Effects of Prothrombin on the Individual Activated Protein C-mediated Cleavages of Coagulation Factor Va*

The factor Va (FVa) inactivation by activated protein C (APC), mediated by cleavages at Arg306 and Arg506 in FVa, is inhibited by both factor Xa (FXa) and prothrombin. Although FXa is known to specifically inhibit the Arg506 cleavage, the effect of prothrombin has not been confined to one cleavage site. We used recombinant FV variants, FV:R506Q/R679Q and FV:R306Q/R679Q, to investigate the effect of prothrombin on the individual cleavage sites. The APC-mediated FVa inhibition was monitored by a prothrombinase-based FVa assay, and apparent first order rate constants were calculated for each of the cleavage sites both in the presence and absence of prothrombin. Prothrombin impaired cleavages at both Arg306 and Arg506 and the inhibition correlated with a delayed appearance of proteolytic products on Western blots. Almost complete inhibition was obtained at around 3 μm prothrombin, whereas half-maximal inhibition was obtained at 0.7 μm prothrombin. After cleavage of prothrombin by thrombin, the inhibitory activity was lost. The inhibitory effect of prothrombin on APC-mediated inhibition of FVa was seen both in the presence and absence of protein S, but in particular for the Arg306 sites, it was more pronounced in the presence of protein S. Thus, prothrombin inhibition of APC inactivation of FVa appears to be due to both impaired APC function and decreased APC cofactor function of protein S. In conclusion, FVa, being part of the prothrombinase complex, is protected from APC by both FXa and prothrombin. Release of products of prothrombin activation from the prothrombinase complex would alleviate the protection, allowing APC-mediated inactivation of FVa.

the protein C anticoagulant system by activating protein C. Prothrombin (PT), the precursor of thrombin, is the most abundant of the vitamin K-dependent zymogens, circulating in blood at a concentration of 1.4 M. PT is activated by factor Xa (FXa) by sequential cleavages of two peptide bonds. During activation of PT, activated FV (FVa) plays a crucial role, since it binds to FXa on negatively charged membrane surfaces, thus forming the prothrombinase (PTase) complex (4,6). In this complex, the catalytic efficiency of FXa is increased 10 5 -10 6fold as compared with FXa alone. FVa promotes the binding of FXa to the membrane surface and in addition increases the k cat of FXa for PT (7,8). Moreover, FVa promotes the binding of PT to the PTase complex, thus lowering the K m for the reaction (9). PT can also be cleaved by thrombin, at Arg 155 , yielding fragment-1 and prethrombin-1. Further cleavage of prethrombin-1 by FXa results in thrombin formation (4).
The interaction between FVa and PT has not been fully characterized. The two proteins bind to each other with a K d of about 8 M in the absence of negatively charged phospholipids (10,11), but the affinity is probably enhanced by the presence of a negatively charged phospholipid surface. In PT, the Gla domain, both kringle domains, and the serine protease domain participate in the binding of FVa (12)(13)(14)(15)(16)(17)(18). In FVa, the heavy chain is involved in the binding to PT, but a more detailed mapping of the binding site in FVa has not been accomplished (14). However, it has been reported that the C-terminal part of the heavy chain is involved in the interaction with PT (19,20).
The procoagulant function of FVa is efficiently down-regulated by activated protein C (APC), which cleaves FVa at three sites, at Arg 306 , Arg 506 , and Arg 679 (21). The cleavage at Arg 506 is kinetically favored over the other two sites but only results in partial loss of FVa function (22,23). This loss in FVa activity is caused by a 40-fold reduction in FXa binding affinity (24). The Arg 306 cleavage is required for full loss of FVa function. After cleavages at both Arg 306 and Arg 506 , the 306 -506 fragment dissociates, which results in complete loss of FVa activity (25). The importance of the cleavage at Arg 679 has not been fully elucidated, but it is believed that this cleavage plays a minor role in the inactivation of FVa.
The physiological importance of the APC cleavage sites in FVa is demonstrated by naturally occurring point mutations. The most common is a substitution of Arg 506 with a Gln (FV Leiden), resulting in the APC resistance phenotype (26 -28). This is associated with a 5-10-fold increased risk for venous thrombosis in heterozygous carriers and 50 -100-fold in homozygous cases. Two naturally occurring mutations affect the Arg 306 site, FV Hong Kong (R306G) and FV Cambridge (R306T) (29,30). However, although the Arg 306 cleavage is believed to be of crucial importance for the FVa inactivation, these mutations are not associated with an increased risk for venous thrombosis and APC resistance in heterozygous carriers (31).
Several plasma proteins affect the rate of FVa inactivation by APC. Protein S is an important cofactor to APC, mainly enhancing the cleavage at Arg 306 (23,32,33). In contrast, FXa inhibits the APC-mediated cleavage at Arg 506 (23,33). PT has also been reported to inhibit the APC-mediated inactivation of FVa (34 -36). It was originally suggested that PT inhibited the APC cleavage by serving as a competitive inhibitor to protein S (34). However, more recent studies demonstrated inhibition of APC-mediated FVa inactivation also in the absence of protein S (35,36). The mechanism of the PT-dependent inhibition of APC and/or protein S is not known. Moreover, it is not known which of the APC cleavage sites in FVa are affected by the PT inhibition.
The aim of the present study was to further elucidate the mechanisms of the PT-mediated inhibition of the FVa inactivation. In particular, we wanted to determine the effect of PT on the individual APC cleavage sites in FVa. For this purpose, recombinant FV variants, FV:R506Q/R679Q and FV:R306Q/ R679Q, that were only cleaved at Arg 306 or Arg 506 , respectively, were used. PT was found to inhibit both Arg 306 and Arg 506 cleavages to similar extents. The inhibition did not require the presence of protein S but was more prominent when protein S was present.
Expression and Quantification of Recombinant Factor V Variants-The recombinant FV variants R306Q/R679Q and R506Q/R679Q have been utilized in previous studies and were constructed and transiently expressed as previously described (40). The proteins were harvested in serum-free medium (Opti-MEM) 60 -70 h after transfection and concentrated in Vivaspin concentrator with a molecular weight cut-off of 100,000 (Vivaspin). Aliquots were frozen at Ϫ80°C. The concentrations of the recombinant proteins were determined with both an enzyme-linked immunosorbent assay and the PTase assay. The ELISA for factor V was performed as described (40).
Affinity Purification of Recombinant FV Variants-The recombinant FV variants were purified as previously described (41). In brief, a biotinylated monoclonal antibody against the B-chain of FV (MK30) was bound to streptavidin-coated magnetic beads. Thereafter, the recombinant variants were incubated with the beads and subjected to a series of washing steps. To release FV from the beads and to activate FV, the beads were incubated with thrombin. Because the epitope of MK30 is in the B-domain of FV, the activated form of FV is released from the beads, whereas the B-domain remains associated with the MK-30 coated beads. Our earlier results have shown that the affinity-purified FVa yielded similar APC degradation curves and kinetic constants as the unpurified FV in the condition medium (33,41). For the Western blotting experiments, the affinity-purified FVa was used.
Phospholipid Vesicle Preparation-The phospholipid stocks were dissolved in 10/90 (v/v) methanol/chloroform solution, and the concentrations were determined by phosphate analysis (42). Mixtures of the lipids (weight-based) were prepared in 10/90 methanol/chloroform and kept at Ϫ20°C. Aliquots were drawn from the stocks, dried under N 2 , and then resuspended in Hepes buffer at room temperature. Phospholipids for the PTase assay were sonicated in an XL 2020 sonicator (Misonix, Farmingdale, NY) at amplitude 3 for 10 min. Fresh phospholipid vesicles were prepared every day. For the FVa inactivation assay, extruded phospholipid vesicles were used. The extrusion was performed using LiposoFast basic extruder (Armatis, Germany), as described (43). The phospholipid mixtures dissolved in buffer were subjected to freeze-thaw cycles and subsequently extruded through a membrane with 100 nm pore size 19 times. The extruded phospholipids were used for 2 days. If not otherwise mentioned, the natural phospholipids were used.
PTase-based FVa Assay-To determine the procoagulant activity of FVa, a PTase-based assay was used, as described (40). Briefly, a mixture of 0.5 M PT and 50 M phospholipid vesicles (10:90 PS/PC) was prepared in HNO buffer (25 mM Hepes, 150 mM NaCl, 2 mM CaCl 2 , pH 7.7, containing 0.5 mg/ml ovalbumin). FV was activated by thrombin (final concentration 0.5 units/ml) at 37°C for 10 min. The FVa concentration was 0.8 nM in the FVa inactivation, and it was diluted in total 1:125 when present in the PTase assay, yielding a final concentration of 64 pM. FXa (10 l, yielding final concentrations of 5 nM for FVa:WT and FVa:R506Q/R679Q and 0.5 nM for FVa:R306Q/ R679Q) and the FVa samples (10 l of the 1:5 diluted aliquot; see below) were added to the PTase mix (230 l), and after 2 min, the PT activation was stopped by 40-fold dilution in icecold EDTA buffer. The low FXa concentration (0.5 nM) used with FVa:R306Q/R679Q was chosen because the Arg 506cleaved FVa had much lower remaining activity at this FXa concentration than at the standard 5 nM concentration that was used for the FVa:WT and FVa:R506Q/R679Q variants. The EDTA buffer contained 50 mM Tris, 100 mM NaCl, 20 mM EDTA, 1% polyethylene glycol 6000, pH 7.9. The amount of thrombin formed was measured kinetically with the chromogenic substrate, S2238.
FVa Inactivation Assay-To study the time course of FVa inactivation, recombinant FV variants FV:R506Q/R679Q and FV:R306Q/R679Q as well as wild-type FV were used in a FVa inactivation assay. The FV variants (final concentrations 0.8 nM) were incubated with thrombin (0.5 units/ml) for 10 min at 37°C to activate FV, and the reaction was stopped by the addition of hirudin (final concentration 5 units/ml). Thereafter, phospholipids (final concentration in most experiments was 25 M, but in control experiments 100 and 200 M were used), protein S (final concentration 100 nM) and PT (final concentration 1.4 M) were added. Aliquots were drawn from the mixtures and diluted 1:5 in HNBSACa buffer (25 mM Hepes, 150 mM NaCl, pH 7.7, with 5 mg/ml BSA and 5 mM CaCl 2 ). The inactivation was started with the addition of APC (concentration indicated for each individual experiment). At different time points, subsamples were drawn from the inactivation mixture and diluted 1:5 in ice-cold HNBCACa buffer and then kept on ice until measured. The remaining FVa activity of each subsample was measured using the PTase assay, as described above.
PT and protein S titrations were also performed in the FVa inactivation assay to determine the dose response of PT and protein S. In these experiments, the FV:R306Q/R679Q was activated as described above, and phospholipids (10:20:70 1-palmitoyl-2-oleoyl-PS/PE/PC synthetic, final concentration 25 M) were added. Aliquots of the FVa were added to parallel tubes, and different concentrations of PT and or protein S were added. The APC-mediated inactivation was started by the addition of APC (final concentration 0.05 nM), and the reaction was stopped after 10 min by 1:5 dilution in ice-cold HNBSCa. The remaining FVa activity was measured using the PTase assay as described above.
Western Blot Analysis of Recombinant FV Variants-Affinity-purified FVa variants (final concentration 0.8 nM) in HNBSACa buffer were incubated with APC (concentration indicated for each individual experiment) with or without 1.4 M PT in the presence of 25 M phospholipids (10:20:70, PS/PE/PC). The assay was performed both in the presence and absence of 100 nM protein S. At different time points, the inactivation was stopped by the addition of denaturing sample preparation solution. The samples were run under reducing conditions in a 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. To detect the FVa fragments, a monoclonal antibody against the heavy chain of FVa, AH5146, was used together with horseradish peroxidase-conjugated goat anti-mouse IgG and the Supersignal West Dura extended duration chemiluminescence substrate. The chemiluminescence was traced with a Fuji LAS 3000IR CDD camera, and the signals were quantified with the Image-Gauge program.
Equations Used for Curve Fitting-To calculate pseudo-first order rate constants for APC-mediated cleavages at Arg 306 and Arg 506 , inactivation of FVa R506Q/R679Q and FVa:R306Q/ R679Q, respectively, was followed in time. The inactivation curves were fitted to an equation, as reported (23). The equation was modified because only one cleavage occurs in our FV variants at the APC concentrations used. For calculation of the rate constant of the Arg 506 cleavage, the time curves obtained for FVa:R306Q/R679Q were fitted to Equation 1, in which Va t represents 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.
For calculation of the rate constant of the Arg 306 cleavage, the Equation 2 was used, to fit the time curve obtained with FVa R506Q/R679Q. 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. It is only valid to use the equation when the inactivation rates are independent of the FVa concentration (percentage of FVa inactivation versus time), and the inactivation rates are linear for the concentration of APC. Control experiments were performed, and the inactivation curves were found to fulfill these criteria.
Purification of PT-It was crucially important that the PT preparation be completely free of contaminating FXa, since traces of FXa would be sufficient to bind to FVa and cleave PT. We found this to be a problem with the PT batches that were tried. Therefore, the human PT was subjected to an additional purification step to remove contaminating traces of FXa. Soybean trypsin inhibitor (20 mg) was coupled to a 1-ml NHS-activated Hitrap column (Amersham Biosciences). The column was equilibrated with HN buffer (25 mM Hepes, 150 mM NaCl, pH 7.7) and 1 ml of 50 M human PT with 10 M PPACK was added to the column and allowed to incubate for 2 h in 37°C. Thereafter, the PT was eluted in 1-ml fractions with HN buffer. The buffer was changed to TBS using a PD-10 column. The concentration of PT was determined with S2238 after activation of PT with the snake venom Ecarin (0.025 units/ml Ecarin for 15 min at 37°C). To ensure that the PT preparation was free of any traces of FXa, the PT was incubated with factor Va and phospholipid (like in a PTase assay), and thrombin generation was tested as in a PTase assay.
Cleavage of Prothrombin by Thrombin-Human PT (2.0 mg/ml) was incubated with 10 units/ml thrombin for 2 h at 37°C to generate fragment 1 and prethrombin 1, and the fragmentation was followed by SDS-PAGE. The thrombin was inhibited by Hirudin (30 units/ml).

RESULTS
PT Inhibits the APC-mediated Inactivation of FVa-To elucidate the effect of PT on the APC-mediated inactivation of FVa, the inactivation of FVa:WT was followed over time in a PTase-based assay (Fig. 1A). The inactivation was performed in the presence of protein S. Under these conditions, both cleavages at Arg 306 and Arg 506 occur early, since protein S enhances the rate of the slow Arg 306 cleavage. In the absence of PT, the inactivation of FVa:WT was rapid, with a loss of 60% activity within the first few minutes. This was followed by a slower phase of inactivation. After 20 min, only 20% of the FVa activity remained. In the presence of PT, the inactivation of FVa was inhibited. To correlate the loss of FVa activity with the appearance of proteolytic products, aliquots from the inactivation assay were withdrawn and subjected to Western blotting (Fig. 1,  B and C). The monoclonal antibody recognized an epitope located between Arg 306 and Arg 506 . Incubation of FVa:WT with APC resulted in the appearance of two fragments. Initially, after only 30 s, a fragment of about 75 kDa appeared, corresponding to the 1-506 fragment. This coincided with the rapid loss of intensity of the heavy chain band. Within the first minute, a fragment of about 30 kDa appeared, corresponding to fragment 307-506. At the same time, the intensity of the 1-506 fragment declined. Thus, in the presence of protein S, the cleavage at Arg 506 occurs first, followed shortly after by cleavage at Arg 306 , resulting in the 307-506 fragment. In the presence of PT, the appearances of both fragments were delayed, but the presence of PT did not change the order of cleavage, indicating that both the Arg 306 and the Arg 506 cleavage were inhibited by PT.
Effect of PT on the APC-mediated Cleavage at Arg 306 -The FVa:R506Q/R679Q variant was used to investigate the effect of PT on the APC-mediated cleavage at Arg 306 . When no protein S was added, a high APC concentration (3.3 nM) was used, and the FVa-inactivation reaction was followed for 20 min. The rate of inactivation of FVa:R506Q/R679Q was low, consistent with a slow cleavage at Arg 306 in the absence of protein S ( Fig. 2A). In the presence of PT, the inactivation of FVa:R506Q/R679Q was inhibited. Based on the FVa inactivation time curves, apparent second order rate constants were calculated ( Table 1). The addition of PT resulted in an approximately 3.5-fold decrease in inactivation rate (k 306 ϭ 1.3 ϫ 10 6 M Ϫ1 s Ϫ1 in the absence of PT versus k 306 ϭ 3.9 ϫ 10 5 M Ϫ1 s Ϫ1 in the presence of PT). When the FVa inactivation was performed in the presence of protein S, the APC concentration was lowered to 0.2 nM , since protein S enhances the rate of Arg 306 cleavage. The rate of inactivation of FVa:R506Q/R679Q was greatly stimulated by the addition of protein S when no PT was present (k 306 ϭ 1.3 ϫ 10 6 M Ϫ1 s Ϫ1 in the absence of protein S versus k 306 ϭ 3.1 ϫ 10 7 M Ϫ1 s Ϫ1 in the presence of protein S). The addition of PT resulted in a 10-fold decreased inactivation rate (k 306 ϭ 3.1 ϫ 10 6 M Ϫ1 s Ϫ1 in the presence of PT versus k 306 ϭ 3.1 ϫ 10 7 M Ϫ1 s Ϫ1 in the absence of PT).
To analyze the appearance of proteolytic products, aliquots were drawn from the inactivation mixture and subjected to Western blotting (Fig. 2B). As mentioned earlier, the monoclonal antibody used recognizes an epitope between Arg 306 and Arg 506 . The expected band for APC-cleaved FVa:R506Q/ R679Q is 60 kDa, corresponding to the fragment 307-709. The appearance of this fragment and the loss of heavy chain were quantified using chemiluminescence (Fig. 2C). In the absence of PT and protein S, the 307-709 fragment appeared within the first few minutes (Fig. 2B). This coincided with the loss of intensity of the heavy chain band (Fig. 2C). The cleavage of the 306 site in the absence of protein S was very slow, and the addition of PT did not significantly change the pattern on the Western blotting. Incubation of FVa:R506Q/ R679Q in the presence of protein S resulted in the appearance of the 307-709 fragment within the first 30 s (Fig. 2, B and C), which coincided with the rapid loss of intensity of the heavy chain band. The appearance of the 307-709 fragment, as well as the decay of heavy chain, was delayed when PT was added to the inactivation assay. This is consistent with PT-mediated inhibition of APC cleavage at Arg 306 .
Effect of PT on the APC-mediated Cleavage at Arg 506 -The FVa:R306Q/R679Q variant was used to investigate the effect of PT on the APC-mediated cleavage at Arg 506 (Fig. 3). Since the Arg 506 site is very sensitive to APC, the APC concentration was much lower (0.05 nM) than for the Arg 306 cleavage. In the absence of protein S, PT inhibited the Arg 506 cleavage about 3-fold (k 506 ϭ 3.3 ϫ 10 7 M Ϫ1 s Ϫ1 in the absence of PT versus k 506 ϭ 1.2 ϫ 10 7 M Ϫ1 s Ϫ1 in the presence of PT). As already reported, . APC (0.05 nM final concentration) was added to start the inactivation, and at intervals, samples were drawn, and the FVa inactivation was stopped by a 1:5 dilution in ice-cold HNBSACa. The FVa activity was measured in the PTase assay using 0.5 nM FXa. The FVa activity was related to the activity observed without the addition of APC. The plotted values represent means of three individual experiments, and error bars represent S.D. The lines indicate curve fits performed as described under "Experimental Procedures." B, samples drawn from inactivation reactions using affinity-purified FVa:R306Q/R679Q were mixed with SDS-containing sample preparation buffer (reducing conditions) and analyzed by Western blotting (10% SDS-PAGE) using monoclonal antibody AHV5146, the epitope of which is situated in the 307-506 fragment. C, FVa heavy chain fragments were visualized with SuperSignalWest chemiluminescent substrate using a chemiluminescence reader (Fujifilm LAS-3000 IR), and the bands were quantified by using the ImageGauge program (Fujifilm); closed squares indicate the intact heavy chain, and closed triangles show the 1-506 fragment.  Aliquots from the inactivation assay were subjected to Western blotting (Fig. 3, B and C). In the presence of protein S, incubation with APC resulted in the immediate appearance of a fragment of about 75 kDa, corresponding to the 1-506 fragment. This coincided with the rapid loss of intensity of the heavy chain band. The addition of PT resulted in a delay of appearance of the 1-506 fragment and a delay of the heavy chain decay. In the absence of protein S, the incubation of FVa: R306Q/R679Q with APC also resulted in a quick appearance of the 75-kDa fragment and a rapid loss of heavy chain. The addition of PT gave a delay in the appearance of the fragment and a slower degradation of the heavy chain. Thus, the hampered inactivation of FVa:R306Q/R679Q in the presence of PT is caused by an inhibition of the cleavage at Arg 506 .
PT Titration-To elucidate how much PT was needed to inhibit FVa inactivation, the inactivation of FVa:R306Q/R679Q was performed at increasing concentrations of PT (Fig. 4). The PT-dependent inhibition reached a plateau level at around 3 M PT with half-maximum inhibition at 0.7 M. The experiment was also done for FVa-WT in the presence of protein S with a similar result being obtained (data not shown).
Only Intact PT Inhibits the APC-mediated Inactivation of FVa-To elucidate if only intact PT mediated the inhibition of FVa inactivation, the inactivations of FVa:R306Q/R679Q and FVa:R506Q/R679Q were performed in the presence of prothrombin that had been cleaved by thrombin (Fig. 5). Cleaved prothrombin had no inhibitory effect, showing that the inhibition is mediated by the intact molecule and that the inhibition was lost upon cleavage of prothrombin. To ensure that the inhibitory activity of prothrombin was specific, the experiments were performed at both the standard 25 M phospholipid concentration and at two much higher concentrations of phospholipids (100 and 200 M) with similar results. Moreover, factor IX was unable to inhibit the APCmediated cleavage of FVa further, showing the effect of prothrombin to be specific.
Protein S Titration-To examine if protein S could counteract the inhibitory effect of PT, FVa inactivation was performed at increasing concentrations of protein S (Fig. 6). FVa:R306Q/ R679Q was incubated in the absence or presence of PT with APC and increasing amounts of protein S for 10 min, after which remaining FVa activity was measured. Protein S was found to be a less efficient APC cofactor in the presence of PT, but the addition of high concentrations of protein S could overcome the inhibitory effect of PT.  . APC (0.05 nM final concentration) was added to start the inactivation, and after 5 min, samples were drawn, and the FVa inactivation was stopped by a 1:5 dilution in ice-cold HNBSACa. The FVa activity was measured in the PTase assay using 0.5 nM Xa. The FVa activity was related to the activity observed without the addition of APC. In addition, a control experiment with increasing concentrations of FIX was performed (closed circles).

DISCUSSION
The interpretation of how the individual APC cleavage sites in FVa occur in vivo is intricate, since numerous plasma components influence the individual cleavage rates. Also, some of these components counteract each other; e.g. protein S hampers the FXa inhibition of the Arg 506 site (33,44). The inhibition of FVa inactivation by PT was initially reported by Mitchell et al. (34). They claimed that the inhibitory effect required the presence of protein S, indicating that PT interferes with the protein S activity rather than the APC activity directly. In contrast, Smirnov et al. (35) reported that the effect of PT also occurred in the absence of protein S and that the presence of protein S did not change the inhibitory activity of PT. It has not yet been elucidated whether the PT inhibition only impairs one of the APC-mediated cleavages or if it gives a general inhibition of the FVa inactivation. Such a characterization would, however, be interesting, since it should increase our understanding of the importance of the individual cleavages in normal states as well as in pathological situations.
To study the influence of PT on the Arg 306 and Arg 506 cleavages, we used FV variants FV:R506Q/R679Q and FV:R306Q/ R679Q, respectively (32,33,40). These mutants enabled us to investigate each cleavage site separately and to calculate cleavage rates in the presence and absence of PT. Our results show that the PT inhibits both Arg 306 and Arg 506 cleavages. In the absence of protein S, the inhibition was similar in amplitude for each of the two sites. Although Smirnov et al. (35) were not able to study the cleavages separately, they performed Western blot analysis, where they observed that the order of cleavages was not changed by the presence of PT, although the cleavages were slower. This is consistent with our results, since we see that both cleavages are affected to the same extent. In our experiments, the inhibition of FVa inactivation by PT was observed in the absence of protein S, but the inhibitory effect of PT was more prominent in the presence of protein S. This was especially true for the Arg 306 cleavage, which is normally greatly enhanced by protein S, and thus an impairment of the protein S stimulation will severely hamper the cleavage at this site. Our results suggest that PT impairs the APC-mediated inactivation directly as well as indirectly by interfering with the cofactor function of protein S. This may explain why the PT inhibition first was claimed to be protein S-dependent (30). Our present results agree well with those of Smirnov et al. (35) (i.e. the inhibitory activity of prothrombin is observed both in the absence and presence of protein S).
A polymorphism in the PT gene has been found, which is associated with an increased risk of thrombosis (45). The polymorphism is in the 3Ј-untranslated region and leads to an increase in PT plasma levels. The mechanism behind the increased concentration of PT, as well as the mechanism for the increase in risk for venous thrombosis, is not known. However, a high PT level in the absence of the polymorphism is also associated with increased risk of venous thrombosis. It has been suggested that the PT inhibition of the FVa inactivation might contribute to the increased risk for thrombosis. Interestingly, the PT polymorphism has been reported to interact with FV Leiden as risk factors for venous thrombosis (46).
In individuals carrying the PT polymorphism, the PT levels are above 1.15 units/ml, which corresponds to a molar concentration of above 1.6 M, whereas the normal concentration is around 1.4 M. In the PT titration that we performed, halfmaximum inhibition was observed at 0.7 M PT. This means that even at the normal plasma concentration, the inhibitory effect of PT is almost complete, and a further increase in PT concentration would not be expected to yield more inhibition. It is therefore unlikely that the small increase in PT-mediated inhibition of APC pathway should explain why the PT polymorphism is associated with increased risk of venous thrombosis. It is also unlikely that the PT inhibition decreases the difference in APC response between plasma with normal FV and FV Leiden, as was previously proposed (35), since PT inhibits the cleavages at both Arg 306 and Arg 506 . In fact, in the presence of protein S, the PT-mediated inhibition is most prominent at the Arg 306 site, which would rather increase the gap between FV Leiden and normal FV.
In patients receiving vitamin K inhibitors, the levels of APC, protein S, and PT are substantially decreased. At PT levels of around 20% of normal, the inhibition by PT is almost completely abolished. This suggests that APC under these conditions would be expected to be more efficient in inactivating FVa. This may explain why the decrease in APC during oral anticoagulant therapy is well tolerated. It also emphasizes the importance of supplement treatment with heparin in the beginning of coumarin treatment, since APC and protein S levels decrease before the PT level. Also, it indicates that a combination of coumarin and a low dose of APC might be an efficient therapy of thrombosis.
The mechanism of PT inhibition is probably different from that of FXa, since PT inhibits both the Arg 306 site and the Arg 506 site. FXa inhibits only the Arg 506 site (23,33). The FXa binding site has been localized close to the Arg 506 site (47,48). Consistent with this, the cleavage at Arg 506 gives a substantial loss of FXa binding affinity (22). The binding of FXa to FVa probably sterically hinders the docking of the active site of APC to the Arg 506 site. The binding site of PT on FVa has not been clearly identified. Moreover, neither the cleavage at Arg 306 nor that at Arg 506 gives the same substantial loss of PT binding as was reported for the FXa binding (24). It is therefore reasonable to assume that the mechanism by which PT interferes with the APC activity is not by blocking the binding of the active site of APC to FVa. It has been reported that an APC chimera containing the Gla domain of PT is hardly inhibited by PT (35,36). The Gla domain of PT is known to contribute to the FVa binding (12). In addition, both kringle domains and several sites in the serine protease domain of prothrombin have been shown to be involved in the PT-FVa interaction (11,13,15,17,18). To date, there are no reports suggesting that APC interacts directly with PT, and it appears valid to conclude that the PT-mediated inhibition of FVa degradation depends on interactions between FVa and PT rather than interactions between APC and PT. The titrations of the rate enhancement with protein S in the absence and presence of prothrombin showed that protein S can reverse inhibition by prothrombin. This suggests that there are overlapping, mutually exclusive binding sites for protein S and prothrombin on factor Va.
During initiation of coagulation, trace amounts of FVa are formed. During this stage, it is of crucial importance to protect the FVa from inactivation by APC and protein S. FVa, when in the PTase complex, seems to be protected from APC-mediated inactivation. Both PT and FXa shield FVa from proteolysis by APC. PT seems to also diminish the protein S cofactor activity. However, during the coagulation process, PT will be cleaved by FXa and/or thrombin, which leads to a drop in the local PT concentration and loss of FVa protection. Our results indicate that even a minor decrease in the PT levels results in a dramatic decrease in the PT inhibition of APC inactivation. Since protein S counteracts the FXa inhibition, the inactivation of FVa can then occur efficiently. The balance between procoagulant and anticoagulant events is delicate, and small changes in any of the components may disrupt the normal hemostasis.