Role for the Kunitz-3 Domain of Tissue Factor Pathway Inhibitor-α in Cell Surface Binding
Background— Tissue factor pathway inhibitor (TFPI)-α, a key regulator of tissue factor–induced coagulation, contains 3 tandem Kunitz-type inhibitory domains. Kunitz-1 binds and inhibits factor VIIa in the factor VIIa/tissue factor complex, and Kunitz-2 binds and inhibits factor Xa. The role of the Kunitz-3 domain of TFPI-α, however, has remained an enigma.
Methods and Results— To determine the structures within TFPI-α involved in its binding to cell surface, altered forms of TFPI-α were expressed in C127 (mouse mammary) cells: C-terminal truncated forms TFPI-α (252), TFPI-α (242), and TFPI-α (181), which also lacks the third Kunitz domain (K3); TFPI-α (desK3), which lacks only the K3 domain; and TFPI-α (R199L), in which the putative P1 site in K3 is changed from arginine to leucine. By flow cytometry (fluorescence-activated cell sorting), the altered forms 252, 242, and R199L showed significantly reduced binding, whereas the forms 181 and desK3 completely failed to bind to the cell surface. Transient expression of WT-, desK3-, and K3/K2-TFPI-α (in which K3 is replaced with K2) in another cell line (b-end3, mouse endothelial) produced comparable results. Exogenously added C-terminal truncated and R199L forms of TFPI-α bound poorly and desK3 did not bind at all to the surface of ECV304 cells in which TFPI-α expression had been “knocked down” by RNA interference.
Conclusions— Optimal cell binding of endogenously expressed TFPI-α requires its K3 and C-terminal domains, and within the K3 domain, the P1 (R199) residue plays an important role. Thus, one role of the K3 domain involves the cell surface localization of TFPI-α.
Received June 9, 2004; revision received August 17, 2004; accepted August 23, 2004.
Tissue factor pathway inhibitor (TFPI) is the key regulator of the tissue factor (TF)–induced coagulation. It exerts its function by neutralizing the catalytic activity of factor Xa (FXa) and by producing feedback inhibition of the TF factor VIIa (FVIIa) complex in presence of FXa.1,2 Two different forms of TFPI are produced through alternative mRNA splicing: TFPI-α and TFPI-β. The soluble form TFPI-α is an ≈46-kDa protein containing an acidic N-terminal region followed by 3 tandem, Kunitz-type protease inhibitor domains and a basic C-terminal region.3 TFPI-β is an alternatively spliced form of TFPI in which the Kunitz-3 (K3) and C-terminus of TFPI-α are replaced with an unrelated C-terminal domain that directs the attachment of a GPI anchor.4,5 The Kunitz-1 and 2 domains are responsible for TFPI binding and inhibition of TF-FVIIa complex and FXa, respectively, whereas the K3 domain in TFPI-α appears to lack proteinase inhibitory activity, and its physiological function is not known. The C-terminus of TFPI-α has been thought to mediate TFPI surface binding by interacting with anionic membrane structures.6 Furthermore, the C-terminus of TFPI-α appears to play an important role in the activity of soluble TFPI-α, because C-terminal truncated forms exert much less anticoagulant activity.7
In the blood, TFPI-α circulates as both free and lipoprotein (LDL, HDL)–associated forms.8 The endothelium is presumed to be the major source of TFPI in vivo,9 and a small fraction of the TFPI produced by cultured endothelial cells and present in placental tissue remains membrane associated.10–14 Treatment of cultured endothelial cells with heparin induces a significant release of TFPI into the medium10,12–14 without affecting its surface concentration,5 and in vivo heparin infusion increases TFPI plasma levels by 1.5- to 2.5-fold.15 Phosphatidylinositol (PI)-specific phospholipase C (PI-PLC), which cleaves glycosyl-PI (GPI) membrane anchors, releases ≈80% of TFPI from the surface of cultured endothelial cells, and the remaining ≈20% is released by subsequent heparin treatment.5 These results suggest that TFPI is released from intracellular storages in response to heparin and that the cell surface localization of TFPI-α involves GPI anchorage.
Recombinant (r) TFPI-β expressed by transfected Chinese hamster ovary (CHO) cells resides at the cell surface and is completely released by PI-PLC treatment.5 rTFPI-α produced by transfected CHO cells is secreted predominantly. The small fraction of rTFPI-α that is bound at the CHO cell surface is partially released (≈75%) by PI-PLC treatment, and the remainder (≈25%) is released by heparin treatment after PI-PLC pretreatment (suggesting that only TFPI-α is responsible for this phenomenon).5 Thus, both TFPI-α and TFPI-β are GPI-linked at the membrane surface, but through different mechanisms: TFPI-β, which contains a specific signal for a GPI attachment in its C-terminus, is directly GPI-anchored; TFPI-α is apparently indirectly GPI-anchored to the membrane surface by binding to a yet to be identified GPI-linked protein(s). The aim of the present study was to identify the domain(s) within TFPI-α that are involved in its localization on the membrane surface.
The human cell line ECV304, which possesses some endothelium-like properties; the mouse mammary fibroblast-like cell line C127; and the mouse brain endothelial cell line b-end3 were obtained from American Tissue Culture Collection (ATCC). Lumi-Light Western blotting substrate was from Roche. PI-specific phospholipase C (PI-PLC), horseradish peroxidase–labeled goat anti-rabbit IgG antibodies, Tris, and EDTA were from Sigma Chemical. Prestained molecular weight standards were from Bio-Rad. Oligonucleotides and DNA sequencing were commissioned from the Protein Chemistry Laboratory at Washington University School of Medicine. Asserachrom FREE TFPI was from Diagnostica Stago. The production and characterization of monoclonal anti-TFPI Mab2H8 and Mab2B12, rabbit polyclonal anti-TFPI antisera, and rabbit anti–TFPI-α C-terminal peptide have been described previously.8
C127 cells were cultured in DMEM (BioWhittaker) supplemented with 10% FBS. B-end3 cells were cultured in DMEM supplemented with 10% FBS and 1% sodium bicarbonate (BioWhittaker). ECV304 cells were cultured in M199 (BioWhittaker) supplemented with 10% FBS and 1 mmol/L pyruvate. All cells were cultured at 37°C with 5% CO2, and all media contained penicillin (50 U/mL) and streptomycin (50 μg/mL).
Construction of rTFPI Mutants for C-127 Cells
A plasmid, pUCGB9R1, containing the TFPI cDNA modified for insertion into a mammalian expression vector16 was used to construct the mutants. To determine the structures within TFPI-α required for its binding to the membrane surface, different mutated forms were generated: the C-terminal truncated forms TFPI-α (252) (lacking the C-terminus from aa 253), TFPI-α (242) (lacking the C-terminus from aa 243), and TFPI-α (181) (lacking the K3 as well as the C-terminus domain); TFPI-α (desK3), lacking only the K3 domain; and TFPI-α (R199L), in which the putative P1 site in K3 was changed from arginine to leucine (Figure 1). All the mutants were constructed by use of standard techniques as described previously.17,18 For expression, the wild-type (WT) and mutant forms of TFPI-α cDNA were cloned in the mammalian expression vector pMON1123 and cotransfected with pSV2neo into C127 mouse mammary tumor cells by calcium phosphate precipitation (Stratagene). After G418 (Invitrogen) selection (1000 μg/mL), stable clones were isolated and expanded.
Characterization of the TFPI-α WT Expressed in C-127 Cells
In cultured endothelial cells, TFPI that associates with the cell surface represents only a small fraction of the total TFPI secreted in the media.5 To determine whether TFPI expressed in C-127 cells shows a similar distribution, C-127 cells transfected with TFPI-α were grown in T-75 flasks to 90% confluence. After 16 hours in their appropriate serum-free medium, medium was collected, and secreted rTFPI-α was determined by immunoassay as described previously.5 Cells were harvested in unicellular suspension by a PBS-based dissociation buffer (Invitrogen), and cell suspensions (2×106/mL, 0.5 mL) were pretreated in serum-free medium at 37°C with heparin (1 U mL, 1 hour). After heparin preincubation, cells were collected by centrifugation, washed, resuspended in serum-free medium, and treated with heparin (5 U/mL) for 20 minutes after the cells had been incubated for 1 hour with PI-PLC. At the end of the treatment, cells were centrifuged, supernatants were collected, and rTFPI-α was determined by immunoassay, as described previously.5 Results are expressed as amount of TFPI-α (ng)/106 cells.
Characterization of the Mutants Expressed in C-127 Cells
The amount of rTFPI secreted by C-127 cells in the medium was evaluated by immunoassay, as described previously.5 Stable clones secreting comparable amounts of the WT and the various mutant forms were selected and used for further experiments. Cells at 90% confluence in T-75 flasks were allowed to grow overnight in their appropriate serum-free medium. The following day, media were collected, centrifuged (250g for 5 minutes), concentrated 5-fold (Centricon YM10, Millipore), and stored at −20°C for subsequent analysis of TFPI structure (Western blot) and activity (end-point chromogenic assay); cells were harvested in unicellular suspension, and surface TFPI was evaluated by flow cytometry (fluorescence-activated cell sorting, FACS).
Western Blot and Activity Assay Analyses
Comparable quantities (30 μL) of the cleared, concentrated media from C-127 cells transfected with the WT and the various mutated forms of rTFPI-α were evaluated by Western blotting with chemiluminescent detection using rabbit polyclonal anti–whole TFPI and rabbit polyclonal anti–C-terminal peptide (KIAYEEIFVKNM) antibodies. TFPI anti-TF/FVIIa activity was evaluated by an end-point chromogenic assay (Actichrome TFPI activity kit, American Diagnostica) according to the manufacturer’s instructions.
Effects of Heparin and PI-PLC
For rTFPI-α(WT), rTFPI-α(252), and rTFPI-α(242), the effects of heparin and PI-PLC on surface TFPI were also evaluated. Cell suspensions (1×106 cells/mL, 0.5 mL) were treated at 37°C in different ways: (1) incubated in serum-free medium for 1 hour; (2) with heparin (5 U/mL) the last 20 minutes of a 1-hour incubation; (3) with PI-PLC (1 U/mL) for 1 hour; and (4) with heparin (5 U/mL) for 20 minutes after the cells had been incubated for 1 hour with PI-PLC. At the end of the incubations, cells were collected by centrifugation and surface TFPI was evaluated by flow cytometry (FACS).
Transient Transfection of b-end3 cells
Two mutants were generated by use of standard molecular biology techniques: TFPI-α (desK3), lacking the K3 domain from H184 to F242; and TFPI-α (K3/K2), in which the K3 domain (from F183 to G242) was replaced with the K2 domain (from C96 to E148). The WT and the mutant forms of TFPI-α cDNA were cloned in the mammalian expression vector pcDNA 3.1 hygro (+) (Invitrogen), and cells were transfected in 6-well plates with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. After 48 hours, cells whose medium expressed comparable amounts of rTFPI by immunoassay were evaluated by flow cytometry for surface TFPI. Only results from transfection that achieved >50% efficiency, as determined by FACS, were analyzed.
Binding of Exogenously Added TFPI-α to ECV304 Cells
TFPI-α expression by ECV304 cells was knocked down by use of a plasmid-based RNA interference technique. Briefly, a small DNA insert (71 bp) encoding a short hairpin RNA targeting the sequence -CAGAUUCUACUACAAUUCA- in TFPI-α mRNA was cloned in the vector pRNA-H1.1/Neo (GenScript) under the control of the H1 promoter for RNA polymerase III. Transfection of ECV304 cells was performed by use of Lipofectamine 2000, and stable clones were selected by G418 (600 μg/mL). To test the “knock-down” efficiency, stable clones resistant to G418 were treated as follows: heparin 5 U/mL, 2 hours, 37°C; washed (PBS, 3 times); PI-PLC 1 U/mL, 1 hour, 37°C; followed by heparin 5 U/mL, 20 minutes, 37°C. The PI-PLC/heparin releasates were collected, and 2 assays were used to determine the amount of TFPI-α and TFPI-β. The Asserachrom TFPI(free) immunoassay kit (Stago) was used to identify TFPI-α: this assay uses a polyclonal anti-TFPI for capture and a specific anti-K3 monoclonal antibody for detection and does not recognize TFPI-β. The TFPI(total) immunoassay, performed by use of our monoclonal antibodies 2H8 (anti-K1) and 2B12 (anti-K2) as described previously,5 was used to detect both TFPI-α and TFPI-β. TFPI-β was determined as the difference between the TFPI(total) and TFPI(free) immunoassays. A clone with >85% knockdown of TFPI-α (ECV↓TFPI-α) was used for subsequent experiments. Confluent ECV↓TFPI-α cells were cultured overnight in serum-free medium in 6-well plates. The following day, the cells were washed and incubated for 2 hours at 37°C with the conditioned medium (2 mL) of transfected C127 cells expressing comparable amounts of the WT and the mutant forms of TFPI-α. Binding of exogenously added TFPI-α was evaluated by flow cytometry (FACS).
Cells (5×105) were resuspended in 200 μL of ice-cold PBS containing 0.08% (wt/vol) sodium azide, 25 mmol/L EDTA, and 3% (wt/vol) BSA (Sigma) and incubated with mouse monoclonal anti-TFPI Kunitz-1 (Mab2H8, 5 μg/mL) for 30 minutes. After washing 3 times with the same buffer, the cells were incubated for an additional 30 minutes on ice in the dark with phycoerythrin-conjugated F(ab)2 sheep anti-mouse IgG (Sigma). The labeled cells were washed twice and then analyzed by FACScan flow cytometry (Becton Dickinson). Regions appropriate for each of the tested cell types were defined by use of forward light scatter × side-angle scatter light-intensity plot. All measurements were performed with the same instrument settings, and at least 0.5×105 events were analyzed by use of CellQuest software (Becton Dickinson). Results were expressed as mean fluorescence intensity. Cells incubated with isotype control IgG, with the primary antibody alone, with the secondary antibody alone, and with neither were used as negative controls.
All the experiments were repeated at least 3 times. Values were compared by a 2-way ANOVA, followed, if the F value was significant, by a t test with Bonferroni’s correction. A probability value of P<0.05 identifies significant differences among tested populations.
Binding of Endogenously Expressed TFPI-α to Cells
Stable clones of mouse C127 cells expressing various forms of rTFPI-α were produced to determine which domains within TFPI-α are required for its interaction with the cell surface. rTFPI-α(252), rTFPI-α(242), and rTFPI-α(181) are rTFPI-α forms with progressive C-terminal truncations; rTFPI-α(desK3) lacks the K3 domain; and in rTFPI-α(R199L), the putative P1 residue in K3 has been changed from arginine to leucine (Figure 1). In preliminary experiments, the relationship between the levels of secreted and surface-bound rTFPI-α(WT) expressed by C127 was determined. After 16 hours of culture, rTFPI-α numbers in the media and on cell surfaces were 182±19 ng/106 cells and 2.47±0.21 ng/106 cells, respectively (see Methods). Thus, like cultured endothelial cells, only a small fraction of the synthesized TFPI-α remains associated with the C127 cell surface.5
Western blot analysis of the conditioned media from cells producing WT and the altered forms of TFPI-α by use of polyclonal antibodies against whole and the C-terminal peptide of TFPI-α showed rTFPIs of the anticipated molecular weight, without significant proteolytic degradation (Figure 2A). Moreover, the FXa-dependent inactivation of FVIIa/TF produced by the various forms of rTFPI-α was comparable, excluding gross alterations in protein structure (Figure 2B). Surface-bound rTFPI-α was determined by FACS in stable clones secreting comparable amounts of the recombinant proteins. The C-terminal truncated forms rTFPI-α(252) and rTFPI-α(242) and rTFPI-α(R199L) showed significantly reduced cell surface binding compared with rTFPI-α(WT) (P<0.05); rTFPI-α(desK3), which lacks only the K3 domain, and rTFPI-α(181), which lacks K3 and the C-terminus, failed to bind the membrane surface (Figure 3). Surface TFPI was also evaluated in transiently transfected mouse b-end3 cells expressing comparable amounts of rTFPI-α(WT), rTFPI-α(desK3), and rTFPI-α(K3/K2), in which K3 is replaced with K2. The altered forms of rTFPI-α failed to bind to the membrane surface (P<0.05 versus WT), confirming the importance of the K3 domain in the binding of TFPI-α to this cell line as well (Figure 4).
Effects of Heparin and PI-PLC on Certain Forms of TFPI-α Expressed in C127 Cells
For certain forms [rTFPI-α(WT), rTFPI-α(252), and rTFPI-α(242)], the effects of heparin, PI-PLC, and PI-PLC/heparin on surface TFPI-α were also evaluated. All the forms were insensitive to heparin (P=NS); PI-PLC could release the entire fraction of rTFPI-α(242) and ≈75% of rTFPI-α(WT) and rTFPI-α(252) from the membrane surface (P<0.05): the remaining ≈25% of the 2 latter forms was released by heparin only after PI-PLC pretreatment (P<0.05 versus PI-PLC) (Figure 5). The latter data suggest that rTFPI-α(WT) behaves on the membrane surface of C-127 like TFPI on the membrane surface of cultured endothelial cells5 and that the fraction of TFPI-α remaining bound after PI-PLC treatment is released by heparin in a process that involves the entire C-terminus domain.
Binding of Exogenously Added TFPI-α to EVC304 Cells
To confirm the importance of K3 and C-terminal domains in the cell surface localization of TFPI-α, the binding of exogenously added TFPI-α to human ECV cells, which possess some endothelium-like properties, was evaluated. Binding was determined after 2 hours of incubation of ECV↓TFPI-α, knocked-down for TFPI-α expression, with the conditioned media of C127 cells expressing rTFPI-α(WT) and the mutated forms of rTFPI-α. The C-terminal truncated forms rTFPI-α(252), rTFPI-α(242), and rTFPI-α(R199L) bound significantly less well than rTFPI-α(WT); rTFPI-α(181) and rTFPI-α(desK3) failed to localize at the membrane surface (Figure 6).
TFPI is the physiological inhibitor of the extrinsic, tissue factor–dependent, coagulation pathway. TFPI is expressed by different cell types, such as platelets, monocytes, macrophages, and smooth muscle cells, but endothelial cells are thought to represent the principal source of this protein.9,19,20 A small proportion of TFPI expressed by cultured endothelial cells remains associated with the cell surface. Previous studies have shown that endothelium-associated TFPI is released by PI-PLC treatment and colocalizes with detergent-resistant membrane containing glycosphingolipid rafts and caveolae,10,12,21 suggesting that TFPI binds to the membrane surface via a GPI anchorage.
We have recently demonstrated that 2 different forms of TFPI are produced, TFPI-α and TFPI-β, and that both are GPI-linked to the membrane surface.5 Because only TFPI-β carries a signal for a direct GPI linkage at its C-terminus, TFPI-α appears to be indirectly GPI-linked through a still unidentified GPI-linked protein(s). To further analyze the structures in TFPI-α required for cell surface binding, altered forms were tested in C127 cells because, unlike many others cell lines, surface binding of human TFPI-α is easily detected on this mouse cell. Furthermore, the C127-expressed rTFPI-α binds to C127 and ECV cells in a similar manner, suggesting that C127 cells are a reasonable system for further studies.
Progressive C-terminal truncation of the TFPI-α molecule leads to a progressive reduction in surface binding, suggesting the importance of TFPI C-terminus in the cell binding. Removal of K3 completely abrogates TFPI-α surface binding, and, within K3, mutation of P1 residue dramatically reduces the interaction of TFPI-α with the cell surface, demonstrating for the first time a potentially important physiological role for this domain. Although the K3 domain is not thought to possess proteinase inhibitory activity,22–24 its putative P1 residue is important for TFPI-α surface binding, at least in C-127 cells and ECV cells. This could imply a potentially reversible K3-proteinase interaction, but we have been unable to release TFPI-α from the surface with aprotinin, leupeptin, soybean trypsin inhibitor, or low-molecular-weight proteinase inhibitors of the chloromethyl ketone class (data not shown). These experiments, however, do not exclude a K3-proteinase interaction. Our studies show that optimal binding of TFPI-α to cells requires both the K3 and C-terminus of TFPI-α. This could be because a direct interaction of both the K3 and C-terminus with the binding protein(s) is required for high-affinity binding; because 1 domain is involved in maintaining the appropriate conformation of TFPI-α for the interaction of the other domain with the binding protein(s); because both domains play conformational roles to permit the interaction of a third TFPI-α domain with the binding protein(s); or any combination of these possibilities. The physiological function of the K3 domain of TFPI-α has been an enigma. Our work suggests that it plays an important role in localizing TFPI-α at the cell surface.
This work was supported in part by a grant (HL-60782) from the National Heart, Lung, and Blood Institute.
Broze GJ Jr, Warren LA, Novotny WF, et al. The lipoprotein associated coagulation inhibitor that inhibits the factor VII-tissue factor complex also inhibits factor Xa: insight into its possible mechanism of action. Blood. 1988; 71: 335–343.
Wun T-C, Kretzmer KK, Girard TJ, et al. Cloning and characterization of a cDNA coding for the lipoprotein-associated coagulation inhibitor shows that it consists of three tandem Kunitz-type inhibitory domains. J Biol Chem. 1988; 263: 6001–6004.
Zhang J, Piro O, Lu L, et al. Glycosyl phosphatidylinositol anchorage of tissue factor pathway inhibitor. Circulation. 2003; 108: 623–627.
Wesselschmidt R, Likert K, Girard T, et al. Tissue factor pathway inhibitor: the carboxy-terminus is required for optimal inhibition of factor Xa. Blood. 1992; 79: 2004–2010.
Sevinsky JR, Rao LV, Ruff W. Ligand induced protease receptor translocation into caveolae: a mechanism for regulating cell surface proteolysis of the tissue factor-dependent coagulation pathway. J Cell Biol. 1996; 133: 293–304.
Lupu C, Lupu F, Dennehy U, et al. Thrombin induces the redistribution and acute release of tissue factor pathway inhibitor from specific granules within human endothelial cells in culture. Arterioscler Thromb. 1995; 15: 2055–2062.
Ott I, Miyagi Y, Miyazaki K, et al. Reversible regulation of tissue factor–induced coagulation by glycosyl phosphatidylinositol-anchored tissue factor pathway inhibitor. Arterioscler Thromb Vasc Biol. 2000; 20: 874–882.
Crawley J, Lupu F, Westmuckett AD, et al. Expression, localization, and activity of tissue factor pathway inhibitor in normal and atherosclerotic human vessels. Arterioscler Thromb Vasc Biol. 2000; 20: 1362–1373.
Mast AE, Acharya N, Malecha MJ, et al. Characterization of the association of tissue factor pathway inhibitor with human placenta. Arterioscler Thromb Vasc Biol. 2002; 22: 2099–2104.
Caplice NM, Mueske CS, Kleppe LS, et al. Expression of tissue factor pathway inhibitor in vascular smooth muscle cells and its regulation by growth factors. Circ Res. 1998; 83: 1264–1270.
Bajaj M, Kuppuswamy M, Saito H, et al. Cultured normal human hepatocytes do not synthesize lipoprotein associated coagulation inhibitor: evidence that endothelium is the principal site of its synthesis. Proc Natl Acad Sci U S A. 1990; 87: 8869–8873.
Lupu C, Goodwin CA, Westmuckett AD, et al. Tissue factor pathway inhibitor in endothelial cells colocalizes with glycolipid microdomains/caveolae: regulatory mechanism(s) of the anticoagulant properties of the endothelium. Arterioscler Thromb Vasc Biol. 2000; 17: 2964–2974.
Kato H. Regulation of functions of vascular wall cells by tissue factor pathway inhibitor: basic and clinical aspects. Arterioscler Thromb Vasc Biol. 2002; 22: 539–548.