Ca 2ϩ . The Michaelis-Menten curve for activation of the mutant GD-PC by the thrombin mutant D189S in the presence of TM456 also indicates a dras-tically increased K m relative to wild-type (Fig. 3B). In the absence of TM, the thrombin mutant showed negligible catalytic activity toward the mutant GD-PC. The value of k cat /K m toward the mutant GD-PC was ϳ250-fold slower compared with wild type for the thrombin mutant and also ϳ310-fold slower compared with chymotrypsin. Hence, the thrombin mutant D189S in the presence of TM shows significant activity toward the mutant GD-PC, as opposed to the complete absence of activity of the wild-type, but it is not as active as chymotrypsin. This observation strongly suggests that thrombin, unlike chymotrypsin, has structural determinants that undermine its ability to interact efficiently with protein C and that part of this defect is the inability of protein C to penetrate the primary specificity pocket of thrombin. The defect is corrected by inter-  Fig. 1, A and B. b ND, not determined; the activation reaction remained linear for up to 10 M GD-PC.

FIG. 2. Concentration dependence of GD-PC R169W activation by chymotrypsin in the absence and presence of metal ions.
Upper panel, the initial rate of activation of increasing concentrations of the GD-PC mutant by chymotrypsin (1 nM) was measured at room temperature in 20 mM Tris-HCl buffer containing 200 mM NaCl, 0.1% PEG 8000, and 1 mg/ml bovine serum albumin in the presence (q) or absence of 1 mM Ca 2ϩ (E). Following 5 min activation, the rate of activated GD-PC generation was measured as described under "Experimental Procedures." Lower panel, the same as above, except that instead of NaCl, the activation buffer contained 200 mM choline chloride. Solid lines in both panels are best fit of data to the Michaelis-Menten equation. APC, activated protein C.

FIG. 3. Ca 2؉ concentration and substrate concentration dependence of GD-PC R169W activation by thrombin D189S.
A, the initial rate of activation of the GD-PC mutant (2 M) by the thrombin mutant (50 nM) in complex with a saturating concentration of TM456 (500 nM) was monitored in TBS containing 0.1% PEG 8000, 1 mg/ml bovine serum albumin, and increasing concentrations of Ca 2ϩ . Following an 80-min incubation at room temperature, the activity of the thrombin mutant inhibited by hirudin and the rate of activation were measured by an amidolytic activity assay described under "Experimental Procedures." B, the initial rate of activation of increasing concentrations of the GD-PC mutant by thrombin D189S (50 nM) in complex with TM456 (500 nM) was measured in the same buffer. Following a 10 -20 min activation at room temperature, the rate of activation was measured as described above. The solid lines in A and B are best fit of data to non-linear and linear equations, respectively. APC, activated protein C. action with TM through a mechanism that remains to be identified.
Fluorescence Measurements-Binding of Ca 2ϩ to GD-PC caused an ϳ5% decrease in intrinsic fluorescence, as previously reported (15), with a K d of 26.6 M (Fig. 4). On the other hand, binding of Ca 2ϩ to the mutant GD-PC decreased the intrinsic fluorescence ϳ30%, with a K d of 111 M. The more pronounced change in intrinsic fluorescence can be assigned to the introduction of Trp 169 in the mutant protein C. An important conclusion from this observation is that Ca 2ϩ binding to GD-PC induces a conformational change in the activation peptide. The weaker Ca 2ϩ affinity of the mutant indicates that the P1 residue of the activation peptide is linked energetically to the Ca 2ϩ binding site in the zymogen. These results are in agreement with previous mutagenesis data where substitution of the P3, P2, or P3Ј residues of the activation peptide also impaired the Ca 2ϩ affinity of the mutant proteins (18,26).
Interestingly, the conformation of the activation peptide was also influenced by Na ϩ binding. Na ϩ enhanced the intrinsic fluorescence of the mutant GD-PC ϳ9% in the absence of Ca 2ϩ , with a K d of ϳ285 mM (Fig. 5A). The Na ϩ effect was abrogated by replacement of the 220-loop of protein C with that of trypsin, as seen for activated protein C (8). K ϩ binding showed an even higher affinity (K d ϭ 104 mM) and a similar change in intrinsic fluorescence (Fig. 5A). On the other hand, Li ϩ did not elicit any fluorescence change. The presence of Ca 2ϩ increased the affinity for Na ϩ ϳ5-fold and for K ϩ ϳ2-fold, and both monovalent cations caused a decrease in intrinsic fluorescence (Fig. 5B, Table II). These results suggest that the Ca 2ϩ and Na ϩ binding sites are energetically linked not only in activated protein C but also in the zymogen.
Similar to Ca 2ϩ , the effect of the monovalent cations can be directly attributed to changes in the environment/conformation of the P1 Trp 169 in the activation peptide of the mutant protein because the metal ions exhibited opposite effects with the wildtype GD-PC, thus decreasing the emission intensity in the absence of Ca 2ϩ 10 -35% depending on the metal ion (Fig. 6A).
In the presence of Ca 2ϩ , the decrease in protein fluorescence of GD-PC varied from ϳ5 to 25% for the monovalent cations (Fig.  6B). The K d values for interaction with the metal ions under different conditions are presented in Table II. DISCUSSION The results presented here shed light on how Ca 2ϩ enhances activation of protein C by thrombin in the presence of TM. Ca 2ϩ influences the conformation of the activation peptide of protein C and perhaps makes it more complementary to the active site of thrombin. There are several lines of evidence in support of this hypothesis: 1) the molecular modeling data based on the crystal structures of the protease domain of activated protein C and chymotrypsinogen have predicted electrostatic interactions between the charged residues of the activation peptide and those of the 70-loop (27,28), 2) the mutagenesis of residues surrounding the scissile bond is associated with changes in the affinity of mutant zymogens for binding Ca 2ϩ (18,26), and 3) it is known that the GD-PC zymogen but not activated GD-PC undergoes a 5 Ϯ 1% quenching in the intrinsic protein fluorescence upon binding Ca 2ϩ (15). These results together with kinetic data have supported the hypothesis that the conformations of the activation peptide and the Ca 2ϩ -binding 70-loop of protein C are allosterically linked. Nevertheless, direct evidence for this hypothesis was missing until now because the Trp residues responsible for reporting the intrinsic fluorescence of GD-PC upon binding Ca 2ϩ have been mapped to the two Trp residues 231 and 234 (76 and 79 in chymotrypsin numbering), which are both located in the Ca 2ϩ loop of the zymogen (15). Thus, the 5% fluorescence quenching of GD-PC upon interaction with Ca 2ϩ could not have been directly attributed to changes in the conformation of the activation peptide of the zymogen. The observation of this study that the binding of Ca 2ϩ is associated with a markedly higher quenching of the intrinsic fluorescence (ϳ30%) of the P1 R169W mutant directly establishes for the first time that the conformation of the activation peptide of protein C is altered upon interaction with Ca 2ϩ . The observation that the affinity of the R169W mutant for binding Ca 2ϩ (111 M) is impaired ϳ4-fold shows that these two sites are energetically linked.
This study also demonstrates that the Na ϩ binding site of activated protein C is partially folded in the zymogen form and retains the cation selectivity (K ϩ Ͼ Na ϩ Ͼ Li ϩ ) of the mature protease (29). Similar to Ca 2ϩ , binding of Na ϩ to the 220-loop of protein C elicits changes in the activation peptide. Furthermore, binding of Na ϩ to protein C, though weaker than to activated protein C, is positively linked to the Ca 2ϩ site as in the mature protease. It is well established that the interaction of Na ϩ with thrombin promotes the procoagulant and signaling functions of the enzyme, but has no effect on the activation of protein C (24). Consistent with this key property of thrombin, activation of GD-PC in the absence or presence of saturating TM456 occurred with similar catalytic efficiency in either Na ϩor choline-containing buffer, suggesting that the Na ϩ -induced conformational change of the activation peptide may be functionally irrelevant to protein C activation. The role of Na ϩ binding to protein C seems to be related to the positive linkage with Ca 2ϩ binding to the 70-loop of the zymogen, which in turn ensures efficient activation of protein C by the thrombin-TM complex. Hence, although both Ca 2ϩ and Na ϩ influence the conformation of the activation peptide of protein C, only Ca 2ϩ produces a better substrate for cleavage by the thrombin-TM complex. Therefore the two cations produce distinct functional changes in the conformation of the activation peptide, although their binding interactions are energetically linked.
A noteworthy observation of this study was that chymotrypsin activates the P1 Trp 169 mutant of GD-PC with essentially the same efficiency as the thrombin-TM456 complex activates wild-type GD-PC. It is unclear whether the chymotrypsin specificity toward the mutant GD-PC arises exclusively from the P1 site of cleavage. Future mutagenesis studies of chymotrypsin should clarify whether the protease exploits additional interacting sites in its interaction with the mutant GD-PC. In the case of thrombin, a conversion of the primary specificity site to that of chymotrypsin with the D189S mutation produces a derivative capable of activating the mutant GD-PC but with a specificity Ͼ300-fold lower compared than that of chymotrypsin. In addition, efficient cleavage of the mutant GD-PC still requires saturating amounts of TM and Ca 2ϩ . These observations have the obvious implication that the conformation of protein C is potentially suitable for efficient activation and that no action is likely required of TM on protein C to optimize interaction with the thrombin active site. It is still possible, however, that TM provides facilitated diffusion of protein C into the thrombin active site (30), which is hindered by structural constraints not present in chymotrypsin. Possible structural determinants preventing thrombin from cleaving protein C efficiently in the absence of TM are the extended binding pocket residues like Glu 192 , Arg 35 , and other residues that are not conserved in chymotrypsin. In agreement with this hypothesis, mutagenesis of either Glu 192 (31) or Arg 35 (22) has dramatically improved the rate of protein C activation by mutant thrombins independent of TM. It would be of interest to assess whether introduction of these residues into chymotrypsin produces a derivative that mimics thrombin in its interaction with the mutant GD-PC.