Importance of protein S and phospholipid for activated protein C-mediated cleavages in factor Va.

The procoagulant function of activated factor V (FVa) is inhibited by activated protein C (APC) through proteolytic cleavages at Arg306, Arg506, and Arg679. The effect of APC is potentiated by negatively charged phospholipid membranes and the APC cofactor protein S. Protein S has been reported to selectively stimulate cleavage at Arg306, an effect hypothesized to be related to reorientation of the active site of APC closer to the phospholipid membrane. To investigate the importance of protein S and phospholipid in the APC-mediated cleavages of individual sites, recombinant FV variants FV(R306Q/R679Q) and FV(R506Q/R679Q) (can be cleaved only at Arg506 and Arg306, respectively) were created. The cleavage rate was determined for each cleavage site in the presence of varied protein S concentrations and phospholipid compositions. In contrast to results on record, we found that protein S stimulated both APC cleavages in a phospholipid composition-dependent manner. Thus, on vesicles containing both phosphatidylserine and phosphatidylethanolamine, protein S increased the rate of Arg306 cleavage 27-fold and that of Arg506 cleavage 5-fold. Half-maximal stimulation was obtained at approximately 30 nm protein S for both cleavages. In conclusion, we demonstrate that APC-mediated cleavages at both Arg306 and Arg506 in FVa are stimulated by protein S in a phospholipid composition-dependent manner. These results provide new insights into the mechanism of APC cofactor activity of protein S and the importance of phospholipid composition.

Activated factor V (FVa) 1 is a procoagulant cofactor to activated factor X (FXa) in the conversion of prothrombin to thrombin (1)(2)(3). The procoagulant function of FVa is down-regulated by the anticoagulant serine protease activated protein C (APC) (2,4). APC cleaves FVa at three sites, Arg 306 , Arg 506 , and Arg 679 (5). The Arg 506 site is kinetically most favored, and cleavage results in decreased affinity for FXa and partial loss of procoagulant activity (6). Cleavage at Arg 306 impairs the procoagulant function more severely, and the combined cleavage at Arg 306 and Arg 506 results in the dissociation of the A2 fragment and complete loss of FVa activity (7). The role of the Arg 679 cleavage in the regulation of FVa activity is not fully elucidated, but appears to be of minor importance. Recently, we found that, in the presence of protein S, the Arg 679 cleavage might play a role in the degradation of naturally occurring FVa variants that are mutated at Arg 306 , such as FV Cambridge and FV Hong Kong (8).
Efficient inactivation of FVa by APC requires the presence of negatively charged phospholipids, with phosphatidylserine (PS) being the most potent stimulator of the APC cleavage of FVa (9). In addition, incorporation of phosphatidylethanolamine (PE) into the phospholipid membrane has been found to enhance the inactivation of FVa by APC (10). This was proposed to be a selective effect on the FVa inactivation because the effect of PE on the APC cleavage of FVa was more pronounced than what had been reported for other coagulation complexes (11,12). Other membrane components that have been reported to increase the efficiency of APC cleavage of FVa include cardiolipin and glycolipids (13,14). The phospholipid membrane fluidity, which is determined by the presence of polyunsaturated fatty acids, also affects FVa inactivation, with increased degradation being seen at higher degrees of polyunsaturation (15). Lipid oxidation is also associated with an increased rate of FVa inactivation (16). So far, studies on how the individual FVa cleavage sites are affected by the membrane composition have not been performed.
Protein S works as a cofactor to APC in the degradation of FVa (17). It has been reported that protein S selectively enhances the APC-mediated cleavage at Arg 306 , whereas it has no effect on cleavage at Arg 506 (18). The mechanism by which protein S works as a cofactor for APC and the molecular background for the selective stimulation of the Arg 306 cleavage are not fully understood. Protein S has been shown to stimulate APC binding to the phospholipid surface (19). Moreover, protein S has been shown to decrease the distance of the active site of APC from the membrane surface, an effect that is speculated to be related to the selective stimulation of the Arg 306 cleavage (20). Another suggested function of protein S is to annihilate the FXa-mediated protection of FVa against APC degradation (21), but this has been challenged by other groups (18). Initially, the protein S effect was reported to be dependent on the phospholipid concentration, with protein S having no effect at saturating concentrations of phospholipids (19). Later studies have claimed that, although protein S requires the presence of negatively charged phospholipids, its effect is independent of phospholipid concentration and composition (9).
The aim of this investigation was to elucidate how phospholipid membrane composition and protein S affect the individual APC cleavage sites. For this purpose, we constructed recombinant FV(R306Q/R679Q) and FV(R506Q/R679Q) variants, which can be cleaved only at Arg 506 and Arg 306 , respectively. The FVa inactivation was monitored by a prothrombinase-based assay, and apparent second-order rate constants for the Arg 306 and Arg 506 cleavages were calculated. Surprisingly, we found that protein S enhanced cleavage not only at Arg 306 , but also at Arg 506 . The protein S effect was dependent on the presence of PS and stimulated by the inclusion of PE in the membrane. These results provide new insights into the importance of membrane composition for APC-mediated degradation of FVa and the mechanism of protein S activity.
Phospholipid Vesicle Preparation-The phospholipid stocks were dissolved in 10:90 methanol/chloroform solution, and the concentrations were determined by phosphate analysis (27). Mixtures of the lipids (weight-based) were prepared in 10:90 methanol/chloroform and kept at Ϫ20°C. Aliquots were taken from the stocks, dried under N 2 , and then resuspended in Hepes at room temperature. Phospholipids for the prothrombinase assay were sonicated in a Misonix XL 2020 sonicator at amplitude 3 for 10 min. Fresh phospholipid vesicles were prepared every day. For the inactivation assay, extruded phospholipid vesicles were used to avoid heterogeneity in the size of the vesicles (10). The extrusion was performed using a LiposoFast basic extruder (Armatis, Mannheim, Germany) as described (28). The phospholipid mixtures dissolved in buffer were subjected to freeze-thaw circles and subsequently extruded 19 times through a membrane with a 100-nm pore size. The extruded phospholipids were used for 2 days.
Expression and Quantification of Recombinant FV Variants-The recombinant FV(R306Q/R679Q) and FV(R506Q/R679Q) variants were constructed as previously described (8). The recombinant proteins were transiently expressed in COS-1 cells using the DEAE-dextran transfection method (29) with minor modifications. Briefly, FV cDNA in the pMT2 vector was mixed with Tris (pH 7.3), 0.1 mM chloroquine, and DEAE-dextran in Dulbecco's modified essential medium (Invitrogen, Paisley, Scotland) and incubated for 4 h. The cells were thereafter shocked with 10% Me 2 SO for 2 min. The proteins were harvested in serum-free medium (Opti-MEM) 60 -70 h after transfection and concentrated in Vivaspin with a molecular weight cutoff of 100,000. Aliquots were frozen at Ϫ80°C. The concentrations of the recombinant proteins were determined by both enzyme-linked immunosorbent assay and prothrombinase assay. The FV-specific enzyme-linked immunosorbent assay was performed as described (8).
Partial Purification of Recombinant FVa-Conditioned medium (400 ml) from FV transfections was applied to a 25-ml Q-Sepharose fast flow column equilibrated with 25 mM Hepes and 5 mM CaCl 2 (pH 7.5) (buffer A) at 4°C. Prior to the application, the medium was first diluted three times in buffer A to lower the ionic strength, and 10 mM benzamidine was added to inhibit degradation during the purification procedure. After loading, the column was washed with 50 ml of buffer A supplemented with 0.1% Triton X-100 at a flow rate of 2 ml/min, followed by a wash with 70 ml of buffer A without the detergent. A 25-ml 0 -1 M NH 4 Cl gradient in buffer A was applied at a flow rate of 1 ml/min, and the fractions were screened for FVa activity with the prothrombinase assay. The buffer in the peak fractions was exchanged with 25 mM Hepes and 150 mM NaCl (pH 7.7) with 0.5 mg/ml bovine serum albumin and 5 mM CaCl 2 (buffer B) using a PD-10 column. The phospholipid content before and after purification was measured with a phospholipid B kit (Wako Bioproducts), which detects choline-containing phospholipids.
FVa Inactivation Assay-FV was incubated with thrombin (0.5 units/ ml) for 10 min at 37°C in buffer B. After activation of FV, phospholipid vesicles (final concentration of 25 M) and protein S were added, and a subsample was taken from the mixture and diluted 1:5 in ice-cold buffer B. APC was subsequently added. The amounts of APC and protein S added varied in the different experiments as indicated in the description of the individual experiments. At different time points, samples were taken from the inactivation mixture and diluted 1:5 in ice-cold buffer B to stop the reaction. The FVa activities in the diluted samples were then measured with the prothrombinase assay for the remaining FVa activity.
Equations Used for Curve Fitting-Inactivation of recombinant FVa(R506Q/R679Q) and FVa(R306Q/R679Q) was followed in time to calculate pseudo first-order rates for Arg 306 and Arg 506 cleavages by APC, respectively. The inactivation curves obtained were fitted to an equation already reported (18). The equation was modified due to the fact that only one cleavage occurs in our FV variants. For calculation of cleavage at position 506, the time curves obtained for FVa(R306Q/ R679Q) were fitted to Equation 1, in which Va t is the cofactor activity determined at time t, Va 0 is the cofactor activity determined before APC is added, B is the remaining procoagulant cofactor activity of FVa cleaved at position 506, and k 506 is the rate constant of cleavage at position 506. The residual activity B was determined from time curves where complete cleavage at Arg 506 had been accomplished, i.e. in the presence of 10:20:70 PS/PE/PC and protein S. The complete inactivation was obvious from the fact that prolonged incubation with APC did not lead to more loss of activity, i.e. the time curve leveled off at the residual FVa activity. Additionally, the validity of the B value was investigated in experiments using increasing concentrations of APC. For calculation of cleavage at position 306, Equation 2 was used, and fitted to the time curve of FVa(R506Q/R679Q). Here, C is the remaining procoagulant cofactor activity of FVa cleaved at position 306, and k 306 is the rate constant of cleavage at position 306. The residual activity C was determined in a manner similar to that described above for cleavage at position 506. Use of the equations requires that the inactivation curves are independent of FVa concentration (percent FVa inactivation versus time) and that the rates are linear for APC concentration. Control experiments were performed, and the inactivation curves fulfilled these criteria. Western Blot Analysis of Recombinant FV Variants-Recombinant and plasma-derived human FV were incubated with 0.3 units/ml thrombin for 10 min at 37°C to activate FV to FVa. FVa (final concentration of 0.8 nM) was subsequently incubated with APC (at the concentration indicated for each individual experiment) with or without 100 nM protein S in the presence of 25 M phospholipids (10:20:70 PS/PE/PC). At different time points, the inactivation was stopped by addition of denaturing solution. The samples were run under reducing conditions on a 10% SDS-polyacrylamide gel and transferred to polyvinylidene difluoride membranes. A monoclonal antibody against the heavy chain of FVa (AHV-5146) was used to detect the FVa fragments, and the Western blots were developed using a Vectastain Elite ABC kit (Vector Labs, Inc., Burlingame, CA) according to the manufacturer's instructions.
FXa Titration of APC-cleaved FVa(R306Q/R679Q) and FVa(R506Q/ R679Q)-To determine the FXa concentration dependence of the residual activities of Arg 306 -and Arg 506 -cleaved FVa, a FXa titration experiment was performed using the prothrombinase assay as described (30). Briefly, wild-type FV, FV(R506Q/R679Q), and FV(R306Q/R679Q) were activated by thrombin. FVa(R506Q/R679Q) and FVa(R306Q/R679Q) were subsequently incubated for 10 min at 37°C with 25 M phospholipid vesicles (10:20:70 PS/PE/PC, natural extracts) and 0.05 nM APC for FVa(R306Q/R679Q) and 0.2 nM APC plus 100 nM protein S for FVa(R506Q/R679Q). The FVa variants were subsequently diluted to a final concentration of 32 pM, and FXa was added at concentrations ranging from 0.1 to 50,000 pM together with 50 M phospholipid vesicles (10:90 PS/PC, natural extracts). The mixtures were incubated for 4 min before preheated prothrombin (0.5 M) was added to start the thrombin generation. The reaction was stopped after 1 min by a 40-fold dilution with ice-cold EDTA buffer, and the thrombin formed was measured using S2238. 306 and Arg 506 -The FVa(R506Q/R679Q) and FVa(R306Q/R679Q) variants were cleaved with APC, and their residual FVa activities were determined at increasing concentrations of FXa ( Fig.  1). In comparison, a FXa titration was performed using wildtype FVa. The Arg 506 cleavage resulted in a molecule expressing decreased maximal FVa activity (ϳ70% at 20 nM FXa) as well as decreased FXa affinity (ϳ10-fold). After cleavage at Arg 306 , the FVa activity was severely impaired; and even at the highest FXa concentration, the activity was only 10 -20% of the activity of uncleaved FVa.

Residual FVa Activity after APC Cleavage of FVa at Arg
Protein S Stimulates APC-mediated Cleavages at Both Arg 306 and Arg 506 -The time courses of APC-mediated inactivation of the recombinant FVa variants were determined using the prothrombinase-based assay to measure the remaining FVa activity (Figs. 2 and 3). As the residual FVa activities after APC cleavage of the FVa(R506Q/R679Q) and FVa(R306Q/R679Q) variants were distinct (Fig. 1), different FXa concentrations were used in the prothrombinase assays for the two FVa variants. In the case of FVa(R506Q/R679Q) (Arg 306 cleavage), the FXa concentration was 5 nM, whereas the FXa concentration was reduced to 0.05 nM for FVa(R306Q/R679Q) (Arg 506 cleavage). At this low FXa concentration, the remaining activity of the FVa(R306Q/R679Q) variant was ϳ15%, whereas much higher activities (50 -60%) were obtained at higher FXa concentrations (5 nM). The lowering of the FXa concentration made it easier to monitor the APC-mediated cleavage at Arg 506 due to the larger difference in activity between cleaved and intact FVa. For the FVa(R506Q/R679Q) variant, the remaining activity after APC cleavage was quite low also at the high FXa concentration.
To study the effect of protein S on the Arg 306 cleavage, time courses of APC inactivation of the FVa(R506Q/R679Q) variant performed in the presence and absence of protein S at different phospholipid compositions were followed (Fig. 2). Because cleavage at position 306 constitutes the slow phase of FVa inactivation, a high concentration of APC (0.8 nM) was used in these experiments. The inactivation was performed in the presence of phospholipid vesicles with three different compositions, 10:90 PS/PC (Fig. 2, upper panel)  To study the effect of protein S on the APC-mediated cleavage at Arg 506 , the same set of phospholipid vesicles was used as for cleavage at position 306, but the APC concentration was lowered to 0.05 nM because the Arg 506 cleavage is much faster than the Arg 306 cleavage (Fig. 3). In contrast to results on record, protein S was found to enhance the Arg 506 cleavage by APC. The protein S-dependent stimulation of the APC-mediated Arg 506 cleavage was ϳ3-5-fold at all three phospholipid compositions (Table II), which was lower than the stimulation observed for the Arg 306 cleavage. As described for cleavage at position 306, the residual FVa activity of APC-cleaved FVa can be determined from the experiment using 10:20:70 PS/PE/PC in the presence of protein S, and this was used as the B value in Equation 1.
In the experiments described, we used well defined synthetic phospholipids with a 1-palmitoyl-2-oleoyl composition. The phospholipids that have been used to investigate FVa inactivation vary a lot between the different studies. Because no other group has reported any stimulation of the Arg 506 cleavage by protein S, inactivation of FVa(R306Q/R679Q) was also performed in the presence of 10:90 PS/PC prepared from natural extracts (Fig. 4 and Table II). On these vesicles, the APCmediated cleavage at Arg 506 occurred much faster both in the presence and absence of protein S, but the potentiation exerted by protein S was still detectable. Inactivation was also followed on 10:20:70 PS/PE/PC prepared from natural phospholipid extracts for both Arg 506 and Arg 306 (Tables I and II). Also on these vesicles, the effect of protein S on the Arg 506 cleavage was detectable, but was less prominent than on synthetic phospholipids (Table II). However, cleavage at Arg 306 was still greatly enhanced by the presence of protein S (ϳ25-fold) ( Table I).
Western Blot Analysis of APC-mediated FVa Cleavage-To correlate the loss of activity according to the FVa inactivation assay with the formation of proteolytic products, aliquots were taken from the FVa inactivation mixtures and analyzed by Western blotting (Figs. 5 and Fig. 6). A monoclonal antibody against the heavy chain of FVa (epitope located between positions 307 and 506) was used in combination with a highly sensitive detection system. This allowed detection of the low amount of APC-generated fragments directly from the FVa degradation incubation mixture, even though it contained only 0.8 nM FVa. The high sensitivity of the Western blot technique also allowed detection of the small amount of uncleaved heavy chain that remained after the APC digestion. According to titration experiments in which decreasing amounts of FVa were applied to the Western blot, ϳ5% of the uncleaved heavy chain (corresponding to 0.04 nM FVa) could be detected. Because the remaining uncleaved heavy chain did not express FVa activity (as determined by FXa titration as in Fig. 1), it probably had dissociated from its light chain, which would explain its resistance to APC. As the calculation of rate constants is insensitive to the concentration of FVa, the presence of this FVa species did not affect the calculated constants. In the FVa(R506Q/R679Q)-containing incubation mixtures, APC-mediated proteolysis resulted in the generation of a 60-kDa band, which corresponds to the Arg 306 -Arg 709 fragment generated after cleavage at Arg 306 (Fig. 5). In the presence of protein S, this band was clearly detected already after 0.5 min of incubation. In contrast, in the absence of protein S, this band was significantly weaker on the Western blot and occurred only at later time points.

Protein S in FVa degradation
The APC cleavage of the FVa(R306Q/R679Q) variant yielded a 75-kDa band (Fig. 6, upper panel). This is consistent with cleavage at Arg 506 and the generation of a fragment comprising residues 1-506. When protein S was present in the incubation mixture, the 75-kDa band appeared at earlier time points and with greater intensity. The same band was observed when normal plasma-derived FVa was used (Fig. 6, middle panel), but not when FVa purified from an individual homozygous for the FV Leiden mutation (human FVa(R506Q)) was used (lower panel), confirming that the fragment is indeed the result of cleavage at Arg 506 . When normal plasma-derived FVa was degraded in the presence of protein S, a barely detectable 30-kDa band was also observed, corresponding to the fragment after cleavage both at Arg 306 and Arg 506 (Fig. 6, middle panel). Also a faint band was observed in the presence of protein S with a size of ϳ60 kDa, corresponding to the fragment after isolated cleavage at Arg 306 . This band could also be observed faintly in human FVa(R506Q) (Fig. 6, lower panel).
Protein S Concentration Dependence of the APC-mediated Cleavage at Arg 306 and Arg 506 -To further elucidate the effect of protein S on the individual APC cleavage sites in FVa, the initial rate of FVa inactivation (first 20% loss of activity) was determined at different concentrations of protein S. Pseudo first-order rates were calculated, and the protein S concentra-tion was plotted as a function of the rate constants obtained (Fig. 7). At saturating levels of protein S, the Arg 306 cleavage proceeded ϳ25-fold faster in the presence of protein S, whereas the corresponding number for the Arg 506 cleavage was only 4-fold. The concentration of protein S giving half-maximal stimulation of APC cleavage was similar for the two cleavages (ϳ30 nM).
Phospholipid Composition Dependence of Arg 306 and Arg 506 Cleavages in the Absence and Presence of Protein S-Time courses of FVa inactivation were performed in the presence of numerous different phospholipid compositions, and the results are summarized in Fig. 8 and Tables I and II. Even though the rates of Arg 506 cleavage under all conditions were at least 10-fold higher than the those of Arg 306 cleavage, the two cleavages demonstrated similar responses to changes in phospholipid composition. In the absence of protein S, the rate of Inactivation assays were also performed in the absence of phospholipids to determine the absolute requirement of phospholipid for each cleavage. Because the recombinant FVa preparations were present in expression medium from COS-1 cells, we were concerned that there could be phospholipid remnants in the medium. Using the phospholipid B kit, we estimated that the medium contributed 3-4 M phospholipids to the final FVa inactivation mixtures. To eliminate these phospholipids, the recombinant FV variants were absorbed to a Q-Sepharose column, which was then washed with buffer containing the detergent Triton X-100 before elution. After the Q-Sepharose column step, no phospholipids were detected in the FV preparation. Inactivation of purified FVa in the presence of 10:20:70 PS/ PE/PC was similar to that of non-purified FVa (Fig. 9). Using purified FV, the APC-mediated inactivation of FVa(R306Q/ R679Q) was performed at 0.8 nM APC in the presence and absence of 100 nM protein S. In the absence of detectable phospholipids, the rate of cleavage at Arg 506 was ϳ7 ϫ 10 5 M Ϫ1 s Ϫ1 , with no significant influence of protein S. The APC-mediated inactivation of FVa(R306Q/R679Q) was also performed in the presence of vesicles composed of 100% PC. No stimulation of the Arg 506 cleavage could be observed upon addition of these vesicles. Moreover, protein S did not stimulate the cleavage rate in the presence of 100% PC vesicles. Inactivation assays were also performed in the presence of 20:80 PE/PC. No significant differences in the rates of APC-mediated cleavage could be observed in the presence and absence of such vesicles, regardless of whether protein S was present or not. The APCmediated cleavage of FVa(R506Q/R679Q) was also followed in the absence of phospholipid vesicles. Because very little cleavage was expected under these conditions, an APC titration was performed instead of a time course. FVa(R506Q/R679Q) was incubated with increasing concentration of APC for 10 min, and the FVa activity was measured with the prothrombinase assay. No inactivation of FVa(R506Q/R679Q) could be observed even at 6.4 nM APC, which is 8-fold higher than the FVa concentration (data not shown).

DISCUSSION
The regulation of FVa activity by APC is a complicated phospholipid membrane-bound process involving the three cleavages at Arg 306 , Arg 506 , and Arg 679 (5). These cleavages have been reported to proceed at different kinetics, to have different protein S and phospholipid composition requirements, to yield products with different levels of FVa activity, and to be differently influenced by other components of the prothrombinase complex (18,31). Despite a wealth of investigations, our knowledge is limited because it has not been possible to investigate each individual cleavage site separately. To study the influence of phospholipid composition and protein S on individ- ual APC cleavage sites, we constructed the recombinant FV(R306Q/R679Q) and FV(R506Q/R679Q) variants, which can be cleaved only at Arg 506 and Arg 306 , respectively (8). These FV variants have similar procoagulant activity/antigen ratios as wild-type FV and yield the expected cleavage patterns upon Western blotting after incubation with thrombin and APC (8).
In the present investigation, we used these FV variants for detailed studies on the influence of protein S and the phospholipid composition on the rate of APC-mediated cleavages.
It has previously been reported by Rosing et al. (18) that protein S selectively enhances the APC-mediated cleavage at Arg 306 , whereas Arg 506 is unaffected by protein S. In contrast to this, we now demonstrate a protein S-dependent acceleration of the APC-mediated cleavage at Arg 506 . Several experimental differences between our study and that of Rosing et al. explain the difference in the results. Rosing et al. based their conclusions on studies of plasma-derived FVa from individuals homozygous for normal FV or for FV(R506Q) (FV Leiden) (18). As the Arg 679 cleavage was negligible under the conditions used, they could selectively investigate the Arg 306 cleavage. However, the conclusions on the Arg 506 cleavage were indirect, being based on comparisons of normal FVa and FVa Leiden. The APC-mediated inactivation of normal FVa in the absence of protein S occurs in a biphasic manner. The initial rapid loss of FVa activity corresponds to cleavage at Arg 506 , whereas the slower phase is caused by cleavage at Arg 306 . When protein S was added to the inactivation, Rosing et al. observed that the slow phase was accelerated and that full loss of activity was achieved in a much shorter period of time. Therefore, they concluded that protein S preferentially enhances cleavage at Arg 306 . This is consistent with our results because we also observed that the APC-mediated cleavage at Arg 306 is enhanced to a much greater extent than that at Arg 506 . However, their results did not exclude that Arg 506 could also be stimulated by protein S. Because the protein S effect on the Arg 306 cleavage is much larger than that on the Arg 506 cleavage, the protein S-dependent stimulation is hard to detect by simply comparing the inactivation of purified human FVa and FVa(R506Q) in the presence and absence of protein S. The use of recombinant FV variants that can be cleaved only at one of the APC cleavage sites enabled us to specifically investigate the protein S effects on each cleavage site under optimized conditions for the particular cleavage site.
Another important difference between our study and that of Rosing et al. (18) is related to the phospholipids. In this study, we used synthetic phospholipids with monounsaturated fatty acid chains, whereas Rosing et al. used synthetic phospholipids with a higher degree of polyunsaturated chains. It is known that the inactivation of FVa by APC is enhanced on vesicles composed of polyunsaturated chains compared with unsaturated chains (15). One of the proposed effects of protein S is that it enhances the binding of APC to the phospholipid membrane (19). Therefore, it is conceivable that the specific stimulation of the Arg 506 cleavage by protein S has remained undetected because of the use of phospholipid vesicles with optimal binding properties. For this reason, the inactivation of FVa(R306Q/R679Q) was also performed on vesicles containing phospholipids from natural extracts, which have a high degree of polyunsaturation. Even though the Arg 506 cleavage was faster on these vesicles than on those prepared from synthetic phospholipids, it was still possible to detect a specific protein S-dependent stimulatory effect, which was, however, smaller than that observed using the synthetic vesicles.
Even though the rate of cleavage at Arg 306 was enhanced to a greater extent than that at Arg 506 by protein S, the concentration giving half-maximal stimulation was the same for the two cleavages. This indicates that the mechanism of protein S stimulation is similar for the two cleavage sites. Taken together with the previously observed enhancement of the Arg 679 cleavage by protein S (8), the data suggest that protein S gives a general enhancement of the efficiency of APC on the membrane surface. The mechanism by which protein S enhances the APC-mediated inactivation of FVa is not fully elucidated. One proposed mechanism is that protein S enhances the binding of APC to the membrane surface (19), with the two proteins forming a membrane-bound complex, which is consistent with a general increase in cleavage rates. Protein S has also been shown to decrease the distance between the active site of APC and the membrane surface (20). It has been hypothesized that this change would bring the active site of APC closer to Arg 306 than to Arg 506 . However, the lack of inhibition of the Arg 506 cleavage by protein S argues against this hypothesis; in fact, in this study, we saw the opposite, i.e. stimulation by protein S. Thus, our results do not argue against protein S having a specific effect on the orientation of the active site of APC, but do not support the idea that this would result in a specific enhancement of the Arg 306 cleavage.
In the study, we have also analyzed the effect of membrane composition on the individual APC cleavage sites in FVa. It has been suggested that the Arg 306 cleavage site is more dependent on the phospholipid membrane compared with the other cleavage sites (32). We observed that the Arg 306 and Arg 506 cleavages were affected to the same extent by changes in the membrane composition. The effect of protein S and changes in phospholipid composition were synergistic, which is reasonable because a membrane with a high content of negatively charged phospholipids binds protein S with greater affinity, thereby increasing the effective pool of cofactor. Appreciably high rates of Arg 506 cleavage could also be observed in the absence of phospholipids, whereas no Arg 306 cleavage was observed under these conditions. Thus, even though both cleavages are dependent on the presence of phospholipids, cleavage at position 506 can occur in their absence.
The effect of PE on the APC cleavage rates that we observed was smaller than that reported by Smirnov et al. (15). Moreover, they reported that the PE effect on the FVa inactivation could not be overcome by increasing the PS content, which stands in contrast to what has been observed for the FVIIatissue factor complex, the prothrombinase complex, and FVIIIa binding. Based on this, Smirnov et al. proposed that PE selectively enhances the FVa inactivation. Our present results do not support this because we observed the same cleavage rates on 5:20:75 PS/PE/PC as on 10:90 PS/PE. This indicates that the effect of PE is more general, increasing the affinity of the phospholipid membrane for many of the coagulation proteins. Probably, the differences in the observed effects of PE are explained by the different phospholipids used in the two studies. Whereas we used PE containing the same monounsaturated side chain as PS and PC, Smirnov et al. used a polyunsaturated phospholipid. Also, their vesicles contained more PE than ours (50% compared with 20%).
In conclusion, our results indicate that the two APC-mediated cleavages at Arg 306 and Arg 506 in FVa are more similar in their dependence on phospholipid composition and the presence of protein S than previously reported. Thus, both cleavages are enhanced by the presence of protein S, even though the effect is more pronounced for the Arg 306 cleavage site, and changes in membrane composition similarly influence both cleavages. The differences in kinetics between the Arg 506 and Arg 306 cleavages are presumably due to differences in specific interactions between exosites in APC and the FVa surfaces surrounding the cleavage sites. In agreement with this hypothesis is the demonstration of a positively charged exosite in APC, which is important for cleavage at Arg 506 , but not for cleavage at Arg 306 (33,34). Thus, elimination of the positive cluster in APC by site-directed mutagenesis resulted in specific loss of Arg 506 cleavage potential, but did not affect the ability of APC to cleave Arg 306 . Therefore, the efficiency of each APCmediated cleavage in FVa is determined by multiple specific molecular interactions between APC, FVa, protein S, and the phospholipid membrane.