Relocating the Active Site of Activated Protein C Eliminates the Need for Its Protein S Cofactor

The effect of replacing the γ-carboxyglutamic acid domain of activated protein C (APC) with that of prothrombin on the topography of the membrane-bound enzyme was examined using fluorescence resonance energy transfer. The average distance of closest approach (assuming κ2 = 2/3) between a fluorescein in the active site of the chimera and octadecylrhodamine at the membrane surface was 89 Å, compared with 94 Å for wild-type APC. The γ-carboxyglutamic acid domain substitution therefore lowered and/or reoriented the active site, repositioning it close to the 84 Å observed for the APC·protein S complex. Protein S enhances wild-type APC cleavage of factor Va at Arg306, but the inactivation rate of factor Va Leiden by the chimera alone is essentially equal to that by wild-type APC plus protein S. These data suggest that the activities of the chimera and of the APC·protein S complex are equivalent because the active site of the chimeric protein is already positioned near the optimal location above the membrane surface to cleave Arg306. Thus, one mechanism by which protein S regulates APC activity is by relocating its active site to the proper position above the membrane surface to optimize factor Va cleavage.

at Arg 506 and slowly at Arg 306 , and cleavage at Arg 306 totally inactivates factor Va (4 -7). Protein S is the cognate cofactor for membrane-bound APC (8), and protein S stimulates the rate of factor Va cleavage at Arg 306 by 20-fold (9).
APC is a vitamin K-dependent anticoagulant enzyme that has extensive structural homology to other vitamin K-dependent enzymes (for reviews see Refs. [1][2][3]. The N-terminal module (amino acids 1-38) of human APC contains 9 ␥-carboxyglutamic acid (Gla) residues and is termed the Gla domain. The Gla domain is followed by a domain rich in aromatic residues (also known as the aromatic stack), by two domains that are homologous to the epidermal growth factor, and then by a serine protease domain that contains the active site.
As is true for other vitamin K-dependent plasma proteins, APC binds via the Gla domain to membranes containing negatively charged phospholipids in the presence of calcium ions (1)(2)(3)10). Light scattering experiments indicated that two elongated vitamin K-dependent proteins, prothrombin (PT) and factor X, project radially from the surface when bound to the membrane (11), and our fluorescence resonance energy transfer (FRET) experiments showed that the active site of each of the vitamin K-dependent enzymes is located far (Ͼ70 Å) above the membrane surface (12)(13)(14)(15)(16), thereby indicating that they project approximately perpendicularly from the membrane surface. In the case of membrane-bound APC, its active site is located an average of 94 Å above the surface (assuming 2 ϭ 2/3; Ref. 12). Our FRET study also revealed that protein S relocates the active site of membrane-bound APC to a unique position above the membrane surface (84 Å, assuming 2 ϭ 2/3). These FRET results therefore provide a possible structural explanation for the protein S-dependent alteration in the APC cleavage site on factor Va from Arg 506 to Arg 306 .
Because of the sequence similarity of the Gla domains in different vitamin K-dependent proteins, it has been assumed that the Gla domains of all vitamin K-dependent enzymes must be structurally and functionally similar (17). Consistent with this view, exchanging the Gla domain of factor VIIa for that of APC had no affect on APC plasma anticoagulant activity (18). On the other hand, replacing the Gla domain of factor IXa with that of factor VIIa decreased the V max for factor X activation (19). Furthermore, the membrane binding affinities of vitamin K-dependent plasma proteins differ (e.g. Ref. 10). In addition, unique protein-protein interactions sometimes involve the Gla domain. For example, factor IXa, factor VIIa, and protein C have been shown to bind specifically to collagen IV (20), tissue factor (21), and endothelial protein C receptor (22), respec-tively, through the Gla domain.
Recently, Smirnov et al. replaced the Gla domain and the aromatic stack of APC with the corresponding domains of prothrombin to form the APC-PTGla chimera (23). Exchange of the Gla domains did not alter the affinity for phosphatidylcholine/ phosphatidylserine vesicles significantly but did increase the rate of factor Va inactivation on these vesicles. Furthermore, the activity of the chimera was not increased by protein S. In this study, we test the hypothesis that the location of the active site in the chimera may be similar to that in the APC⅐protein S complex and thereby explain the increased activity and protein S-independence exhibited by the chimera.
Activation of Protein C-PTGla-Human protein C-PTGla (3 mg) was activated in buffer A (50 mM HEPES (pH 7.5), 150 mM NaCl) plus 5 mM EDTA by incubation with 150 g of human thrombin at 37°C in 10.5 ml of final volume. Complete activation of protein C-PTGla was determined by measuring chromogenic activity as a function of time until the activity plateaued (24). The chromogenic activity of the chimera has previously been shown to be equivalent to native APC (23).
Active Site-directed Labeling of Human APC-PTGla-APC-PTGla was generated and isolated as described previously (23). Fl-FPR-labeled APC-PTGla and human APC were then generated essentially as described for bovine APC (12).
Factor Va Leiden Isolation and Inactivation-Factor V Leiden was purified from a patient homozygous for this dimorphism by a minor modification of the published procedure (23) in which the barium precipitation step was omitted and cryoprecipitate was removed before direct application to the monoclonal antibody column. The purified factor V Leiden was then activated with thrombin as described for normal factor V (23).
Factor Va inactivation with APC or the APC-PTGla chimera was studied as a function of enzyme concentration as described (23). Factor Va activity was monitored with purified factor Xa, prothrombin, and PC/PS vesicles. Residual factor Va activity was then determined by reference to a standard curve of prothrombin activation rate versus factor Va concentration (23).
Phospholipid Vesicles-PC/PS (the molar ratio of PC to PS was 4:1) and 100% PC vesicles were prepared by sonication and centrifugation as described previously (13). Samples containing OR were prepared in the same way except that the desired amount of OR (in ethyl acetate) was added to the phospholipid before lyophilization and sonication (13). The concentrations of phospholipid and OR in a purified vesicle sample were determined as before (12), as was , the OR acceptor density at the vesicle surface.
Gel Filtration Chromatography-Fl-FPR-APC-PTGla binding to vesicles was evaluated by incubating 15 nM Fl-FPR-APC-PTGla with 600 M PC/PS containing a trace amount of L-3-phosphatidylcholine-1,2di[1-14 C]oleoyl in buffer A plus 2 mM CaCl 2 for 15 min at room temperature and then analyzing the sample by gel filtration over a Superdex 200 FPLC column (Amersham Pharmacia Biotech). The phospholipid vesicles and any bound Fl-FPR-APC-PTGla eluted in the void volume, whereas unbound protein eluted later. The phospholipid concentration was quantified by liquid scintillation counting, whereas Fl-FPR-APC-PTGla elution was detected by measuring fluorescein emission intensity. The calcium dependence of Fl-FPR-APC-PTGla binding to vesicles was examined by performing a parallel incubation and gel filtration in buffer A plus 5 mM EDTA.
Spectral Measurements-All spectral measurements, including determinations of Q, J DA , and R 0 were performed as before (12).
FRET Measurements-FRET experiments were performed as before (12), except that the D (donor-containing) and DA (containing donor and acceptor) microcells initially received 15 nM Fl-FPR-APC-PTGla (the donor), whereas microcells B (blank) and A (acceptor-containing) received 15 nM unmodified APC-PTGla. The initial net emission intensity (F o ) was obtained by the subtraction of the signal of B from the signals of DA, A, and D. Samples D and B were then titrated with phospholipid vesicles lacking the OR acceptor, whereas samples DA and A were titrated with an equivalent amount of phospholipid vesicles containing OR. The emission intensity of a sample was measured 5 min after each addition of phospholipid, a time that was found to be sufficient to reach equilibrium (i.e. a stable signal). The net intensity of D, DA, or A (F D , F DA , and F A , respectively) was obtained by subtracting the signal of the background B and then correcting for dilution. The blank signal never exceeded 0.5% of the fluorescent signal of the D or DA samples. To compensate for any signal in the DA sample caused by direct excitation of the acceptor, the net dilution-corrected emission intensity of the A sample was subtracted from that of the DA sample. The intensity of DA was then normalized by comparison with its own initial intensity as shown below, as was that of D. Making the reasonable assumption that the absorbance of the donor dye in the active site is not altered by the presence of the OR at the membrane surface, the ratio of the donor quantum yields in the D and DA samples is given by where F is the net dilution-corrected emission intensity of a sample at some point in the titration, and the subscript o is used to denote the initial intensity of the sample. At the end of the phospholipid titration, the membrane-bound Fl-FPR-APC-PTGla was released from the membrane surface by the addition of 5 mM EDTA. After donor release from the membrane, the spectral measurements were repeated to determine what fraction of the acceptor-dependent reduction in donor emission intensity was due to Fl-FPR-APC-PTGla binding to the membrane. The Q D /Q DA value used in Equation 2 below was calculated by dividing the Q D /Q DA value before EDTA addition by the Q D /Q DA value after EDTA addition. This normalization procedure corrects for the contribution of OR inner filter effects and membrane-binding independent energy transfer to the observed total reduction in donor emission intensity.
For experiments with protein S, 15 nM Fl-FPR-APC-PTGla was first titrated with PC/PS(OR) vesicles until the FRET efficiency reached a constant value. Protein S was then titrated into membrane-bound Fl-FPR-APC-PTGla up to a final concentration of 300 nM. Identical procedures were used while performing control experiments with human Fl-FPR-APC, except that excess DTT instead of EDTA was used to release the fluorescein-labeled heavy chain of Fl-FPR-APC from the membrane surface (12).
Distance of Closest Approach-When the extent of energy transfer between randomly and uniformly distributed donor dyes in one infinite plane and randomly and uniformly distributed acceptor dyes in a parallel infinite plane is small, the first term in the approximate series solution of Dewey and Hammes (31) can be used to solve for L, the distance of closest approach between the donor and acceptor dyes in Å, where is the density of acceptor chromophores at the membrane surface (in OR/Å 2 ), and R o is the distance between donor and acceptor dyes at which FRET efficiency is 50% (in Å). This approach is justified here because L Ͼ 1.5 R o (32). The extent of FRET to the acceptor dyes at the inner surface of the 50-Å-thick phospholipid bilayer is negligible and has not been included in our calculations.

Extent of Labeling-Human
Fl-FPR-APC-PTGla was prepared as described under "Experimental Procedures." When the fluorescein concentration was determined as described by Bock (33), the number of dyes per protein averaged 0.7 in our preparations of both Fl-FPR-APC-PTGla and wild-type human Fl-FPR-APC, the same yield that we obtained previously with bovine Fl-FPR-APC (12). For the experiments described in this paper, the presence of nonfluorescein-labeled APC-PTGla or APC molecules in the sample does not interfere with our interpretation of the spectroscopic data.
Spectral Properties of Fluorescein-labeled Proteins-The corrected wavelength of maximum emission and the average val-ues of quantum yield and of steady-state anisotropy were 520 nm, 0.30, and 0.20 for human Fl-FPR-APC and Fl-FPR-APC-PTGla, the same as previously published for bovine Fl-FPR-APC (12). Thus, there is no significant difference in probe environment in the active sites of human and bovine APC. Furthermore, replacing the Gla domain of APC with the Gla domain of PT did not alter the spectral properties of the fluorescein dye and, hence, did not detectably alter the conformation of the active site of APC.
When PC/PS vesicles were added to either human Fl-FPR-APC or Fl-FPR-APC-PTGla, no significant changes in fluorescein spectral properties were detected. In addition, the fluorescence lifetime of the fluorescein (3.9 ns) was unaltered when Fl-FPR-APC-PTGla bound to PC/PS, so the quantum yield of the fluorescein was unaffected by membrane binding. Thus, the binding of the protein to a membrane surface did not elicit a detectable alteration in the environment of the fluorescein dye in the active site of human APC.
Active Site to Membrane Surface Energy Transfer-In our FRET experiments, the fluorescein dye in the active site of the protein is the FRET donor, whereas the rhodamine in OR is the FRET acceptor. The rhodamine dye is positively charged at pH 7.5 and remains in the aqueous phase, whereas the hydrophobic octadecyl aliphatic chain partitions into the lipid bilayer, thereby anchoring the rhodamine moiety at the aqueous-lipid interface.
When human Fl-FPR-APC-PTGla was titrated with PC/PS vesicles, only a very small decrease in fluorescein emission intensity was detected (see Fig. 1, -OR). However, when Fl-FPR-APC-PTGla was titrated with PC/PS vesicles containing OR, the fluorescein intensity decreased until the phospholipid added was sufficient to bind all of the Fl-FPR-APC-PTGla (Fig.  1, ϩOR). The association of all of the Fl-FPR-APC-PTGla molecules with vesicles was confirmed by gel filtration (see below). This OR-dependent decrease in fluorescein intensity results largely from FRET from the fluorescein dyes in the active site of the protein to the rhodamine dyes localized at the phospholipid membrane surface. To facilitate analysis, the data in Fig.  1 were normalized and expressed in Fig. 2 as the ratio of donor quantum yields in the presence and absence of acceptor using Equation 1. The OR-dependent decrease in fluorescein intensity evident in Figs. 1 and 2 shows that the fluorescein dyes are close enough to the rhodamine dyes for FRET to occur.
For comparison, wild-type human Fl-FPR-APC was titrated in parallel with the same stock of PC/PS vesicles used for the Fl-FPR-APC-PTGla titrations. When Fl-FPR-APC was titrated with PC/PS vesicles, no significant change in fluorescein emission intensity was observed. However, when titrated with PC/PS vesicles containing OR (PC/PS(OR)), the Fl-FPR-APC emission decreased because of FRET, as shown by the reduction in Q DA /Q D (Fig. 2).
As a control, Fl-FPR-APC-PTGla was also titrated with 100% PC vesicles with or without OR because the PT Gla domain requires negatively charged phospholipids to bind to a membrane surface (34). As expected, no FRET was observed (Fig. 2). The small reduction in Q DA /Q D observed with PC vesicles (Fig.  2) results from an inner filter effect (a reduction in detected fluorescein emission caused by the absorption of excitation and emission light by rhodamine), not from FRET, as we have documented elsewhere (12).
The data in Fig. 2 also show that the Q DA /Q D values for wild-type human Fl-FPR-APC titrations were always higher than those for Fl-FPR-APC-PTGla at each point in the titration. Because the proteins were titrated with the same PC/ PS(OR) vesicles, the extent of FRET between the fluorescein dye in the active site of membrane-bound Fl-FPR-APC-PTGla and OR dyes on the membrane surface was greater than that between membrane-bound Fl-FPR-APC and OR. The increased efficiency of energy transfer in the chimeric APC-PTGla relative to wild-type APC shows that the probe in the active site of the membrane-bound chimera is not in the same position as that in membrane-bound APC. Thus, the active sites of membrane-bound wild-type APC and APC-PTGla are positioned at different locations above the membrane, with the active site of the chimera closer to the surface and/or rotated so that the relative orientation of the donor and acceptor transition dipoles is more parallel.
Reversibility of Energy Transfer-At the low concentrations of fluorescent-labeled protein and OR used in our experiments, the average separation between free protein and OR molecules is too large for detectable FRET to occur. Thus, if PC/PS-bound Fl-FPR-APC-PTGla or Fl-FPR-APC is released from the mem-  brane surface at the end of the experiment, no FRET should occur, and the Q DA /Q D value should return to 1.0. Because vitamin K-dependent proteins require calcium ions to bind to negatively charged phospholipid surfaces, an excess of EDTA is commonly used to chelate the calcium ions and dissociate the protein⅐membrane complex (e.g. Refs. [13][14][15][16]. However, as documented in the case of bovine Fl-FPR-APC (12), we observed that the EDTA-stimulated dissociation of the human Fl-FPR-APC⅐PC/PS complex was too slow and incomplete to allow us to use this approach for examining the reversibility of Fl-FPR-APC-to-OR FRET (data not shown). We therefore used an excess of DTT to reduce the disulfide bond between the two chains of Fl-FPR-APC and thereby release the fluorescein-labeled heavy chain from the vesicle surface. Upon addition of excess DTT, the value of Q DA /Q D increased to a value close to 1.0 (0.94 -0.99 depending on the acceptor density). We have shown earlier that this small residual OR-dependent decrease in donor intensity (Fig. 2, open triangle) that cannot be reversed by DTT (or EDTA for the chimera; see below) is caused by an inner filter effect (12). Thus, only the changes in Q DA /Q D resulting from membrane binding were used to calculate the distance of closest approach between the fluorescein and rhodamine dyes (i.e. the DTT-or EDTA-reversible Q DA /Q D ).
Interestingly, in contrast to wild-type APC, when excess EDTA was added to PC/PS-bound Fl-FPR-APC-PTGla, the chelation of the calcium ions resulted in an immediate increase in fluorescein emission intensity in the DA sample cuvette such that the value of Q DA /Q D returned to a value close to 1.0 (Fig.  2, open circle). This spectral change results from the dissociation of Fl-FPR-APC-PTGla from the membrane surface. The Fl-FPR-APC-PTGla release was confirmed by gel filtration chromatography (see below). Thus, the replacement of the Gla domain of wild-type human APC with that of the Gla domain of PT yields a chimeric protein with the membrane-binding properties that correspond, as expected, to PT, not APC.
Phospholipid Dependence of Fl-FPR-APC-PTGla to OR Energy Transfer-Because prothrombin does not bind to membrane surfaces that lack acidic phospholipids (10), one would not expect to see FRET if Fl-FPR-APC-PTGla was titrated with PC vesicles containing OR. As shown in Fig. 2 (open squares), only a very small decrease in fluorescein intensity was observed when Fl-FPR-APC-PTGla was titrated with PC(OR). This OR-dependent decrease in donor intensity was due to an inner filter effect rather than to membrane-binding-dependent FRET, as evidenced by the fact that the magnitude of this decrease was nearly equivalent to that observed when Fl-FPR-APC-PTGla and Fl-FPR-APC were dissociated from the membranes with excess EDTA and DTT, respectively (Fig. 2).

Association of Fl-FPR-APC-PTGla with Membranes Detected by Gel Filtration-
The magnitude of FRET can be determined accurately only if the fraction of membrane-bound Fl-FPR-APC-PTGla molecules is known. To address this issue, Fl-FPR-APC-PTGla was incubated with a large excess of PC/PS vesicles in the presence of Ca 2ϩ , and the distribution of free and membrane-bound chimera was then determined using gel filtration. Fl-FPR-APC-PTGla bound to PC/PS vesicles will elute in the excluded volume, whereas unbound Fl-FPR-APC-PTGla will elute in the included volume. More than 98% of the Fl-FPR-APC-PTGla fluorescence co-eluted with the radioactive vesicles (Fig. 3A), thereby demonstrating that essentially all of the Fl-FPR-APC-PTGla molecules can bind to the PC/PS vesicles and participate in FRET. In contrast, no fluorescence was detected co-eluting with the radioactive vesicle peak when the incubation lacked Ca 2ϩ (Fig. 3B).
Distance of Closest Approach: Active Site to Membrane Surface-R o , the distance at which the efficiency of FRET is 50% efficient, was determined as before (12), assuming that the refractive index of the medium between the donor and acceptor is 1.4 and that the transition dipoles of the donor and acceptor dyes are oriented randomly during the excited state lifetime of the donor (i.e. 2 Table I. L, the distance of closest approach between the plane of donor dyes in the active sites of membrane-bound enzymes and the plane of OR acceptor dyes at the membrane surface, averaged 88.7 Å for human Fl-FPR-APC-PTGla, whereas the average L value of the human Fl-FPR-APC was 94.3 Å. The uncertainty noted in Table I reflects Table I are also uncertain because the relative orientation of the donor and acceptor transition dipoles cannot be determined experimentally, and we have assumed 2 ϭ 2/3. As discussed previously, the uncertainty in R o in our experiments because of this assumption is approximately Ϯ 10% (Ref. 12 and references therein). However, for the purposes of this study, the absolute value of L is not important, whereas the relative efficiencies of FRET for the chimera and wild-type proteins are very important. There was no dependence of L on the method used to release Fl-FPR-APC-PTGla from the membrane, because the same average L value was obtained for titrations reversed by the addition of excess EDTA or DTT.
The average L value obtained for human Fl-FPR-APC-PT-Gla, assuming a 2 value of 2/3, was 5.6 Å shorter than either human or bovine Fl-FPR-APC. Because the spectral properties of the fluorescein dye in Fl-FPR-APC-PTGla are the same as those in both human and bovine Fl-FPR-APC, the difference in FRET efficiency between the chimera and the wild-type proteins must arise from a difference in the heights of the active sites of the membrane-bound proteins above the membrane surface and/or from a difference in fluorescein orientation ( 2 ). A difference in 2 could arise either from different orientations of the protease domains of the chimeric and wild-type APC relative to the membrane surface or from different rotational freedoms of the donor dyes in the active sites of the two proteins. Because fluorescein anisotropy was the same for the wild-type APC species and the chimera, we conclude that the active-site probe has the same rotational freedom in each enzyme.
When the FRET results obtained with human Fl-FPR-APC-PTGla are compared with those obtained with human Fl-FPR-APC, the S.D. values for the average L values appear to overlap (Table I). However a rigorous statistical analysis of these data using the Tukey HSD method to compare the means (35) reveals that the average L values for these membrane-bound proteins are different at the 98% confidence level (p Ͻ 0.02). (The critical criterion is not whether the limit on individual means overlap but rather whether the limit on the differences between the means includes zero. Stated another way, the S.E. of the differences is more important than the differences in the S.E. values.) Therefore the difference between the locations of the active sites of APC and APC-PTGla above the membrane is statistically significant.
Protein S Dependence of FRET-The active site of membrane-bound bovine APC moves upon binding to bovine protein S (12). Because the effect of protein S on APC has been shown to be species-specific (36), we wanted to determine whether the protein S-dependent alteration in the topography of membrane-bound APC observed with the bovine proteins also occurs in the human system. Thus, we have here examined the effect of human protein S on the location of the active sites of mem-brane-bound human Fl-FPR-APC and human Fl-FPR-APC-PT-Gla. Samples of human Fl-FPR-APC were first titrated with sufficient PC/PS vesicles (ϮOR) to bind all of the Fl-FPR-APC and an excess of protein S cofactor. When human protein S was titrated into these samples, little change in fluorescein emission intensity was observed in the D sample lacking OR (data not shown). However, when human protein S was titrated into the human Fl-FPR-APC⅐PC/PS sample containing PC/PS⅐OR (the DA cuvette), the donor intensity decreased until all of the Fl-FPR-APC was bound to protein S. This protein S-dependent change in human Fl-FPR-APC emission is expressed in Fig. 4 as the relative Q DA /Q D ; specifically the ratio (Q DA /Q D ) ϩprotein S / (Q DA /Q D ) -protein S . This spectral change was saturatable with respect to protein S concentration, suggesting that it reflects APC⅐protein S complex formation. The protein S-dependent change in FRET efficiency seen in Fig. 4 can occur either because of translational (closer to the membrane) and/or rotational (more parallel alignment of donor-acceptor transition dipoles) movement of the active site of membrane-bound Fl-FPR-APC relative to the membrane surface. Assuming that this movement is solely translational, the average height of the fluorescein in this membrane-bound APC⅐protein S complex would be 84 Å above the membrane surface. Thus, the human protein S relocates the active site of human APC to the same extent observed with the bovine proteins.
Although human and bovine protein S appear to elicit the same topographical change upon binding to their cognate membrane-bound enzymes, these changes are species-specific. When human protein S was titrated into a sample of bovine Fl-FPR-APC, no change in FRET efficiency was observed (Fig.  4). Thus, the absence of a cofactor-dependent structural change correlates with the inability of human protein S to stimulate  Strikingly, when the human Fl-FPR-APC-PTGla⅐PC/ PS(ϮOR) complex was titrated with human protein S, no protein S-dependent increase in the efficiency of energy transfer was observed, even when high concentrations of protein S were added to the complex (Fig. 4). Because human protein S does not elicit any fluorescein spectral changes upon association with human Fl-FPR-APC and because protein S does not stimulate APC-PTGla activity (23), we cannot ascertain whether or not protein S is binding to Fl-FPR-APC-PTGla in these experiments. However, it is clear that protein S does not elicit the same change in the topographies of APC-PTGla⅐PC/PS and wild-type APC⅐PC/PS. Given the absence of any protein S-dependent stimulation of APC-PTGla activity (23), it is interesting that protein S also does not have any influence on the location of the active site of membrane-bound Fl-FPR-APC-PTGla.
Inactivation of Factor Va Leiden-Because the active sites of the chimera and of APC in the presence of protein S have similar locations, it raised the possibility that the chimera and the APC⅐protein S complex might be functionally similar. If the protein S-dependent enhancement of factor Va cleavage at Arg 306 by APC (9) is caused by the movement of the APC active site, then the chimera might cleave factor Va at Arg 306 at a rate similar to that of APC⅐protein S. In factor Va Leiden, the most common source of APC resistance, Arg 506 is mutated to Gln, leaving only Arg 306 available for APC inactivation of factor Va (9,38). We therefore compared the dose-response curves of factor Va Leiden inactivation by APC in the presence and absence of protein S to that of the chimera in the absence of protein S (Fig. 5). Consistent with the above proposal, the dose-response curve for the inactivation of factor Va by the chimera is very similar to that of APC plus protein S, and both curves are shifted to the left relative to APC in the absence of protein S. Consistent with earlier findings of the Rosing group (9), protein S had a much greater effect on factor Va Leiden inactivation than we or others had observed on normal factor Va (23).

DISCUSSION
Our previous FRET studies have revealed that a feature shared by most coagulation enzyme⅐cofactor complexes is an alteration in the topography of the membrane-bound enzyme upon binding to the cofactor. For example, tissue factor was found to alter the location of the active site of membrane-bound factor VIIa relative to the membrane surface (15), and this position was dictated by the cofactor, not by the membrane binding Gla domain of factor VIIa (39). Also, factor Va binding to factor Xa on a membrane surface causes translocation and/or rotation of the active site of the enzyme relative to the membrane surface (13). In addition, the cofactor may alter the conformation of the zymogen and/or the location of the scissile bond in the substrate above the membrane surface, as is the case with factor Va and the prothrombin activation intermediate, meizothrombin (16). The alignment of enzyme active sites with the scissile bond in the substrates cannot account for all cofactor effects, because most cofactors can stimulate zymogen activation by their cognate enzyme in the absence of membrane surfaces (40). Protein S differs from other cofactors in that it cannot enhance factor Va inactivation by APC in the absence of phospholipid (41). This observation makes the protein S system ideally suited to examine the importance of aligning the active site of the enzyme with the scissile bond of the substrate, particularly because we recently observed that protein S moves the active site of APC as much as 10 Å closer to the membrane surface (12).
An approach to testing the functional significance of the latter protein S-dependent change in membrane topography is to develop an APC homologue in which the location of the active site above the membrane surface in the absence of protein S is similar to that of the APC⅐protein S complex. The FRET measurements reported here indicate that the topography of the active site of the chimera and the APC⅐protein S complex are similar. That this change in topography has functional consequences is borne out by the observation that factor Va Leiden is inactivated on PC/PS vesicles at comparable rates by the APC⅐protein S complex and by the chimera in the absence of protein S. This is true despite the fact that the chimera and APC bind PS/PC vesicles with comparable affinity (23). This suggests that protein S functions, at least in part, by aligning the APC active site with the Arg 306 cleavage site in factor Va.
The fact that the PT Gla domain can be substituted for the APC Gla domain without reducing the rate of factor Va inactivation argues strongly that the Gla domain is not directly involved in substrate recognition by APC. It is extremely unlikely that the prothrombin Gla domain improves the activity of the chimera in the absence of protein S over wild-type APC because of direct interactions between the prothrombin Gla domain and factor Va. For example, in the absence of membranes, the heavy chain of factor Va binds intact prothrombin and prethrombin 1, a derivative of prothrombin that lacks the Gla domain, with the same affinity (42). The observation that APC-PTGla inactivates membrane-bound factor Va Leiden at the same rate as the APC⅐protein S complex demonstrates that replacing the Gla domain of APC with that of prothrombin is functionally equivalent to binding protein S to APC.
The nature of the FRET-detected difference in active-site locations between membrane-bound APC and APC-PTGla cannot be determined unambiguously. Assuming that the observed increase in FRET efficiency for Fl-FPR-APC-PTGla relative to Fl-FPR-APC was due solely to translational movement, the average height of the active site above the membrane was about 6 Å less with APC-PTGla than with APC (Table I). If the exchange of the Gla domains also caused the active site to rotate relative to the planar bilayer surface, then the actual FIG. 5. Inactivation of factor Va Leiden by APC and APC-PtGla. Factor Va Leiden (0.2 nM) was inactivated by APC (q, E) or the chimera (Ⅺ) at the concentrations indicated for 30 min at room temperature. When present (q), the protein S concentration was 70 nM. The reaction mixtures contained 10 g/ml PC/PS vesicles. Remaining factor Va activity was determined by prothrombinase assays containing 1 nM factor Xa, 1.4 M prothrombin, 10 g/ml PC/PS. change in height caused by the domain swap could be more or less than 6 Å. Although the FRET measurements cannot tell us the exact magnitude of the structural change, they do show that exchanging the Gla domains caused the active site of the enzyme to move significantly relative to the membrane surface, which, as discussed above, probably accounts for the selective enhancement of cleavage at Arg 306 in factor Va.
Although the position of the active site above the membrane surface has been postulated to be functionally important (12,13,15,16,43), the data reported here constitute the first direct demonstration that modulation of active site location can alter enzyme activity and that the cofactor can be rendered irrelevant by the appropriate repositioning of the active site. Thus, we conclude that protein-dependent modulation of topography (specifically, structure relative to the membrane surface) is an effective means to regulate the activity of membrane-bound enzymes involved in hemostasis and likely in other membranedependent processes.