Comparative platelet binding and kinetic studies with normal and variant factor IXa molecules.

We have recently shown that thrombin-stimulated human platelets have specific, saturable receptors for factor IXa, occupancy of which promotes factor X activation (Ahmad, S. S., Rawala-Sheikh, R., and Walsh, P. N. (1989) J. Biol. Chem. 264: 3244-3251, 20012-20016; Rawala-Sheikh, R., Ahmad, S. S., and Walsh, P. N. (1990) Biochemistry 29, 2606-2611). To study the structural requirements for factor IXa binding to platelets, equilibrium binding studies and kinetic studies of factor X activation were carried out with normal factor IXa and with two variant proteins: factor IXaAlabama (FIXaAL; Asp47----Gly substitution) and factor IXaChapel Hill (FIXaCH; Arg145----His substitution). In the absence of factors VIIIa and X, there were 331 binding sites/platelet for FIXaCH (Kdapp = 2.8 nM), and 540 sites/platelet for FIXaAL (Kdapp = 3.2 nM), compared with 540 sites/platelet (Kdapp = 2.3 nM) for normal factor IXa. The addition of factors VIIIa and X, both at saturating concentrations, had no effect on the number of binding sites for either normal or variant factor IXa, resulted in a decrease in the Kd for normal factor IXa to 0.67 nM, resulted in a suboptimal decrease in Kd for FIXaAL (1.4 nM), and had no effect on the Kd for FIXaCH. Kinetic studies of factor X activation at variable factor IXa concentration confirmed these values of Kd in the presence of factors VIIIa and X. Determination of rates of factor X activation at variable substrate concentrations yielded normal values of catalytic efficiency (kcat/Km) for the variant proteins, thereby indicating that the abnormally low rates of factor X activation obtained were a consequence of the low affinity binding of FIXaAL and FIXaCH to thrombin-activated platelets in the presence of factors VIIIa and X. These studies suggest that the presence of Asp47 and the cleavage of factor IX at Arg145-Ala146 are important structural features required for specific, high affinity factor IXa binding to platelets in the presence of factors VIIIa and X.

We have recently shown that thrombin-stimulated human platelets have specific, saturable receptors for factor IXa, occupancy of which promotes factor X activation (Ahmad, S. S., Rawala-Sheikh, R., and Walsh, P. N. (1989)  Activated human platelets promote the activation of factor X by factor IXa (1,2). Previous studies from our laboratory, aimed at elucidating the mechanisms by which platelets and factor VIII contribute to this coagulation reaction, have shown that thrombin-stimulated human platelets have specific, saturable binding sites for factor IXa and that the presence of factor VIII and factor X increases the binding affinity 5-fold (3). We have also shown that platelet receptor occupancy with factor IXa is closely correlated with rates of factor X activation (4,5).
To study the structural requirements for factor IXa binding to platelets, we have now carried out detailed comparative platelet binding and kinetic studies with normal and variant factor IXa molecules. One of these proteins, faCtOr IX,&hrne (factor IXAL),l can be activated by factor XIa in the presence of calcium ions to a factor IXaB form with about 10% of the clotting activity of the normal factor IXaO (6, 7). An adenine to guanine transition in the first nucleotide of exon d causes the substitution of a glycine codon (GGT) for the normal aspartic acid codon (GAT). This point mutation results in a single amino acid substitution at residue 47 of the zymogen in the first epidermal growth factor-like domain of factor IXAL (8). The factor IX defects previously reported in the first epidermal growth factor-like domain are mostly associated with mild hemophilia B. In addition to factor IXaAL, these include factor IXLondon7 (Pro55 + Ala, 10% of normal activity) (9), factor IXDurham (Gly6' + Ser, 14% of normal activity) (lo), factor IXLondon,j (P-OH Asp@ + Gly, 8% of normal activity), and factor IXN,, b,,don (G1u5' + Pro, Cl% of normal activity) (11).
The second variant protein we have studied is factor I&hap, Hill (factor IX&. The molecular defect in factor IXcH is the substitution of histidine for arginine at position 145 (12)(13)(14). This is the first cleavage site in the normal pathway of factor IX activation.
Thus, factor IXcH is not activated normally either by factor XIa or by factor VIIa-tissue factor. Only the Arg'80-Va1'81 bond is cleaved giving rise to factor IXa,, in which the activation peptide remains covalently attached to the light chain. Factor IXacucH has 20% of the clotting activity of normal factor IXa@. Mutations in certain abnormal factor IX proteins have been demonstrated to cause abnormally slow or incomplete activation of factor IX by factor XIa. These include defects such as that in factor IXHilo in which cleavage of the Arg'80-Va1'81 bond is prevented by substitution of glutamine for arginine at position 180. The dysfunction of factor IXNewLondon has also been attributed to '  Binding and Kinetic Studies of Variant Factor IXa defective cleavage at the Arg'*"-Val'*' bond (11). The lipid binding and kinetic properties of normal factor IXa (factor IXaN) have been examined and compared in detail with the two variant proteins, namely factor IXAL and factor IXcH (15). In this paper we have carried out detailed studies to compare the platelet binding and kinetic studies of these abnormal variant factor IXa molecules with factor IXaN. electrophoresis in NaDodS04 was carried out according to the procedure of Laemmli (17).
Other materials were the same as reported previously (3)(4)(5). Proteins-Details of the purification, assay, and characterization of human coagulation proteins, including factor IX, factor IXa, factor VIII, factor X, and cu-thrombin, were previously published (3). The conditions used for activation of factor VIII were identical with those previously published (3)(4)(5).
The variant factor IX molecules were isolated from human plasma as described previously (15). Both the normal and variant factor IX molecules were radiolabeled with "'1 by the Iodo-Gen method as previously described (3), and specific radioactivities of all proteins were in the range of 2.0-2.5 x lo6 cpm/ pg. Activation of purified factor IXb, factor IXA,., and factor IX,.-" by purified human factor XIa were carried out as previously described (3). Autoradiograms of normal and variant factor IX and factor IXa were developed to provide structural characterization of "Y-labeled proteins.
Both l'SI-labeled factor IXA~ and factor IXc" appeared as single bands at M, = 57,000 ( One of the two major bands migrated with a mobility identical with that of the heavy chain band from factor IXN, whereas the other band had an apparent M, of 45,000 which corresponds to a cleavage product consisting of the light chain and the activation peptide similar to that found on cleavage of either factor IXN or factor IXCH by Russell's viper venom (13,16

Specific Binding of "'I-Labeled
Factor IXaN, Factor IXaal., and Factor IXacH to Thrombin-activated Normal Human Platelets-In the present work, we have compared binding of '2"I-labeled factor IXa with the binding of factor IXaAL and factor IXacH to normal human platelets. Scatchard analysis of the binding data ( Fig. 2) gave straight lines indicating the presence of a single class of binding sites for both the normal and variant factor IXa molecules both in the presence and absence of factor VIIIa and factor X. The affinity and stoichiometry of binding for these ligands under both experimental conditions was determined in six separate experiments, the means (+S.E.) of which are given in Table I. In addition, the stoichiometry and affinity of factor IX binding was determined as previously reported (3), and the results were recorded in Table I. In the absence of factor VIIIa and factor X, there were 331 binding sites/platelet for factor IXacH (Kd spp = 2.8 nM) and 540 sites/platelet for factor IXaAL (& app = 3.2 nM), compared with 540 sites/platelet (& app = 2.5 nM) for factor IXaN, and 306 sites/platelet (Kdapp = 2.6 nM) for factor IXN. The addition of factor VIIIa and factor X, both at saturating concentrations, had no effect on the number of binding sites for either normal or variant factor IXa molecules or for factor IXN, resulted in a decrease in the & for factor IXaN to 0.67 nM, resulted in a suboptimal decrease in & for factor IXaAL (1.4 nM), and bad no effect on the & for either factor IXacH or factor IXN. The number of binding sites for factor IXaAL was not significantly different from that for factor IXaN. The number of binding sites for factor IXacH was significantly lower than that for factor IXaN (p < 0.01) and was not significantly different from that for factor IXN.  in the presence of saturating concentrations of factor X and factor VIIIa (Fig. 3). The kinetic approach gave similar results to the binding studies (3,4), and its use is justified in our previous studies (3)(4)(5). The apparent Kd was determined as 0.61 nM for factor IXan, 1.4 nM for FIXaAL and 2.0 nM for FIXacH ( Fig. 3 and Table I).
We also determined the kinetic parameters for factor X activation by normal and variant factor IXa molecules in the presence of thrombin-stimulated platelets and factor VIIIa (Fig. 4). Studies were carried out at a factor IXa concentration of 0.01 nM, well below the apparent dissociation constant for binding of factor IXa to platelets. The values of K,,,, V,.,, k,,,, and catalytic efficiency (kJK,,,) for factor IXaN, factor IXaAL and factor IXacH obtained at saturating concentrations of 100 factor VIIIa are summarized in Table II. Catalytic efficiency can be assessed in two ways: 1) as a function of the total amount of factor IXa added, or 2) as a function of the amount of factor IXa bound to the platelet. If catalytic efficiency is assessed as a function of the amount of factor IXa bound, there is essentially no difference in the rate at which factor X is activated (the last column of Table II). So the decrease in the rate at which a given amount of factor IXaAL or factor IXacu can catalyze factor X activation is solely due to the reduced affinity of these two proteins for the platelet-factor VIIIa-factor X complex. ' k,., expressed as moles of factor Xa formed per min per mol of total added factor IXa. b k,., expressed as moles of factor Xa formed per min per mol of total platelet-bound factor IXa. DISCUSSION The purpose of the studies reported here was to begin an analysis of the structural features of the factor IXa molecule that are important for factor IXa binding to platelets and for the assembly of the factor X activating complex on the platelet surface. Previously we have shown that factor IXa binds reversibly to 500-600 sites per platelet and that platelet activation and the presence of calcium ions are required for this interaction (3). The dissociation constant (I&) for factor IXa binding to activated platelets is -2.5 nM in the absence of factor VIIIa and factor X and -0.5 nM in the presence of these proteins at saturating concentrations (3). Similar findings have been reported for bovine aortic endothelial cells by Stern et al. (19). We have also shown that zymogen factor IX binds to 250-300 sites per activated platelet in the presence of calcium ions with a & of -2.5 nM either in the presence or absence of factors VIIIa and X and that factor IX competes with factor IXa for about one-half its low affinity sites in the presence of factors VIIIa and X (3). This suggests that the zymogen contains a domain important for binding of the enzyme to its platelet receptor and that the conversion of factor IX to factor IXa involves either a conformational alteration or the exposure of domains in factor IXa that allow it to interact with twice the number of platelet receptors as factor IX and also allow it to interact with factors VIIIa and X. We have further demonstrated that factor IXa binding to its high affinity site is closely correlated with rate enhancements of factor X activation by activated platelets (4), which can decrease the K, for factor X activation and can permit factor VIIIa to increase the kCat with a consequent increase of catalytic efficiency (k,,t/K,,,) of (17.4 x 106)-fold (5). Finally, we have studied the role of the active site of factor IXa in the binding of the enzyme to platelets by examining the interaction with platelets of factor IXa active site inhibited with dansyl-L-glutamyl-glycyl-L-arginyl chloromethyl ketone (DEGR-CK).
Since DEGR-factor IXa was shown to be a competitive inhibitor both of factor IXa binding and of factor X activation, with a Ki almost identical with the & for factor IXa binding, we concluded that the active site of factor IXa is not involved in binding to the high affinity site in the presence of factor VIIIa and factor X (4).
To determine the structural features of factor IXa that are required for interaction with platelet receptors and for assembly of the factor X activating complex, we studied the binding of factor IXaAL and factor IXacn to activated platelets. It has previously been suggested (16,20) that the abnormal coagulant activities of these two proteins could be a consequence of deficient binding to charged membrane surfaces. However, it was subsequently shown (15) that the abnormal rates of factor X activation observed with both factor IXaAL and factor IXacn in the presence of small, unilamellar vesicles composed of 30% phosphatidylserine and 70% phosphatidylcholine were not a consequence of abnormal binding. Thus, the zymogen and activated forms of factor IXN, factor IXAL, and factor IXcn were shown to bind with similar affinities to small, unilamellar vesicles as determined by 90" light scattering (15). It was therefore concluded that the normal function of factor IXa must entail interactions between the light and heavy chains on the phospholipid surface. The present studies addressed the structural requirements for binding of factor IXa to what must be presumed the physiologic locus of factor X activation, i.e. the platelet membrane. Factor IXaAL was shown to bind with normal affinity (I& = 3.2 nM) to a normal number (540) of sites per platelet in the absence of factors VIIIa and X. This implies that the aspartic acid residue at position 47 in the normal protein (and mutated to a glycine in factor IX.& is not required for normal binding of factor IXa to its platelet receptor. This demonstration of a normal affinity of factor IXaAL binding to platelet membranes in the absence of factors VIII and X confirms the results of Jones et al. (15) showing normal binding of factor IXaAL to phospholipid membranes. However, in the presence of factors VIIIa and X, the binding affinity of factor IXaAL was reduced (& = 1.4 nM) as determined either by equilibrium binding or by kinetic determinations of rates of factor X activation (Table I). This result is not in conflict with those of Jones et al. (15) since their phospholipid binding studies were carried out only in the absence of factors VIII and X.
The results of our kinetic analysis of factor X activation by factor IXaAL (Table II) help to clarify the interpretation of our binding studies and are also in general agreement with the previous factor X activation studies of Jones et al. (15). Thus, we found the I& for factor X activation by factor IXaAL to be normal, suggesting that factor X binds with normal affinity to the factor IXaAL-factor VIIIa-platelet membrane complex. The Vmax (7.4 nM . min-') with factor IXaAL was 40% of normal, i.e. when compared with normal factor IXa (18.7 nM . min-'), a result consistent with the reduction in relative rate of factor X activation by factor IXaAL (43% of normal) obtained by Jones et al. (15) with phospholipids.
In the present study, however, since we carried out both equilibrium binding studies in parallel with factor X activation studies, we were able to calculate true catalytic constants ( kC,,J expressing the maximal rate of factor Xa formation as a function of the amount of factor IXa bound (Table II). The result of this analysis demonstrates a normal turnover number (k,,J and catalytic efficiency (k,,/K,) for factor IXaAL compared with factor IXaN, indicating that the defect in factor X activation by factor IXaAL is entirely a consequence of its decreased affinity for platelet receptors in the presence of factor VIIIa. Therefore, although the interpretation of the molecular basis for the defect is open to further study, it would appear that the Asp47 ---f Gly mutation in factor IXAL results in a decreased rate of factor X activation solely as a consequence of decreased affinity of the mutant enzyme for its platelet receptor in the presence of factor VIIIa and factor X.
A recent paper published by McCord et al. (21) reporting studies with factor IXAL provides evidence for a conformational change in factor IX due to high affinity calcium binding in the first epidermal growth factor domain. These authors found that although factor IXAL binds calcium ions normally to a high affinity site in the first epidermal growth factor domain, the variant enzyme fails to undergo a calcium-induced conformational change that occurs in normal factor IXa, thereby permitting it to interact properly with factor VIIIa and factor X. Thus, whereas factor IXAL was activated normally by factor XIa and factor IXaAi. had 52-60% of normal activity in a calcium/phospholipid vesicle system, the addition of factor VIIIa decreased the relative rate of factor X activation by factor IXaAL to 18-19% of normal. These observations are consistent with our suggestion that the defect in factor IXaAL is a consequence of its failure to bind with normal affinity to membranes in the presence of factor VIIIa.
Our studies with factor IXacu show that this variant enzyme binds to platelets in a manner indistinguishable from normal zymogen factor IX (Table I). Thus, the number of binding sites per activated platelet appears to be similar to that for normal factor IX and about half the number for normal factor IXa. Moreover, the affinity of binding, which is 5-fold enhanced for normal factor IXa in the presence of factors VIIIa and X, is unaffected by factors VIIIa and X in the case of factor IXacH. This very interesting result suggests that cleavage of factor IX at Arg'45-Ala'46 (defective in factor IXcH because of the Arg'45 -+ His substitution), as well as at Arg'80-Val'8', with consequent formation of an activation peptide, is required for binding to the normal complement of receptors. In attempting to explain why the number of binding sites for factor IXa is almost exactly double that for zymogen factor IX, it is tempting to speculate that the receptor is bivalent (possibly homodimeric) and can accommodate two factor IXa molecules but only one factor IX molecule. It is possible that the presence of the heavily glycosylated activation peptide region of factor IX prevents access of a second factor IX molecule to the receptor complex, whereas formation of the activation peptide or its release from covalent attachment allows the resultant factor IXa access to both binding sites on the putative dimeric receptor. The fact that the affinity of factor IXacH for platelets is unaffected by the presence of factors VIIIa and X suggests that cleavage of factor IXa at Arg'45-Ala'46 is essential for the exposure in factor IXa of a binding site for the factor VIIIa-factor X complex that is involved in high affinity binding of factor IXa to platelets.
The present studies that demonstrate a normal affinity of factor IXacu binding to activated platelets in the absence of factors VIII and X are in agreement with the demonstration by Jones et al. (15) of normal binding of factor IXacH to phospholipid (phosphatidylserine/phosphatidylcholine) vesicles by 90" light scattering.
Our demonstration that factor IXacH fails to bind to activated platelets with high affinity in the presence of factors VIIIa and X is not inconsistent with the phospholipid binding studies of Jones et al. (15) which were not done in the presence of factors VIII and X. Finally, the present studies demonstrating abnormally low rates of factor X activation by factor IXacH are entirely consistent with the studies of Jones et al. (15). Thus, as shown in Table  II, the catalytic efficiency (k,,&,,) calculated for factor IXacH in the presence of activated platelets and factor VIIa was 6,944 @K' .min-' or 37% of normal (18,700 PM-'. mini) when the k,., was based on the total amount of enzyme added, compared with a relative rate of factor X activation by factor IXacH shown to be 36% of normal by Jones et al. (15). However, when we calculated k,,, as moles of factor Xa formed per mol of factor IXa bound, the catalytic efficiency proved to be essentially normal (36,783 min-' compared with 26,714 min-' for normal factor IXa). This indicates that the defect in factor X activation observed with factor IXacH is attributable solely to the decreased amount and affinity of factor IXacH binding to platelets in the presence of factors VIIIa and X.