Activation of the Factor VIIIa-Factor IXa Enzyme Complex of Blood Coagulation by Membranes Containing Phosphatidyl-L-serine*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 271, No. 19, Issue of May 10, pp , by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Activation of the Factor VIIIa-Factor IXa Enzyme Complex of Blood Coagulation by Membranes Containing Phosphatidyl-L-serine* (Received for publication, August 22, 1995, and in revised form, February 8, 1996) Gary E. Gilbert and Andrew A. Arena From the Department of Medicine of the Brockton-West Roxbury VA Medical Center, the Department of Medicine of Brigham and Women s Hospital, and the Department of Medicine of Harvard Medical School, Boston, Massachusetts Factor IXa, a serine protease of blood coagulation, functions at least 100,000 times more efficiently when bound to factor VIIIa on a phospholipid membrane than when free in solution. We have utilized the catalytic activity of the factor VIIIa-factor IXa complex to report the effect of phospholipid membranes on binding of factor IXa to factor VIIIa and on enzymatic cleavage of the product. The apparent affinity of factor IXa for factor VIIIa was 10-fold lower in the absence of phospholipid membranes with a K D of 46 nm versus 4.3 nm with phospholipid membranes. The K m for activation of factor X by the factor VIIIa-factor IXa complex was 1700 nm in solution, 70-fold higher than the value of 28 nm when bound to membranes containing phosphatidyl-l-serine, phosphatidylethanolamine, and phosphatidylcholine at a ratio of 4:20:76. The largest effect of phosphatidyl-lserine-containing membranes on the factor VIIIa-factor IXa complex was the accelerated rate of peptide bond cleavage, with the k cat increased by 1,500-fold from to 33 min 1. Membranes in which phosphatidyl-l-serine was replaced by phosphatidyl-d-serine, phosphatidic acid, or phosphatidylglycerol were at least 10-fold less effective for enhancing the k cat. Thus, while membranes containing phosphatidyl-l-serine enhance condensation of the enzyme with its cofactor and substrate, their largest effect is activation of the assembled factor VIIIafactor IXa enzyme complex. Factor VIII is a phosphatidyl-l-serine (Ptd-L-Ser) 1 binding cofactor (1, 2) for the vitamin K-dependent serine protease, factor IXa, that also binds to Ptd-L-Ser containing membranes (3, 4). The membrane-bound factor VIIIa-factor IXa complex cleaves the zymogen, factor X, to factor Xa which is then responsible for catalyzing prothrombin activation (5). The importance of this enzyme complex is illustrated by hemophilia, a disease in which a deficiency of either factor VIII or factor IX leads to life-threatening bleeding. Factor IXa gains more than 100,000-fold greater efficiency in activating factor X by assembling with factor VIIIa on a Ptd-L-Ser containing membrane * This work was supported by National Institutes of Health, NHLBI Grant P01 HL42443I and the Medical Research Service of the Department of Veterans Affairs. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Recipient of National Institutes of Health Clinical Investigator Award HL To whom correspondence should be addressed: Brockton-West Roxbury VA Medical Center, 1400 VFW Parkway, West Roxbury, MA Tel.: ; Fax: ; 1 The abbreviations used are: Ptd-L-Ser, phosphatidyl-l-serine; PC, phosphatidylcholine; PE, phosphatidylethanolamine; dansyl, 5-dimethylaminonapthalene-1-sufonyl than when free in solution (6). While prior reports indicate that Ptd-L-Ser containing membranes decrease the K m of factor IXa for factor X (6, 7), they do not indicate what effect these membranes have upon the enzymatic parameters of the factor VIIIa-factor IXa complex. Therefore, we compared the enzymatic parameters of the factor VIIIa-factor IXa complex in solution to the parameters of the membrane-bound complex to determine whether Ptd-L-Ser containing membranes function to enhance assembly of the enzyme and cofactor, binding of the substrate to the enzyme, or catalysis of the substrate. EXPERIMENTAL PROCEDURES Bovine brain Ptd-L-ser, phosphatidylethanolamine (PE) synthesized by transphosphatidylation of egg PC, dimyristoyl PE, egg PC, dioleoyl PC, dioleoylphosphatidic acid, and dioleoyl phosphatidylglycerol were from Avanti Polar Lipids (Alabaster, AL). Cholesterol was from Aldrich. Recombinant human factor VIII was a gift from D. Pittman of Genetics Institute, Cambridge, MA. Factor IXa, factor X, and factor Xa were from Enzyme Research Laboratories (Southbend, IN). Thrombin was from Sigma. Activation of Factor VIII by Thrombin Recombinant human factor VIII was activated by thrombin as described elsewhere (8). The 0.5-ml sample was then diluted 1:1 with 0.3 M NaCl, 5 mm CaCl 2,25mM sodium acetate, acetic acid, ph 5.5. Inactivated thrombin and free D-Phe-Pro-Arg-CClH 2 (Calbiochem) were removed by ultrafiltration using a Centricon 100 micro concentrator (Amicon Inc., Beverly, MA) at 4 C. The residual volume, 100 l, was diluted to 2 ml in the same buffer, the procedure repeated twice, and the factor VIIIa diluted to 0.5 ml. The product contained only the three bands with migration correlating to the A1, A2, and A3C1C2 domain constituents of heterotrimeric factor VIIIa as evaluated by SDS-polyacrylamide gel electrophoresis with silver staining. Concentration was determined by absorbance at 280 nm using E 0.1% 1 cm (8). The factor VIIIa prepared in this way had factor X activating activity equivalent to factor VIII from the same lot which was activated at the same final concentration by 0.02 unit/ml thrombin at the time of assay initiation indicating that the factor VIIIa remained in tact through the ultrafiltration step of preparation. Preparation and Evaluation of Proteins The purity of proteins was evaluated by SDS-polyacrylamide gel electrophoresis with silver staining. All preparations used in these studies exhibited only the bands corresponding to those previously attributed to the respective proteins with the exception of a single contaminant band that co-migrated with bovine albumin standards. This contaminant was present as approximately 5% of total protein for the factor IXa, factor X, and factor Xa preparations. Factor X was contaminated by factor Xa at a level of approximately 1 part per 2,000 as judged by the rate of development of chromogenic substrate, S-2765 (Helena, Beaumont, TX). Therefore, stock factor X was incubated with 19 M dansyl-glu-gly-arg-cclh 2 (Calbiochem, San Diego, CA) for 80 min. at RT and dialyzed against 0.14 M NaCl, 0.05 M Tris/HCl, ph 7.5, yielding a product that did not cleave S-2765 at a rate above buffer at the highest concentrations used in these experiments. All proteins were aliquoted into fractions for single usage, flash-frozen in liquid nitrogen, and stored at 80 C until use. Phospholipid Vesicles and Phospholipid Synthesis Phospholipid vesicles were synthesized by extrusion through two stacked polycarbonate membranes of 0.1- m pore size (Nucleopore Corp.) under argon as described previously (9). Cholesterol was included in all vesicles in a 2:10 ratio to phospholipid to enhance membrane strength and ensure

2 that lipids were in the liquid crystalline state at room temperature. Phospholipid concentration was determined by phosphorous assay (10). Dioleoyl Ptd-L-Ser and dioleoyl Ptd-D-Ser were synthesized by transphosphatidylation of dioleoyl PC (11) and purified as described (2, 12). Factor Xase Assay Factor Xase activity was measured with a two step amidolytic substrate assay. Phospholipid vesicles were mixed with a reaction mixture containing factor IXa and factor X. The reaction was started by adding Ca 2 and factor VIIIa at final concentrations specified for each experiment taking care that 1 min. elapsed from dilution of concentrated factor VIIIa until reaction initiation. After 5 min. at 25 C for experiments with phospholipid vesicles, or 30 min for experiments without, the reaction was stopped by diluting the mixture 1:0.8 with 16 mm EDTA and factor Xa activity was determined immediately in a thermostatted kinetic microtiter plate reader (Molecular Devices, Menlo Pk, CA) at 25 C using 0.1 mm S A standard curve was prepared using pure factor Xa. The results displayed in the figures are duplicates from a representative experiment. Affinity data were analyzed according to the standard equilibrium binding model, using the quadratic equation when the apparent K D was within 5-fold of the limiting reactant and steady state kinetics according to the Michaelis- Menton model with curve fitting by nonlinear least squares analysis (FitAll, MTR Software, Toronto, Quebec, Canada). Activation of Factor VIIIa-Factor IXa Complex FIG. 1.Assembly of the factor VIIIa-factor IXa complex in the absence and presence of phospholipid membranes. A, when increasing quantities of factor IXa were added to 5 nm factor VIIIa ( ) the quantity of factor Xa formed increased saturably indicating formation of the factor VIIIa-factor IXa complex. When the factor VIIIa concentration was increased to 20 nm (f) the maximum rate of factor Xa formation increased 4-fold. Results were corrected for the quantity of factor Xa produced by factor IXa in the absence of VIII and the correction did not exceed forty percent of total factor Xa formed at the highest concentration of factor IXa. A representative experiment is depicted and the line represents a K D of 42 nm for the lower curve and 71 nm for the upper curve obtained by nonlinear least squares analysis of data. When varying concentrations of factor IXa were added to 1 nm factor VIIIa in the presence of phospholipid vesicles of 4% Ptd-L-Ser, 20% PE (B) or 25% Ptd-L-Ser (C) saturable binding of factor IXa to factor VIIIa was detected with apparent K D values of 6.6 nm and 4.0 nm, respectively. The reactions were allowed to proceed for 5 min in the presence of 25 M phospholipid (B and C) or 30 min in the absence of phospholipid (A). Preliminary experiments indicated that production of factor Xa was proportional to elapsed reaction time for 5 min in the presence of phospholipid and for at least 40 min in the absence of phospholipid. RESULTS We wished to determine whether catalytic enhancement of the factor VIIIa-factor IXa enzyme complex by Ptd-L-Ser containing membranes primarily reflects enhanced assembly of factor IXa with factor VIIIa versus enhanced binding of the substrate, factor X to the enzyme, factor IXa versus acceleration of peptide bond cleavage. We first asked whether the apparent affinity of factor IXa for factor VIIIa is influenced by Ptd-L-Ser-containing membranes. Saturable binding of factor IXa to 5 nm factor VIIIa was detected by increased catalytic efficiency of factor IXa toward factor X (Fig. 1A, lower curve). The apparent K D was 42 nm, approximately 10-fold higher than prior measurements in the presence of phospholipid membranes. When the factor VIIIa concentration was increased 4-fold to 20 nm the maximum catalytic rate increased 4-fold confirming that factor IXa assembles with available factor VIIIa to form a complex with enhanced catalytic activity in the absence of phospholipid. The average K D obtained from five experiments was 46 nm (Table I). For comparison we examined the binding of factor IXa to 1 nm factor VIIIa in the presence of vesicles of 4% Ptd-L-Ser, 20% PE and 25% Ptd-L-Ser (Fig. 1, B and C). The apparent K D was 6.6 nm for vesicles with 4% Ptd-L-Ser and 4 nm for vesicles with 25% Ptd-L-Ser. The mean dissociation constants were 4.3 nm for 5 such experiments using vesicles of 4% Ptd-L-Ser and 2.3 for 3 experiments with vesicles of 25% Ptd-L-Ser (Table I) in agreement with prior measurements under similar conditions (13). These results indicate that Ptd-L-Ser-containing membranes enhance the affinity of factor IXa for factor VIIIa by fold in the presence of the substrate, factor X. We next determined the concentration of phospholipid vesicles that would enhance cleavage of factor X by the factor VIIIa-factor IXa complex (Fig. 2). We have recently observed that PE induces binding sites for factor VIII in membranes with low mole fractions of Ptd-L-Ser (2)and we compared these membranes to those containing 25% Ptd-L-Ser without PE, similar to those used in prior studies of the enzymatic parameters for this complex. In contrast to our prior studies we utilized extruded vesicles rather than sonicated vesicles and 1 mm Ca 2 rather than 5 mm Ca 2, attempting to better simulate the curvature of the platelet membrane versus maximally curved sonicated vesicles and to better approximate the Ca 2 concentration in plasma. Vesicles of 25% Ptd-L-Ser were as much as 10-fold more effective than those with 4% Ptd-L-Ser and PE when the phospholipid concentrations less than 20 M and the increased activity of the factor VIIIa-factor IXa complex plateaued at a concentration of 32 M. When the phospholipid concentration was increased to 125 M the difference between membrane types decreased to 2-fold and the activity related to vesicles of 4% Ptd-L-Ser with PE had not yet plateaued. When the Ca 2 was increased to 5 mm and vesicles were prepared by sonication the two vesicle type were equivalent at phospholipid concentrations of 2 M or above as we previously reported (2). These results indicate that the activity of the factor VIIIa-factor IXa complex is greatly increased by membranes containing Ptd-L-Ser and is modestly influenced by the curvature of the membranes. They are consistent with prior results indicating that membrane binding of factors IXa and factor X to membranes containing Ptd-L-Ser is enhanced by

3 11122 Activation of Factor VIIIa-Factor IXa Complex TABLE I Effect of phosphatidylserine-containing membranes on assembly and enzymatic function of the factor VIIIa-factor IXa complex Phospolipid type K D App a K M b k cat k cat /K M nm min 1 X10 7 M 1 min 1 No Phospholipid c d Ptd-L-Ser:PC, e 23 1 f :75 Ptd-L-Ser:PE:PC, 4:20: g 28 6 h a The rate of factor Xa formation was measured with a chromogenic substrate as a function of [factor IXa] in the presence of factor VIIIa, 65 nm factor X, and either zero or 25 M phospholipid. Rate values were fitted to a standard binding model to obtain the apparent K D. b The rate of factor Xa formation was measured as a function of [factor X] in the presence of 5 nm factor IXa (no phospholipid) or 0.01 nm factor IXa, 25 nm factor VIIIa, and zero or 25 M phospholipid. c Mean S.D. for five experiments, each performed in duplicate. In three experiments the factor VIIIa concentration was 5 nm and in two experiments it was 20 nm. One experiment with 20 nm factor VIIIa yielded an apparent K D of 140 nm and was excluded from analysis as an outlier. d Mean S.D. for three experiments, each performed in duplicate. e Mean S.D. for three experiments, each performed in duplicate. f Fitted K M standard error of fit for one experiment performed in duplicate. g Mean S.D. for five experiments, each performed in duplicate. h Mean standard error of fit for one experiment performed in duplicate. FIG. 3.Steady state kinetics of the factor VIIIa-factor IXa complex in the absence and presence of phospholipid membranes comparable. Factor IXa was incubated with factor VIIIa, varying concentrations of factor X and no phospholipid (A) or25 Mphospholipid (B) of 25% Ptd-L-Ser ( ) or 4% Ptd-L-Ser and 20% PE (å). Lines indicate best fits of the data sets. Duplicates were averaged and shown as single points for clarity of presentation in B. The factor IXa concentrations were 5 nm (A) and 0.01 nm (B); factor VIIIa was 120 nm. The reactions proceeded for 30 min in the absence of phospholipid (A) or 5 min in the presence (B). FIG. 2. Activation of the factor VIIIa-factor IXa complex by Ptd-L-Ser-containing membranes. Various quantities of phospholipid vesicles were added to a mixture of factor IXa, factor VIIIa and factor X and the quantity of factor X formed was evaluated. Vesicles contained either 25% Ptd-L-Ser ( ) or 4% Ptd-L-Ser, 20% PE (f) with the balance as PC. Protein concentrations were as described in the legend for Fig. 1B. elevating the Ca 2 concentrations above the plasma concentration to 5 mm (6, 14, 15). We next performed steady state kinetic experiments to determine whether Ptd-L-Ser-containing membranes primarily affect the K m versus the k cat. Conditions were chosen such that more than 70% of factor IXa would be bound to factor VIIIa. The rate at which the factor VIIIa-factor IXa complex activated factor X increased saturably with the factor X concentration in the presence and the absence of phospholipid vesicles (Fig. 3). The Michaelis constant for the factor VIIIa-factor IXa complex was 1700 nm in the absence of phospholipid vesicles (Fig. 3A), versus 23 nm in the presence of vesicles containing 4% Ptd-L-Ser and PE (Fig. 3B and Table I). In contrast, the k cat increased 1,500-fold from min 1 in the absence of phospholipid to 33 min 1 in the presence of these vesicles (Table I). Similarly, the k cat increased to 136 min 1 in the presence of membranes containing 25% Ptd-L-Ser and no PE. These results indicate that Ptd-L-Ser-containing membranes enhance the efficiency of the factor VIIIa-factor IX enzyme complex by increasing the affinity for the substrate and by increasing the rate of peptide bond cleavage. To place our results in context with prior evaluations of the kinetics of the factor VIIIa-factor IXa complex (6, 7) we also performed steady state kinetics at a Ca 2 concentration of 5 mm (data not shown). The apparent affinity of factor IXa for factor VIIIa was affected less than 2-fold in the absence or presence of both types of vesicles. Likewise, the K m was affected less than 2-fold in the presence or absence of vesicles. The k cat was approximately 10-fold higher at 5 mm Ca 2 for vesicles of 4% Ptd-L-Ser, 20% PE and 3-fold higher for vesicles of 25% Ptd-L-Ser with values comparable to prior reports (6, 7). The difference between the enhancement with these two vesicle types may be rationalized by noting that concentrations of Ca 2 above 1 mm enhance binding of factor IXa and factor X to vesicles of low PS content more than to those with high PS content (14, 16). In the absence of phospholipid vesicles the k cat was approximately 3-fold higher at 5 mm Ca 2 than at 1 mm Ca 2. Thus, elevation of the Ca 2 concentration from 1 to 5 mm has modest effects on the kinetic parameters of the factor VIIIa-factor IXa complex bound to a membrane site. To determine whether Ptd-L-Ser is a specific activator of the factor VIIIa-factor IXa complex or whether activation could be

4 Activation of Factor VIIIa-Factor IXa Complex TABLE II Effect of Ptd-L-Ser containing membranes on the factor Xase and prothrombinase enzyme complexes -Fold increase due to membrane Parameter binding Prothrombinase Factor Xase Affinity, enzyme for cofactor 800 (19) a 10 (Table I) Affinity, substrate for enzyme alone 240 (31) 100 (6) Affinity, substrate for enzyme complex 170 (31) 70 (Table I) k cat enzyme alone 6 (31) 25 (6) k cat enzyme-cofactor complex 7 (31) 1,500 (Table I) a Numbers in parentheses are references. FIG. 4. Ptd-L-Ser is a specific activator of the factor VIIIafactor IXa complex. The capacity of membranes containing 4% of the indicated negatively charged dioleoyl phospholipid, 20% dimyristoyl PE, with the balance as phosphatidylcholine were compared for their ability to activate the factor VIIIa-factor IXa complex at 5 mm Ca 2.(f) Membranes containing 4% dioleoyl Ptd-L-Ser activated the complex efficiently at 120 nm. (å) Dioleoyl phosphatidic acid, with a negative valence twice that of Ptd-L-Ser, was at least 25-fold less effective. ( ) Dioleoyl Ptd-D-Ser and ( ) dioleoyl phosphatidylglycerol were at least 50-fold less effective than dioleoyl Ptd-L-Ser. When the Ca 2 concentration was 1 mm (inset) the effective phospholipid concentrations were apparently 5-fold higher. Displayed results are the mean of duplicate samples from two experiments, representative of six experiments. induced equivalently by other negatively charged phospholipids we compared membranes containing 4% Ptd-L-Ser to those containing 4% Ptd-D-Ser, 4% Phosphatidic acid, or 4% phosphatidylglycerol (Fig. 4). We compared activation at 5 mm Ca 2 (main graph) and1mmca 2 (inset). While Ptd-L-Ser containing membranes activated the factor VIIIa-factor IXa complex, the diastereomer, Ptd-D-ser was at least 10-fold less effective. Likewise, phosphatidic acid and phosphatidylglycerol-containing membranes were at least 10-fold less effective than Ptd-L- Ser. Activation by Ptd-L-Ser occurred at a 5-fold lower phospholipid concentration in the presence of 5 mm Ca 2 compared to1mmca 2. Unfortunately, the quantity of dioleoyl Ptd-L-Ser and dioleoyl Ptd-D-Ser available from our synthesis prevented experiments at 5 10 fold higher concentrations apparently necessary to activate all of the factor VIIIa-factor IXa complexes. These results indicate that when negatively charged phospholipids are present as a low mole fraction of the membrane the factor VIIIa-factor IXa complex is activated by a stereoselective interaction with Ptd-L-ser. DISCUSSION Our results indicate that Ptd-L-Ser-containing membranes primarily activate the assembled factor VIIIa-factor IXa complex, enhancing the chemical step of peptide bond cleavage more than the physical steps of enzyme-cofactor or enzymesubstrate binding. The specificity with which Ptd-L-Ser containing membranes activate the factor VIIIa-factor IXa complex parallels the specificity with which Ptd-L-Ser containing membranes bind factor VIII (1, 17) which suggests that membrane binding of factor VIIIa is a primary determinant of enzyme complex activity. The studies described in this report are the first to elucidate the kinetic properties of the factor VIIIa-factor IXa complex in the absence of phospholipid membranes or cells. We find that the k cat for the complex in solution is min 1 compared with values ranging from 0.01 and 0.07 min 1 previously reported for factor IXa in the absence of factor VIII (6, 7). Therefore, formation of the factor VIIIa-factor IXa complex enhances the k cat not more than 2-fold. However the K m for factor IXa alone is M (6, 7) versus 1.7 M for the factor VIIIafactor IXa complex. Thus, binding to factor VIIIa increases the affinity of factor IXa for its substrate, factor X, by at least 40-fold. We note that binding of factor IXa to either factor VIIIa or a Ptd-L-Ser containing membrane affects the K m for factor X to approximately the same degree (Table I) (6, 7). This raises the possibility that the mechanism for decreasing the K m in both cases may be a similar conformational change in factor IXa that increases the affinity for factor X. The effect of Ptd-L-Ser-containing membranes on the prothrombin-activating factor Va-factor Xa complex contrasts with the effects on the factor VIIIa-factor IXa complex (Table II). Ptd-L-Ser-containing membranes do enhance the affinity of the enzyme, factor Xa, for the cofactor, factor Va by fold while a 10-fold effect is detected for factor IXa in the factor VIIIa-factor IXa complex (18, 19). The k cat for the factor Vafactor Xa complex is increased by only about 5-fold when bound to a Ptd-L-Ser-containing membrane compared with 1,500-fold for the factor VIIIa-factor IXa complex. Thus, although Ptd-L- Ser-containing membranes increase the catalytic efficiency of the prothrombinase complex and the factor VIIIa-factor IXa complex to a similar degree the primary effect on the prothrombinase complex is condensation of the enzyme, the cofactor, and the substrate on the same site. In contrast, the primary effect of Ptd-L-Ser-containing membranes on the factor VIIIa-factor IXa complex is to increase the k cat of the enzyme-cofactorsubstrate complex which assembles with high affinity in the absence of membranes. This report helps to explain the importance of Ptd-L-Ser as a component of binding sites for the factor VIIIa-factor IXa complex in that binding to this lipid alters the complex by a 1,500- fold enhancement in catalytic activity. Platelets develop procoagulant activity in parallel with the re-orientation of Ptd-L-ser and PE from the inner to the outer bilayer of the plasma membrane (20). Under the same conditions that lead to Ptd-Lser and PE re-orientation, platelets express specific receptors/ binding sites for factor VIII and support function of factor VIII in the factor Xase complex (21, 22). When platelets are stimulated by agonists that induce procoagulant activity, they release small vesicles derived from the plasma membrane (23 25). These vesicles, also referred to as microparticles, have a high density of membrane receptors/binding sites for factor VIII (22). In vitro data support the hypothesis that Ptd-L-Ser on the platelet membrane is a constituent of binding sites/receptors for factor VIII. The affinity of factor VIII binding to activated platelets and to microparticles is equivalent to the affinity of binding to synthetic membranes containing Ptd-L-ser (21, 22, 26). Binding sites containing Ptd-L-ser, like those of activated platelets, are highly specific for factor VIII (17). The specificity is mediated by a stereoselective interaction of factor

5 11124 Activation of Factor VIIIa-Factor IXa Complex VIII with O-phospho-L-serine, the head group of Ptd-L-ser (1). Furthermore, inclusion of PE induces high affinity binding sites for factor VIII in synthetic membranes with a low mol fraction of Ptd-L-Ser, comparable to platelet membranes (2). High affinity binding of factor IXa to factor VIII in the absence of a Ptd-L-Ser-containing membrane was recently reported by Lenting et al. (27). The binding energy for this interaction comes primarily from interaction of the light chain of factor VIII with factor IXa. Prior studies (13) indicated that the affinity of factor IXa for factor VIII is not affected by proteolytic activation of factor VIII. However, proteolytic activation releases factor VIII from von Willebrand factor in plasma, making the binding to factor IXa possible. The studies in the present report imply a 4-fold lower affinity of factor IXa for factor VIIIa in the absence of phospholipids than previously reported; a K D of 46 nm versus 11 nm. The difference in the absence of phospholipid may reflect the different techniques employed or it may indicate that the interaction of factor X with both factor IXa and factor VIIIa decreases the apparent affinity of each protein for the other. The enzyme constants in this report are in agreement with prior measurements for the membrane-bound factor VIIIa-factor IXa complex (6, 7, 28). Of the three prior reports, two did not include experiments for the factor VIIIa-factor IXa complex in the absence of phospholipid membranes (6, 28) and the other reported that factor VIIIa and factor IXa, in the absence of phospholipid, had parameters equivalent to factor IXa alone (7). The experimental design employed in that study may explain why no significant effect of factor VIIIa on factor IXa was detected. First, factor VIII, which was used at a 1:5 ratio to factor IXa, was activated by thrombin prior to mixing with factor IXa. This allowed some time for the unstable factor VIIIa molecule to decompose prior to mixing and probably further reduced the factor VIIIa:factor IXa ratio. Second, the lowest concentration of factor X examined was 5 M, more than 3-fold greater than the K m for the factor VIIIa-factor IXa complex. At this substrate concentration the quantity of factor Xa generated by free factor IXa which was present at a 4:1 ratio to the factor VIIIa-factor IXa complex would be substantial. Thus, in the presence of these factor X concentrations the effect of the factor VIIIa-factor IXa complex may have been small compared to the effect of free factor IXa. We report that the apparent affinity of human factor IXa for human factor VIIIa in the presence of Ptd-L-Ser containing membranes is 5 10-fold higher than previously reported (8). The discrepancy likely relates to the differences in experimental design and is rationalized by considering the instability of free factor VIIIa which dissociates into inactive subunits with a half-life of two minutes versus factor IXa-bound factor VIIIa which decays with a half-life of 20 min (8). Our reported values are derived from experiments in which varying factor IXa was added to a fixed concentration of factor VIIIa. This design allows for factor IXa to stabilize factor VIIIa at all concentrations where it is in excess over factor VIIIa. The prior report utilized fixed concentrations of factor IXa and increasing factor VIIIa. In this design all factor VIIIa that is in excess over factor IXa decays rapidly so that the equilibrium process favors a large fraction of dissociated factor VIIIa. The plausibility of this explanation is supported by the reported stoichiometry of human factor VIIIa to factor IXa, 3:1 versus the stoichiometry of the more stable porcine factor VIIIa to factor IXa 0.8:1 associated with a 10-fold higher apparent affinity. To confirm the adequacy of this explanation we performed experiments in which increasing concentrations of factor VIIIa were added to a fixed concentration of factor IXa. This approach yielded an apparent affinity 3 5-fold lower than the results in Table I and within 2-fold of the previously reported value (data now shown). There are four possible mechanisms through which the Ptd- L-Ser containing membranes may activate the factor VIIIafactor IXa complex. First, the membranes may have an allosteric effect on factor IXa. Second, they may work by altering the configuration of the factor IXa-factor VIIIa complex. Third they may affect the conformation of the substrate, factor X, which is also a membrane-binding protein. Finally, membrane lipids may alter the geometry of the interaction between factor X and the factor VIIIa-factor IXa complex. The first mechanism, an allosteric effect upon factor IXa is consistent with two prior reports indicating that Ptd-L-Ser-containing membranes enhance the k cat with which factor IXa cleaves factor X in the absence of factor VIIIa by fold (6, 7). Because factor VIIIa is also required to achieve the 1,500-fold enhancement in k cat it may provide additional constraints upon the factor IXa conformation. The plausibility of this explanation is enhanced by a report indicating that modified factor X with an abbreviated activation peptide has a normal K m for interaction with the membrane-bound factor VIIIa-factor IXa complex but a 100-fold reduced k cat (28). The difference in k cat was only 3-fold for membrane-bound factor IXa in the absence of factor VIIIa. These data indicate that factor X interacts with a factor IXa exosite which influences the k cat but not the K m and that interaction with the exosite is modified when factor IXa binds to both factor VIIIa and Ptd-L-Ser containing membranes. A conformational change of this nature would resemble the change in factor Xa when bound to factor Va on Ptd-L-Ser containing membranes (29) that occurs only if the membrane lipids contain unsaturated acyl chains (30). Thus, the activation peptide-binding exosite is probably the best candidate site to undergo a conformational change enhancing the k cat. If this interpretation is correct then the factor VIIIa-factor IXa complex should also be activated by Ptd-L-Ser and possibly other phospholipids in micelles as opposed to bilayers, and possibly by soluble, short-chain Ptd-L-Ser molecules. We are currently investigating these possibilities. This explanation may be further probed by examining the effect of Ptd-L-Ser-containing membranes upon cleavage of substrates that do not interact with the membrane. Acknowledgments We are grateful to Dr. Kenneth Bauer and to Dr. Arnaldo Arbini for reading the manuscript. REFERENCES 1. Gilbert, G. E., and Drinkwater, D. (1993) Biochemistry 32, Gilbert, G. E., and Arena, A. A. (1995) J. Biol. Chem. 270, Mertens, K., Cupers, R., Van Wijngaarden, A., and Bertina, R. M. (1984) Biochem. J. 223, Beals, J., and Castellino, F. (1986) Biochem. J. 236, Mann, K. G., Nesheim, M. E., Church, W. R., Haley, P., and Krishnaswamy, S. (1990) Blood 76, van Dieijen, G., Tans, G., Rosing, J., and Hemker, H. (1981) J. Biol. Chem. 256, Rawala-Sheikh, R., Ahmad, S., Ashby, B., and Walsh, P. (1990) Biochemistry 29, Curtis, J. E., Helgerson, S. L., Parker, E. T., and Lollar, P. (1994) J. Biol. Chem. 269, Hope, M. J., Bally, M. B., Webb, G., and Cullis, P. R. (1985) Biochim. Biophys. Acta 812, Chen, P., Toribara, T., and Warner, H. (1956) Anal. Chem. 28, Comfurius, P., Bevers, E. M., and Zwaal, R. F. A. (1990) J. Lipid Res. 31, Comfurius, P., and Zwaal, R. (1977) Biochim. Biophys. Acta 488, Duffy, E., Parker, E., Mutucumarana, V., Johnson, A., and Lollar, P. (1992) J. Biol. Chem. 267, Nelsestuen, G. L., and Broderius, M. (1977) Biochemistry 16, Nelsestuen, G. (1976) J. Biol. Chem. 251, Nelsestuen, G. L., Kisiel, K., and Di Scipio, R. G. (1978) Biochemistry 17, Gilbert, G. E., Drinkwater, D., Barter, S., and Clouse, S. B. (1992) J. Biol. Chem. 267, Skogen, W. F., Esmon, C. T., and Cox, A. C. (1984) J. Biol. Chem. 259,

6 Activation of Factor VIIIa-Factor IXa Complex Pryzdial, E. L., and Mann, K. G. (1991) J. Biol. Chem. 266, Bevers, E., Comfurius, P., and Zwaal, R. (1983) Biochim. Biophys. Acta 736, Nesheim, M. E., Pittman, D. D., Wang, J. H., Slonosky, D., Giles, A. R., and Kaufman, R. J. (1988) J. Biol. Chem. 263, Gilbert, G. E., Sims, P. J., Wiedmer, T., Furie, B., Furie, B. C., and Shattil, S. J. (1991) J. Biol. Chem. 266, Wolf, P. (1967) Br. J. Haematol. 13, Sandberg, H., Bode, A., Dombrose, F., Hoechli, M., and Lentz, B. (1985) Thromb. Res. 39, Sims, P., and Wiedmer, T. (1986) Blood 68, Gilbert, G. E., Furie, B. C., and Furie, B. (1990) J. Biol. Chem. 265, Lenting, P. J., Donath, M., Vanmourik, J. A., and Mertens, K. (1994) J. Biol. Chem. 269, Duffy, E., and Lollar, P. (1992) J. Biol. Chem. 267, Krishnaswamy, S., Vlasuk, G. P., and Bergum, P. W. (1994) Biochemistry 33, Kung, C., Hayes, E., and Mann, K. G. (1994) J. Biol. Chem. 269, Rosing, J., Tans, G., Govers-Riemslag, J., Zwaal, R., and Hemker, H. (1980) J. Biol. Chem. 255,