of factor X. The chimeric protein, factor IX(Xegf2), and the wild- type, factor IXWt9 ... deficient binding of the chimeric protein to the relevant protein or receptor.
427
Biochem. J. (1995) 310, 427-431 (Printed in Great Britain)
The role of the second growth-factor domain of human factor to platelets and in factor-X activation Syed S. AHMAD,*t Razia RAWALA,* Wing-Fai
CHEUNG,§11
IXa
in binding
Darrel W. STAFFORD§ and Peter N. WALSH*tt
*The Sol Sherry Thrombosis Research Center, tDepartment of Biochemistry and tDepartment of Medicine, Temple University School of Medicine, Philadelphia, PA 19140, U.S.A., and §Department of Biology, University of North Carolina, Chapel Hill, NC 27599, U.S.A.
To study the structural requirements for factor IXa binding to platelets, we have carried out equilibrium binding studies with human factor IXa after replacing the second epidermal growth factor (EGF) domain by the corresponding polypeptide region of factor X. The chimeric protein, factor IX(Xegf2), and the wildtype, factor IXWt9 produced in embryonic kidney cells 293 were radiolabelled with 1251 and activated with factor XIa. Direct binding studies with thrombin-activated platelets showed normal stoichiometry and affinity of binding of factor IXawt in the presence of factor Vllla (2 units/ml) and factor X (1.5 ,uM). However, under similar experimental conditions, factor IXa(xegf2) was bound to a smaller number of sites (396 sites/platelet) with decreased affinity, i.e. a dissociation constant (Kd) of 1.4 nM,
compared with normal factor IXa, factor IXaN (558 sites/platelet; Kd 0.67 nM), or factor IXa,t (590 sites/platelet; Kd 0.61 nM). The concentrations of factor IXaN and factor IXawt required for half-maximal rates of factor-X activation were 0.63 nM and 0.7 nM, indicating a close correspondence of the Kdapp for binding of factor IXawt to the factor-X activating complex on activated platelets to the Kd obtained in equilibrium binding studies. In contrast, kinetic parameters for factor-X activation by factor IXa(xegf2) showed a decreased affinity (Kd 1.5 nM), in agreement with results of binding studies. These studies with factor IX(Xegf2) suggest that the EGF-2 domain may be important for specific high-affinity factor IXa binding to platelets in the presence of factor Vllla and factor X.
INTRODUCTION
for factor X substituted for the second EGF-like domain. We have assumed that any amino acid sequences within the EGF-2 domain of factor IXa that are essential for binding to platelet receptors, to factor X, or to factor VIII should be revealed by deficient binding of the chimeric protein to the relevant protein or receptor.
Factor IX is a member of a class of vitamin K-dependent bloodcoagulation proteins that share a common domain structure, consisting of an N-terminal domain containing y-carboxyglutamic acid (Gla) residues, linked to a short hydrophobic sequence, two epidermal growth factor (EGF)-like domains, an activation peptide sequence and a C-terminal serine protease domain [1]. In an effort to understand the structural features of factor IXa that are important for its binding to platelets and for the assembly of the factor-X activating complex on the platelet surface, we have recently carried out a series of investigations focused on the contributions of these molecular domains [2]. We have previously shown that the N-terminal EGF-1 domain does not contain amino acid sequences involved in platelet binding, since a chimeric factor IXa molecule in which the EGF-1 domain of factor X replaced that of factor IX was shown to bind with normal affinity and stoichiometry to the activated platelet surface [3]. In contrast, a major role for the Gla-containing domain in binding of factor IXa to the platelet surface has been demonstrated in our experiments with both Gla-modified and Gladomainless factor IXa molecules [4]. In support of this hypothesis is our recent demonstration that residues 3-11 at the N-terminus of the Gla domain forms a surface-exposed loop structure containing amino acid residues essential for high-affinity factor IXa binding to the platelet surface either in the presence or in the absence of factor VIIIa [5]. The present studies are focused on the possible contribution of the second EGF-like domain of factor IXa in the assembly of the factor-X activating complex on the platelet surface. To explore this possibility, we utilized recombinant factor IX molecules including a chimeric protein containing the amino acid sequence
EXPERIMENTAL Materials The chromogenic substrate S2337 [Bz-Ile-Glu-(y-piperidyl)-GlyArg-p-nitroanilide] was purchased from AB Kabi Diagnostica (Stockholm, Sweden). p-Aminobenzamidine was obtained from Sigma Chemical Co. (St. Louis, MO, U.S.A.). D-Phenylalanylprolyl-arginylchloromethane (PPACK) was purchased from Calbiochem-Behring Corp. (San Diego, CA, U.S.A.). All other reagents and chemicals used were the same as previously reported [6], and were obtained from Sigma Chemical Co., Aldrich Chemical Co. (Milwaukee, WI, U.S.A.), and CalbiochemBehring Corp., and were of the highest grade commercially available. Proteins Human coagulation proteins, including factor IX, factor IXa, factor VIII, factor X and a-thrombin, were purified, assayed and characterized as previously published [6]. The conditions used for activation of factor VIII with human a-thrombin were identical with those previously published [6]. All proteins were
Abbreviations used: Gla, y-carboxyglutamic acid; EGF, epidermal growth factor; PPACK, D-phenylalanyl-L-prolyl-L-arginylchloromethane. IlPresent address: Biomedical Center, Box 575, 751 23 Uppsala, Sweden. §To whom correspondence should be addressed, at: The Sol Sherry Thrombosis Research Center, Temple University School of Medicine, 3400 North Broad Street, Philadelphia, PA 19140, U.S.A.
428
S. S. Ahmad and others
0
= 0.10
\
|
X
f
f
0.05-
0
100
200
300 0 Bound (pM)
Figure 1 Scatchard analysis of specfic binding of I'll-factor IXa,, 151-factor In the absence (a) and the presence (b) of factor Vil and factor X
IXaa,,,
100
and I'l-factor
200
300
IXa", to thrombin-stimulated normal human platelets
Gel-filtered platelets (3.5 x 108/ml) were incubated at 37 °C with mixtures of human cc-thrombin (0.1 unit/ml) and factor X (1.5 ,M). Binding was determined as detailed in the Experimental section. Non-specific binding was determined in the presence of excess unlabelled factor IXa (0.44 1sM; 25 ,ug/ml) and was subtracted from total binding to obtain specific binding. The results in (a) represent a Scatchard plot of specific binding data for factor IXaN in the absence (0) of factor Villa and factor X compared with specific binding data for either factor IXawt (A) or factor MXave, (EO) in the absence of factor Vllla and factor X. The results shown in (b) represent a Scatchard plot of specific binding data for factor IXaN in the presence (-) of factor VIII and factor X compared with specific binding data for either factor lXaw, (A) or factor IXa(x,f2) (*) in the presence of both thrombin-activated factor VIII and factor X. The plotted results represent mean values of three experiments with normal and chimeric factor IXa molecules, each done in duplicate.
> 98 % pure as determined by polyacrylamide-slab-gel electrophoresis in SDS [7]. Protein concentrations were determined by the Bio-Rad dye-binding assay according to instructions provided by the manufacturer (Bio-Rad, Richmond, CA, U.S.A.). Chimeric factor IX molecules were obtained by replacing the EGF-2 coding sequence in the factor-IX cDNA with the corresponding factor X EGF-domain coding sequence by sitedirected mutagenesis as reported previously [8]. The chimeric and wild-type proteins, expressed in human embryo kidney cells, were purified from cell supernatants, and analysis for Gla was carried out as reported previously [8,9]. The normal plasmaderived, the wild-type and the chimeric factor IX molecules were radiolabelled with 1251 by the lodogen method as previously described [6], and specific radioactivities of all proteins were in the range of(2.0-2.5) x 106 c.p.m./mg. Activation ofboth normal and chimeric factor IX molecules by purified factor XIa was carried out as previously described [6]. We also utilized the paminobenzamidine fluorescence assay to examine quantitatively the activation of factor IXN and recombinant factor IX molecules as previously reported [8,10]. After gel electrophoresis, autoradiograms of normal and chimeric factor IXa were developed to provide structural characterization of 125I-labelled proteins. Both recombinant proteins, factor IXa(Xeg 2) and factor IXawt, migrated under reducing conditions as two polypeptide chains of Mr 27000 and 17000, representing the heavy and light chains, and were indistinguishable from plasma-derived factor IXaN (results not shown).
Binding experiments In a typical binding experiment, gel-filtered platelets [(3-4) x 108/ml] in Ca2+-free Hepes Tyrode's buffer, pH 7.4, were incubated at 37 °C in a 1.5 ml Eppendorf plastic centrifuge tube with mixtures of unlabelled and radiolabelled factor IXa, CaC12, platelet stimuli, and other proteins. Platelets were separated from
unbound proteins as previously described [6]. The data were analysed, and the number of binding sites and Kd were calculated from the means of three independent determinations, each done in duplicate, as previously described [6] by using a Macintosh Quadra 900 Computer (Apple Computer, Cupertino, CA, U.S.A.) and the LIGAND program as modified by G. A. McPherson (Elsevier Science Publishers BV, The Netherlands, 1985).
Measurements of rates of factor Xa formation The activation of factor X by normal and variant factor IXa was carried out at 37 °C in the presence of thrombin-stimulated gelfiltered platelets, factor Vllla and CaCl2 as described previously [11,12]. The details of experimental conditions and concentrations of reactants are given in the Results section and in the Figure legends.
Calculations of kinetic constants The derivation of kinetic constants of factor-X activation by factor IXa was based on a one-enzyme one-substrate model. The Michaelis constant (K.) and the maximum velocity (Vmax ) were calculated from the mean + S.E.M. of three independent determinations each done in duplicate, of factor-X activation rates at variable factor X concentrations as described previously [12]. Values of Kd were obtained from experiments in which rates of factor-X activation were determined at variable factor IXa concentration as previously described [11]. The values of turnover numbers (kcat.) were calculated by dividing Vmax. values by the amount of enzyme (factor IXa) bound under the conditions of the experiment. This latter value was obtained from the equation: Amount bound = Bmax.E/(Kd + E)
Second epidermal-growth-factor domain in factor IXa binding to platelets
429
Table 1 Binding constants for normal and chimeric factor IXa molecules and for Scott-syndrome platelets Data are shown for Scott-syndrome platelets [13] for comparison. Kd (nM) No. of sites/platelet
Platelets
Ligand
Factor VIII
(±S.E.M.)
Equilibrium
Normal
Factor IXa(N)
Normal
Factor
lXa(Xef)
Normal
Factor
MXa()
Scott-syndrome
Factor IXa(N)
Absent Present Absent Present Absent Present Absent Present
576 + 61 558 + 55 381 + 48 396 + 50 600 + 65 590 + 70 461 + 60 412 + 44
2.70 + 0.40 0.67 + 0.07 3.11 + 0.52 1.40 + 0.09 2.80 + 0.36 0.61 + 0.04 3.2 + 0.33 2.5+ 0.19
Kinetic
0.61 + 0.03
1.70 + 0.12
0.67 + 0.06 2.5
-r 80
E
E E 60
x
2
E m
Cu
0
X 40
a-r-
0 co U-
20
m
O'
0.01
2 0.1
1.0
10
100
33 Factor IXa (nM)
Concn. of competing protein (nM)
Figure 3 Rates of factor Xa formation by normal and chimeric factor IXa molecules In the presence of thrombin-activated platelets and factor Villa Figure 2 Competition by unlabelled factor IXa, factor IXa,xm and factor lXa, for I'1l-labelled factor IXa binding sites on thrombin-activated platelets In the presence of factor VIII and factor X .tP0Pt n1M atot'31a7 toru IIIin a ncntrations n,idulduLckllill nrc fnctP nrf wTrh riatormintatlin t[IIUUlIuIl-dailVdlBU Ul D X 1u- thrnmhiursc t[lU P16MAtUC WUMe UMlUMMillU 111 tho mM NaCI, 5 mM CaCI2, 175 mM 7.9), Tris (pH containing 50 reaction volume of 100,1 Gel-filtered platelets (3.8 x 108/ml) were incubated for 20 min at 37 0C with human a-thrombin Platelets were human serum albumin. mg/ml Vllla and 0.5 units/ml factor ,sM factor 5 X, 1.5 of concentrations and various nM), IXaN (3.3 1251-labelled factor mm), CaCI2 (5 (0.1 unit/ml), stimulated with 0.1 unit/ml thrombin in the presence of CaCI2 (5 mM), and factor IXa was unlabelled factor IXaN (@), factor IXa(xef2) (U) or factor IXaw, (A), in the presence of preincubated with platelets for 10 min at 37 °C. Excess thrombin was neutralized with 50 nM thrombin-activated factor Vil (5 units/ml) and factor X (0.15 ,uM). Binding was determined as PPACK before addition of factor Vllla and performance of the assay. The plotted results for factor described in the Experimental section. Maximum binding (100%) was determined by iXaN (0) are shown, compared with factor lXawt (A) and factor IXa(xegt2) (U). These subtracting the non-specific binding, i.e. the binding determined in the presence of excess observations are means+S.E.M. of duplicate observations from three separate experiments. unlabelled factor IXaN, factor IXa(X,f2) or factor IXawt from total binding. The results shown represent residual factor IXa binding in the presence of factor VIII and factor X. The results presented are means+S.E.M. of three separate experiments, each done in triplicate.
of factor IX with the second EGF-like domain of factor X, factor where Bmax is the maximum amount of factor IXa bound or the total receptor concentration, E is total factor IXa concentration, and Kd is the dissociation constant. The details of this calculation are provided in previous papers [11,13].
RESULTS Specific binding of 1251-labelled factor IXaN, factor lXa, and factor lXa(X,f2) to thrombin-activated normal human platelets To study the structural requirements for factor IXa binding to platelets, equilibrium binding studies were carried out in the presence of saturating concentrations of factor X and thrombinactivated factor VIII, with the chimeric protein or with wild-type recombinant or plasma-derived factor IXa molecules. The chimeric factor IX molecule consisted of the Gla domain, the EGF- 1 domain, the activation peptide and the catalytic domain
IXa(xegf2). The wild-type and the chimeric factor IX molecules were fully carboxylated, with 10.1 mol of Gla/mol of factor IXa(xegf2), 10.4 mol of Gla/mol for factor IXW,, and 10.5 mol of
Gla/mol for factor IXaN. The factor IXa molecules, after activation by factor XIa, were shown to be fully active by activesite titration with p-aminobenzamidine when compared with plasma-derived factor IXa, and both consisted of completely activated factor IXa, as shown by SDS/PAGE under reducing conditions (results not shown). Scatchard analysis of the binding data (Figure 1) gave straight lines, indicating the presence of a single class of binding sites for both the normal and chimeric factor IXa molecules. The affinity and stoichiometry of binding for these ligands under both experimental conditions were determined in six separate experiments, the means ( ± S.E.M.) of which are given in Table 1. Similar studies were carried out with factor IXawt and the results are recorded in Table 1. In the absence of factor Vllla and factor X there were 381+ 48
430
S. S. Ahmad and others Kinetic studies of normal and recombinant factor IXa molecules In this study we also determined the apparent K4 for normal and recombinant factor IXa binding to platelets by kinetic studies of factor-Xa formation in the presence of saturating concentrations of factor X and factor VIIIa (Figure 3). The Kd values obtained are presented in Table 1. The use of this kinetic approach for the determination of binding affinity is justified in our previous studies [1 1,13]. The apparent Kd values obtained for factor IXa" (0.61 + 0.03 nM), factor IXa(xegf2) (1.70 + 0.12 nM) and factor IXawt (0.67 + 0.06 nM) respectively (Figure 3 and Table 1) are similar to those obtained from the equilibrium binding studies (Table 1). The kinetic parameters for factor-X activation by normal and recombinant factor IXa molecules were determined in the presence of thrombin-stimulated platelets and factor VIlla (Figure 4). Studies were carried out at a factor IXa concentration of 0.01 nM, well below the apparent Kd for binding of factor IXa to platelets. The values of K., Vnax., kcat. and kcat./Km for normal, wild-type and chimeric proteins, obtained at saturating concentrations of factor VIIIa, are summarized in Table 2. From the results it is apparent that Km values for factor IXa(xegf2) are entirely normal. Similarly, Vmax values for factor IXa", were normal compared with factor IXaN (i.e., plasma-derived factor IXa). In contrast, Vmax values for factor IXa(xegf2) were about 25 % of normal. However, turnover numbers were calculated from values of Vm.. and amounts of factor IXa bound to thrombin-activated platelets and expressed as mol of factor Xa formed/min per mol of total platelet-bound factor IXa. The details of this calculation and its justification are provided in the Experimental seetion and in previous publications [11,13]. It is clear from this analysis that the values for kcat and kcat /Km for factor IXa(Xegf are entirely normal (Table 2). This indicates that the decreased rates of factor-X activation (i.e. the Vma. values) observed with factor IXa(xegf2) are solely a consequence of the decreased affinity of these molecules for platelet binding sites in the presence of factor VIIIa and factor X.
.E 15 E
'10 x 0
5 IL
Factor X (nM)
Figure 4 Factor X activation by factor IXa,, factor IXa,, and factor IXaa,,, Initial rates of factor X activation were determined as described in the Experimental section at various concentrations of factor X as indicated in the graph. The experiment was carried out essentially as described in the legend to Figure 3, except that factor X concentration was varied and the concentration of factor IXaN, factor IXat and factor lXa(XW) was 10 pM. The plotted results for factor IXaN (@), factor IXaw, (A) and factor IXa(xegf2) (X) are the means+ S.E.M. of duplicate observations from five separate experiments.
sites/platelet (Kd 3.11 + 0.52 nM) for factor IXa(xegfta, compared with 576+61 sites/platelet (Kd 2.70+0.40 nM) for factor IXaN, and 600 + 65 sites/platelet (Kd 2.8+0.36 nM) for factor IXawt. The presence of factor Vllla and factor X, both at saturating concentrations, had no effect on the number of binding sites for factor IXaN or recombinant factor IXawt, and resulted in a decrease in the Kd for these proteins to 0.67 nM and 0.61 nM respectively. In contrast, under similar experimental conditions, factor IXa(xeg,2) was bound to a smaller number (P < 0.05) of sites (396 + 50 sites/platelet), with a sub-optimal decrease in affinity (Kd 1.4+ 0.09 nM).
2)
=
Competftlon studies with factor IXaN, factor
1Xa(X"
IXat and factor
We carried out competition studies with unlabelled factor IXaN, factor IXawt and factor IXa(xegf2) by incubating thrombinstimulated platelets in the presence of CaCl2 for 20 min at 37 °C with 1251I-labelled factor IXa and various concentrations of unlabelled proteins in the presence of factor X and thrombinactivated factor VIII. When the residual binding of 12m1-labelled factor IXa was determined (Figure 2), it was apparent that excess factor IXaN9 factor IXawt and factor IXa(Xegf2) all prevented > 95% of 125I-labelled factor IXa binding. From the results presented in Figure 2, it is estimated that the concentrations of factor IXaN, factor IXawt and factor IXa(xeV2) required for halfmaximal inhibition of factor IXa binding in the presence of factor VIIIa and factor X were 0.5 nM, 0.65 nM and 1.9 nM respectively.
DISCUSSION
The purpose of the studies reported here was to examine the possible contributions of the second EGF-like domain of factor IX in the binding of factor IXa to activated human platelets and in the assembly of the factor-X activating complex on the platelet membrane. To accomplish this goal we utilized a recombinant chimeric factor IX molecule in which amino acid sequences for factor X were substituted for the second EGF-like domain [8]. The assumption underlying this experimental approach is that the three-dimensional structure of the chimeric protein should be similar to the native factor IX molecule, whereas the cofactor and substrate specificity should be dependent on specific amino acid residues exposed on the surface of the protein. There is
Table 2 Factor X activation by normal, wild-type and chimeric factor IXa molecules: kinetic analysis kct. is expressed as mol of factor Xa formed/min per mol of total platelet-bound factor lXa. Data are shown for Scott-syndrome platelets [13] for comparison. Km
V,
kmt
kIt/Km
Platelets
Enzyme
(AM)
(nM min-)
(min-')
(A -1
Normal Normal Normal Scott-syndrome
Factor IXa(N) Factor lXa(X,e) Factor lXa(,) Factor IXa(N)
0.11 0.10 0.09 0.11
16.7 4.2 15.9 5.5
2319 2397 2078 2361
21 082
-
23970 23089 21 463
-
min-r)
Second epidermal-growth-factor domain in factor IXa binding to platelets
considerable support for this assumption, since the crystal structures of evolutionarily related proteins with conserved amino acid sequences have been shown to be similar [14,15] and the backbone structures of many serine proteases are nearly identical with sequence differences residing mainly in surface structures that mediate ligand interactions [16-18]. Therefore, we reasoned that any amino acid sequences within the EGF-2 domain of factor IX that are essential for binding to platelet receptors should be revealed by deficient binding of the chimeric
protein. The results of our previous studies with factor IXa(xegfl ) in which the first EGF-like domain of factor X replaced that of factor IXa, revealed entirely normal binding of the chimeric protein to thrombin-activated platelets in either the absence or the presence of factor VIlla and entirely normal kinetics of factor-X activation [3]. This result suggests either that the EGF1 domain of factor IX does not contain a platelet binding site, or that the EGF-l domain of factor X, when inserted into the factor IX molecule, suffices to support the assembly of factor IXa into the factor-X activating complex. The present results with factor IX(Xegf2), in which the second EGF-like domain of factor X replaced that of factor IXa, revealed abnormal binding of the chimeric protein to thrombinactivated platelets. Several interesting abnormalities were observed. First, the number of factor IXa binding sites was significantly decreased (P < 0.05) compared with normal or wild-type factor IXa. Second, although the affinity of binding was normal in the absence of factor Vllla (Kd = 3.11 nM; Table 1), the presence of factor VIIla had a sub-optimal effect on binding affinity (Kd = 1.4-1.7 nM, compared with 0.6-0.67 nM for wild-type or normal factor IXa; Table 1). Therefore, it would appear that the EGF-2 domain is essential for normal, highaffinity, factor IXa binding to activated platelets in the presence of factor Vllla and factor X. Third, the kinetic studies reported in Table 2 demonstrate that, although both chimeric and wildtype factor IXa molecules bind with normal Km to factor X, the V.,X of factor-X activation observed for factor IXa(xegf2) is only 25 % of normal, compared with 100 % for factor IXa(xegfl) [3]. However, since the kcat and kcat IKm for factor IXa(xegf2) are entirely normal (when kcat is expressed as mol of factor Xa formed/min per mol of platelet-bound factor IXa), the decreased rates of factor-X activation can be attributed entirely to the decreased amount and affinity of factor IXa bound to platelets in the presence of factor Vllla. All these results taken together are consistent with the interpretation that the EGF-2 domain of factor IXa is required for binding of factor IXa to a normal number of receptors exposed on activated platelets in the absence of factors VIIIa and X, and additionally is essential for the highaffinity (Kd 0.5 nM), factor-IXa binding and consequent factor-X activation observed in the presence of factor VIIIa. Based on these findings, we suggest that the EGF-2 domain of factor IXa contains an accessory binding site that recognizes a platelet receptor in addition to the platelet receptor with which residues 3-11 in the Gla domain interact [5]. The presence of factor Vllla failed to induce in factor IXa(xegf2) the high-affinity (Kd 0.5 nM) binding observed with factor IXaN or factor IXawt, although the kcat of the bound enzyme was entirely normal (Table 2), suggesting that, although the interaction of factor VIlla with factor IXa(Xegf1) was normal [3], the enhanced affinity of factor IXa binding through the putative EGF-2domain accessory binding site was deficient. Platelets in Scott syndrome are defective in their ability to convert factor X into Xa, which is due to the defect in mechanisms by which -
Received 17 January 1995/6 April 1995; accepted 24 April 1995
431
phosphatidylserine becomes exposed at the outer surface of stimulated platelets [19] and microvesicles [20]. Our results with the factor IXa(Xegf 2) chimera are strikingly similar to those observed previously with Scott-syndrome platelets [13], in which a bleeding disorder due to an isolated deficiency of platelet procoagulant activity was shown to be associated with a decreased number of factor IXa receptors in the absence or presence of factor Vllla and a deficient enhancement of factor IXa binding affinity in the presence of factor Vllla. The relevant binding and kinetic parameters are shown for Scott syndrome platelets in Tables 1 and 2. Taken together, these results are consistent with, although not proof of, the hypothesis that the EGF-2 domain contains a binding site that interacts with a platelet receptor that is deficient or absent in Scott-syndrome [13]. The enhanced high-affinity binding of factor IXa normally induced by the binding of factor Vllla to factor IXa and to activated platelets [6] may be deficient both with factor IXa(xegf 2) and with Scott-syndrome platelets, because factor VIlla promotes high-affinity factor IXa binding through the putative EGF-2 platelet (Scott) receptor interaction. This study was supported by research grants to P.N.W. from the National Institutes of Health (HL36579, HL46213, HL45486 and HL25661) and from the W. W. Smith Charitable Trust, to S.S.A. from The Council For Tobacco Research (Grant 3190), and to D.W.S. from the National Institutes of Health (HL38973). We are grateful to Frances S. Seaman and Stephanie Zhang for technical assistance and to Patricia Pileggi for her assistance in manuscript preparation.
REFERENCES 1 2 3 4
5 6 7
8 9
10 11
12
13 14 15 16 17
18 19
20
Hedner, U. and Davie, E. W. (1987) Hemostasis and Thrombosis: Basic Principles and Clinical Practice, 2nd edn., pp. 39-47, Lippincott, Philadelphia Ahmad, S. S., Rawala-Sheikh, R. and Walsh, P. N. (1992) Semin. Thromb. Hemostasis 18, 311-323 Ahmad, S. S., Rawala-Sheikh, R., Cheung, W-F., Stafford, D. W. and Walsh, P. N. (1992) J. Biol. Chem. 267, 8571-8576 Rawala-Sheikh, R., Ahmad, S. S., Monroe, D. M., Roberts, H. R. and Walsh, P. N. (1992) Blood 79, 398-405 Ahmad, S. S., Rawala-Sheikh, R., Cheung, W.-F., Stafford, D. W. and Walsh, P. N. (1994) Biochemistry 33, 12048-12055 Ahmad, S. S., Rawala-Sheikh, R. and Walsh, P. N. (1989) J. Biol. Chem. 264, 3244-3251 Ahmad, S. S., Rawala-Sheikh, R., Monroe, D. M., Roberts, H. R. and Walsh, P. N. (1990) J. Biol. Chem. 265, 20907-20911 Lin, S.-W., Smith, K. J., Welsch, D. and Stafford, D. W. (1990) J. Biol. Chem. 265, 144-150 Cheung, W.-F., Straight, D. L., Smith, K. J., Lin, S.-W., Roberts, H. R. and Stafford, D. W. (1991) J. Biol. Chem. 266, 8797-8800 Monroe, D. M., Sherrill, G. B. and Roberts, H. R. (1988) Anal. Biochem. 172, 427-435 Ahmad, S. S., Rawala-Sheikh, R. and Walsh, P. N. (1989) J. Biol. Chem. 264, 2001 2-20016 Rawala-Sheikh, R., Ahmad, S. S., Ashby, B. and Walsh, P. N. (1990) Biochemistry 29, 2606-2611 Ahmad, S. S., Rawala-Sheikh, R., Ashby, B. and Walsh, P. N. (1989) J. Clin. Invest. 84, 824-828 Barker, W. C. and Dayhoff, M. 0. (1976) Atlas of Protein Sequences and Structure, vol. 5, pp. 105-110, National Biomedical Research Foundation, Silver Spring, MD Craik, C. S., Rutter, W. J. and Fletterick, R. (1983) Science 220, 1125-1129 Rogers, J. (1985) Nature (London) 315, 458-459 Fersht, A. (1985) in Enzyme Structure and Mechanism (Fersht, A., ed.), pp. 405-426, W. H. Freeman and Co., New York Olsson, G., Lindqvist, 0., Peterson, T. E., Magnusson, S. and Sottrup-Jenson, L. (1980) in Vitamin K Metabolism and Vitamin K Dependent Proteins (Suttie, J. W., ed.), pp. 8-12, University Park Press, Baltimore Rosing, J., Bevers, E. M., Comfurius, P., Hemker, H. C., Dieijen, V., Weiss, H. J. and Zwaal, R. F. A. (1985) Blood 65, 1557-1561 Sims, P. J., Faioni, E. M., Wiedmer, T. and Shattil, S. J. (1988) J. Biol. Chem. 263, 18205-1 8209
