Cooperative activation of human factor IX by the human extrinsic pathway of blood coagulation.

The activation of human coagulation factor IX by human tissue factor.factor VIIa.PCPS.Ca2+ (TF.VIIa.PCPS.Ca2+) and factor Xa.PCPS.Ca2+ enzyme complexes was investigated. Reactions were performed in a highly purified system consisting of isolated human plasma proteins and recombinant human tissue factor with synthetic phospholipid vesicles (PCPS: 75% phosphatidylcholine (PC), 25% phosphatidylserine (PS)). Factor IX activation was evaluated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, [3H]factor IX activation peptide assay, colorimetric substrate thiobenzyl benzyloxycarbonyl-L-lysinate (Z-Lys-SBzl) hydrolysis, and specific incorporation of a fluorescent peptidyl chloromethyl ketone. Factor IX activation by the TF.VIIa.PCPS.Ca2+ enzyme complex was observed to proceed through the obligate non-enzymatic intermediate species factor IX alpha. The simultaneous activation of human coagulation factors IX and X by the TF.VIIa.PCPS.Ca2+ enzyme complex were investigated. When factors IX and X were presented to the TF.VIIa complex, at equal concentrations, it was observed that the rate of factor IX activation remained unchanged while the rate of factor X activation slowed by 45%. When the proteolytic cleavage products of this reaction were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, it was observed that the intermediate species factor IX alpha was generated more rapidly when factor X was present in the reaction mixture. When factor IX was treated with factor Xa.PCPS in the presence of Ca2+, it was observed that factor IX was rapidly converted to factor IX alpha. The activation of factor IX alpha by the TF.VIIa.PCPS.Ca2+ complex was evaluated, and it was observed that factor IX alpha was activated more rapidly by the TF.VIIa.PCPS.Ca2+ complex than was factor IX itself. These data suggest that factors IX and X, when presented to the TF.VIIa.PCPS.Ca2+ enzyme complex, are both rapidly activated and that factor Xa, which is generated in the initial stages of the extrinsic pathway, participates in the first proteolytic step in the activation of factor IX, the generation of factor IX alpha.

Human factor IX (Christmas factor)' is a 57,000-dalton single chain glycoprotein that is essential for normal blood coagulation (1). The biologic importance of factor IX in hemostasis is illustrated by the disease hemophilia B in which patients with congenital defects in factor IX synthesis or function develop a severe bleeding diathesis. Factor IXa participates in the middle phase of the blood coagulation pathway by activating factor X to Xa. Factor XIa in the presence of Ca2+ activates factor IX to factor IXaP (2). Factor IX can also be activated by the extrinsic coagulation pathway (tissue factor and factor VIIa) (3). In spite of in vivo evidence demonstrating the importance of factor IX, enigmatic data from in uitro studies have not satisfactorily explained how defects in the function of factor IX, or factor VIII, inhibits the propagation of blood coagulation. It has been shown that the lack of plasma activators of factor IX by the intrinsic pathway (factors XI and XII) results in mild or no bleeding disorders (4). Furthermore, a recent report by Bauer et al. (5) suggested that patients with hereditary factor VI1 deficiency activated less factor IX in uiuo than either normal controls or patients with factor XI deficiency. These reports would suggest that the physiologically important factor IX activation may proceed via the extrinsic pathway. However, when the relative rates of human (6) and bovine (7) factors IX and X activation by the extrinsic pathway were studied, it was observed that factor X was activated between two and seven times more rapidly than bovine factor IX using in uitro activation peptide release assays. From these data it was generally concluded that factor X activation by the extrinsic pathway could rapidly bypass the need for factor IX activation. In contrast to these studies, a recent report by Komiyama et al. (8) has suggested that factor IX may be the preferred substrate for the TF2 .VIIa enzyme complex in a purified reaction system. Thus, there appears to be little agreement on the physiological role of factor IX activation by either coagulation pathway.
The role of factor Xa in the activation of factor IX is ' The nomenclature used for factor IX and its activation products is that of Davie and co-workers (1, 2): factor IX, single-chain native factor IX ( M , = 57,000); factor IXa, two-chain inactive intermediate consisting of a heavy chain ( M , = 39,000) and a light chain ( M , unclear. It has been shown that bovine factor Xa can slowly activate bovine factor IX to IXaP (2, 9). However, it has been reported that human factor Xa does not activate factor IX to IXaP (3). This has led to the general conclusion that if factor Xa activates factor IX to IXaP, this reaction proceeds so slowly that it is of little physiologic significance (2). During the activation of factor IX, two peptide bonds must be hydrolyzed to generate the active enzyme species IXaP ( Fig. 1) (1, 10). The mechanism of factor IX activation by the TF .VIIa .phospholipid. Ca2+ enzyme complex has been an area of dispute. Zur and Nemerson (11) reported that during the activation of bovine factor IX by the bovine TF.bovine VIIa. mixed brain phospholipids complex that the Arg'45and Arg'so-Val's' bonds were cleaved so rapidly that no intermediate species was observed upon analysis of the reaction. They concluded that the rate-limiting step in the activation of bovine factor IX was the cleavage of the first peptide bond Arg'45-Ala'46. Data reported by Bajaj et al. (10) suggested that the activation of human factor IX by human TF. human VIIa e mixed brain phospholipid complex occurs in a two-step process. In their model the first step of factor IX activation occurs with the cleavage of the Arg'45-Ala146 peptide bond, which forms the two-chain inactive intermediate IXa. Factor IXa is then cleaved by a second proteolytic step at the ArglsO-Val''' peptide bond resulting in the formation of the active enzyme species IXaP. During this second step, a 35-amino acid activation peptide (AP) is released containing residues 146-180 (Fig. 1). From these two reports, it is unclear whether the difference in activation mechanisms of factor IX can be explained by species differences between the bovine and human molecules or if different methodologies lead to different conclusions about the basic activation mechanisms of factor IX by the extrinsic pathway.
In this report, we clearly establish the order of the proteolytic steps involved in the activation of human factor IX by the human extrinsic pathway (TF e VIIa. PCPs. Ca2+). We have also investigated the relative rates of human factor IX and human factor X activation by the extrinsic pathway when these proteins are presented to the TF VIIa. PCPs. Ca2+ enzyme complex as simultaneous substrates. Lastly, the role of human factor Xa. PCPs. Ca2+ in the activation of human factor IX was investigated.

Proteins
Human coagulation factors IX and X were isolated from fresh frozen plasma obtained from the Vermont Red Cross using the general methods of Bajaj et al. (15) with modifications described by Krishnaswamy et a1.3 Generally proteins were isolated from fresh frozen plasma by precipitation with barium citrate. The precipitated proteins were subjected to anion-exchange chromatography using DEAE-Sepharose (Pharmacia LKB Biotechnology Inc.). The protein peak which contained factors IX and X was pooled and subjected to a second chromatographic step on heparin-Sepharose which allowed for distinct resolution of factors IX and X. Factor IX and X were concentrated by ultrafiltration followed by precipitation in 80% saturated (NH,),SOI. Precipitated proteins were collected by centrifugation and resuspended in 50% (v/v) glycerol and water and stored at -20 "C. From seven liters of human fresh frozen plasma, 12 mg of factor IX and 15.6 mg of human factor X were isolated. Factor Xa was prepared by activating factor X with purified Russell's viper venom as described previously (16). Purified human factor VI1 was prepared by the methods of Jenny et al. (14) and was obtained from Haematologic Technologies, Essex Junction, VT. Recombinant human tissue factor (TF) was provided as a gift from Dr. Hissids of Genentech, Inc. Recombinant human factor VIIa (rVIIa) was provided as a gift from Dr. Ulla Hedner, Novo Pharmaceuticals.
For the proteins used in this study concentrations were calculated using the following molecular weights (M,) and extinction coefficients (Go"): human factor VIIa/VIIa, 50,000, 1.39 (17); human factor IX, 57,000, 1.33 (18); human factor X, 58,900, 1.16 (18); human factor Xa 45,000, 1.16 (18). The concentration of recombinant human TF was determined by amino acid analysis provided by Genentech Inc. (19). 3H-Labeled Factor IX was prepared by the methods of Van Lenten and Ashwell (20) with modifications described by Zur and Nemerson (11). After the labeling procedure the molar ratio of tritium to factor IX was calculated to be 0.42 mol of 'H/mol of factor IX (8.62 X lo9 cpm/mol of 3H factor IX). The labeled protein possessed 74% of the biological activity of the nonlabeled control factor IX when tested in coagulation assays generally described by Proctor and Rapaport (211, using factor IX deficient plasma supplied by George King Co. and aPTT reagent provided by Sigma.

Methods
Relipidation of Recombinant Human Tissue Factor-Human recombinant tissue factor was solubilized in 0.04% (v/v) n-dodecyl octaethylene glycol ether (C12E8, M, = 539). The solubilized tissue factor was mixed with PCPs and incubated at 37 "C for 30 min in 0.02 M Hepes, 0.15 M NaCI, pH 7.4 (HBS). The TF.n-dodecyl octaethylene glycol ether. PCPs mixture was then diluted to a final concentration of 14 nM tissue factor, 1 mM PCPs, and 1.7 X lo-'% n-dodecyl octaethylene glycol ether. In control experiments where recombinant tissue factor was relipidated with 0.8% octyl glucoside followed by exhaustive dialysis using the general methods of Bach et al. (22), no difference in cofactor activity was observed when compared with tissue factor.PCPS preparations which contained a final S. Krishnaswamy, K. A. Field, T. S. Edgington, J. H. Morrissey, and K. G . Mann, manuscript in preparation. concentration of 1.7 X n-dodecyl octaethylene glycol ether (Fig.  2). These data illustrate that the small amount of detergent in our reaction system had little effect on the progress curves reported in this study.
The concentration of PCPs used in this study was chosen from the original work of Broze et al. (23), where optimal tissue factor activity was observed with a 1000-fold excess of lipid to tissue factor (w/w). The ratio of PCPs to tissue factor used in our studies was 1200:l. In preliminary experiments it was observed that concentration of PCPs (between 0.5 and 1 mM) provided maximal tissue factor activity (data not shown). This is in general agreement with the work of Broze et al. (23). Thus, the standard lipid concentration of 1 mM was used for all TF. VIIa experiments in this report.
Activation of Human Factor ZX, Factor X, or Factors ZX and X by the TF. VIIa.PCPS.Ca2+ Complex-After relipidation of the recombinant tissue factor, natural factor VII/VIIa or rVIIa (5 nM) was added to the reaction mixture and incubated at 37 "C for 15 min. The activation of human factor IX, factor X, or factors IX and X were carried out under the following reaction conditions: tissue factor (14 nM), PCPs (1 mM), and CaClz (5 mM) at 37 "c in 0.02 M Hepes, 0.15 M NaCl, pH 7.4 (HBS). After the TF'VIIa. PCPs. CaZ+ enzyme complex was assembled, substrates of factor IX, factor X, or a mixture of factors IX and X (2 p~ for each zymogen) were added to the reaction mixture, and the reaction was started. A progress curve of zymogen activation was generated by removing aliquots from the reaction mixture at designated time points and quenching the reaction in 25 mM EDTA (final concentration). The progress of various reactions was monitored by SDS-PAGE, colorimetric substrate hydrolysis, [3H]factor IX activation peptide release assay, and incorporation of Ir-FPR-ck into polypeptides which contain serine protease active sites. Colorimetric Substrate Assays-Factor IX activation was monitored by two step colorimetric substrate hydrolysis as generally described by Green and Shaw (24). Factor IX was added to the assembled TF. VIIa. PCPs. Ca2+ enzyme complex, and reaction subsamples were removed from the activation reaction at designated time points and quenched in EDTA (25 mM) as described above. A 100-p1 aliquot from each reaction time point was added to 60 p~ thiobenzyl benzyloxycarbonyl-L-lysinate (Z-Lys-SBzl) and 60 p~ dithiolnitrobenzoic acid diluted in HBS, pH 7.4. The rate of Z-Lys-SBzl hydrolysis was followed spectrophotometrically by allowing the released benzyl mercaptan to react directly with the dithiolnitrobenzoic. The increase in absorbance at 405 nm was due to the formation of the 3-carboxy-4nitro-phenoxide ion (Elll 13,600 M" cm"), which was FIG. 2. The effect of n-dodecyl octaethylene glycol ether on the tissue factor activity. Tissue factor activity was measured in the presence (0) or absence (A) of 1.7 X n-dodecyl octaethylene glycol ether. Tissue factor activity was determined by factor X activation. The enzymatic activity of factor Xa generation was measured by removing subsamples from an ongoing reaction of factor X activation. Reaction subsamples were removed and quenched in 25 mM EDTA at designated time points. Each reaction time point was then mixed with 200 p~ Spectrozyme Xa, and the rate of absorbance change at 405 nm was monitored by a Molecular Devices Vmax spectrophotometer. The concentration of components of the reaction were: 500 nM factor x, 1 nM W a , 14 nM TF, 1 mM PCPs, and 5 mM CaCL in HBS buffer, pH 7.4, at 37 'C. Under these conditions it was observed that n-dodecyl octaethylene glycol ether had no effect on the progress curve of factor X activation when compared with a detergent-free system of tissue factor which had been relipidated using octyl glucoside followed by extensive dialysis.
continuously in a Molecular Devices Vmax spectrophotometer. Factor X activation was monitored by chromogenic substrate hydrolysis using the substrate Spectrozyme Xa. Reaction subsamples of factor X activation were removed and quenched in EDTA (25 mM) as described above. Each subsample was mixed with 200 pM Spectrozyme Xa (final concentration) in HBS, pH 7.4. The rate of substrate hydrolysis was monitored by an absorbance change at 405 nm in a Molecular Devices Vmax spectrophotometer.
3H-Labeled Factor IX Actiuatwn Peptide Release A~say-[~H]Factor IX activation by the TF. VIIa. PCPs. Ca2+ enzyme complex was monitored by the method of Zur and Nemerson (11) with the following modifications.
[3H]factor IX (2 p~) was added to the assembled TF. VIIa.PCPS.CaZ+ complex as described above. At designated time points, a 60-p1 aliquot of the reaction mixture was removed and immediately treated with an equal volume of 50 mM EDTA. Thirty microliters of the EDTA-treated sample were added to 200 pl of 0.1% bovine serum albumin in 0.02 M Tris, 0.15 M NaCl, pH 7.4 (TBS), and the remaining 90 pl of the EDTA-treated sample from each time point was treated with SDS, 2-mercaptoethanol, and prepared for SDS-PAGE. This mixture was then added to 250 p1 of ice-cold 10% trichloroacetic acid and allowed to incubate on ice for 3 min. The sample was then centrifuged for 5 min at 9000 X g in a Beckman Microfuge at 22 "C. Duplicate 100-pl aliquots of the supernatant were added to 5 ml of aqueous scintillation fluid and counted for tritium in a liquid scintillation counter. The rate of 'H activation peptide release, as measured by an increase in trichloroacetic acid-soluble tritium, was compared with the rate of [3H]factor IXaP generation as monitored by SDS-PAGE. Approximately 35% of the total ['Hlfactor counts remained soluble in the trichloroacetic acid supernatant when a plateau was observed in the reaction progress curve. This plateau corresponded well with the complete conversion of [3H]factor IX to ['Hlfactor IXaP as documented by SDS-PAGE analysis of the reaction as described above. Control experiments were performed with no factor VIIa and/or tissue factor in the activation reaction mixture to evaluate for possible enzyme contamination of the tissue factor, factor VIIa, and factor IX preparations.
Electrophoresis-Factor IX activation was monitored by SDS-PAGE using 10% polyacrylamide gels. Subsamples (50 pl) of activation reactions were brought to final concentration of 2% (w/v) SDS, 0.01 M Tris-HC1, pH 6.8, 10% glycerol, and 2% (v/v) 2-mercaptoethanol, heated for 5 min at 90 "C, and analyzed by SDS-PAGE as described by Laemmli (25). Protein bands were visualized by staining the gels with Coomassie Brilliant Blue and destaining by diffusion.
Treatment of Reaction Samples with lr-FPR-&-Activation reactions of both factors IX and X were monitored by treatment of reaction subsamples with Ir-FPR-ck. lr-FPR-ck was used to identify polypeptides which contain serine protease active sites. Subsamples from activation reactions (described above) were treated with lr-FPRck as described by Williams et al. (13) with the following modifications. EDTA-treated subsamples from activation reactions were incubated with Tris-HC1 (50 mM), pH 7.4, and lr-FPR-ck (30 p~) for 30 min at 37 "C prior to the addition of SDS and 2-mercaptoethanol. Reaction subsamples were then subjected to SDS-PAGE as described above. Protein bands which specifically incorporated lr-FPR-ck were visualized by excitation with a long-wave UV light box and photographed through a 540-nm cut-off filter. Total protein from each reaction time point was visualized on the gel by staining with Coomassie Brilliant Blue.
Brilliant Blue-stained SDS-PAGE gel of factor IX and X activation Demitometric Analyses-Densitometric analyses of the Coomassie reactions were performed with a Microscan 1000 scanning densitometer (TRI Inc.) equipped with a solid-state linear diode array camera to digitize images through a photographic lens. Digitized images contain approximately 0.25 X lo6 pixels with each pixel measuring 0.16 mm X 0.16 mm. Linearity of the system is 0-2.5 absorbance units. Data were analyzed with a 80826-based computer equipped with a math coprocessor and software which allows for either automatic or manual background subtraction and full editing capability. Data are expressed as integrated volumes for each protein band using the arbitrary density units of the scanning system.
Activation of Factor I X by Factor Xu-Factor IX was converted to IXa by factor Xa by mixing PCPs (400 p~) with factor Xa (at various concentrations ranging from 80 to 400 nM) and CaClz (5 mM) in HBS, pH 7.4. This mixture was incubated at 37 "C for 5 min prior to the addition of factor IX (2 p~ final) to the reaction. Aliquots were removed from the reaction at designated time points, quenched in EDTA (25 mM), and rapidly denatured in SDS and 2-mercaptoethanol. Protein samples were then subjected to SDS-PAGE analysis as of Human Factor IX described above, followed by scanning densitometry of the Coomassie Brilliant Blue-stained gel.
Isolation of I X a by Zmmunoaffinity Chromatography-Factor IX (6 p M ) was converted to 1x0 by factor Xa (100 nM), PCPs (400 pM), and CaC1, (5 mM) in HBS, pH 7.4, at 37 "C. The reaction was allowed to proceed for 30 min and was stopped by the addition of EDTA (final 25 mM). Two milliliters of the reaction mixture were loaded onto an anti-factor IX-Sepharose antibody column (10-ml bed volume equilibrated in the same buffer at 4 "C) and 1-ml fractions were collected. The column was washed with the same buffer until the absorbance at 280 nm was less than 0.05. Proteins that remained bound to the column were eluted with 3 M sodium thiocyanate (NaSCN), pH 7.4, in HBS. The corrected absorbance was measured, the eluted protein pooled, and then dialyzed exhaustively in HBS, pH 7.4, at 4C. Factor IX was also subjected to immunoaffinity chromatography on the anti-factor IX Sepharose column. Factor IX was eluted from the anti-factor IX Sepharose with 3 M NaSCN. The pooled factor IX peak was then exhaustively dialyzed into HBS at pH 7.4 at 4 "C. This factor IX was then used as a control in the factor IXa activation experiments described below.
Immunoaffinity isolated factor IX or factor IX (final 1.4 p M ) was activated to IXaP by the addition of these substrates to the TF. VIIa. PCPs. Ca2+ enzyme complex as described above. A progress curve of factor IXa or factor IX activation was generated by removing aliquots from the reaction mixture at designated time points and quenching each aliquot immediately in EDTA (25 mM). Proteins were then treated with SDS, 2-mercaptoethanol, and prepared for SDS-PAGE analysis as described above. Polyacrylamide gels were then stained with Coomassie Brilliant Blue and subjected to scanning densitometry. The density of each protein band visualized on the gel was recorded and plotted as a function of total protein detected in each electrophoresis lane.
The Activation of Factor IX by Both the Xa. PCPS.Ca2+ and TF.
Vlla.PCPS.Ca2+ Enzyme Complexes-Factor IX (2 p~) was added to a mixture of factor Xa (400 nM), PCPs (400 pM), and CaC1, (5 mM) in HBS at 37 "C. At designated time points, after the addition of factor IX to the Xa.PCPS.Ca*+ complex, subsamples from the reaction mixture were removed and quenched in 25 mM EDTA. Twenty-nine minutes and forty-five seconds after the initiation of the reaction a mixture of TF (14 nM), factor VIIa (5 nM), PCPs (1 mM final), and CaCI2 (5 mM final) was added to the reaction. At designated time points after the addition of the T F .VIIa .PCPs. Ca2+ complex to the reaction system, aliquots from the reaction were removed and rapidly quenched in 25 mM EDTA. Reaction subsamples were then treated with SDS, 2-mercaptoethanol, and analyzed by SDS-PAGE as described above. The Coomassie Brilliant Blue-stained gels were then analyzed by scanning densitometry. The density of each protein band visualized on the gel was recorded and plotted as a function of the total protein detected in each electrophoresis lane.
Data Analysis-Progress curves of factor IX, factor X, and factor IXa activation were analyzed using a nonlinear least squares regression analysis (NLLSQ; Cet Research Corp., Normand, OK) using the Marquardt algorithm (26) on an IBM PS-30 computer. Curve fit analysis has been described in detail by Krishnaswamy et al. (27). Progress curves of factor IX activation were modeled through Equation l, where factors IX, IXa, and IXaP represent the zymogen factor IX, the non-enzymatic intermediate factor IXa, and the enzymatic product of the reaction factor IXaP (see Fig. 1) and where k, and k z represent the rate constants for the two reactions.
The solutions for IX, IXa, and IXap as a function of time are given below in Equations 2-4.
where IX, is the initial concentration of factor IX, k, and kZ represent the rate constants for independent steps 1 and 2, and IX,, IXa,, and IXa& represent the relative concentration of factor IX, IXa, and IXap at time t. Rate constants k, and k, reported in the results section were determined by fitting data for the formation of factor IXaP to Equation 4 using the IX, concentration of 2 p~. The overall rate of factor IX activation was calculated from the slope of the product factor IXaP formed per unit time as plotted in the results section. Progress curve fitting for factor X activation was analyzed using to the integrated form of the Michaelis-Menten equation assuming the conversion of a single substrate to product.

The Activation of Factor ZX by the T F . VZZa. PCPs. Ca2+
Enzyme Complex-Purified human factor IX (2 PM final concentration) was added to the TF-VIIa. PCPs. Ca2+ enzyme complex as described under "Experimental Procedures." Factor IX activation by the TF. VIIa. PCPs. Ca2+ complex was evaluated by four different methods: SDS-PAGE, colorimetric substrate hydrolysis, [3H]factor IX activation peptide assay, and specific incorporation of the fluorescent-labeled lr-FPR-ck. SDS-PAGE analysis of the cleavage products from the activation of the human factor IX activation reaction illustrates that when factor IX i s activated by the TF. VIIa. PCPS.Ca*' complex, the reaction proceeds through the intermediate species factor IXa (Fig. 3). The Coomassie Bluestained gel represented in Fig. 3A was evaluated by scanning densitometry to quantify the factor IX activation reaction sequence and generate a progress curve of factor IX activation. Relative staining intensities of each protein band were plotted versus reaction time as shown in Fig. 3B. Factor IXaP formation as a function of time was fit to Equation 4 as described under "Experimental Procedures" using the IX, concentration of 2 PM, and the rate constants kl and kz for the two reactions were determined by nonlinear least squares regression analysis. The values for kl and kz were 5.6 X s" and 4.7 X s-', respectively. The overall rate of factor IX conversion to factor IXaP was calculated as the rate of IXaP formed per min/mol of TF. VIIa complex. Using the known dissociation constant of factor VIIa for tissue factor it was calculated that the effective enzyme concentration in our reaction system was 4.5 nM TF-VIIa complex under equilibrium conditions. From this the overall rate of factor IXaP formation was determined to be 49 mol of factor IX min"/mol of VIIa. The activation pathway observed here is consistent with the report of Bajaj et al. (10) in which an intermediate species was observed in the reaction sequence of human factor IX by the proteolytic cleavage of the Arg'45-Ala'46 bond. This reaction step is then followed by the cleavage of a second peptide bond at residues Arg'80-Va11s1. This proteolytic step then causes the release of the activation peptide and generates the active enzyme species IXaP (10). The other possible intermediate in the factor IX activation pathway, factor IXaa (steps 3 and 4, Fig. 1) was not observed in any activation reactions analyzed in this report suggesting that factor IXa is the sole intermediate in factor IX activation by the TF-VIIa. PCPs. Ca2+ complex.
To identify which polypeptide in the factor IX activation sequence contained a serine protease active site, subsamples from the factor IX activation reaction were treated with lr-FPR-ck. From these experiments it was observed that only the polypeptide which migrated as the heavy chain of activated factor IX (ZXaP hc) M, = 28,000 specifically incorporated the fluorescent label (Fig. 4). This provides strong evidence that only the heavy chain of activated factor IXaP expresses the serine protease active site. At no time was the IXa band observed to incorporate lr-FPR-ck, which is consistent with the conclusion that the IXa species is a nonenzymatic intermediate of the factor IX activation process.
A progress curve of factor IX activation was also generated by removing aliquots from the reaction mixture and assaying each reaction sample for the ability to hydrolyze the thiob- enzyl ester Z-Lys-SBzl. This enzymatic hydrolysis was compared to the rate of factor IXaP hc generation as measured by scanning densitometry of a reduced Coomassie Brilliant Blue-stained gel (Fig. 5). The progress curve from this reaction showed that factor IX was completely converted to IXaP after 15 min and that the initial rate of factor IXaP generation under these reaction conditions was 46 mol min"/mol of VIIa. These data illustrate that enzymatic activity of factor IXa& as reflected by Z-Lys-SBzl hydrolysis, parallels the generation of the polypeptide species which corresponds to The activation of ['Hlfactor IX by the TF-VIIa. PCPs Ca2+ complex was investigated. The rate of ['Hlfactor IX activation was measured using both the activation peptide release assay and scanning densitometry of a Coomassie Brilliant Blue-stained gel, which contained reaction subsamples from a ["Hlfactor IX activation reaction. These data were compared with the rate of unlabeled factor IX activation as measured by scanning densitometry of a Coomassie Brilliant Blue-stained gel and colorimetric substrate hydrolysis (Fig.  5). Data in this figure illustrate that the rate of ["Hlfactor IX activation by the TF-VIIa. PCPs. Ca2+ complex was 45% slower than the rate of unlabeled factor IX activation under identical experimental conditions. However, as noted under "Experimental Procedures," the ['Hlfactor IX that was prepared by the methods of Zur and Nemerson (11) had only 74% of the bioactivity of unlabeled factor IX when tested in factor IX-deficient coagulation assays.
When the rate of [:3H]factor IX activation, as measured by release of the ["Hlfactor IX activation peptide, was compared with the rate of ["H]factor IX activation, as measured by scanning densitometry of Coomassie Brilliant Blue-stained gel, it was observed that the release of the 'H activation peptide paralleled the appearance of the ['Hlfactor IXaP heavy chain M , = 28,000 (Fig. 5). These data support the work of Zur and Nemerson (11) that the activation peptide assay is useful for evaluating the progress of ['Hlfactor IX activation. However, the significant loss in bioactivity of the ['HI factor IX after labeling the protein with sodium ['HI borohydride causes the usefulness of this assay for the evaluation of factor IX activation kinetics to be questioned. In control experiments, when no factor VIIa was added to the enzymatic complex (TF.PCPS.Ca2+), no activation of [3H] factor IX was detected by either an increase in soluble ' H activation peptide in 10% trichloroacetic acid or in reaction samples analyzed by SDS-PAGE over a 120-min reaction period. This illustrates that no contaminating protease was activating factor IX in our TF preparation. Furthermore, when no TF was added to the reaction system (VIIa. P C P s . Ca2+), no activation of [3H]factor IX was observed over a 120min reaction period. This demonstrates that our preparation of factor VIIa contained no contaminating protease responsible for the activation of factor IX that functioned independently of TF and further illustrates the importance of the entire TF VIIa. P C P s . Ca2+ complex to be assembled for the rapid activation of factor IX (data not shown).
When factor IX activation reactions were evaluated using rVIIa, no difference in the activation progress curve of factor IX was observed whether purified human factor VII/VIIa or rVIIa was used as the enzymatic cofactor in the above reactions (data not shown). Both the purified human factor VII/ VIIa and rVIIa were added to the TF.PCPS.Ca2+ complex 15 min prior to the addition of factor IX. These data suggest that all of the factor VI1 in the factor VII/VIIa preparation was converted to factor VIIa prior to the addition of factor IX, when the former is complexed to TF. PCPS.Ca2'.
The Activation of Factor X by the TF, VIIa. P C P s . Ca2+ Complex-The rate of factor X activation by the TF. VIIa. PCPS-Ca2+ complex was investigated. Human factor X was added to the enzyme complex of TF. VIIa as described under "Experimental Procedures." A progress curve of factor X activation was generated by removing aliquots of the reaction mixture and assaying each reaction sample for the amount of factor Xa generated. Factor Xa generation was monitored by hydrolysis of the chromogenic substrate Spectrozyme Xa and by densitometric analysis of a Coomassie Brilliant Bluestained gel (Fig. 6). It can be seen from these data that the densitometric analysis of the generation of the Xaa and XaP bands is consistent with the generation of the active enzyme species as reflected by chromogenic substrate hydrolysis. A progress curve from this reaction showed that factor X was completely converted to factor Xa after 10 min, and the initial rate of factor X conversion to factor Xa was calculated to be 78 mol of factor Xa min"/mol of factor VIIa.
From these data we have observed that when human factors IX and X were activated independently by the TF.VIIa. PCPSaCa2+ complex that human factor X was activated nearly two times more rapidly than human factor IX. This is in reasonably good agreement with the report of Morrison and Jesty (6), where it was observed that human factor X was activated about 1.5 times more rapidly than human factor IX in a human plasma system using 'H-labeled factors IX and X.
The Simultaneous Activation of Human Factors IX and X by the TF. VIIa. P C P s . Ca2+ Enzyme Complex-The activation of human factors IX and X was measured simultaneously when presented to the TF. VIIa . P C P s . Ca2+ enzyme complex.
Factor IX activation was measured by densitometric analysis of the Coomassie Brilliant Blue-stained gel shown in Fig. 7. In this reaction system we were unable to use the colorimetric substrate Z-Lys-SBzl to follow the generation of factor IXaP, because the factor Xa that was simultaneously generated in the reaction mixture could also hydrolyze the factor IX substrate. However, as illustrated by Figs. 2-4, scanning densitometry of Coomassie Brilliant Blue-stained gels is an effective means of following the progress of factor IX activation and corresponds well with an increase of enzymatic activity of factor IXaP. Factor Xa formation in this complex reaction was measured both by chromogenic substrate hydrolysis and densitometric analysis of the Coomassie Brilliant Bluestained gel. We were able to use colorimetric substrate hydrolysis to follow factor Xa generation in this reaction because the factor IXaP that was generated simultaneously in this reaction system was unable to hydrolyze the factor Xa substrate Spectrozyme Xa. When the rates of factor IX and X activation were measured simultaneously, the overall rate of factor IXaP generation remained unchanged when compared with the rate of factor IXaP generation measured without factor X present in the reaction system (Fig. 7C). Interestingly, when the rate constants kl and kz for factor IX activation in the presence of factor X were determined, it was observed that lzl increased to 15.7 X lo-' s-', whereas kp decreased slightly to 3.2 X lo-' s-'. The rate of factor Xa formation in the presence of factor IX decreased by 46% to 42 mol min-' of factor Xa generated/mol of factor VIIa when compared with the rate of Xa generation observed when no factor IX was present in the reaction system (Fig.  7 0 ) . Analysis of the cleavage products of this complex reaction by SDS-PAGE (7a) suggested that the first bond in the activation of factor IX (Arg'45-Ala'46) was hydrolyzed at a faster rate than was observed when no factor X was present in the reaction (Fig. 7C). Furthermore the increase in kl of factor IX activation in the presence of factor X suggested that factor X/Xa in the reaction system may be participating in the first proteolytic step in the activation of factor IX. Thus, the role of factor Xa in the activation process of factor IX was investigated.  1, 2, 4,8, 12, 16, 30, and 60 min after the addition of factor X to the established TF.VIIa.PCPS .Ca2+ enzyme complex. Protein bands are identified as factor X, factor Xaa, factor Xa& and factor X IC, respectively. B illustrates the relative concentration of factor X and factor Xa as determined by densitometric analysis of Coomassie Bluestained gels. Gels depicted in A were scanned to determine the relative amounts of each protein a t various points in the reaction. Staining intensities were assigned arbitrary density units as described under "Experimental Procedures." The relative concentration of factor X is depicted with open squares, those of factor Xa (the sum of Xaa and Xap) with open circles. The enzymatic activity of factor Xa generation was also measured during the progress of the reaction. Subsamples from the factor X activation reaction were removed and quenched in EDTA a t designated time points as described above. Each reaction time point was then mixed with 200 PM Spectrozyme Xa, and the rate of absorbance change at 405 nm was monitored by a Molecular Devices Vmax spectrophotometer. The relative amount of factor Xa generation as monitored by chromogenic substrate hydrolysis is depicted with open triangles.

Proteolysis of Factor IX by Factor Xu-The role of factor
Xa. PCPs. Ca'+ activation of factor IX was investigated. Treatment of factor IX with factor Xa e P C P s a Ca'+ was performed as described under "Experimental Procedures." The proteolytic products formed by the treatment of factor IX (2 p M ) with the Xa (80 nM) . P C P s .Ca2+ are shown in Fig. 8A.
This gel illustrates that the Xa-PCPS. Ca'+ complex can rapidly convert factor IX to factor IXa in one proteolytic step. It was observed that factor IX generation increased in a dose-dependent manner with increasing concentrations of factor Xa. Furthermore, this reaction was observed to be absolutely phospholipid-and calcium-dependent. These data illustrate that factor Xa can rapidly generate the intermediate species IXa from factor IX by cleavage of the Arg'45-Ala'4fi bond. We also observed at high concentration of factor Xa (400 nM) that the second Arg'X"-Val'R' bond is cleaved, but at a much slower rate than the first bond cleavage. From these observations we have concluded that factor Xa can rapidly generate the species IXa but is much less efficient in cleaving the second peptide bond (Arg'""-Val'"') and converting factor IXa to IXaP. It is unlikely that this observation was the result of a contaminating protease from the Russell's viper venom that was used in the preparation of factor Xa, because in control experiments, when factor Xa was added to the factor IX reaction without PCPs, no proteolysis of factor IX was observed over a 90-min period. Furthermore it has been reported that Russell's viper venom proteolytically cleaves human factor IX at only the Arg'""-Val'"' peptide bond and generates the product IXaa (1). This product was never observed in any of reactions where human factor Xa was added to the preparation of human factor IX.

Isolation of IXa after Activation with Xu-It was of interest
to find out if the factor IXa that was generated by factor Xa could function as an intermediate in the generation of factor IXaP. In order to investigate this, factor IXa was isolated by immunoaffinity chromatography (Fig. 9). The isolated material, when analyzed by SDS-PAGE under reducing conditions, migrated as a mixture of three protein bands. The two major bands migrated with apparent M , = 39,000 and M , = 18,000, which is consistent with the polypeptides IXa hc and IX IC, respectively. The third protein band migrated with an apparent M , = 28,000 which corresponds to the heavy chain of factor IXaP (IXaP hc). At no time was the M, = 39,000 polypeptide able to incorporate lr-FPR-ck, suggesting that this polypeptide did not contain a serine protease active site which is also consistent with identification of factor IXa.
The isolated factor IXa was tested to evaluate if it could function as a substrate for the TF. VIIa. P C P s . Ca'+ complex. Isolated IXa was added to the TF. VIIa. PCPs. Ca" as described under "Experimental Procedures." Subsamples of the activation mixture were removed from the reaction and analyzed by SDS-PAGE. The Coomassie Brilliant Blue-stained gel was then analyzed by scanning densitometry. From these data, the rate of factor IXaP generation was determined and plotted in Fig. 10. It was observed that the initial rate of factor IXa conversion to IXaP was 40% faster than the initial rate of immunoaffinity-isolated factor IX conversion to IXaP under identical experimental conditions. These data illustrate that the factor IXa species which is generated by proteolytic cleavage by factor Xa can function as a substrate for the TF VIIa enzyme complex. Furthermore, an increase in the rate of product formation when the reaction is started with factor IXa is consistent with the notion that factor IXa only requires one proteolytic cleavage event (Arg'"-Val'"') to become activated to the IXaP species, whereas factor IX requires two proteolytic steps (cleavage of both the ArgI4'-AlaI4' and Arg"-Val181 peptide bonds) to become activated to factor IXaP.
Last, it was of interest to test whether factor Xa. PCPs -Ca'+ and TF. VIIa. P C P s . Ca'+ could work in concert in the same reaction system in the activation of factor IX. To test this, the activation of factor IX was evaluated by separating the addition of Xa. PCPs Ca'+ from the addition of the TF- In A, subsamples of an ongoing reaction were withdrawn a t designated time points and subjected to analysis by SDS-PACE in the presence of 2-mercaptoethanol. Components of the reaction were: 2 p M factor IX, 2 pM factor X, 14 nM TF, 5 nM VIIa, 1 mM PCPs, and 5 mM CaCI,, in HBS, pH 7.4, a t 37 "C. Lane 1, molecular weight markers with indicated molecular weights. Lanes 2-11, samples obtained by quenching each time point in 25 mM EDTA a t the designated intervals 0, 0.5, 1, 2, 4, 8, 12, 16, 30, and 60 min after the simultaneous addition of factors IX and X to the TF.VIIa-PCPS.Ca'+ enzyme complex. The indicated bands in A are identified as factors IX, X, IXn hc, Xan, Xap, IXaB hc, IX IC, and X IC, respectively. In I?, a subsample from each reaction time point was treated with Ir-FPR-ck as described above. Each reaction sample was then subjected to SDS-PAGE analysis. Prior to staining the gel with Coomassie Brilliant Blue, the polypeptides which specifically incorporated the lr-FPR-ck were visualized by excitation with a long-wave UV light box and photographed through a 540-nm cut-off filter. Bands which incorporated lr-FPR-ck are identified with arrows and correspond with the polypept.ide hands noted on the Coomassie Brilliant Blue-stained gel in A. C illustrates the relative concentrations of factor IX, factor IXn, and factor IXap when factor X is and is not present in the reaction mixture. The relative amount of factor IX activation was determined by densitometric analysis of the Coomassie Brilliant Blue-stained gel. Staining intensities were assigned arbitrary density units as described under "Experimental Procedures." The activation of factor IX by the TF.VIIa.PCPS.Ca" enzyme complex, when no factor X is present in the reaction, is depicted by a solid line with open symbols where factor IX is represented with open squares, factor IXn open triangles, and factor IXap open circles. Factor IX activation when an equal concentration of factor X is present in the reaction system is depicted by a dotted line with filled symbols where factor IX is represented with filled squares, factor IXn filled triangles, and factor IXaP filled circles. D illustrates the rate of factor X activation when an equal amount of factor IX is and is not present in the reaction mixture. Factor X activation was evaluated by scanning densitometry of a Coomassie Brilliant Blue-stained gel shown in Fig. 6 (no factor IX present in the reactor) and A (equal molar amounts of factor IX in the activation reaction). The relative concentrations of factor X activation when no factor IX was present in the reaction system are depicted by a solid line and open symbols where factor X is represented by open squares and factor Xa by open circles. The relative concentrations of factor X activation when factor IX was present in the reaction system are depicted by a dofred line with clospd symbols where factor X is represented by closed squares and factor Xa by closed circles. VIIa PCPs. Ca2+ as shown in Fig. 11. In this experiment TF. VIIa . P C P s . Cap+ was added to the reaction system and factor IX was first added to a mixture of factor Xa. P C P s . the rapid conversion of the IXa species to IXaP was observed. Ca2+, as described under "Experimental Procedures,'' and the These data clearly show that factor IXa generated by treatrapid generation of factor IXa was observed. Fifteen seconds ment with factor Xa is able to function as an efficient subprior to the removal of the 30-min time point of this reaction, strate for the TF. VIIa. P C P s . Ca'+ enzyme complex, further  suggesting that human factors Xa and VIIa may work together in the generation of human factor IXaP.

DISCUSSION
In this report we have evaluated the activation of human factor IX by the human TF. VIIa. P C P s -Ca'+ complex. Data presented here illustrate that human factor IX, when activated in an exclusively human reaction system consisting of recombinant human tissue factor, purified human factor VIIa, and synthetic phospholipid vesicles (PCPs) in the presence of calcium, takes place in a two step process (Fig. 1). The first step of factor IX activation is the cleavage of the Arg'45-Ala'4" peptide bond with the generation of the intermediated species IXa (step 1). This step is followed by the subsequent cleavage  of the second peptide bond Arg'X"-Val'R' and the generation of the active enzyme IXaP (step 2). This two-step reaction mechanism is identical to the mechanism of bovine factor IX activation by bovine factor XIa that has been reported by Di Scipio et al. (1). This mechanism is also consistent with the reports of Bajaj and co-workers (10) for the activation of human factor IX by the extrinsic pathway complex. These data do not agree with the reports of Zur and Nemerson (11) who stated that no reaction intermediate was observed when bovine factor IX was treated with a mixture of bovine factor VIIa and bovine tissue factor.
The clear establishment of this intermediate species in the activation of factor IX is important for a proper mechanistic evaluation of the factor IX activation. A large body of literature exists on factor IX activation using the [3H]factor IX activation peptide release assay as described by Zur and Nemerson (11). Although this assay has proven useful for following the release of the activation peptide from the 3Hlabeled factor IX molecule, data presented in this report illustrate that this assay does not detect potentially important proteolytic events in the activation of factor IX and provides only a quantitative approximation of the kinetics of the factor IX to factor IXaP reaction. This point is illustrated by both the reactions of factor IX with factor Xa' PCPs. Ca2+ and the simultaneous activation of factors IX and X by the TF . VIIa.
PCPS.Ca2+ complex. In both of these reactions the rapid generation of factor IXa was observed. Using the activation peptide release assay, these reaction steps proceed completely unnoticed because they take place before the release of the factor IX activation peptide from the factor IX molecule. These data illustrate the need to study all of the proteolytic steps involved in a reaction process. Furthermore, when factor IX is treated with [3H]factor borohydride using commonly reported methods (ll), we routinely lose >25% of the bioactivity of factor IX. This observation is consistent with other reports (10) where it was also observed that biological activity of the 3H-labeled factor IX was only 82% of unlabeled factor IX. When the activation of 3H-labeled factor IX was compared with the activation of unlabeled factor IX by the TF.VIIa. PCPs. Ca2+ complex, it was observed that 3H-labeled factor IX was activated 45% slower than unlabeled factor IX ( first activates small amounts of both factors IX and X. The factor Xa that is generated by the TF.VIIa can rapidly assist in the conversion of factor IX to factor IXa, generating more of the intermediate species IXa which is then ready to be converted to factor IXaP. Ultimately, factor IXaP when complexed with factor VIIIa accelerates factor X activation more than 50-fold. This pulse of factor Xa generation exceeds the levels of circulating factor Xa inhibitors (antithrombin I11 and lipoprotein-associated coagulation inhibitor) and allows for the formation of the prothrombinase enzyme complex leading to the generation of thrombin and the physiologic coagulation response. Published kinetic constants for the reactions shown above were from the following literature: 5). These observations demonstrate the need to use caution in the analysis of kinetic data generated by the [3H]factor activation peptide release assay in the activation of factor IX.
The relative importance of human factor IX and factor X activation by the extrinsic pathway is unknown. In this report we have evaluated the activation of human factors IX and X by the human TF. VIIa. PCPs. Ca2+ complex. Initially, we expected to observe that human factors IX and X would function as competitive substrates for the TF VIIa. PCPs. Ca2+ enzyme complex in view of the report by Jesty and Silverberg (7). However when this hypothesis was tested under the experimental conditions described in this report, it was surprising to observe that the rate of factor IX activation in the presence of factor X did not change significantly while the rate of factor X activation in the presence of factor IX decreased by 46%. Furthermore, when the proteolytic steps involved in the activation of factor IX were analyzed by SDS-PAGE, it appeared that the conversion of factor IX to IXa (cleavage of Arg'45-Ala'46 peptide bond) occurred faster when factor X was present in the reaction mixture. From these data we hypothesized a reaction model in which the factor Xa, which was generated simultaneously in the reaction, could play a role in the first proteolytic step in the activation of factor IX.
When the role of human factor Xa.PCPS.Ca2+ in the activation of human factor IX was investigated, it was observed that human factor Xa. PCPs. Ca2+ was able to rapidly cleave factor IX to the intermediate species factor IXa. The proteolysis of factor IX by factor Xa has been an area of significant confusion. Lindquist et ul. (2) and Kalousek et ul. (9) have reported that bovine factor Xa could activate bovine factor IX to factor IXaP. However, Osterud and Rapaport (3) have stated that when human factor IX was reacted with human factor Xa that no activation of factor IX was observed. In this report we have clearly observed the generation of the IXa species after treatment of factor IX with factor Xa. P C P s . Ca'+. Furthermore, we have observed that both PCPs and Ca'+ are essential for this reaction to proceed. We have been unable to perform a complete kinetic study of the factor Xa cleavage of factor IX because factor IXa is a nonenzymatic intermediate in the reaction process and its formation is undetectable by activation peptide release assays, clotting studies, or chromogenic substrate hydrolysis. Thus, we have limited the analysis of this reaction to concentrations of factor IXa which can be visualized and quantitated in a Coomassie Brilliant Blue gel system. However, because the activation of factor IX takes place in a two-step process, with the first step being the generation of factor IXa, the generation of the factor IXa species by factor Xa is potentially a significant step in the activation of factor IX.
To test if factor IXa, which was generated by treatment of factor IX with Xa. PCPs. Ca'+ could function as a substrate for the TF. VIIa. P C P s . Ca'+ complex, factor IXa was isolated by immunoaffinity chromatography. As Fig. 10 illustrates, the isolated factor IXa can function as a substrate for the TF. VIIa. PCPS.Ca'+ enzyme complex. Furthermore, as Fig. 11 illustrates, the factor IXa generated by Xa.PCPS. Ca'+ was able to act directly as a substrate for the TF.VIIa.PCPS. Ca2+ complex in the generation of factor IXaP. These data suggest that TF . VIIa. P C P s . Ca'+ and Xa. PCPs. Ca" work in concert to produce the first proteolytic cleavage of factor IX, the generation of factor IXa.
We propose a model of human factor IX activation by the extrinsic pathway where in step 1 factor IX is converted to IXa by either TF . VIIa. P C P s . Ca2+ or Xa. PCPs. Ca2+ by the proteolytic cleavage of the Arg'45-Ala'46 peptide bond. In step 2 the factor IXa is converted to IXaP primarily by the TF. VIIa. PCPs. Ca2+ complex (Fig. 1). These data do not rule out the possibility of a large enzymatic complex consisting of TF. VIIa. Xa. PCPs. Ca2+ being a rapid activator of factor IX. We hypothesize that factor Xa rapidly generates factor IXa, whereas factor VIIa takes the generated factor IXa and rapidly converts it to factor IXaP. In this model, the TF. VIIa.
Xa. PCPs. Ca'+ complex provides a pivotal point in the regulation of blood coagulation. As Rapaport and others (28,29) have shown, extrinsic pathway inhibitor (EPI) or lipoproteinassociated coagulation inhibitor (LACI) is an inhibitor of both factor Xa in solution and the assembled TF . VIIa. Xa. PCPs. Ca2+ complex. In our model, the rapid activation of factor IX by this enzyme complex provides a reasonable pathway around the extrinsic pathway inhibition by LACI and provides a direct link between intrinsic and extrinsic coagulation pathways.
The role of factor IX in the propagation of blood coagulation has remained puzzling. It is obvious that factor IXa is essential for normal hemostasis in vivo as illustrated by the diseases hemophilia A and B. However, current in vitro reaction models have not satisfactorily described how the lack of factor IX inhibits the propagation of the blood coagulation reaction. We suggest a model (Fig. 12) that may add insight into the important role that factor IX plays in blood coagulation. In this model, TF.VIIa initially generates a small amount of both factors IXa and Xa. The factor Xa that is generated then works in concert with the TF.VIIa complex to rapidly convert factor IX to IXa and then to IXafi. Ulti-mately the factor IXaP, when complexed with factor VIIIa, propagates the activation of factor X to factor Xa. This notion is supported by published kinetic constants shown in Fig. 12, where the activation of factor X by the VIIIa. IXaP enzyme complex is at least 50-fold more efficient (kcat/K,,,) than the activation of factor X by the TF. VIIa enzyme complex. This model provides a reasonable activation mechanism for factor IX in blood and may explain the role of human factor IX in the propagation of the coagulation cascade as illustrated by the disease hemophilia B.