Ca2+ binding to the first epidermal growth factor-like domain of human blood coagulation factor IX promotes enzyme activity and factor VIII light chain binding.

Ca2+ binding to the first epidermal growth factor (EGF)-like domain of factor IX is known to be required for biological activity, but the mechanism by which Ca2+ contributes to factor IX function has remained unclear. We have studied recombinant factor IX mutants which lack Ca2+ binding to the first EGF-like domain, due to a replacement of Asp64 by Glu, Lys, or Val. The purified mutants (factors IX D64E, D64K, and D64V), were compared to plasma-derived and recombinant wild-type factor IX with regard to a number of metal-ion dependent functional parameters. In the presence of Mg2+, the activated mutants were indistinguishable from normal factor IXa in hydrolyzing the synthetic substrate CH3-SO2-Leu-Gly-Arg-p-nitroanilide. Replacing Mg2+ by Ca2+ further stimulated the activity of normal factor IXa but not of mutant factor IXa. In factor VIII-independent factor X activation, factor IXa D64K and D64E displayed reduced catalytic activity compared to normal factor IXa (apparent kcat/Km ≈ 1, 2, and 4 × 103 M−1 s−1, respectively). In the presence of factor VIIIa, factor X activation rates by normal and mutant factor IXa were stimulated by factor VIIIa to a different extent (≈700- and 200-fold, respectively), indicating that Asp64 replacements affect the interaction with factor VIIIa. This possibility was addressed in inhibition studies employing synthetic peptides comprising the factor IXa-binding motifs of factor VIII heavy or light chains. Whereas the heavy chain peptide (Ser558-Gln565) inhibited factor VIII-dependent factor X activation by normal and mutant factor IXa with similar efficiency, the light chain peptide (Lys1804-Lys1818) inhibited normal factor IXa 2-3-fold more efficiently than did mutant factor IXa. This indicates that the reduced response to factor VIIIa may be due to impaired binding of mutant factor IXa to the factor VIII light chain. This was further explored in direct binding studies. In the presence of Mg2+, normal and mutant factor IXa were similar in binding to the factor VIII light chain. However, in the presence of Ca2+, factor IXa mutants were less efficient than normal factor IXa, which was illustrated by a 4-5-fold lower affinity than normal factor IXa for factor VIII light chain. Collectively, our data demonstrate that a number of factor IXa functions, including enzymatic activity and assembly into the factor IXa-factor VIIIa complex, are dependent on Ca2+ binding to the first EGF-like domain of factor IX.

Factor IX (FIX) 1 is a vitamin K-dependent serine protease precursor which participates in the blood coagulation process (1). The physiological importance of FIX is apparent from the notion that its deficiency or dysfunction is associated with the severe bleeding disorder hemophilia B (2). In plasma, FIX circulates as a single chain polypeptide (3). Activation of the zymogen is achieved by limited proteolysis of the Arg 145 -Ala 146 and Arg 180 -Val 181 bonds by factor XIa (FXIa) or factor VIIa (FVIIa) resulting in the enzyme factor IXa␤ (FIXa) (4,5). FIXa activates factor X (FX) in a complex which requires the presence of phospholipids, Ca 2ϩ ions, and a nonenzymatic cofactor, activated factor VIII (FVIIIa) (6). The presence of FVIIIa in this complex markedly enhances proteolytic activity of the enzyme FIXa (7,8). FIXa and FVIIIa assemble into the FX-activating complex through a number of interactive sites (9,10). Binding sites for FIXa have been reported to be located in the FVIII heavy chain region Ser 558 -Gln 565 (9) and in the FVIII light chain region Glu 1811 -Lys 1818 (11). Binding of FIX to the latter site requires the presence of Ca 2ϩ ions (10) and cleavage of the FIX Arg 145 -Ala 146 bond (12).
FIXa consists of a light and heavy chain (M r 18,000 and 28,000, respectively) which are covalently linked. Furthermore, FIXa comprises a number of discrete domains arranged in a structure which is shared with a number of other vitamin K-dependent serine proteases (13). FIXa heavy chain comprises the trypsin-like protease domain with the catalytic center (14). The light chain consists of an amino-terminal domain containing several ␥-carboxylated glutamic acid residues (15), a short hydrophobic stack (16), and two domains that share considerable homology with the epidermal growth factor (EGF) (17). Like other vitamin K-dependent serine proteases, FIX comprises a number of Ca 2ϩ binding sites, with affinities in the micromolar and millimolar range (18,19). One high affinity binding site is located in the FIXa protease domain (20), while another is located in the first EGF-like domain (21). Recent analysis of the crystal structure of the Ca 2ϩ -containing EGF-like domain revealed that residues Asp 47 , Gln 50 , and Asp 64 are crucial for Ca 2ϩ binding (22). Recombinant as well as natural variants of FIX comprising mutations at these positions have been reported to display impaired FIX activity (23,24). For instance, the FIX mutant with the Asp 64 residue replaced by a Lys residue (FIX D64K) lacks Ca 2ϩ binding to its first EGF-like domain (25) and displays less than 5% biological activity compared to normal FIX (23). This demonstrates that Ca 2ϩ binding to the first EGF-like domain indeed is essential for appropriate FIX function. However, it remains unclear by which mechanism Ca 2ϩ binding to the first EGF-like domain contributes to FIX activity.
In the present study, the effect of Ca 2ϩ binding to the first EGF-like domain was investigated employing recombinant FIX mutants, in which the Asp 64 had been substituted by Glu, Lys, or Val. These mutants were compared to normal FIX with regard to a number of functional parameters. These included reactivity toward synthetic and natural substrates and the interaction with the cofactor FVIII. This approach revealed that binding of Ca 2ϩ to the first EGF-like domain contributes to enzyme activity and assembly of the FIXa⅐FVIIIa complex.
Proteins-The monoclonal anti-FIX antibodies CLB-FIX 2 and CLB-FIX 14 have been described previously (10,12). Polyclonal antibodies against FIX were obtained as described previously (10). Antibodies were purified employing protein A-Sepharose as recommended by the manufacturer. Antibodies were conjugated with horseradish peroxidase as described elsewhere (26). FVIII and FVIII light chain were purified as outlined previously (10). FX was purified as described elsewhere (27). FXIa was obtained from Enzyme Research Laboratories. Purified antithrombin III and human serum albumin (HSA) were from the Division of Products of our institute.
Synthetic Peptides-Peptide Lys 1804 -Lys 1818 represents the FIXa binding site on the corresponding FVIII light chain region (i.e. KNFVK-PNETKTYFWK) and was prepared as described previously (11). Peptide Ser 558 -Gln 565 comprises the FIXa binding site on the corresponding FVIII heavy chain region (i.e. SVDQRGNQ) (9). This peptide was synthesized by Eurosequence B.V. (Groningen, the Netherlands). Both peptides were more than 90% pure as determined by high performance liquid chromatography analysis, and their identity was confirmed by mass spectrometry.
Recombinant FIX-Stable cell lines expressing wt-FIX or the substitution mutants D64E, D64K, or D64V were obtained by transfecting Madin-Darby canine kidney cells with previously described plasmids (23,28). Cell lines producing appropriate levels of FIX antigen (120 -300 ng/(10 6 cells⅐24 h)) were selected for large scale production. Each cell line was analyzed to verify that it contained the appropriate substitution in the cDNA encoding the FIX molecule. Cells were maintained in 1-liter cell factories in Dulbecco's modified Eagle's medium supplemented with 2.5% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, 0.4% (v/v) Fungizone, 5 g/ml vitamin K 1 . Culture medium was harvested every 48 h. The medium was subsequently filtered to remove cell debris and concentrated approximately 10-fold employing a hollow fiber cartridge (Hemoflow 5; Fresenius, Bad Homburg, Germany). Benzamidine was added to a final concentration of 10 mM, and the concentrates were stored at Ϫ20°C.
Purification of FIX-In order to purify recombinant FIX, the concentrated medium was subjected to immunoaffinity chromatography employing the anti-FIX antibody CLB-FIX 14 (5 mg/ml CNBr-Sepharose). After two extensive washing steps with 0.01 M benzamidine, 5% (v/v) glycerol, 0.05 M HEPES (pH 7.4) supplemented with 0.1 M NaCl and 1 M NaCl, respectively, FIX was eluted with 3 M KSCN in the same buffer.
FIX-containing fractions were pooled and stored at Ϫ20°C in 5% (v/v) glycerol, 0.1 M NaCl, 50 mM HEPES (pH 7.4). The specific antigen of the purified recombinant proteins varied between 160 and 200 units/mg. Like plasma-derived FIX (pd-FIX), recombinant wt-FIX displayed full biological activity, whereas the mutants displayed activities of 5% or less compared to wt-FIX and pd-FIX (not shown). Analysis by SDSpolyacrylamide electrophoresis and silver staining revealed one single band for recombinant wt-FIX or mutants thereof, which migrated with a similar M r as pd-FIX (not shown). This demonstrates that no propeptide-containing, unprocessed FIX was present. Furthermore, in preliminary experiments recombinant FIX proteins proved indistinguishable from pd-FIX with regard to binding to barium citrate and to the Ca 2ϩdependent anti-FIX antibody CLB-FIX 2 (not shown). These data are in agreement with previous reports that established normal ␥-carboxylation in the FIX expression system employed (23,29,30). pd-FIX was purified as described elsewhere (12). Recombinant and pd-FIX were converted into FIXa employing FXIa as described (12). FIXa and FXIa were separated by anion exchange chromatography as outlined previously (31). Active site titrations employing the active site titrants Antithrombin III (12) or NPGB (32) revealed the presence of 0.8 -0.9 mol of active sites per mol of pd-FIXa, wt-FIXa, FIXa D64K, FIXa D64V, or FIXa D64E.
Protein Concentrations-Protein was quantified by the method of Bradford (33), using HSA as a standard. FVIII activity was assayed employing bovine coagulation factors and a synthetic substrate for FXa in a spectrophotometric assay (Coatest FVIII, Chromogenix AB, Mölndal, Sweden). The amount of FVIII present in 1 ml of human plasma (1 unit/ml) was assumed to correspond to 0.35 nM. FIX biological activity was assessed employing a one-stage clotting assay as described previously (34). FIX antigen was quantified employing a previously described method (12).

Hydrolysis of CH 3 SO 2 -
LGR-pNA-Cleavage of CH 3 SO 2 -LGR-pNA was assayed as described previously (12). Briefly, 50 l of a 6 mM solution of CH 3 SO 2 -LGR-pNA was added to a 50-l sample in a microtiter plate (Costar, flat bottom type). Initial rates of substrate hydrolysis were measured by monitoring absorbance at 405 nm in time. Absorbance values were converted into molar concentrations using a molar extinction coefficient of 9.65 ϫ 10 3 M Ϫ1 cm Ϫ1 for p-nitroanilide (pNA) and a path length of 0.35 cm for a 100-l volume.
Binding Assays-Binding of normal or mutant FIXa to the immobilized FVIII light chain and calculation of binding parameters were performed as described elsewhere (10).

RESULTS
Amidolytic Activity-The role of Ca 2ϩ binding to the first EGF-like domain of FIX was investigated using recombinant FIX mutants with various replacements of Asp 64 . For comparison, pd-FIX and recombinant wt-FIX were employed. First, the mutant FIXa D64K was tested for its reactivity toward the synthetic substrate CH 3 -SO 2 -LGR-pNA in the presence of various concentrations of divalent cations (Fig. 1). While some amidolytic activity was observed in the absence of metal ions, the presence of divalent cations markedly augmented the rate of substrate hydrolysis. Experiments employing pd-FIXa demonstrated that Ca 2ϩ and Mg 2ϩ stimulate amidolytic activity by increasing k cat , whereas K m remains unchanged (not shown). The Ca 2ϩ dependence of wt-FIXa was indistinguishable from that of pd-FIXa (Fig. 1A), which suggests that these proteins are similar in Ca 2ϩ binding to sites involved in substrate hydrolysis. In the presence of Ca 2ϩ , the mutant FIXa D64K displayed lower activity than wt-FIXa and pd-FIXa (Fig. 1A). When Ca 2ϩ was replaced by Mg 2ϩ the activity of mutant FIXa D64K remained unchanged. However, pd-FIXa and wt-FIXa exhibited reduced activity, which now was indistinguishable from FIXa D64K (Fig. 1B). This demonstrates that replacement of Asp 64 by Lys results in reduced amidolytic activity. The same experiments were performed using mutants with other substitutions at position 64: FIXa D64E and FIXa D64V. For both of these mutants, data were obtained (not shown) that are identical to those for FIXa D64K (Fig. 1). Collectively, these data demonstrate that Asp 64 is associated with a Ca 2ϩ -specific contribution to amidolytic activity.
FX Activation in the Absence of FVIIIa-Because in the presence of Ca 2ϩ the reactivity toward the synthetic substrate CH 3 -SO 2 -LGR-pNA of FIXa mutants was dissimilar to that of normal FIXa, also the reactivity toward the physiological substrate FX was addressed. Therefore, the rate of FXa formation by normal and mutant FIXa was assessed in the presence of phospholipids, Ca 2ϩ ions, and various concentrations of FX. FXa formation by wt-FIXa was indistinguishable from that by pd-FIXa, with catalytic efficiencies (apparent k cat /K m ) of 4.0 and 4.1 ϫ 10 3 M Ϫ1 s Ϫ1 (Fig. 2). In contrast, the catalytic efficiency of FIXa D64K was 5-fold lower (Fig. 2). Similar experiments using FIXa D64E (data not shown) revealed a catalytic efficiency of 2.0 ϫ 10 3 M Ϫ1 s Ϫ1 . FIXa D64V showed a further reduced FX activation, the kinetics of which were not studied in detail. These data demonstrate that Asp 64 is involved in modulating proteolytic activity of FIXa toward FX.
FX Activation in the Presence of FVIIIa-The effect of the Asp 64 substitutions was further explored in FX activation studies in the presence of various concentrations of the cofactor FVIIIa. As expected, FXa generation by wt-FIXa and pd-FIXa was enhanced in the presence of FVIIIa to the same extent and in a saturable and dose-dependent manner (Fig. 3). In contrast, FX activation by FIXa D64K was considerably less enhanced by FVIIIa (Fig. 3). As is listed in Table I, FVIIIa stimulated FX activation by normal FIXa approximately 700-fold, whereas FIXa D64K was stimulated approximately 200-fold. The mutant FIXa D64E displayed the same suboptimal stimulation by FVIIIa (Table I). Apparently, the Asp 64 substitutions result in FIXa molecules that display a reduced response to FVIIIa. This raises the possibility that these FIXa mutants are less efficient than normal FIXa in their interaction with FVIIIa.
Effect of Synthetic FVIII Peptides on FX Activation-Because both FVIII heavy and light chain regions are involved in complex assembly with FIXa, we addressed the possibility that binding of FIX to one or both of these regions is affected by the Asp 64 substitutions. For this purpose, FX activation studies were performed in the presence of the synthetic peptides Lys 1804 -Lys 1818 and Ser 558 -Gln 565 . These peptides encompass the primary structure of FIXa interactive sites in the light and heavy chain of FVIII, respectively, and interfere with FX activation in a noncompetitive manner (9, 11). Indeed, FXa formation catalyzed by wt-FIXa was inhibited by both peptides (Fig.   FIG. 2. Effect of  4A). Peptide Ser 558 -Gln 565 inhibited FXa formation by FIXa D64K or wt-FIXa to the same extent (Fig. 4, A and B). This shows that FIXa D64K and wt-FIXa are similar in their interaction with peptide Ser 558 -Gln 565 , suggesting that the interaction between FIXa D64K and this FVIII heavy chain region is unaffected by the Asp 64 3 Lys mutation. In contrast, FIXa D64K was inhibited by the FVIII light chain peptide Lys 1804 -Lys 1818 2-3-fold less efficiently than wt-FIXa (Fig. 4, A and B), suggesting that the Asp 64 mutation is associated with abnormal binding to the FVIII light chain.
Interaction with FVIII Light Chain-To investigate the effect of the Asp 64 mutation on the interaction with the FVIII light chain, equilibrium binding studies were performed employing the immobilized FVIII light chain. The FVIII light chain bound wt-FIXa and pd-FIXa to the same extent, which was illustrated by similar binding parameters for wt-FIXa and pd-FIXa (Fig. 5) (10). The mutants FIXa D64K and FIXa D64E were similar to normal FIXa in their interaction with the FVIII light chain with regard to stoichiometry, whereas affinity was 4 -5-fold lower (Fig. 5). Apparently, both Asp 64 mutants combine a reduced stimulation of FX activation by FVIIIa with a reduced affinity for the FVIII light chain.

Effect of Divalent Cations on the FIXa-FVIII Light Chain
Interaction-As Asp 64 is part of the high affinity Ca 2ϩ binding site in the first EGF-like domain (22,23,25), it was of interest to address the interaction with the FVIII light chain in the presence of various concentrations of divalent cations. While some binding of FIXa to the FVIII light chain was observed in the absence of metal ions, the presence of metal ions markedly enhanced binding (Fig. 6). The Ca 2ϩ dependence of wt-FIXa was indistinguishable from that of pd-FIXa (Fig. 6A), which suggests that these proteins are similar in Ca 2ϩ binding to sites that contribute to FVIII light chain binding. In the presence of Ca 2ϩ , FIXa D64K was less efficient than normal FIXa in binding the FVIII light chain (Fig. 6A). When Ca 2ϩ was replaced by Mg 2ϩ , FVIII light chain binding by FIXa D64K remained unchanged (Fig. 6B). However, pd-FIXa and wt-FIXa exhibited reduced binding, which now was indistinguishable from FIXa D64K. These data suggest that Asp 64 is involved in a Ca 2ϩ -specific contribution to assembly of the FIXa⅐FVIII light chain complex.

DISCUSSION
Blood coagulation factor IX is among the many extracellular proteins that contain EGF-like domains (17). A subset of these domains is characterized by the ability to bind Ca 2ϩ with high affinity (23,25). Because of the widespread distribution of Ca 2ϩ -binding EGF-like domains, it has been assumed that these domains are associated with a general biological role, such as ligand-cell receptor or protein-protein interactions (36). For some proteins of the hemostatic system, these include interactions between enzyme and cofactor. For instance, Ca 2ϩbinding EGF-like domains of thrombomodulin and FVIIa have been proposed to be involved in binding of thrombin and tissue factor, respectively (37,38). In FIX two consecutive EGF-like domains are present, the first of which is capable of high affinity Ca 2ϩ binding (17,21). The role of the Ca 2ϩ -binding EGF-like domain of FIX seems to be more complicated, because this domain has been associated with a variety of FIX specific functions (17,23,39,40). The physiological significance of the Ca 2ϩ -binding EGF-like domain is apparent from the notion that mutations in this domain are associated with low biological activity and the bleeding disorder hemophilia B (24,41).
We investigated the contribution of Ca 2ϩ binding to the first EGF-like domain of FIX with regard to a number of FIXspecific functions. In this respect, it was of interest to address in particular those properties which have been established as

TABLE I Proteolytic activity and FVIII light chain binding by normal and mutant FIXa
Activation of FX (0.2 M) in the absence or presence of FVIIIa (0.35 nM) was performed as described under "Experimental Procedures." All FX activation rates are compared to FX activation by pd-FIXa in the absence of FVIIIa (i.e. 3.2 ϫ 10 Ϫ2 mol FXa/(mol FIXa⅐min)), which is referred to as 1. Stimulation represents the ratio of FX activation rates in the presence and absence of FVIIIa. The interaction with FVIII light chain was assessed as described elsewhere (10). Data represent the mean of at least three experiments. All data refer to experiments in the presence of Ca 2ϩ ions. The possibility was considered that replacement of Ca 2ϩ by Mg 2ϩ would abolish the differences between normal and mutant FIXa with respect to FVIII stimulation. However, the relative rate of FX activation by FIXa in the presence of Mg 2ϩ appeared to be extremely inefficient (Ͻ0.01). No stimulation by FVIII could be observed under these conditions. being Ca 2ϩ -dependent. These include reactivity toward synthetic substrates (32) and binding to FVIII light chain (10). To distinguish between Ca 2ϩ binding to the first EGF-like domain and to other Ca 2ϩ binding sites within the FIXa molecule (18,19), we employed mutants with substitutions of Asp 64 which are known to abolish Ca 2ϩ binding to the first EGF-like domain (25). We found the Asp 64 mutants to display amidolytic activity and FVIII light chain binding to the same extent as normal FIXa in the presence of Mg 2ϩ (Figs. 1 and 6). This is in agreement with the notion that Mg 2ϩ may replace for Ca 2ϩ in the majority of the metal-ion binding sites but not that of the first EGF-like domain (18,42). Our observation that in the presence of Ca 2ϩ the Asp 64 mutation results in reduced substrate hydrolysis ( Fig. 1) (12). However, whereas FIXa␣ displays 50% activity compared to fully activated FIXa (4,12), the mutants display much lower activity (Fig. 3, Table  I). Apparently, defective FX activation by FIXa D64K and FIXa D64E is only partially explained by suboptimal FVIII light chain binding. We considered the possibility that FIXa D64K has a defect in binding the FVIII heavy chain which might be more pronounced than its defect in FVIII light chain binding (Fig. 5). This was investigated employing synthetic peptides that comprise the FIXa-binding sequences Ser 558 -Gln 565 and Gln 1811 -Lys 1818 of the heavy and light chain of FVIII, respectively (9,11). In agreement with previous observations (9,11), both peptides interfere with FVIII-dependent FX activation by wt-FIXa (Fig. 4A). Peptide Lys 1804 -Lys 1818 also inhibits FX activation by FIXa D64K, albeit with a 2-3-fold lower efficiency (Fig. 4B). This is in support of our conclusion that FIXa D64K displays impaired interaction with FVIII light chain (Fig. 5). In contrast, peptide Ser 558 -Gln 565 interferes with FX activation by FIXa D64K or normal FIXa to the same extent (Fig. 4). Although we cannot exclude the possibility that the mutant FIXa D64K displays impaired binding to another, so far unidentified portion of the FVIII heavy chain, our observations indicate that FIXa D64K and wt-FIXa are similar in their interaction with the FVIII heavy chain region Ser 558 -Gln 565 . This suggests that other functional abnormalities also contribute to the reduced FX activation. In this regard it is important to note that Asp 64 mutants which have identical amidolytic activity display significantly different proteolytic activity (Table I). Such differences in enzymatic activity are in line with the diverge levels of biological activity which have been reported for hemophilia B patients with various substitutions at position 64 (41).
Collectively, our data demonstrate that the Ca 2ϩ -binding EGF-like domain of FIX contributes to a variety of functional properties. These include enzymatic activity toward synthetic substrates and the natural substrate FX, both of which are intrinsic properties of the heavy chain of FIXa. Furthermore, binding of Ca 2ϩ to the first EGF-like domain favors binding of FVIII light chain to the light chain of FIXa (12), although we cannot completely exclude that the FIXa heavy chain also contributes to this process. Apparently, the FIXa light chain serves a crucial role in FIXa functions that involve portions of the FIXa heavy or light chain. This view is supported by our observation that a monoclonal antibody directed against the FIXa light chain completely inhibits amidolytic activity (12). Thus, interdomain interactions within the FIXa molecule are essential for FIXa to display optimal enzyme function. Clearly, occupancy of the Ca 2ϩ binding site in the first EGF-like domain contributes to these interactions to a significant extent. Strikingly, FIXa molecules in which the Ca 2ϩ -binding domain is substituted by the homologous region of FX have been reported to display normal biological activity (43,44). As the Ca 2ϩ - binding EGF-like domains of FIX and FX share considerable structural homology (22,45), these experiments suggest that the conserved tertiary structure of the Ca 2ϩ -containing EGFlike domain is more important than the protein-specific primary structure in supporting enzyme function. The importance of the conserved, Ca 2ϩ -bound structure of the EGF-like domain is supported by the notion that impaired Ca 2ϩ binding to distinct EGF-like domains severely affects biological activity of for instance FIXa, thrombomodulin, protein C, and fibrillin-1 (22, 37, 46 -48). With regard to fibrillin-1, it has been reported that mutations of the Ca 2ϩ -binding residues are associated with impaired microfibril formation, which is clinically manifested by the severe connective tissue disorder known as the Marfan syndrome (46,47). Similarly, mutations of the Ca 2ϩ -binding residues in the sixth EGF-like domain of thrombomodulin severely affect thrombomodulin cofactor activity (residual activity Ͻ 10%) (37). Interestingly, like the FIXa-FVIII light chain interaction, the interaction of these mutated thrombomodulin molecules with the enzyme thrombin was affected less severely (no more than 5-fold) (37).
In conclusion, we propose that the Ca 2ϩ -bound structure of EGF-like domains serves an important role in supporting intra-and intermolecular protein-protein interactions. We suggest that mutations of the Ca 2ϩ -binding residues in the EGFlike domains of other hemostatic proteins, such as FVII, FX, protein C, and protein S, should also result in proteins that display reduced activity. The defects underlying the reduced activity may, as in FIX, be associated with more than one functional parameter.