Activation-dependent exposure of the inter-EGF sequence Leu83-Leu88 in factor Xa mediates ligand binding to effector cell protease receptor-1.

Binding of factor Xa to human umbilical vein endothelial cells (HUVEC) is contributed by effector cell protease receptor-1 (EPR-1). The structural requirements of this recognition were investigated. Factor Xa or catalytically inactive 5-dimethylaminonaphthalene-1sulfonyl (dansyl) Glu-Gly-Arg-(DEGR)-chloromethylketone-factor Xa bound indistinguishably to HUVEC and EPR-1 transfectants, and inhibited equally well the binding of 125I-factor Xa to these cells. Similarly, factor Xa active site inhibitors TAP or NAP5 did not reduce ligand binding to EPR-1. A factor X peptide duplicating the inter-EGF sequence Leu83-Phe84-Thr85-Arg86-Lys87-Leu88-(Gly) inhibited factor V/Va-independent prothrombin activation by HUVEC and blocked binding of 125I-factor Xa to these cells in a dose-dependent manner (IC50 ~ 20–40 μM). In contrast, none of the other factor X peptides tested or a control peptide with the inter-EGF sequence in scrambled order was effective. A recombinant chimeric molecule expressing the factor X sequence Leu83-Leu88 within a factor IX backbone inhibited binding of 125I-factor Xa to HUVEC and EPR-1 transfectants in a dose-dependent fashion, while recombinant factor IX or plasma IXa had no effect. An antibody generated against the factor X peptide 83–88, and designated JC15, inhibited 125I-factor Xa binding to HUVEC. The JC15 antibody bound to factor Xa and the recombinant IX/X83-88 chimera in a concentration dependent manner, while no specific reactivity with factors X or IXa was observed. Furthermore, binding of 125I-factor Xa to immobilized JC15 was inhibited by molar excess of unlabeled factor Xa, but not by comparable concentrations of factors X or IXa. These findings identify the inter-EGF sequence Leu83-Leu88 in factor Xa as a novel recognition site for EPR-1, and suggest its potential role as a protease activation-dependent neo-epitope. This interacting motif may help elucidate the contribution of factor Xa to cellular assembly of coagulation and vascular injury.

lar excess of unlabeled factor Xa, but not by comparable concentrations of factors X or IXa. These findings identify the inter-EGF sequence Leu 83 -Leu 88 in factor Xa as a novel recognition site for EPR-1, and suggest its potential role as a protease activation-dependent neo-epitope. This interacting motif may help elucidate the contribution of factor Xa to cellular assembly of coagulation and vascular injury.
Effector cell protease receptor-1 (EPR-1) 1 functions as a receptor for factor Xa on leukocytes (9) and endothelial cells (19), thus enhancing factor V/Va-independent prothrombin activation and leukocyte costimulation (17). On activated platelets, EPR-1-factor Xa interaction contributes to membrane assembly of the prothrombinase complex (20). For the procoagulant potential of factor Xa in vivo (21) and its mitogenic activity on endothelium and smooth muscle cells (15,16), factor Xa-cellular interactions may directly contribute to the pathogenesis of vascular injury (22).
In this study, we sought to investigate the structure-function requirements of EPR-1-factor Xa interaction and the receptor specificity for the active protease versus the zymogen factor X. Using synthetic peptides, a recombinant factor IX/X chimera, and a sequence-specific antibody, we found that the interconnecting EGF sequence Leu 83 -Phe 84 -Thr 85 -Arg 86 -Lys 87 -Leu 88 -(Gly) in factor Xa mediates ligand binding to EPR-1 and becomes surface exposed only after zymogen activation.

EXPERIMENTAL PROCEDURES
Proteins and Protein Labeling-The experimental procedures for the isolation and characterization of human plasma factors IX and X and the generation of the corresponding active proteases IXa and Xa have been reported (8). Aliquots of factor Xa purchased from Calbiochem or Haematologic Technologies Inc. (Essex Junction, VT) gave indistinguishable results in binding assays and thrombin generation experiments. Dansyl-Glu-Gly-Arg (DEGR)-chlomethylketone factor Xa and human prothrombin were purchased from Haematologic Technologies and Calbiochem, respectively. Factor Xa active site inhibitors TAP and NAP5 (23), and NAPc2, which recognizes a factor X exosite involved in zymogen activation by tissue factor-factor VIIa, were generously provided by Dr. G. Vlasuk (Corvas International, San Diego, CA). A library of factor X peptides, including the inter-EGF sequence Leu 83 -Phe 84 -Thr 85 -Arg 86 -Lys 87 -Leu 88 -(Gly) and its control scrambled variant Lys-Phe-Thr-(Gly)-Arg-Leu-Leu (residues in parentheses added to the natural sequence), was synthesized and characterized previously (24). Factor X numbering was according to Fung et al. (25). Aliquots (2.1-5.4 M) of factor Xa or DEGR-factor Xa were radiolabeled with 125 I-Na (Amersham Corp.) by the IODO-GEN method (26) to a specific activity of 0.4 -1 Ci/g of protein, with separation of free from protein-bound radioactivity by gel filtration on a Sephadex G-25 column pre-equilibrated with phosphate buffered saline, pH 7.4, and collection of 0.5 ml fractions. Both unlabeled factor Xa and 125 I-factor Xa indistinguishably cleaved the factor Xa-sensitive chromogenic substrate S-2222.
Construction of a Recombinant Factor IX/X83-88 Chimera-A recombinant chimeric factor IX/X molecule was genetically engineered in which the interconnecting EGF sequence Leu-Asp-Val-Thr of factor IX was exchanged for the corresponding region in factor X (25), containing the sequence Leu-Phe-Thr-Arg-Lys-Leu. Preliminary experiments demonstrated that factor IX or IXa did not inhibit binding of 125 I-factor Xa to EPR-1 transfectants, thus demonstrating the suitability of factor IX as an unrelated frame to characterize EPR-1 ligand recognition. The cDNA encoding the factor IX/X 83-88 chimera was produced by overlapping polymerase chain reaction using Vent DNA polymerase (New England Biolabs, Beverly, MA), as described previously (27), using the full-length sequence of human factor IX in the pCMV5 vector (28) as a template. In a first round of amplification, two DNA fragments were generated by polymerase chain reaction, one containing the nucleotide sequence of the factor X 83-88 peptide at the 3Ј end (452 base pairs), and a second one containing the same sequence at the 5Ј end (539 base pairs). The two fragments were annealed at the overlapping site corresponding to the factor X sequence 83-88, and used as a template for a second round of amplification. The resulting polymerase chain reaction product of 973 base pairs was digested with BglII and MluI and directionally subcloned into the pCMV5 vector, containing the factor IX sequence. The correct insertion of the factor X 83-88 sequence into the factor IX framework was confirmed by DNA sequencing.
Transfection Experiments and Purification of Recombinant Coagulation Proteins-Subconfluent cultures of Chinese hamster ovary (CHO) cells were transfected with 15 g of an EPR-1 cDNA clone in the mammalian cell expression vector pcDNA3 (Invitrogen, San Diego, CA) by electroporation, as described elsewhere (29). After a 48-h culture at 37°C, cells were washed, suspended in serum-free Dulbecco's modified Eagle's medium (DMEM) (BioWhittaker, Walkersville, MD) to a final concentration of 5 ϫ 10 6 /ml, and analyzed for 125 I-factor Xa binding or factor V/Va-independent prothrombin activation (29), as described below. Clonal CHO cells stably transfected with the EPR-1 cDNA in pRC/CMV were characterized previously (9). For production of chimeric coagulation proteins, human embryonic kidney 293 cells (American Type Culture Collection, Rockville, MD) were grown in DMEM (Bio-Whittaker) supplemented with 10% fetal bovine serum. Six g of pCMV5 vector containing the wild-type factor IX cDNA or the factor IX/X83-88 chimera cDNA and 0.6 g of pSV2-neo were used to cotransfect subconfluent cultures of 293 cells, using a LipofectAMINE reagent (Life Technologies, Inc.). After a 48-h culture at 37°C, cells were diluted 1:10 and 1:20 and grown in 1:1 ratio of DMEM and F-12 tissue culture medium in the presence of 0.5 mg/ml Geneticin (G418, Life Technologies, Inc.). Two weeks later, the surviving colonies were transferred to 24-well tissue culture plates, grown to confluency at 37°C, and screened for factor IX expression by enzyme-linked immunosorbent assay, as described previously (30). A single clone expressing high levels of recombinant protein was expanded into roller bottles for large scale production of recombinant factor IX in DMEM-F-12 medium, supplemented with insulin/transferrin/sodium selenite and 6 g/ml vitamin K, with collection of the conditioned medium every 24 h. For purification of recombinant factor IX or the factor IX/X83-88 chimera, conditioned medium from the various cultures was filtered through a 0.22-m sterile filter (Millipore Corp.) to remove cellular debris, diluted in buffer containing 20 mM Tris-HCl, pH 7.2, 5 mM benzamidine, 5 mM EDTA, and applied to a Q-Sepharose column (Pharmacia Biotech, Inc.). After washes in 20 mM Tris-HCl, pH 7.2, 60 mM NaCl, and 1 mM benzamidine, bound proteins were eluted in 20 mM Tris-HCl, pH 7.2, 700 mM NaCl, 1 mM benzamidine. The eluted material was then applied to a second column containing the factor IX conformation-dependent monoclonal antibody A-7 (31). The affinity matrix was washed in 20 mM Tris-HCl, pH 7.2, 150 mM NaCl, and 20 mM MgCl 2 , before elution of recombinant wild-type factor IX or the factor IX/X83-88 chimera with 20 mM Tris-HCl, pH 7.2, 150 mM NaCl, and 20 mM EDTA. Purity of the eluted material was assessed by SDS-gel electrophoresis followed by silver staining. Both recombinant factor IX and the factor IX/X83-88 chimera were tested in competition experiments of 125 I-factor Xa binding to EPR-1-expressing cells.
Gla Analysis of Recombinant Proteins-Analysis of ␥-carboxyglutamic acid content (Gla analysis) was performed essentially as described by Price (32). Ten g of protein were subjected to alkaline hydrolysis in 2 M KOH for 20 h at 110°C. Hydrolyzed amino acids were titrated to a pH of 7.0 with perchloric acid and centrifuged to remove precipitate. Supernatants were separated by high performance liquid chromatography on a DC-4A cation exchange resin column (Dionex, Sunnyvale, CA) using an elution buffer of lithium citrate, pH 2.0. Amino acids were detected by post-column derivatization with o-phthalaldehyde buffer and quantitated by fluorescence spectrophotometry (excitation wavelength 340 nm, 418 nm cut-off filter) using a recording integrator. Purified ␥-carboxyglutamic acid and aspartic acid (Sigma) were used as standards, and samples were compared with a control sample (plasma-derived factor X, Enzyme Research Labs, South Bend, IN). Moles of purified protein were determined from the aspartate/ asparagine peak. Determinations were performed in duplicate, and results are reported as an average.
Binding Studies-HUVEC were isolated by collagenase treatment and maintained in DMEM tissue culture medium (BioWhittaker), supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and endothelial cell growth factor, plated onto gelatinized 48-well tissue culture plates (Costar Corp., Cambridge, MA) at a density of 4 ϫ 10 4 cells/well, and cultured for 2-4 d prior to the assay. Cells were washed twice with serum-free RPMI 1640 and incubated in a total volume of 300 l with 2.5 mM CaCl 2 and 10.8 nM 125 I-factor Xa in the presence or in the absence of increasing concentrations (0.1-1 M) of unlabeled competitors, factors Xa, IXa, or DEGR-Xa. For competition experiments with recombinant factor IX or the IX/X83-88 chimera, 5.43 nM 125 I-factor Xa was used. After a 15-min incubation at 22°C, cells were washed three times in serum-free medium and solubilized in 10% SDS, and radioactivity associated with the cell monolayer under the various conditions tested was determined in a ␥ counter. In parallel experiments, CHO cells transiently transfected with the EPR-1 cDNA (29) were detached by phosphate-buffered saline-EDTA, pH 7.4, washed, diluted to 2.0 ϫ 10 6 /ml in serum-free RPMI 1640 medium, and mixed in a total volume of 300 l with 10.8 nM 125 I-factor Xa in the presence or in the absence of the various competitors, as described above. After a 15-min incubation at 22°C, cell surface-bound radioactivity was separated from free by centrifugation of 200-l aliquots of each incubation reaction through a mixture of silicone oil (Dow Corning, New Bedford, MA) at 14,000 ϫ g for 5 min and counted in a ␥ counter. In another series of experiments, 10.8 nM aliquots of 125 I-factor Xa were preincubated with increasing concentrations (1-500 nM) of the factor Xa-specific inhibitors TAP (23), NAP5, or NAPc2 for 15 min at 22°C, before addition to HUVEC or EPR-1 transfectants and determination of specific binding, as described above. For antibody neutralization experiments, 10.8 nM aliquots of 125 I-factor Xa were preincubated with increasing concentrations of control non immune rabbit IgG (Zymed Laboratories, San Francisco, CA) or sequence-specific JC15 antibody (see below), for 15 min at 22°C before addition to HUVEC monolayers and determination of specific binding. For all experiments nonspecific binding was assessed in the presence of a 100 -150-fold molar excess of unlabeled factor Xa added at the start of the incubation and was subtracted from the total to calculate net specific binding (29). For peptide inhibition experiments, increasing concentrations (0.1-500 M) of the various factor X-derived peptides or their variants (24) were preincubated with HUVEC monolayers in serum-free RPMI 1640 medium for 20 min at 22°C before addition of 10.8 nM 125 I-factor Xa and determination of specific binding, as described above.
Thrombin Generation-HUVEC monolayers in 96-well plates were preincubated in a total volume of 100 l of serum-free RPMI 1640 with 100 M concentrations of the various factor X-derived synthetic peptides for 20 min at 22°C. Cells were mixed with 138 M prothrombin, 1.2 mM CaCl 2 , and 43 nM factor Xa for 5 min at 22°C. Thrombin generation under the various conditions tested was quantitated by hydrolysis of the thrombin-sensitive chromogenic substrate S-2238 (Chromogenix, Molndal, Sweden) at A 405 (29) and converted to thrombin concentrations (nanomolar) using a standard curve constructed with serial increasing concentrations of bovine thrombin. In another series of experiments, stable EPR-1 transfectants or HUVEC were mixed with increasing concentrations (0.1-200 M) of the factor X-derived peptides 83-88, its control scrambled variant, or the -COOH terminus EGF peptide (Gly)-His 101 -Glu 102 -Glu 103 -Gln 104 -Asn 105 -Ser 106 -Val 107 -Val 108 -(Gly), for 20 min at 22°C, before addition of factor Xa (43 nM) and prothrombin (138 nM), and determination of S-2238 hydrolysis. Background hydrolysis of S-2238 in the absence of prothrombin (8 -14%) was subtracted to calculate specific thrombin cleavage (29). The suboptimal concentration of prothrombin of 138 nM used in these ex-periments is nonsaturating for the system, and has been used for comparison with previous data obtained with other EPR-1 ϩ cells (9). No specific thrombin generation was detected in the absence of cells, under the same experimental conditions (not shown).
Production and Characterization of Sequence-specific JC15 Antibody-A sequence-specific antibody was generated in a rabbit by multiple subcutaneous injections in complete Freund's adjuvant of 100 g of the inter-EGF factor X peptide LFTRKL(G) preparatively coupled to keyhole limpet hemocyanin. After a 4-week interval, animals were boosted with subcutaneous injection of 100 g of peptide in incomplete Freund's adjuvant and sequentially boosted and bled at alternate weeks. Rabbit immunoglobulin fractions of the relevant serum, designated JC15, were purified by affinity chromatography on protein A-Sepharose and used for inhibition of 125 I-factor Xa binding to EPR-1expressing cells, as described above. The recognition specificity of JC15 antibody was characterized by enzyme-linked immunosorbent assay. Ninety-six well plastic microtiter plates (Costar Corp., Cambridge, MA) were coated with 0.21 M factors IXa, X, and Xa, or the IX/X83-88 chimera in Tris-buffered saline (TBS), pH 7.4, in a total volume of 100 l for 18 h at 4°C. After washes in TBS, pH 7.4, wells were post-coated with 3% gelatin (Sigma) in TBS, pH 7.4, for 60 min at 37°C, and mixed with increasing concentrations (1.25-50 g/ml) of control non immune rabbit IgG or JC15 antibody in TBS, pH 7.4, containing 0.05% Tween-20 plus 1% bovine serum albumin (Sigma) for 90 min at 37°C. After washes, binding of the primary antibody was revealed by addition of a 1:4000 dilution of biotin-conjugated goat anti-rabbit IgG (Zymed) for 1 h at 37°C followed by washes in TBS, pH 7.4, and addition of a 1:1000 dilution of alkaline phosphatase-conjugated streptavidin reagent (Zymed) and p-nitrophenyl phosphate for 30 min at 37°C, with determination of absorbance at A 405 . In another series of experiments, 96well plastic microtiter plates were coated with 50 g/ml aliquots of JC15 antibody in TBS, pH 7.4. After washes and post-coating with 3% gelatin, wells were further incubated with 5.4 nM 125 I-factor Xa in the presence or in the absence of increasing concentrations of unlabeled factors X, Xa, or IXa. After a 45 min incubation at 37°C, wells were washed in TBS, pH 7.4, and extracted in 10% SDS, and radioactivity was determined in a ␥ counter.

Catalytic Active Site-independent Binding of Factor Xa to EPR-1-Previous
studies demonstrated that EPR-1 bound factor Xa but not the zymogen factor X (9). A potential requirement of factor Xa proteolytic activity in ligand binding to EPR-1 was investigated. Factor Xa active site inhibitors TAP, NAP5 (23), or NAPc2, did not reduce 125 I-factor Xa binding to HUVEC or EPR-1 transfectants (Table I). Similarly, catalytically inactive DEGR-factor Xa inhibited binding of 125 I-factor Xa to HUVEC (19) in a concentration-dependent reaction, quantitatively indistinguishable from that observed with the active protease (Fig. 1A). Consistent with these data, 125 I-DEGR-factor Xa bound specifically to HUVEC, in a reaction inhibited equally well by increasing molar excess of factor Xa or DEGR-factor Xa (Fig. 1B). In control experiments, TAP-or NAP5-treated factor Xa, or DEGR-factor Xa, were completely devoid of catalytic activity by S-2222 hydrolysis (not shown).
Synthetic Peptidyl Mimicry of the EPR-1 Binding Site on Factor Xa-Screening of factor X sequences (24) revealed that a synthetic peptide corresponding to the inter-EGF region Leu 83 -Phe 84 -Thr 85 -Arg 86 -Lys 87 -Leu 88 -(Gly) inhibited HUVEC prothrombin activation in the absence of factor V/Va by ϳ80% (Fig. 2). With the exception of the factor X sequence 366 -375, which produced only a partial and variable reduction in HUVEC prothrombin activation, none of the other factor X peptides tested, including antagonists of factor X binding to Mac-1 (24), were effective under the same experimental conditions (Fig. 2). A potential mimicry of EPR-1-factor Xa interaction by the inter-EGF sequence Leu 83 -Leu 88 was investigated. Increasing concentrations of the factor X peptide 83-88 inhibited prothrombin activation on HUVEC (Fig. 3A) or EPR-1 transfectants (not shown) in a concentration-dependent manner with IC 50 ϳ 20 -40 M (Fig. 3A). In contrast, a control peptide with the 83-88 sequence in scrambled order KFT-(G)RLL, or the factor X -COOH terminus EGF sequence (Gly)-His 101 -Glu 102 -Glu 103 -Gln 104 -Asn 105 -Ser 106 -Val 107 -Val 108 -(Gly), did not reduce HUVEC prothrombin activation (Fig. 3A). Con-  1. Catalytic active site-independent binding of factor Xa to EPR-1. HUVEC were equilibrated in a total volume of 300 l with 10.8 nM 125 I-factor Xa (A) or 125 I-DEGR-factor Xa (B) and 2.5 mM CaCl 2 , in the presence or in the absence of the indicated increasing concentrations (0.1-1 M) of unlabeled factor Xa (q) or DEGR-factor Xa (E). After a 15 min incubation at 22°C, cells were washed three times in serumfree RPMI 1640, solubilized in 10% SDS, and cell-associated radioactivity was determined in a ␥ counter. One hundred percent specific binding was 0.56 Ϯ 0.13 nM (A) or 0.73 Ϯ 0.1 (B). Data are the mean Ϯ S.E. of at least three independent experiments. sistent with the requirement of EPR-1-factor Xa interaction for prothrombin activation (29), binding of 125 I-factor Xa to HUVEC was inhibited in a dose-dependent manner by the factor X peptide 83-88, but not by the 83-88 control scrambled peptide, nor by the -COOH terminus EGF peptide 101-108 (Fig. 3B). The higher peptide concentrations required to inhibit factor Xa-HUVEC interaction as compared with prothrombin activation (Fig. 3) suggests heterogeneity in ligand binding, potentially comprising both functional (EPR-1-mediated) and nonfunctional cellular associations. In control experiments, increasing concentrations (1-200 M) of factor X peptides 83-88 or 101-108 did not reduce factor Xa-or thrombin-dependent hydrolysis of S-2222 and S-2238, respectively, ruling out a potential substrate competition mechanism for inhibition of cell prothrombin activation (not shown).
Molecular Characterization of the EPR-1 Ligand Binding Site-A recombinant chimeric molecule containing the factor X inter-EGF sequence 83-88 was engineered into the framework of factor IX, purified by monoclonal antibody affinity chromatography, and tested in gain-of-function experiments for inhibition of 125 I-factor Xa binding to EPR-1-expressing cells. Gla analysis of the factor IX/X83-88 chimera showed a Gla content similar to that of control plasma-derived protein (9.1 mol of Gla/mol of protein for the recombinant chimeric protein and 7.9 mol of Gla/mol of protein for plasma-derived factor X). Under these experimental conditions, binding of 125 I-factor Xa to HUVEC (Fig. 4A) or EPR-1 transfectants (Fig. 4B) was dosedependently inhibited by increasing molar excess of unlabeled factor IX/X83-88 chimera, in a reaction quantitatively indistinguishable from that observed with unlabeled factor Xa (Fig.  4). In contrast, comparable concentrations of unlabeled recombinant factor IX, or plasma-derived IXa, did not decrease binding of 125 I-factor Xa to EPR-1-expressing cells (Fig. 4). In another series of experiments, a sequence-specific antibody, designated JC15, was generated against the factor X peptide 83-88 and tested for inhibition of EPR-1-factor Xa interaction. As shown in Fig. 5, preincubation of 125 I-factor Xa with JC15 antibody resulted in dose-dependent inhibition of ligand bind-ing to HUVEC, while incubation of factor Xa with comparable concentrations of control non immune rabbit IgG was without effect (Fig. 5).
Activation-dependent Exposure of the EPR-1 Ligand Binding Site-Potential changes in accessibility of the inter-EGF sequence 83-88 during zymogen activation were investigated. First, in enzyme-linked immunosorbent assay, JC15 antibody bound to immobilized factor Xa or the recombinant IX/X83-88 chimera in a concentration-dependent fashion, while no specific reactivity with factor IXa or the zymogen factor X was observed, under the same experimental conditions (Fig. 6A). In control experiments, comparable concentrations of non immune rabbit IgG did not bind to immobilized factor Xa (Fig.  6A). Second, binding of 125 I-factor Xa to immobilized JC15 was competitively inhibited by molar excess of factor Xa in a concentration-dependent manner (Fig. 6B), while comparable doses of factor IXa or the zymogen factor X did not significantly reduce 125 I-factor Xa binding to JC15-coated plates (Fig. 6B).

DISCUSSION
In this study, we have shown that the interconnecting EGF sequence Leu 83 -Leu 88 in factor Xa mediates ligand binding to EPR-1. This conclusion was based on synthetic peptidyl mimicry, gain-of-function of a recombinant factor IX/X chimera, and neutralization experiments with a sequence-specific antibody. Moreover, EPR-1 ligand binding could be recapitulated by catalytically inactive factor Xa, thus ruling out a potential requirement of local proteolysis for receptor recognition. Finally, competition studies with the sequence-specific JC15 antibody suggested that the inter-EGF 83-88 region was not accessible in the zymogen factor X, but became surface-exposed in the active protease.
Binding of factor Xa to EPR-1 is one of several receptormediated associations between coagulation/fibrinolytic proteases and vascular cells (2-4, 6, 33, 34). In addition to the paradigm of protease-activated receptors (4,33), these interactions can also be mediated by specific structural requirements, as exemplified by the receptor-binding sequences in the EGFlike modules of urokinase (35), and factor XII (6). Our observations propose an unexpected role for the short inter-EGF sequence 83-88 in mediating docking of factor Xa to EPR-1 and imparting specificity for ligand binding, at least with two different mechanisms. First, the unique structural features of this inter-EGF region, with two unique charged residues Lys 86 and Arg 87 , and its high degree of flexibility (36), as opposed to its "locked" conformation in factor IXa (37), could determine the ability of EPR-1 to bind factor Xa, but not an homologous protease, i.e. factor IXa. Secondly, the accessibility of this interacting motif in factor Xa but not in factor X could dictate the selective recognition of EPR-1 for the active protease, and not for the zymogen factor X.
The idea that zymogen activation may result in conformational changes with exposure of selective neo-antigenic epitopes has been proposed earlier. Keyt et al. (38) proposed that factor X activation resulted in conformational transitions in the heavy chain, as well as metal ion-dependent transitions in the heavy and light chains. Conformational remodeling of the protease domain upon factor X activation was also demonstrated by Persson et al. (39) using domain-specific F(abЈ) 2 and factor X-derived tryptic fragments. Similar observations were reported for factors IXa␤ and X upon Ca 2ϩ binding to the Gla module (40,41). Our competition experiments with the JC15 antibody provide evidence for an additional activation-dependent conformational transition in the light chain of factor Xa, specifically targeted to the inter-EGF 83-88 sequence. Alternatively, this domain may become physically unmasked by the release of the factor X activation peptide during zymogen activation (1). Interestingly, the factor IX/X83-88 chimera inhibited ligand binding to EPR-1 without requiring zymogen activation. The ability of the JC15 antibody to recognize this recombinant protein indistinguishably from native factor Xa suggested that the chimeric insertion produced per se conformational changes and surface exposure of the inter-EGF region.
Current experimental evidence demonstrates that protease receptors initiate multiple cellular signaling pathways. Factor Xa, in particular, has been implicated in endothelial, vascular smooth muscle cell, and leukocyte activation and proliferation mechanisms (15)(16)(17), of potential relevance for the pathogenesis of atherosclerosis and vascular diseases (22). Although the data presented here identify the inter-EGF sequence in factor Xa as mediating ligand binding to HUVEC, additional requirements may be involved in post-receptor occupancy events of factor Xa-dependent signal transduction. In this context, catalytic inactivation of factor Xa abolished EPR-1-stimulated proliferation of HUVEC and smooth muscle cells (19), and modulation of endothelial cell gene expression. 2 Altogether, these data suggest a cooperative model of factor Xa binding to vascular cell EPR-1, potentially involving an initial Gla-dependent contact stabilized by an high affinity recognition of the inter-EGF sequence 83-88, and followed by an as yet unidentified step of local proteolysis by cell surface-bound factor Xa for downstream signal transduction events and effector responses (19).
In summary, we have identified the inter-EGF peptide in factor Xa as a novel, activation-dependent sequence for receptor-mediated assembly of coagulation on vascular cells. Antagonists with similar specificity may be beneficial at targeting factor Xa-dependent cellular effector functions in vascular injury and atherosclerosis (21,22). FIG. 6. Ligand recognition specificity of JC15 antibody. A, 96well plastic microtiter plates were coated with 0.21 M aliquots of factors IXa, X, and Xa, or the IX/X83-88 chimera for 18 h at 4°C, post-coated with 3% gelatin, and incubated with the indicated increasing concentrations of anti-factor X 83-88 JC15 antibody or control rabbit IgG (Xa). After a 2-h incubation of 22°C, wells were washed, and binding of the primary antibody was determined by addition of alkaline phosphate-conjugated goat anti-rabbit IgG, and quantified by absorbance at A 405 . B, assay plates were coated with 50 g/ml JC15 antibody, post-coated with 3% gelatin, and incubated with 5.43 nM 125 I-factor Xa in the presence or in the absence of the indicated increasing concentrations of unlabeled factors X, Xa, or IXa. After a 45-min incubation at 37°C, wells were washed and extracted in 10% SDS, and radioactivity was determined in a ␥ counter. For both panels, data are the mean Ϯ S.E. of at least four independent experiments.