Factor XSt. Louis II. Identification of a glycine substitution at residue 7 and characterization of the recombinant protein.

A molecular defect in factor X (fX) results from a point mutation that causes glycine substitution for gamma-carboxylated glutamic acid at position 7. The variant (fXSt. Louis II) and wild type (fXWT) proteins were produced in a mammalian expression system and characterized. fXSt. Louis II has <1% and approximately 3% of normal clotting activity in modified prothrombin time and partial thromboplastin time assays, respectively. The rate of activation of fXSt. Louis II by factor VIIa and tissue factor is undetectable under conditions that result in complete activation of fXWT; activation by factors VIIIa and IXa is approximately 30% of normal activation. The X-activating protein from Russell's viper venom activates fXSt. Louis II completely but at a reduced rate. Thrombin generation on phoshopolipid vesicles or activated platelets is approximately 30% or approximately 5%, respectively. Membrane-dependent autolysis is markedly reduced for fXSt. Louis II. In reactions that are not surface-dependent, fXSt. Louis II is nearly identical to that of fXWT. The rate of inhibition by antithrombin is indistiguishable, as is the rate of thrombin formation in the absence of phospholipid, with or without factor Va.

Activated factor X (fXa), 1 a vitamin K-dependent plasma serine protease, holds a pivotal position in blood coagulation as the only known physiological activator of prothrombin. The zymogen precursor, fX, is activated as fXa by cleavage of a single peptide bond (Arg 194 -Ile 195 ) on a phospholipid surface by activated factor VII and tissue factor (fVIIa/TF, extrinsic pathway) or activated factors IX and VIII (fIXa/fVIIIa, intrinsic pathway); the same bond can also be hydrolyzed by the factor X-activating protein from Russell's viper venom (RVV-X) (1)(2)(3)(4)(5)(6). The light chain of fX has an amino-terminal "Gla domain," which includes 11 ␥-carboxylated glutamic acid (Gla) residues, connected by a short stretch of hydrophobic amino acids to two sequential epidermal growth factor-like domains (EGF-1 and EGF-2). The heavy chain is joined to the light chain by a single disulfide bond and contains a 52-amino acid peptide released on activation and the serine protease domain. The gene has been localized to chromosome 13 and contains 8 exons spanning ϳ25 kilobases (7,8).
Commercially available plasmas from fX-deficient donors were screened to discover potentially interesting dysfunctional proteins. A plasma with normal fX antigen but Ͻ1% fX activity by a prothrombin time-based clotting assay was selected for detailed investigation. This variant fX, referred to as fX St. Louis II , does not react with a calcium-dependent monoclonal antibody known to bind to the Gla domain of normal fX. The light chain of fX St. Louis II was purified from donor plasma and sequenced to reveal a glycine substitution for Gla in cycle 7. This was confirmed by analysis of the DNA isolated from whole blood provided by the vendor. To confirm that this single base substitution accounts for the reduced activity, it was engineered into a mammalian expression system. The properties of recombinant wild-type fX (fX WT ) and fX St. Louis II are compared in this report.

MATERIALS AND METHODS
Reagents and Chemicals-Spectrozyme TH TM (H-D-hexahydrotyrosyl-L-alanyl-L-arginine-p-nitroanilide diacetate) and Spectrozyme fXa TM (methoxycarbonyl-D-hexahydrotyrosyl-L-alanyl-L-arginine-p-nitroanilide diacetate) were obtained from American Diagnostica, Greenwich, CT. Organic solvents were high performance liquid chromatography grade. All other reagents and chemicals were of the highest quality commercially available.
Factor X-deficient Plasmas-fX-deficient plasmas were obtained from George King Bio-Medicals, Inc. (Overland Park, KS). The variant fX described in the study was isolated from plasma designated GK1004. Whole blood from this donor, which provided DNA for amplification and subsequent sequencing, was also obtained from George King Bio-Medicals, Inc. Plasma GK1007 was used in experiments requiring a separate fX-deficient plasma.
Immunoglobulins-Affinity-purified caprine anti-mouse IgG conjugated to horseradish peroxidase was obtained from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). Murine monoclonal antibodies specific for human fX were developed by standard methods. Three relevant monoclonal antibodies are briefly described here: 1) 3698.1A8.10 reacts with the normal fX Gla domain, but only in the presence of calcium ions; 2) 3514.5H12.10 binds to the distal end of the Gla domain independent of calcium; and 3) 3448.1D7.20 binds to the second growth factor-like domain. An immunofluorescence assay for fX was devised, as described previously for protein Z (10), using 3448.1D7.20 labeled with fluorescein isothiocyanate and 3514.5H12.10 on polystyrene particles as the capture antibody. This combination yields fluorescence linearly proportional to the total concentration of fX/fXa. These antibodies also enable identification of fX in Western blots following SDS-PAGE, with detection limits of ϳ50 pg; 3448.1D7.20 only recognizes unreduced protein.
Proteins-Prothrombin was purified from plasma by barium citrate adsorption and ammonium sulfate elution, followed by column chromatography using DEAE-Sepharose TM (10), sulfated dextran beads (11), and Mono Q TM FPLC. Following purification, prothrombin was concentrated (to Ͼ1 mM) and immunodepleted of residual fX (Ͻ1 nM final concentration) by passage through a column of immobilized antibody 3448.1D7.20. Antithrombin III was purchased from Kabi Pharmacia Diagnostics (Piscataway, NJ). Factors Va, VIIa, and IXa from human plasma were purchased from Hematologic Technologies (Essex Junction, VT). Thrombin was prepared by cation exchange chromatography after activation of human prothrombin with Taipan snake (Oxyuranus scutellatus) venom. Porcine fVIII was obtained from Porton Products (Agoura Hill, CA) and was further purified (12) using an immobilized antibody (W-3, a generous gift from Dr. David Fass). Innovin TM , a lipidated recombinant tissue factor was obtained from Baxter Diagnostics (Deerfield, IL). The X-activating protein, RVV-X, was purified from Russell's viper venom (13). SDS-PAGE was performed by the method of Laemmli (14). Proteins were stained directly using Coomassie Blue or were transferred to 0.45 M nitrocellulose, probed overnight using antibodies described above (1 g/ml), and visualized with an immunochemiluminescence kit (Amersham Corp., Buckinghamshire, UK).
Both recombinant fXs were partially purified from culture medium as follows. 1.0 M barium chloride was added dropwise (final concentration of 60 mM) to conditioned medium adjusted to 32 mM citrate. After incubation for 60 min on ice, the barium citrate precipitate was washed once with 100 mM sodium chloride, 20 mM Tris buffer, pH 7.5, containing 10 mM barium chloride. The precipitate was collected by centrifugation and resuspended in one-tenth of the initial volume in 20% saturated ammonium sulfate containing 5 mM diisopropyl fluorophosphate. After 1 h at 4°C, the solution was adjusted to 85% saturated ammonium sulfate, and the resulting precipitate was collected by centrifugation. The final pellet was resuspended in 10 mM HEPES, pH 7.0, 100 mM NaCl, 1 mM benzamidine and dialyzed overnight against the same buffer. The solution was clarified by centrifugation at 12,000 ϫ g and loaded onto a Pharmacia Mono Q TM FPLC column for further purification. The fX was eluted with 500 mM NaCl, adjusted to pH 6.5 with MES and 2.5 mM CaCl 2 , sterile filtered, made 20% concentration in sterile glycerol, aliquotted, snap frozen in liquid nitrogen, and stored at Ϫ70°C. Protein concentration was determined by the two-site immunoassay described using fX purified from plasma as the standard.
Mutagenesis, Transfection, and Cell Culture-A 1549-basepair fX cDNA was cloned from a human hepatocyte library. The codon corresponding to amino acid Glu 39 was mutated (GAA3 GAG) to eliminate an internal EcoRI restriction site, and the cDNA was inserted into pUC19. Additional mutations were introduced using a Transformer® mutagenesis kit (Clontech, Palo Alto, CA). The codon corresponding to the Ϫ2 residue of fX was changed (ACG3 AGG, Thr3 Arg) to allow correct cleavage of the propeptide when expressed in 293 cells. The entire wild-type cDNA sequence was verified, as was the mutation at residue 7 (GAG3 GGG). The ZMB3 expression vector was a gift from Dr. Don Foster (ZymoGenetics). Expression is driven by the adenovirus major late promoter; neomycin resistance is oriented in the opposite direction. The fX cDNA was shuttled into the vector through a unique EcoRI site and mapped for correct orientation with informative restriction enzyme digests. Human kidney cells (293 cells, American Type Culture Collection, CRL 1573) were transfected by calcium phosphate precipitation (15). Culture medium, minimum Eagle's medium with Earle's balanced salts (Cellgro®, Mediatech, Herndon, VA), was supplemented with 6.25 mM HEPES, pH 7.2, 1.0 mM L-glutamine, 5 mg/liter vitamin K, and 5% (v/v) fetal calf serum. 500 mg/liter G418 was used as the selection agent. Isolated colonies were screened for fX production by immunoassay before expansion of selected clones for medium conditioning.
Amino Acid Sequence Analysis-fX St. Louis II was immunoprecipitated from 1 ml of donor plasma (ϳ6 g), reduced, and loaded onto a 12.5% SDS-polyacrylamide gel. The resolved peptides were transferred to an Immobilon membrane (Millipore Corp., Bedford, MA) and stained with Coomassie Blue. The light chain was excised and sequenced directly by the Washington University protein sequence laboratory.
Modified Prothrombin Time (PT) and Activated Partial Thromboplastin Time (PTT) Assays-Donor plasma (10% v/v) or pooled normal plasma (0.1-10% v/v) were mixed with fX-deficient plasma from a second donor and incubated at 37°C for 5 min in a fibrometer. Clotting was initiated by the addition of rabbit brain thromboplastin (Ortho, Raritan, NJ) for the PT or by the addition of 0.02 M CaCl 2 after a 3-min preincubation with a partial thromboplastin reagent (Ortho) for the activated PTT. The fX activity of donor plasma was estimated from plots of the clotting times versus the dilutions of pooled normal plasma.
Activation-Recombinant fX St. Louis II was activated by three different enzymes and compared directly with recombinant fX WT in each case. Each fX (150 nM) in assay buffer (10 mM HEPES, pH 7.5, 100 mM NaCl, 5 mM CaCl 2 , 1 mg/ml bovine serum albumin, 1 mg/ml polyethylene glycol 8000) was activated by 1) 100 pM (rate) or 100 nM (end point) RVV-X, 2) 44 pM fVIIa in the presence of 500 pM lipidated recombinant tissue factor, and 3) 2 nM fIXa in the presence of 4 units/ml of thrombinactivated fVIIIa on 20 M PC:PS vesicles. Activation of fX was monitored by quenching samples over time into assay buffer, with 5 mM EDTA substituted for CaCl 2 , and then measuring the rate of hydrolysis of 100 M Spectrozyme fXa TM in a Molecular Devices kinetic microplate reader.
Stability of Activated fX St. Louis II -fX WT and fX St. Louis II (150 nM) were activated by 100 nM RVV-X at 37°C in assay buffer with 10% (v/v) rabbit brain cephalin (Sigma). Samples were removed from the reaction mixture over time (0 -120 min), quenched in SDS-PAGE sample buffer, separated by the Laemmli method (14), transferred to nitrocellulose, and visualized using fX-specific monoclonal antibodies (3514.5H12.10 and 3448.1D7.20).
Inhibition by ATIII-fX WT and fX St. Louis II (150 nM in assay buffer) were fully activated with RVV-X, diluted to 0.5 nM in assay buffer, and then reacted with Spectrozyme Xa TM (100 M) and ATIII (0 -4 M). The appearance of p-nitroanilide acetate was monitored continuously over 15 min at 405 nm in a Molecular Devices kinetic microplate reader. Primary data were fitted, using the Marquardt-Levy algorithm (Sig-maPlot, Jandel Scientific, Corte Madera, CA), to an exponential equation of the form where A is the observed absorbance, A f is the amplitude of the curve, k obs is the apparent first-order rate constant (units of sec Ϫ1 ), t is time in seconds, and A O is the initial absorbance. Each calculated apparent rate constant (k obs ) was then plotted as a linear function of the effective inhibitor concentration, and the resulting slope was calculated as the apparent second-order rate constant.
Thrombin Generation-The rates at which recombinant fXa WT and fXa St. Louis II can activate prothrombin to thrombin were compared for four different reagent compositions: 1) 20 pM fXa with 1.0 M prothrombin, 100 pM fVa, and 20 M PC:PS vesicles, from 0 -3 min; 2) varying fXa (0 to 1 nM) with 1.0 M prothrombin and 10 8 /ml washed platelets (preactivated with 0.5 units/ml thrombin); 3) 10 nM fXa with 100 M prothrombin, from 0 -120 min; 4) 5 nM fXa with 1 M prothrombin and 50 nM fVa, from 0 -3 min. In every assay, aliquots of the reaction mixture were taken at specified times and diluted at least 5-fold into buffer containing 5 mM EDTA to stop further thrombin formation. Thrombin activity at each time was then determined from the initial rate of hydrolysis of 500 M Spectrozyme TH TM . Thrombin generation was then plotted as a function of time, except in the experiments with activated platelets where the initial rate of thrombin generation for each fXa concentration was obtained from the slope of a least squares fit to the initial linear portion of a plot of thrombin activity versus reaction time.

Partial Purification of fX St. Louis II from Plasma and Identification of the Amino Acid Substitution-
The donor plasma has 72% fX antigen compared with pooled normal plasma based on the immunofluoresence assay. Normal fX and fX St. Louis II were immunoadsorbed from pooled normal and donor plasmas, respectively, using monoclonal antibody 3448.1D7.20 coupled to Sepharose. The proteins have equivalent electrophoretic mobility by SDS-PAGE analysis (data not shown). Following transfer to Immobilon, the isolated fX light chains were sequenced, with approximately 85% average yield at each of the first 12-15 cycles. A glycine, proportional in amount to the amino acids detected in other cycles, was found as the seventh residue of fX St. Louis II . Gla residues, including the one in cycle 7 of normal fX, were not specifically identified in the standard sequencing protocol.
Detection of the Point Mutation-All 8 exons and flanking splice junctions of the fX St. Louis II gene were sequenced as described (16). Only a single point mutation (A3 G) at nucleotide 1200 in exon II was found, and it explains the glycine in fX St. Louis II (GAG3 GGG) at amino acid 7. Consistent with the amino acid sequencing data, no other base is detected at nucleotide 1200 for fX St. Louis II (Fig. 1). These results, together with the normal fX antigen level, suggest that the plasma donor is homozygous for the mutation. It is possible, though much less likely, that the donor is a compound fX heterozygote. However, a putative second abnormal allele would have to elude detection by the amplification and sequencing methods and fail to produce detectable protein in the plasma.
Plasma-based Assays of Function-The clotting activity of the donor plasma was estimated in a PT-based assay. Pooled normal plasma or donor plasma was incubated with fX-deficient plasma from a separate donor, and clotting was initiated by the addition of rabbit brain thromboplastin. At 10% (v/v), fX St. Louis II plasma does not shorten the terminal clotting time (Ͼ100 s) of the second fX-deficient plasma, while as little as 0.1% (lowest dilution tested) pooled normal plasma still has an effect (ϳ90 s clotting time, data not shown). In a similar activated PTT-based assay, fX St. Louis II plasma has 3% of the fX activity of pooled normal plasma (data not shown).
Expression of Recombinant fX WT and fX St. Louis II -Recombinant wild-type and mutant fX species were produced in a mammalian cell culture expression system to examine the functional consequences specifically attributable to a glycine substitution engineered at residue 7. Expression levels from multiple clones were similar, consistent with the hypothesis that the mutation has no impact on protein folding or secretion. The recombinant fX species are indistinguishable by SDS-PAGE under reducing and non-reducing conditions. Activation of Recombinant fX WT and fX St. Louis II -At 100 pM RVV-X, 150 nM fX WT is completely activated in 60 min. In contrast, fX St. Louis II is Ͻ10% activated under the same conditions, and the initial rate of activation is less than 5% of the fX WT rate (data not shown). Even at 1000 ϫ the concentration of activator (100 nM RVV-X), it takes more than 1 h to completely activate fX St. Louis II , instead of Ͻ 2 min for fX WT (Fig. 2). Once activated, however, fXa WT and fXa St. Louis II have indistiguishable reaction kinetics with the chromogenic substrate Spectrozyme fXa TM .
Surprisingly, activation of fX St. Louis II by the intrinsic Xase complex (fIXa/fVIIIa) occurs at ϳ30% of the fX WT rate (Fig. 3A). In contrast, activation by the extrinsic Xase complex (fVIIa/TF) is barely detectable under conditions that result in robust activation of fX WT (Ͻ5% in 6 h versus 50% in Ͻ15 min, Fig. 3B). This disparity between the physiologic activation complexes parallels the results of plasma-based assays, which assess both activation and subsequent enzymatic activity. fX St. Louis II has ϳ3% activity in a PTT-based assay but no detectable activity in a PT-based assay (see above).
Prothrombin Activation-The prothrombin used in all exper-iments reported was specially immunodepleted of residual plasma fX as described under "Materials and Methods." This was necessary because the high concentrations of RVV-X needed to completely activate fXa St. Louis II are able to rapidly activate residual plasma fX, even after substantial dilution. A control reaction containing RVV-X but no recombinant fX was carried through all steps of the activation protocols to ensure that the same concentration of RVV-X was always present in the controls. Recombinant fX WT and fX St. Louis II were fully activated by RVV-X and tested at 20 pM for their ability to activate 1.0 M prothrombin on 20 M, 100 nm phosphatidylcholine:phosphatidylserine (3:1) vesicles in the presence of 100 pM fVa (Fig. 4). The fXa St. Louis II cleaves prothrombin to thrombin at 30% of the normal rate under these conditions. The catalytic function of the mutant fXa species was also compared with fXa WT on activated platelets (Fig. 5). The ability of the mutant to activate prothrombin is attenuated but clearly measurable. The apparent K d of fXa WT for the activated platelet surface is ϳ50 pM in this system, consistent with values reported for plasma fXa. In contrast, fXa St. Louis II demonstrates a much lower and linear response out to 1.0 nM, the highest concentration tested. The simplest interpretation of these data is that the apparent K d of fXa St. Louis II for the activated platelet surface is about 20 times higher than the apparent K d of fXa WT .
Thrombin formation by fXa WT and fXa St. Louis II species was also examined in the absence of a phospholipid surface. Without fVa, 10 nM recombinant fXa St. Louis II activates 100 M prothrombin at a very slow but similar rate compared with fXa WT (Fig.  6A). The addition of cofactor, but still no phospholipid, accelerates prothrombin activation by both recombinant molecules to about the same extent. The activities of 5.0 nM fXa St. Louis II or 5.0 nM fXa WT toward 1.0 M prothrombin in the presence of 50 nM fVa are comparable (Fig. 6B). Taken together, these data suggest that fXa St. Louis II has intact catalytic potential compared with fXa WT . The functional loss most likely is a consequence of decreased membrane binding capacity.
Inhibition by Antithrombin-The ability of ATIII to inhibit recombinant fXa WT and fXa St. Louis II was compared in order to probe their catalytic pockets in another surface-independent reaction. They are inhibited at the same rate (Fig. 7), with a derived second-order rate constant for inhibition of 1.66 ϫ 10 3 M Ϫ1 s Ϫ1 for both. The substitution of glycine at position 7 in the Gla domain has no impact on the inhibition of fXa by ATIII.
Decreased Autocatalytic Degradation of fX St. Louis II -Normal fXa proteolytically degrades itself by reactions that are accelerated on phospholipid surfaces (17,18). The fastest of these cleavages releases a small glycopeptide from the carboxyl terminus of the heavy chain, forming fXa␤ from fXa␣. Since fXa␣ and fXa␤ have equivalent catalytic capacities, fXa␤ formation is most easily monitored by SDS-PAGE analysis. The autocatalytic capacity of both recombinant species at 150 nM on rabbit brain cephalin was examined (Fig. 8). In 10 min, fXa␤ WT is the major species. In contrast, 2 h elapse before an increase in fXa␤ St. Louis II is apparent. The increased stability of fXa␣ St. Louis II most likely reflects the reduced membrane binding capacity of this variant molecule. DISCUSSION The concentration of fX St. Louis II antigen circulating in the donor plasma is within the normal range, but the protein has Ͻ1% activity based on its ability to correct the PT of other factor X-deficient plasmas in mixing studies. Since fX St. Louis II does not react with a calcium-dependent monoclonal antibody that recognizes the Gla domain, we isolated the light chain for sequence analysis and detected glycine in cycle 7 with a yield consistent with molar equivalency to residues of surrounding cycles. Amplification and sequence analysis of all 8 exons and splice junctions from the donor DNA revealed a single A3 G transition at position 1200 in exon II, which explains the glycine substitution. A normal allele was not detected. These data in aggregate suggest that the donor is homozygous for this molecular defect. By engineering the mutation into a wild-type fX expression vector and producing proteins that differ only at this residue, we have confirmed that this single alteration causes the functional deficit of fX St. Louis II .
The crystal structure of fVIIa/TF has recently been reported (19). It was concluded that the Gla domain and the first growth factor-like domain of fVIIa act as a cohesive unit joined to the second growth factor-like domain and the catalytic domain, which also function as a unit, via a "hinge" formed by a short helical turn. FX and fVII are highly homologous; the genes encoding them are even within a few kilobases of DNA on chromosome 13. It is reasonable to expect them to have similar structural organization. Our data for fX St. Louis II are all consistent with this view, i.e. the consequences of the glycine substitution at residue 7 in the Gla domain are limited to interactions involving the Gla domain and possibly EGF-1. The catalytic domain is entirely normal as evidenced by unperturbed interactions with a chromogenic substrate, with antithrombin, and with prothrombin in the absence of phospholipid, both with and without the accelerating cofactor, fVa. Moreover, since fX St. Louis II is made and secreted normally, the conformation of the Gla domain must not be critical for intracellular folding, processing, or secretion.
All interactions requiring membrane binding are abnormal, as expected, given the known dependence on an intact Gla domain for this capability. Activation by RVV-X, markedly attenuated for fX St. Louis II , is not membrane-associated but does require an intact Gla domain (20). The most surprising finding of this study is the variability of the impact of the glycine substitution on different membrane-associated reactions. Compared with fX WT tested under identical conditions, activation by fVIIa/TF is reduced more than 200-fold while activation by fIXa/fVIIIa is only 3-fold less. Prothrombin activation in the presence of fVa is about 30% efficient on PC:PS vesicles but only 5% as effective as fX WT on activated platelets. Of course there are assay variables, including the concentrations of reactants and the composition of the membrane surfaces, that preclude definitive conclusions about relative contributions of the Gla domain to these different interactions. Systematic investigation of these variables, including direct measurement of membrane binding under different conditions, is beyond the scope of this report. These results do, however, dramatically underscore that the importance of the Gla domain cannot be generalized from any single model reaction system. Two independent reports describe molecular defects involving substitution of Gla 14 in the fX Gla domain. Substitution with glycine, fX Ketchikan (21), or lysine, fX Vorarlberg (22), causes a mild bleeding tendency. The activity of fX Vorarlberg in a PTbased assay is 5% of normal compared with 25% of normal in a PTT-based assay. fX Vorarlberg is activated by fVIIa/TF or fIXa/ fVIIIa at 17% and 76% of the normal rate, respectively, and activation by RVV-X is normal. Our data for the substitution of Gla 7 generally parallel these findings. However, even though substitution by glycine is more conservative than a change to lysine, substitution of residue 7 causes a more severe phenotype. The crystal structure of the calcium-loaded form of prothrombin fragment 1 (23) provides a likely explanation. Gla 8 , which is homologous to Gla 7 of fX, is coordinated by 4 bonds to 3 of the 7 calcium atoms bound to the Gla domain, whereas Gla 15 , homologous to Gla 14 of fX, has only one bond to a single calcium.
In conclusion, the functional deficits of fX St. Louis II are completely explained by the glycine substitution for Gla at residue 7 discovered for the donor protein and DNA. This substitution has a marked impact on reactions involving the Gla domain, which is understandable in the context of the published crystal structure of the prothrombin Gla domain. However, the relative importance of this domain varies significantly depending on the reaction components studied. Both fX WT (A) and fX St. Louis II (B) were activated by RVV-X and incubated with rabbit brain cephalin as described under "Materials and Methods." Aliquots of the reaction were removed and quenched into SDS buffer at 0, 10, 30, 60, and 120 min (lanes 1-5). The samples were resolved by SDS-PAGE under non-reducing conditions, transferred to nitrocellulose, and visualized with monoclonal antibodies directed against the fX light chain.
FIG. 7. Inhibition of fXa WT and fX St. Louis II by ATIII. ATIII at 0 (E), 0.5 (q), 1.0 (ç), 2.0 (å), 3.0 (Ⅺ), or 4.0 (f) M was incubated with 0.5 nM of fXa WT (A) or fXa St. Louis II (B) and 100 M Spectrozyme fXa TM to generate progress curves for the reaction. Apparent first-order rate constants (Kobs sec Ϫ1 ) were derived from the fitted curves as described under "Materials and Methods" and plotted (C) as a function of ATIII concentration to compare apparent second-order rate constants.