Advertisement

Thrombin-activable Fibrinolysis Inhibitor (TAFI) Zymogen Is an Active Carboxypeptidase*

  • Zuzana Valnickova
    Affiliations
    Center for Insoluble Protein Structure at the Department of Molecular Biology, Science Park, University of Aarhus, Gustav Wieds Vej 10c, 8000 Aarhus C, Denmark
    Search for articles by this author
  • Ida B. Thøgersen
    Affiliations
    Center for Insoluble Protein Structure at the Department of Molecular Biology, Science Park, University of Aarhus, Gustav Wieds Vej 10c, 8000 Aarhus C, Denmark
    Search for articles by this author
  • Jan Potempa
    Footnotes
    Affiliations
    Department of Microbiology, Faculty of Biochemistry, Biophysics, and Biotechnology, Jagiellonian University, 30-387 Krakow, Poland

    Department of Biochemistry and Molecular Biology, University of Georgia, Athens, Georgia 30602
    Search for articles by this author
  • Jan J. Enghild
    Correspondence
    To whom correspondence should be addressed: Dept. of Molecular Biology, University of Aarhus, Gustav Wied's Vej 10C, 8000 Aarhus C, Denmark. Tel.: 45-8942-5062
    Affiliations
    Center for Insoluble Protein Structure at the Department of Molecular Biology, Science Park, University of Aarhus, Gustav Wieds Vej 10c, 8000 Aarhus C, Denmark
    Search for articles by this author
  • Author Footnotes
    * This work was supported by grants from the Danish Natural Science Research Council (to J. J. E.) and National Institutes of Health Grant DE 09761 (to J. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
    1 Recipient of the SUBSYDIUM PROFESORSKIE award from the Foundation for Polish Science (Warszawa, Poland).
Open AccessPublished:November 30, 2006DOI:https://doi.org/10.1074/jbc.M606559200
      Thrombin-activable fibrinolysis inhibitor (TAFI) is a carboxypeptidase found in human plasma, presumably as an inactive zymogen. The current dogma is that proteolytic activation by thrombin/thrombomodulin generates the active enzyme (TAFIa), which down-regulates fibrinolysis by removing C-terminal lysine residues from partially degraded fibrin. In this study, we have shown that the zymogen exhibits continuous and stable carboxypeptidase activity against large peptide substrates, and we suggest that the activity down-regulates fibrinolysis in vivo.
      Human thrombin-activable fibrinolysis inhibitor (TAFI)
      The abbreviations used are: TAFI, zymogen of thrombin-activable fibrinolysis inhibitor; TAFIa, activated form of thrombin-activable fibrinolysis inhibitor; CPA, carboxypeptidase A; CPB, carboxypeptidase B; ELISA, enzyme-linked immunosorbent assay; Hipp, hippuryl; MALDI-TOF MS, matrix-assisted laser desorption ionization time-of-flight mass spectrometry; PCI, potato carboxypeptidase inhibitor; HPLC, high performance liquid chromatography; GEMSA, guanidinoethylmercaptosuccinicacid; tPA, tissue plasminogen activator.
      3The abbreviations used are: TAFI, zymogen of thrombin-activable fibrinolysis inhibitor; TAFIa, activated form of thrombin-activable fibrinolysis inhibitor; CPA, carboxypeptidase A; CPB, carboxypeptidase B; ELISA, enzyme-linked immunosorbent assay; Hipp, hippuryl; MALDI-TOF MS, matrix-assisted laser desorption ionization time-of-flight mass spectrometry; PCI, potato carboxypeptidase inhibitor; HPLC, high performance liquid chromatography; GEMSA, guanidinoethylmercaptosuccinicacid; tPA, tissue plasminogen activator.
      (EC 3.4.17.20; MEROPS, M14.009; UniProt, Q96IY4) is a single chain 60-kDa protein secreted by the liver and found in plasma at an average concentration of 13.4 μg/ml (220 nm) (
      • Stromqvist M.
      • Schatteman K.
      • Leurs J.
      • Verkerk R.
      • Andersson J.O.
      • Johansson T.
      • Scharpe S.
      • Hendriks D.
      ). TAFI is N-glycosylated, and the attached glycans account for ∼20% of the overall size of the protein. All five potential N-linked glycosylation sites are occupied, and four of these are located on the activation peptide and one on the enzyme moiety (
      • Valnickova Z.
      • Christensen T.
      • Skottrup P.
      • Thogersen I.B.
      • Hojrup P.
      • Enghild J.J.
      ). The protein belongs to the metallocarboxypeptidase family and is also known as plasma pro-carboxypeptidases B, R, and U (
      • Eaton D.L.
      • Malloy B.E.
      • Tsai S.P.
      • Henzel W.
      • Drayna D.
      ,
      • Campbell W.
      • Okada H.
      ,
      • Hendriks D.
      • Wang W.
      • Scharpe S.
      • Lommaert M.P.
      • van S
      • e M.
      ). TAFI was discovered by several groups as a plasminogen-binding protein (
      • Eaton D.L.
      • Malloy B.E.
      • Tsai S.P.
      • Henzel W.
      • Drayna D.
      ,
      • Hendriks D.
      • Wang W.
      • Scharpe S.
      • Lommaert M.P.
      • van S
      • e M.
      ,
      • Wang W.
      • Hendriks D.F.
      • Scharpe S.S.
      ,
      • Bajzar L.
      • Manuel R.
      • Nesheim M.E.
      ) leading investigators to explore a role in fibrinolysis. Indeed, the protein appears to be an important regulator of fibrinolysis by removing surface-exposed C-terminal Lys from partially degraded fibrin clots, thus reducing the number of plasminogen-binding sites (
      • Marx P.F.
      ). Activated TAFI (TAFIa) has also been implicated in the regulation of the inflammatory response by inactivating complement-derived inflammatory peptides (
      • Campbell W.
      • Okada N.
      • Okada H.
      ).
      Several different proteases are able to activate TAFI in vitro, but thrombin/thrombomodulin has been suggested as the most likely physiologically relevant activator (
      • Bajzar L.
      • Morser J.
      • Nesheim M.
      ). At the present, the proteolytic activity is exclusively assigned to TAFIa. This form is generated by cleavage of the Arg92–Ala93 peptide bond releasing the heavily glycosylated activation peptide. The dissociation of the activation peptide from the bulk of the molecule causes the isoelectric point of the enzyme moiety to shift from a pH value of ∼5 to 8, rendering the protein significantly more basic and less soluble (
      • Valnickova Z.
      • Christensen T.
      • Skottrup P.
      • Thogersen I.B.
      • Hojrup P.
      • Enghild J.J.
      ). In addition, the enzymatic activity of TAFIa is unstable, with a half-life ranging from 1.4 to 15 min at 37 °C, depending on the isoform (
      • Boffa M.B.
      • Bell R.
      • Stevens W.K.
      • Nesheim M.E.
      ,
      • Marx P.F.
      • Hackeng T.M.
      • Dawson P.E.
      • Griffin J.H.
      • Meijers J.C.
      • Bouma B.N.
      ). This inactivation is highly temperature-dependent and is most likely caused by a change in the structure of TAFIa (
      • Boffa M.B.
      • Bell R.
      • Stevens W.K.
      • Nesheim M.E.
      ,
      • Marx P.F.
      • Hackeng T.M.
      • Dawson P.E.
      • Griffin J.H.
      • Meijers J.C.
      • Bouma B.N.
      ). The structural change appears to expose regions with affinity for α2-macroglobulin and pregnancy zone protein, as these interactions are properties of TAFIa only (
      • Valnickova Z.
      • Thogersen I.B.
      • Christensen S.
      • Chu C.T.
      • Pizzo S.V.
      • Enghild J.J.
      ) and may function as a clearance mechanism to remove TAFIa that escapes covalent cross-linking to the clot catalyzed by transglutaminase (FXIII) during coagulation (
      • Valnickova Z.
      • Enghild J.J.
      ).
      TAFI is homologous to the pancreatic carboxypeptidases human procarboxypeptidase B (pro-CPB) and human procarboxypeptidases A1 and A2 (pro-CPA1 and pro-CPA2). These are characterized by having preformed active sites. The lack of enzymatic activity of the zymogen is controlled by the N-terminal activation peptide. The activation peptide is located on top of the active site cleft and sterically prevents access of substrate molecules to the catalytic site. As far as pro-CPB is concerned, this shielding is complete, and intrinsic enzymatic activity of the zymogen is negligible (
      • Reeck G.R.
      • Neurath H.
      ). This is in contrast to human pro-CPA1 and pro-CPA2, which cleave small peptide substrates. The activity is ∼10% of Vmax for dipeptides and decreases rapidly as the substrates become larger (
      • Reverter D.
      • Garcia-Saez I.
      • Catasus L.
      • Vendrell J.
      • Coll M.
      • Aviles F.X.
      ). In addition, older studies on pro-carboxypeptidase A from spiny Pacific dogfish and pro-carboxypeptidase B from African lung-fish also show intrinsic peptidase and esterase activity (
      • Reeck G.R.
      • Neurath H.
      ,
      • Lacko A.G.
      • Neurath H.
      ). These properties may be explained by structural differences of the activation peptides (
      • Garcia-Saez I.
      • Reverter D.
      • Vendrell J.
      • Aviles F.X.
      • Coll M.
      ). In particular, pro-CPB maintains a salt bridge between the activation peptide (Asp-41) and the S1′ substrate-binding site (Arg-145) in the catalytic moiety. This prevents the C-terminal carboxyl group of the substrate molecule access to the S1′-binding site (Arg-145), required for substrate immobilization. The residue homologous to Asp-41 is lacking in pro-CPA2, obstructing the formation of this salt bridge, which may at least in part explain the intrinsic activity of pro-CPA2. Similarly, the TAFI zymogen lacks the homologous Asp residue in the activation peptide and is most likely not able to form the stabilizing salt bridge (
      • Barbosa Pereira P.J.
      • Segura-Martin S.
      • Oliva B.
      • Ferrer-Orta C.
      • Aviles F.X.
      • Coll M.
      • Gomis-Ruth F.X.
      • Vendrell J.
      ). In addition, the heavy glycosylation of the TAFI activation peptide is unique among homologous pro-carboxypeptidases. The focus of this study was to investigate the influence of these properties on the intrinsic activity of the TAFI zymogen.

      EXPERIMENTAL PROCEDURES

      Materials—Bovine trypsin, o-phenylenediamine dihydrochloride, 1,10-phenanthroline, 4,7-phenanthroline, phenylmethylsulfonyl fluoride, and chromogenic substrates for carboxypeptidases, N-benzoyl-Gly-Arg (hippuryl-Arg), N-benzoyl-Gly-Lys (hippuryl-Lys), 3-(2-furyl)acryloyl-Ala-Arg-OH (FAAR) and 3-(2-furyl)acryloyl-Ala-Lys-OH (FAAK), were all obtained from Sigma. The chromogenic trypsin substrate HD-Pro-Phe-Arg-p-nitroaniline was purchased from Chromogenix. Ortho-methylhippuric acid was from Aldrich, and human plasminogen antiserum was obtained from MP Biomedicals. The Actichrome TAFIa activity kit was purchased from American Diagnostica. The TAFI inhibitor 2-guanidinoethylmercaptosuccinic acid (GEMSA) was from Calbiochem. Human plasma was from Statens Serum Institute, Copenhagen, Denmark.
      Proteins and Peptides—TAFI was purified from normal human plasma using plasminogen-depleted plasma as described previously (
      • Eaton D.L.
      • Malloy B.E.
      • Tsai S.P.
      • Henzel W.
      • Drayna D.
      ,
      • Valnickova Z.
      • Thogersen I.B.
      • Christensen S.
      • Chu C.T.
      • Pizzo S.V.
      • Enghild J.J.
      ). Porcine CPB was purchased from Calbiochem and further purified on arginine-Sepharose 4B (Amersham Biosciences) using protocols supplied by the manufacturer. Human plasminogen was purified by affinity chromatography using ECH-lysine-Sepharose (Amersham Biosciences) (
      • Deutsch D.G.
      • Mertz E.T.
      ). Human fibrinogen was from Sigma. Recombinant potato carboxypeptidase inhibitor (PCI) was a kind gift from Drs. Francesc X. Aviles, Josep Vendrell, and Irantzu G. Palláres. The protein was prepared as previously described (
      • Molina M.A.
      • Aviles F.X.
      • Querol E.
      ). Human tissue plasminogen activator (tPA) was from ProSpec-Tany TechnoGene Ltd., Rehovot, Israel. Synthetic peptides were supplied by Sigma Genosys or assembled in an Applied Biosystems model 433A peptide synthesizer (Applied Biosystems) using Fast-Moc chemistry. The peptides were purified by reverse phase HPLC and verified by mass spectrometry. Recombinant soluble thrombomodulin (Solulin) was a generous gift from Dr. Achim Schuettler (PAION GmbH, Aachen Germany). TAFI polyclonal rabbit antiserum was raised commercially (Pel-Freez). Protein and peptide concentrations were determined by amino acid analysis as described previously (
      • Ozols J.
      ).
      Polyacrylamide Gel Electrophoresis—Proteins were separated on 5–15% SDS-polyacrylamide gels (
      • Bury A.F.
      ). The samples were boiled for 5 min in the presence of 30 mm dithiothreitol and 1% SDS prior to electrophoresis.
      Reduction and Alkylation—The sample was reduced for 20 min in 20 mm NH4HCO3, pH 8.0, containing 15 mm dithiothreitol and subsequently carboxymethylated for 20 min using 15 mm iodoacetamide.
      Generation of Fibrin Peptides—Fibrinogen (30 mg) was reduced and alkylated and then digested with trypsin (1:20 w/w) for 5 h at 37 °C. The trypsin was removed from the sample by affinity chromatography using a benzamidine-Sepharose 4B FF column, connected to an AKTA Prime system (Amersham Biosciences). Briefly, the sample was applied to the column equilibrated in TBS (20 mm Tris-HCl, 137 mm NaCl, pH 7.6), and the flow-through containing the fibrin peptides was collected. The bound trypsin was eluted with 50 mm glycine, pH 2.7, and discarded. Phenylmethylsulfonyl fluoride was added to the purified fibrin peptides, and the lack of trypsin activity was verified by the inability of the sample to hydrolyze the trypsin substrate HD-Pro-Phe-Arg-p-nitroaniline using a plate reader (Molecular Devices THERMOmax™) operating in the kinetic mode (
      • Enghild J.J.
      • Valnickova Z.
      • Thogersen I.B.
      • Pizzo S.V.
      • Salvesen G.
      ).
      Activation of Human TAFI—TAFI was activated by the Solulin-thrombin complex as described previously (
      • Valnickova Z.
      • Christensen T.
      • Skottrup P.
      • Thogersen I.B.
      • Hojrup P.
      • Enghild J.J.
      ). In short, 1 μg of TAFI and a complex of 0.5 μg of Solulin and 0.02 μg of thrombin were incubated for 30 min in 20 mm Tris-HCl, 100 mm NaCl, pH 7.5, at 25 °C.
      TAFI Zymogen Activity Assay Using Hipp-Arg and Hipp-Lys as Substrates—The assay was performed essentially as described previously (
      • Hendriks D.
      • Scharpe S.
      • van S
      • e M.
      ). Samples containing TAFI (1 μg) in 10 μl of TBS were incubated with 40 μl of 30 mm hippuryl-Arg or 80 μl of 100 mm hippuryl-Lys for 40 min. The reactions were stopped by adding 50 μl or 100 μl of 1 m HCl, and 10 μl of 15 mm ortho-methylhippuric acid was added as an internal standard. The reaction products, as well as the internal standard, were extracted using 300 or 600 μl of ethyl acetate. Subsequently, 100 or 200 μl of the extracts were lyophilized and solubilized in 200 μl of mobile phase (10 mm KH2PO4, pH 3.4, containing 15% acetonitrile) and separated on a reverse phase HPLC column (PTH C18, 5 μm, 220 × 2.1 mm, Applied Biosystems) connected to an AKTA Explorer system (Amersham Biosciences). Some TAFI zymogen was pretreated with 1,10-phenanthroline or 4,7-phenanthroline (8 mm final concentration, 20 min, 25 °C) and 10 mm EDTA prior to the addition of the substrate.
      TAFI Zymogen Activity Assay Using the Actichrome TAFI Activity Kit—This assay was performed according to the manufacturer's instructions (American Diagnostica), except that the supplied activation reagent, activation enhancer, and activation stop reagent were omitted. In short, the samples containing titrations of TAFI zymogen, CPB, or the supplied TAFIa standard were assayed in parallel. The samples were added to 96-well plates in a total volume of 150 μl using the supplied assay buffer. The developer was added to the wells (50 μl), and the samples were incubated for 30 min at 37 °C. The developer contains a tetrapeptide, PFGK (447.25 Da), horseradish peroxidase (EC 1.11.1.7), lysine oxidase (EC 1.4.3.14), and o-phenylenediamine dihydrochloride, a substrate for horseradish peroxidase. The lysine oxidase converts the released lysine to 6-amino-2-oxohexanoate, NH3, and H2O2. Horseradish peroxidase utilizes the produced H2O2 and oxidizes the o-phenylenediamine dihydrochloride to produce a color reaction. The reaction was stopped by adding 50 μl of 2 m H2SO4 to the wells, and finally the plate was measured at 490 nm using a plate reader (Molecular Devices THERMOmax™) operated in end point mode.
      Analysis of TAFI Zymogen Activity Using Larger Synthetic Peptides—To investigate the TAFI zymogen activity against larger substrates, we prepared two synthetic peptides, RGDSTFESKSYK (1403.67 Da) and ERREKEAREASHRQKRSCEAGK (2640.34 Da). TAFI zymogen and the peptides were incubated at various molar ratios ranging from 1:4 to 1:50 (TAFI zymogen: peptide) for 10 or 60 min at 37 °C in TBS, pH 7.4. The reaction products were either separated by reverse phase chromatography or desalted by Zip tipping (μC18-ZipTip, Millipore) prior to analyses by mass spectrometry.
      Kinetic Assays Using a Synthetic Fibrin PeptideKm and kcat values were determined for the synthetic peptide RGDSTFESKSYK (1403.67 Da) using TAFI zymogen, TAFIa, or porcine CPB. The final concentrations of the enzymes were 0.17, 0.025, and 0.0025 μm, respectively, using peptide concentrations ranging from 210 to 2100 μm. The reactions were carried out at 37 °C in TBS and continued for 10 min. Aliquots were removed at timed intervals, the reaction stopped by acidification (0.5% trifluoroacetic acid final concentration), and samples were applied to a reverse phase column (Nucleosil 5 C18, 300A, 250 × 2 mm, Phenomenex) connected to an AKTA Explorer system (Amersham Biosciences). The column was eluted using an isocratic gradient composed of 0.1% trifluoroacetic acid containing 16.2% acetonitrile as a mobile phase. The concentration of the truncated peptide was calculated based on the area of the corresponding peak and converted to substrate turnover rates expressed as μm/min. The initial velocities of the product generated at various substrate concentrations were determined by the best-fit line after incubation for various time periods. The values for Km and Vmax were extracted by direct fit of the Michaelis-Menten equation to the experimental data using non-linear curve fitting employing the method of least squares with Taylor expansion (
      • Sakoda M.
      • Hiromi K.
      ). In addition, the data were analyzed by five different graphical methods, including hyperbolic regression and Hanes-Woolf plot. To determine the effect of the released lysine on carboxypeptidase activity of the TAFI zymogen, TAFIa, and CPB, the initial velocities of the RGDSTFESKSYK peptide cleavage at five different substrate concentrations including 0.42, 0.63, 0.84, 1.05, and 1.26 mm were measured in the presence of 0, 2, 5, 10, and 20 mm lysine. The reciprocal of the initial reaction velocity (1/v) for each substrate concentration was plotted against the inhibitor concentration [I] (Dixon plot) and the Ki value calculated. Additionally, types of inhibitions were depicted by the double reciprocal, plotting 1/v against 1/[S] for each inhibitor concentration (Lineweaver-Burk plot).
      Matrix-assisted Laser Desorption Ionization Mass Spectrometry (MALDI-TOF MS)—Peptides were analyzed by MALDI-TOF MS on a Q-TOF Ultima Global mass spectrometer (Micromass) operated in reflectron MALDI-TOF MS mode. All experiments were performed using α-cyano-4-hydroxycinnamic acid (Sigma) as the matrix. The expected mass was calculated using General Protein/Mass Analysis for Windows software (Lighthouse Data, Odense, Denmark).
      Plasminogen-binding Assay—A 96-well plate was coated overnight at 4 °C with ovalbumin-CRGDSTFESKSYK peptide conjugate or with purified fibrin peptides (see above). The CRGDSTFESKSYK peptide was coupled using the N-terminal Cys residue (
      • Kitagawa T.
      • Aikawa T.
      ,
      • Liu F.T.
      • Zinnecker M.
      • Hamaoka T.
      • Katz D.H.
      ). Coating agents were diluted in 20 mm NaHCO3, pH 9.2, and 150 μl was added to the wells. On the next day, the plates were blocked and washed in TBS containing 0.1% Tween (TBS-T). Serial dilutions of TAFI zymogen or porcine CPB were added to the coated wells in 100 μl of TBS/well, and the plates were incubated for 1 h at 37 °C. TAFI zymogen was replaced three times during this incubation period (3 × 20 min). This was done to reduce the observed product inhibition by the released lysine. Although CPB also exhibited product inhibition, the Ki was high, and the activity was not significantly affected by the released lysine. Following the TAFI zymogen or CPB treatment of the wells, a fixed concentration of human plasminogen (10 μg/ml) in 100 μl of TBS-T was added, and the plate was incubated for 1 h at 37 °C. The plates were washed as described above, and bound plasminogen was detected using a standard ELISA protocol including plasminogen antiserum, horseradish peroxidase conjugate secondary antibody (rabbit anti-goat, Sigma), and o-phenylenediamine dihydrochloride as the substrate. The plate was measured at 450 nm in a plate reader (Molecular Devices THERMO-max™) operated in end point mode. Each data point was collected in duplicates. The relative binding capacity of plasminogen was estimated in the linear range of the binding curve. Nonspecific binding was determined by binding of plasminogen to ovalbumin-CRGDSTFESKSYK or fibrin-coated plates pretreated with porcine CPB. The background (<2%) was subtracted from all groups.
      Plasma in Vitro Clot Lysis Assays—Clot lysis assays were performed essentially as described previously (
      • Arolas J.L.
      • Lorenzo J.
      • Rovira A.
      • Castella J.
      • Aviles F.X.
      • Sommerhoff C.P.
      ,
      • Walker J.B.
      • Hughes B.
      • James I.
      • Haddock P.
      • Kluft C.
      • Bajzar L.
      ). In short, human plasma, diluted 1:3 in 20 mm Hepes, 150 mm NaCl, pH 7.4, was added to 96-well microtiter plates. Clotting and fibrinolysis was initiated by adding human thrombin (final concentration 5 nm), CaCl2 (final concentration 10 mm), and tPA (0.02–0.12 μg/ml) in the presence or absence of Solulin (10 nm) and in the presence and absence of GEMSA (1000 μm) or PCI (1600 nm). The absorbance at 405 nm was measured continuously in a plate reader (Molecular Devices THERMOmax™) at 37 °C. The lysis time was determined as the time required for a 50% reduction in absorbance.

      RESULTS

      TAFI Zymogen Cleaves the Substrates Hipp-Arg and Hipp-Lys—When TAFI zymogen was incubated with the carboxypeptidase substrates hippuryl-Lys (Fig. 1A) or hippuryl-Arg (not shown), a low but significant activity was apparent. The separation of the product and the substrate by reverse phase HPLC was critical for detection of the zymogen activity using this system. This activity was eliminated in the presence of 1,10-phenanthroline (Fig. 1B), supporting the hypothesis that TAFI zymogen exhibits authentic carboxypeptidase activity. EDTA also decreased the release of hippuric acid, although to a lesser extent, whereas 4,7-phenanthroline failed to inhibit the zymogen activity (data not shown).
      Figure thumbnail gr1
      FIGURE 1TAFI zymogen activity detected using hippuryl-Lys as substrate. TAFI zymogen (1 μg) was incubated with the substrates Hipp-Lys alone (A) and in the presence of 1,10-phenanthroline (8 mm)(B). Similar results were obtained using Hipp-Arg (not shown). The reaction products and the added internal standard were separated by reverse phase HPLC. The peaks in the chromatogram represent the substrate (
      • Stromqvist M.
      • Schatteman K.
      • Leurs J.
      • Verkerk R.
      • Andersson J.O.
      • Johansson T.
      • Scharpe S.
      • Hendriks D.
      ), the released hippuric acid (
      • Valnickova Z.
      • Christensen T.
      • Skottrup P.
      • Thogersen I.B.
      • Hojrup P.
      • Enghild J.J.
      ), and the internal standard o-methylhippuric acid (
      • Eaton D.L.
      • Malloy B.E.
      • Tsai S.P.
      • Henzel W.
      • Drayna D.
      ). It is evident that both Hipp-Lys (A) and Hipp-Arg (not shown) are cleaved by the TAFI zymogen. The release of hippuric acid (peak 2) was significantly decreased by the addition of 1,10-phenanthroline (B), confirming authentic carboxypeptidase activity.
      TAFI Zymogen Activity Validated Using Synthetic Peptides— The proteolytic activity of the zymogen was further analyzed using the Actichrome TAFI kit. The TAFI zymogen showed carboxypeptidase activity against the NH2-PFGK peptide, although the activity was severalfold lower than that of porcine CPB (Fig. 2A). To further probe access to the catalytic sites, larger peptides were analyzed. These peptides, RGDSTFESKSYK (1403.67 Da) (Fig. 2B) and ERREKEAREASHRQKRSCEAGK (2640.34 Da) (Fig. 2C), were both cleaved by the zymogen. Approximately 50% of the peptides were cleaved within 10 min using an enzyme:substrate ratio of 1:4, and up to 30% of the peptides were cleaved within 60 min using enzyme:substrate ratios of 1:50 (all molar ratios). This shows that the active site of the TAFI zymogen is significantly easier to access than has previously been observed for other pro-carboxypeptidases.
      Figure thumbnail gr2
      FIGURE 2TAFI zymogen activity was verified using various synthetic peptides. Porcine CPB and TAFI zymogen activity were assayed using the synthetic peptide (PFGK, 447.25 Da) (A). The released lysine is detected as described under “Experimental Procedures.” The graph depicts the absorbance readings of wells at 490 nm, where a series of CPB (•) or TAFI zymogen (▪) dilutions were incubated with developer only. A synthetic fibrin peptide (RGDSTFESKSYK, 1403.67 Da) was incubated with TAFI zymogen for 60 min at 37 °C at an enzyme:substrate molar ratio 1:26 (B). The cleaved peptide (MH+ 1276.59 Da) was separated from the intact peptide (MH+ 1404.67 Da) by reverse phase HPLC and identified by MALDI-TOF MS. A random peptide (ERREKEAREASHRQKRSCEAGK, 2640.34 Da) containing a C-terminal lysine was incubated with TAFI zymogen for 10 min at 37 °C at an enzyme:substrate molar ratio of 1:4 and analyzed by MALDI-TOF MS (C). It was apparent that the peptide (MH+ 2641.34 Da) was cleaved (MH+ 2513.24 Da) by the TAFI zymogen. Both the intact and cleaved peptides were detected as smaller variants (MH+ 2496.0 Da and MH+ 2624.1 Da) due to loss of ammonia (16.9 Da) during MALDI-TOF. Only the monoisotopic masses are labeled.
      TAFI Zymogen Exhibits Activity toward Natural Peptide Substrate—The kinetics of the zymogen carboxypeptidase activity were analyzed in more detail using the synthetic fibrinogen-derived peptide RGDSTFESKSYK. The intact (RGDSTFESKSYK) and the cleaved (RGDSTFESKSY) peptides were readily separated by reverse phase HPLC, providing a quantitative method for analyses of the reaction products (Fig. 2B). The kinetic constants for reaction catalyzed by the TAFI zymogen, TAFIa, or porcine CPB were determined using this assay after the identity of the separated products was verified by MALDI-TOF MS. In each case, the peptide cleavage followed Michaelis-Menten kinetics, as illustrated by two different graphical plots, the hyperbolic regression and Hanes-Woolf plots (Fig. 3, A–C and insets, respectively). The obtained kinetic data allowed for the calculation of Km, Vmax, and kcat values (Table 1). The analysis shows that, although the kcat value of the TAFI zymogen is 41-fold lower than the kcat value for TAFIa, the general catalytic potency of the zymogen measured by the kcat/Km ratio is only 18-fold lower than that of the activated protease (Table 1).
      Figure thumbnail gr3
      FIGURE 3Graphical illustrations of TAFI zymogen, TAFIa, and CPB activity. The enzymatic activity of 0.17 μm TAFI zymogen (A), 0.025 μm TAFIa (B), or 0.0025 μm CPB (C) was determined using the synthetic fibrin peptide (RGDSTFESKSYK) as a substrate. The kinetic data are evaluated by a simple hyperbola curve using initial rate (v) as a function of substrate concentration (main graph in each panel) and the straight line form of the Michaelis-Menten equation depicted as a reciprocal plot (Hanes-Woolf plot) of [S]/v against substrate concentration [S](inset in each panel). Each point represents the mean of duplicates. The kinetic parameters are presented in .
      TABLE 1Summary of the obtained kinetic constants using a large peptide substrate (RGDSTFESKSYK) The values for Km and Vmax were extracted by the direct fit of the Michaelis-Menten equation to experimental data using five different methods of computation. The data represent the enzyme-catalyzed reactions for 0.17 μm TAFI zymogen, 0.025 μm TAFIa, and 0.0025 μm CPB.
      EquationTAFI zymogenTAFIaCPB
      VmaxKmVmaxKmVmaxKm
      μm/minμmμm/minμmμm/minμm
      Hanes6.88277.1842.31650.1117.67395.66
      Lineweaver-Burk6.87275.5741.36609.3114.47384.40
      Eadie-Hofstee6.89277.2841.59618.7017.53387.11
      Eisenthal-Cornish-Bowden6.92280.9441.43609.9117.50385.11
      Hyperbolic regression6.92 ± 0.26280.65 ± 30.4141.99 ± 2.88636.35 ± 126.6917.62 ± 0.53393.14 ± 33.85
      Average values6.89278.3241.73624.8716.95389.08
      Kcat min–140.591669.205853.33
      Kcat/Km min–1m0.152.6715.04
      A progress curve of TAFI zymogen-, TAFIa-, and CPB-catalyzed cleavage of peptides with the C-terminal Lys residue declined during the time course of the product formation (Fig. 4). Because one possible reason for this departure from linearity is enzyme product inhibition, we checked the effect of the released lysine on TAFI zymogen, TAFIa, and CPB activity. These analyses showed that lysine reduces the substrate turnover rate in a concentration-dependent manner. By plotting the kinetic data into the Lineweaver-Burk graph, we established that lysine is a simple competitive inhibitor of CPB activity (Fig. 4C, inset), whereas the inhibition of both the TAFI zymogen and TAFIa is mixed but predominantly noncompetitive (Fig. 4, A and B; insets, respectively). Ki values were determined by Dixon plots (data not shown) to be 2.2 mm for TAFI zymogen, 3.8 mm for TAFIa, and 16.0 mm for CPB.
      Figure thumbnail gr4
      FIGURE 4Product inhibition of TAFI zymogen, TAFIa, and CPB. Progress curves of TAFI zymogen, 0.17 μm (A); TAFIa, 0.025 μm (B); or CPB, 0.0025 μm (C) catalyzed cleavage of a peptide substrate (RGDSTFESKSYK) are plotted using initial rate (v) as a function of time (main graph in each panel). The type of inhibition was analyzed in a Lineweaver-Burk double reciprocal plot, 1/v against 1/[S], (inset in each panel) using various inhibitor concentrations as indicated (0–20 mm). The data show that lysine inhibition of TAFI zymogen and TAFIa is mixed and predominantly non-competitive, whereas CPB inhibition is simple competitive. Each point represents the mean of duplicates.
      Implications of TAFI Zymogen Activity on Fibrinolysis—To analyze the significance of the zymogen activity and its implication on fibrinolysis, a plasminogen-binding assay was devised. Microtiter plates were coated with tryptic fibrinogen peptides containing exposed C-terminal arginine or lysine residues (Fig. 5, A and B). In addition, a synthetic fibrin peptide ovalbumin-(CRGDSTFESKSYK) conjugate was used (Fig. 5C). As expected, plasminogen bound to the peptide-coated wells in a dose-dependent manner (Fig. 5A). When the plates were pretreated with an increasing amount of porcine CPB, a decrease in the plasminogen-binding capacity was apparent (Fig. 5, B and C). Similarly, when the wells were pretreated with TAFI zymogen, the plasminogen binding decreased (Fig. 5, B and C). This effect was eliminated when porcine CPB or TAFI zymogen were inhibited by 1,10-phenanthroline prior to addition to the wells (data not shown). Significantly, the decrease in the plasminogen-binding capacity was substantial at a TAFI zymogen concentration similar to the concentration in human plasma. These data suggest that TAFI zymogen has relevant carboxypeptidase activity that may affect fibrinolysis in vivo.
      Figure thumbnail gr5
      FIGURE 5Comparison of TAFI zymogen and CPB activity in a plasminogen-binding assay. ELISA plates were coated with purified fibrin-derived tryptic peptides (A and B) containing C-terminal Lys and Arg residues or ovalbumin-CRGDSTFESKSYK peptide conjugate (C). The plasminogen-binding capacity was determined using a standard ELISA protocol. Each point represents the mean of duplicates. Plasminogen was binding to fibrin-derived tryptic peptides in a dose-dependent manner (A). The preincubation of the peptide-coated plates with TAFI zymogen (•) or CPB (○) prior to plasminogen addition significantly reduced the plasminogen-binding capacity (B and C).
      Effect of TAFI Zymogen in tPA-induced Clot Lysis—The thrombin-thrombomodulin (Solulin) complex increases the catalytic efficiency of the TAFI zymogen activation reaction by >1000-fold as compared with thrombin alone (
      • Bajzar L.
      • Morser J.
      • Nesheim M.
      ). Consequently, it can be presumed that very little if any TAFI zymogen is activated by low thrombin concentrations in the absence of Solulin. We exploited this in an attempt to further investigate the impact of the TAFI zymogen on clot lysis.
      In the presence of Solulin, clot lysis was significantly delayed (Fig. 6A). This is in agreement with activation of endogenous TAFI zymogen and the subsequent depletion of the Lys-binding sites. When the assay was performed in the presence of thrombin-Solulin complex and PCI, clot lysis was accelerated (Fig. 6A). Significantly, identical clot lysis times were achieved in the presence of thrombin alone (Fig. 6A). In other words, when TAFIa was inhibited by PCI, the effect on clot lysis was the same as when Solulin was omitted from the assay. Apparently, Solulin is required to achieve activation of TAFI zymogen. Additionally, Western blot analysis using a polyclonal TAFI antiserum confirmed that TAFI zymogen was not activated in the absence of Solulin (data not shown).
      Figure thumbnail gr6
      FIGURE 6Effect of TAFI zymogen activity on clot lysis experiments. To analyze the impact of the TAFI zymogen on fibrinolysis, plasma clots were formed in ELISA plates, and the change in turbidity was followed over time at 405 nm. Coagulation was initiated by adding thrombin (5 nm) to diluted human plasma in the presence of calcium chloride (20 mm), tPA (0.02–0.12 μg/ml), and in the presence or absence of Solulin (10 nm). Upon the addition of Solulin, the thrombin-Solulin complex activated the endogenous TAFI, and a significant delay of the clot lysis was observed as expected (▪, A). Samples lacking Solulin (Δ, A + B) exhibited a similar lysis time as samples where the generated endogenous TAFIa was inhibited by the addition of PCI (1600 nm) (□ A). Significantly, when TAFI zymogen (5 μg) was added in the absence of Solulin, the clot was sustained (•, A). This effect was abolished when TAFI zymogen was added to the clot lysis experiments in the presence of PCI (1600 nm)(○, A), confirming genuine carboxypeptidase activity. The antifibrinolytic manifestation of the zymogen was further examined by titrating TAFI into the reaction (0.2–12.8 μg) in the absence of Solulin. The results showed a dose-dependent delay in clot lysis times (B). Taken together, these data suggest that the observed inhibition of fibrinolysis is caused by the TAFI zymogen intrinsic proteolytic activity.
      The clot lysis assays were then performed in the absence of Solulin to evaluate the effect of the TAFI zymogen. The addition of TAFI zymogen prolonged clot lysis in a dose-dependent manner (Fig. 6B). This effect was reversed by the addition of PCI (Fig. 6A) and GEMSA (data not shown). In addition, samples where TAFI was added showed no activation of the zymogen when subjected to SDS-PAGE and Western blotting using polyclonal TAFI antiserum (data not shown). It is apparent that the TAFI zymogen is able to function as an inhibitor of fibrinolysis.

      DISCUSSION

      The glycosylation of the TAFI activation peptide is unique among related pro-carboxypeptidases, which do not carry activation peptide-associated carbohydrate moieties (
      • Eaton D.L.
      • Malloy B.E.
      • Tsai S.P.
      • Henzel W.
      • Drayna D.
      ). The attachment of N-linked glycans occurs co-translocationally early during biosynthesis and thus precedes folding of the polypeptide chain (
      • Kowarik M.
      • Kung S.
      • Martoglio B.
      • Helenius A.
      ,
      • Helenius A.
      • Aebi M.
      ). The importance of glycans during folding is variable, but in general, the folding of glycoproteins that carries a larger number of glycans is more glycan-dependent (
      • Helenius A.
      • Aebi M.
      ). The high density of N-linked glycans on the activation peptide is thus likely to affect the secondary structure and conformational freedom, impacting the overall structure of the activation peptide (
      • Petrescu A.J.
      • Milac A.L.
      • Petrescu S.M.
      • Dwek R.A.
      • Wormald M.R.
      ). This is of particular importance, because the activation peptide of pro-carboxypeptidases functions as a gatekeeper controlling access to the preformed active site. The activation peptide-associated glycans may thus be important for the observed intrinsic catalytic activity of TAFI zymogen.
      The TAFI zymogen activity was initially observed in assays using the synthetic substrates Hipp-Arg and Hipp-Lys followed by extraction and reverse phase HPLC separation of the substrate and product. This was performed according to a protocol where the hippuric acid product was detected at 228 nm (
      • Hendriks D.
      • Scharpe S.
      • van S
      • e M.
      ). The sensitivity of this assay is significantly increased as compared with the more common direct measurement of the reaction products in the spectrophotometer at 254 nm (
      • Bajzar L.
      • Manuel R.
      • Nesheim M.E.
      ,
      • Folk J.E.
      • Piez K.A.
      • Carroll W.R.
      • Gladner J.A.
      ). At 254 nm, an overlap in the absorbance of Hipp-Arg or Hipp-Lys and hippuric acid decreases the dynamic range of this assay. In addition, the hydrophobic and bulky nature of the Hipp-Arg or Hipp-Lys substrates might limit access to the active site of TAFI zymogen. Similarly, 3-(2-furyl)-acryloyl-Ala-Arg-OH (FAAR) and 3-(2-Furyl)-acryloyl-Ala-Lys-OH (FAAK) (
      • Plummer Jr., T.H.
      • Kimmel M.T.
      ) appeared to lack sensitivity. This was in contrast to another chromogenic assay using the peptide PFGK (447.25 Da), where TAFI zymogen activity was readily detected. This assay appeared to be significantly more sensitive, although a direct comparison was not attempted.
      To probe the size restriction imposed on the TAFI zymogen, two larger peptides were prepared, including RGDSTFESKSYK (1403.67 Da) and ERREKEAREASHRQKRSCEAGK (2640.34 Da). These are the largest substrates a pro-carboxypeptidase has been shown to cleave, providing evidence that the active site of the TAFI zymogen is significantly more accessible than observed previously with other pro-carboxypeptidases (
      • Reeck G.R.
      • Neurath H.
      ,
      • Reverter D.
      • Garcia-Saez I.
      • Catasus L.
      • Vendrell J.
      • Coll M.
      • Aviles F.X.
      ,
      • Lacko A.G.
      • Neurath H.
      ). Without structural data, however, the nature and geometry of this access remains unclear.
      In comparison to TAFIa, the peptide substrate turnover rate of the TAFI zymogen is 41-fold lower but compensated by a 2-fold higher affinity for the substrate. At the low substrate concentrations likely to exist within the fibrin clot, the zymogen will have an advantage as compared with TAFIa because of tighter substrate-binding ability. Additionally, in contrast to TAFIa, the TAFI zymogen displays sustained carboxypeptidase activity. This allows the TAFI zymogen to react with the fibrin clot for an extended period of time or until its activity is dampened by the increasing concentration of lysine. In this context, it is tempting to hypothesize that the mechanism of the observed TAFI-associated inhibition of fibrinolysis is regulated by the intrinsic lability of TAFIa and the lower but significant activity of the TAFI zymogen. To this end, generation of TAFIa by the action of thrombin-thrombomodulin during the initial clotting event is associated with an initial burst of the carboxypeptidase activity followed be a sustained slow rate cleavage executed by the TAFI zymogen. Collectively, this new mechanism may provide an additional sophisticated tuning of the rate of clot solubilization in vivo.
      The validity of this hypothesis was substantiated by plasminogen-binding assays, where peptides containing C-terminal Lys residues were immobilized in ELISA plates. The ability of plasminogen to bind to this simulated fibrin clot was significantly reduced at physiological TAFI concentrations (10 ng/μl). The observed reduction in plasminogen binding is proportional to a reduction in plasminogen activation, suggesting that the TAFI zymogen is able to significantly impact fibrinolysis.
      It is interesting to note that the TAFI zymogen is a substrate for transglutaminases, and two of the three amine acceptor sites are located near the N-terminal of the activation peptide (
      • Valnickova Z.
      • Enghild J.J.
      ). This may facilitate the immobilization of the TAFI zymogen to the fibrin clot by the action of FXIIIa during the final stages of the coagulation, thus increasing the local TAFI zymogen concentration.
      The potential physiological importance of the observation was further evaluated in tPA-induced clot lysis assays. These experiments showed that the zymogen significantly prolonged the duration of a clot at concentrations similar to what is found in vivo. The activity was abolished by PCI and GEMSA, corroborating that the observed effect was caused by genuine carboxypeptidase activity. Fibrinolysis was inhibited by the addition of increasing amounts of the zymogen, even though no activation of TAFI was observed, suggesting that the zymogen intrinsic activity is relevant in vivo.
      Taken together, we have shown that the TAFI zymogen is active toward large peptide substrates, cleaves exposed Lys residues at a physiologically relevant concentration, and attenuates fibrinolysis in tPA-induced clot lysis in vitro. The genuine, continuous, and stable carboxypeptidase activity of TAFI zymogen offers sustained protection of the fibrin clot in vivo and constitutes a novel, previously unrecognized mechanism with the potential to significantly impact fibrinolysis.

      Acknowledgments

      We gratefully acknowledge helpful comments and the generous gift of PCI from Drs. Francesc X. Aviles, Josep Vendrell, and Irantzu G. Palláres and recombinant soluble thrombomodulin (Solulin) from Dr. Achim Schuettler (PAION GmbH, Aachen Germany).

      References

        • Stromqvist M.
        • Schatteman K.
        • Leurs J.
        • Verkerk R.
        • Andersson J.O.
        • Johansson T.
        • Scharpe S.
        • Hendriks D.
        Thromb. Haemostasis. 2001; 85: 12-17
        • Valnickova Z.
        • Christensen T.
        • Skottrup P.
        • Thogersen I.B.
        • Hojrup P.
        • Enghild J.J.
        Biochemistry. 2006; 45: 1525-1535
        • Eaton D.L.
        • Malloy B.E.
        • Tsai S.P.
        • Henzel W.
        • Drayna D.
        J. Biol. Chem. 1991; 266: 21833-21838
        • Campbell W.
        • Okada H.
        Biochem. Biophys. Res. Commun. 1989; 162: 933-939
        • Hendriks D.
        • Wang W.
        • Scharpe S.
        • Lommaert M.P.
        • van S
        • e M.
        Biochim. Biophys. Acta. 1990; 1034: 86-92
        • Wang W.
        • Hendriks D.F.
        • Scharpe S.S.
        J. Biol. Chem. 1994; 269: 15937-15944
        • Bajzar L.
        • Manuel R.
        • Nesheim M.E.
        J. Biol. Chem. 1995; 270: 14477-14484
        • Marx P.F.
        Curr. Med. Chem. 2004; 11: 2335-2348
        • Campbell W.
        • Okada N.
        • Okada H.
        Immunol. Rev. 2001; 180: 162-167
        • Bajzar L.
        • Morser J.
        • Nesheim M.
        J. Biol. Chem. 1996; 271: 16603-16608
        • Boffa M.B.
        • Bell R.
        • Stevens W.K.
        • Nesheim M.E.
        J. Biol. Chem. 2000; 275: 12868-12878
        • Marx P.F.
        • Hackeng T.M.
        • Dawson P.E.
        • Griffin J.H.
        • Meijers J.C.
        • Bouma B.N.
        J. Biol. Chem. 2000; 275: 12410-12415
        • Valnickova Z.
        • Thogersen I.B.
        • Christensen S.
        • Chu C.T.
        • Pizzo S.V.
        • Enghild J.J.
        J. Biol. Chem. 1996; 271: 12937-12943
        • Valnickova Z.
        • Enghild J.J.
        J. Biol. Chem. 1998; 273: 27220-27224
        • Reeck G.R.
        • Neurath H.
        Biochemistry. 1972; 11: 3947-3955
        • Reverter D.
        • Garcia-Saez I.
        • Catasus L.
        • Vendrell J.
        • Coll M.
        • Aviles F.X.
        FEBS Lett. 1997; 420: 7-10
        • Lacko A.G.
        • Neurath H.
        Biochemistry. 1970; 9: 4680-4690
        • Garcia-Saez I.
        • Reverter D.
        • Vendrell J.
        • Aviles F.X.
        • Coll M.
        EMBO J. 1997; 16: 6906-6913
        • Barbosa Pereira P.J.
        • Segura-Martin S.
        • Oliva B.
        • Ferrer-Orta C.
        • Aviles F.X.
        • Coll M.
        • Gomis-Ruth F.X.
        • Vendrell J.
        J. Mol. Biol. 2002; 321: 537-547
        • Deutsch D.G.
        • Mertz E.T.
        Science. 1970; 170: 1095-1096
        • Molina M.A.
        • Aviles F.X.
        • Querol E.
        Gene. 1992; 116: 129-138
        • Ozols J.
        Methods Enzymol. 1990; 182: 587-601
        • Bury A.F.
        J. Chromatogr. 1981; 213: 491-500
        • Enghild J.J.
        • Valnickova Z.
        • Thogersen I.B.
        • Pizzo S.V.
        • Salvesen G.
        Biochem. J. 1993; 291: 933-938
        • Hendriks D.
        • Scharpe S.
        • van S
        • e M.
        Clin. Chem. 1985; 31: 1936-1939
        • Sakoda M.
        • Hiromi K.
        J. Biochem. (Tokyo). 1976; 80: 547-555
        • Kitagawa T.
        • Aikawa T.
        J. Biochem. (Tokyo). 1976; 79: 233-236
        • Liu F.T.
        • Zinnecker M.
        • Hamaoka T.
        • Katz D.H.
        Biochemistry. 1979; 18: 690-693
        • Arolas J.L.
        • Lorenzo J.
        • Rovira A.
        • Castella J.
        • Aviles F.X.
        • Sommerhoff C.P.
        J. Biol. Chem. 2005; 280: 3441-3448
        • Walker J.B.
        • Hughes B.
        • James I.
        • Haddock P.
        • Kluft C.
        • Bajzar L.
        J. Biol. Chem. 2003; 278: 8913-8921
        • Kowarik M.
        • Kung S.
        • Martoglio B.
        • Helenius A.
        Mol. Cell. 2002; 10: 769-778
        • Helenius A.
        • Aebi M.
        Annu. Rev. Biochem. 2004; 73: 1019-1049
        • Petrescu A.J.
        • Milac A.L.
        • Petrescu S.M.
        • Dwek R.A.
        • Wormald M.R.
        Glycobiology. 2004; 14: 103-114
        • Folk J.E.
        • Piez K.A.
        • Carroll W.R.
        • Gladner J.A.
        J. Biol. Chem. 1960; 235: 2272-2277
        • Plummer Jr., T.H.
        • Kimmel M.T.
        Anal. Biochem. 1980; 108: 348-353