The active site of antithrombin. Release of the same proteolytically cleaved form of the inhibitor from complexes with factor IXa, factor Xa, and thrombin

Monoclonal antibodies against antithrombin III. Identification of their epitopes and effects on antithrombin III activities

Four monoclonal antibodies with distinct epitopes were prepared against antithrombin III. None of them is directed against the heparin-binding region nor the active site, yet two mAb namely A36 and B108, interfere with antithrombin III inhibition of thrombin. The epitope of monoclonal antibody A36 is located within amino acid residues 1-393, at a site different from the active site since it recognizes antithrombin III and antithrombin-III-thrombin complexes with the same affinity. A36 partially prevents the intrinsic antithrombin III activity and has no effect on the heparin-enhanced antithrombin III activity when added to the antithrombin-III--heparin complex. If A36 is first reacted with antithrombin III and then heparin is added to the reaction mixture, A36 fixes the conformation of antithrombin III so that heparin binds to antithrombin III, but is not able to induce the conformational change in the antithrombin III molecule required for the enhanced activity. The epitope for monoclonal antibody B108 is located within residues 282-393, close to the active site. It does not recognize antithrombin-III-thrombin complexes by solid-phase radioimmunoassay. Its binding to antithrombin III induces a conformational change that enhances antithrombin III activity in a manner that resembles the heparin effect, but its effect is additive to the heparin effect, since when it was added to a reaction mixture which contained a saturating amount of heparin, inhibition of thrombin was enhanced. The epitope for monoclonal antibody A5 is located within residues 1-393, and its recognition of antithrombin III or antithrombin-III-thrombin is strongly dependent on the integrity of the disulfide bonds. A5 has no effect on antithrombin III activities. The epitope for monoclonal antibody A10 is well defined within a narrow range of 55 amino acid residues, 339-393, on the antithrombin III molecule, close to the active site, yet it has no effect on antithrombin III inhibitory activity. These monoclonal antibodies may be developed for various diagnostic or clinical purposes and offer a powerful tool for studying the conformational changes and structure/activity relationships in the antithrombin III molecule.

Antithrombin III activity (residual thrombin activity) in plasma from non-medicated or heparinized horses

Two synthetic substrate assays (fluorometric and chromogenic) were used to measure antithrombin-III (AT-III) activity (residual thrombin activity) in non-medicated and heparin (sodium) treated horses. In 18 non-medicated horses the fluorometric substrate assay (FSA) values were similar to previous reports but they reflected inconsistent trends and larger deviations in the heparin-treated groups (Group 2: 40 and 100 U/kg IV, n = 6; Group 3: 240 U/kg IV, n = 5; Group 4: 80 U/kg IV followed by 160 U/kg SC, n = 8) when compared to the chromogenic substrate assay (CSA) values. The CSA values for the 18 non-medicated horses indicated a higher AT-III activity (lower residual thrombin activity) than the FSA. AT-III activity was quantified in 18 non-medicated horses (29 mg/dl) and compared well with values for humans (30 mg/dl) and dogs (40 mg/dl). Plasma heparin concentrations, determined by the FSA, correlated well with the 'therapeutic range' (1.5 fold to 2.5 fold prolongation of the activated partial thromboplastin time (APTT) normal value) and values reported for humans. The effect of heparin therapy on AT-III activity in four treatment regimens was evaluated. AT-III activity was not significantly affected (with one exception) by a single dose of intravenous (IV) heparin (40 and 100 U/kg) nor by repeated subcutaneous (SC) injections of heparin (240 U/kg). A transient increase in residual thrombin activity was measured 12 h after an intravenous (80 U/kg) injection of heparin. Large doses of heparin (80 U/kg IV followed by 160 U/kg SC) given every 12 h produced a progressive prolongation of the APTT. In this group the APTT remained prolonged 48 h after the last treatment.

Antithrombin Sheffield: amino acid substitution at the reactive site (Arg393 to His) causing thrombosis

A Sheffield family with a predisposition towards thrombosis has been shown to have a functional abnormality of antithrombin. The abnormality was detected as reduced heparin cofactor activity, with normal antigenic levels of antithrombin. Crossed immunoelectrophoresis performed in the absence and presence of heparin was normal. The antithrombin was isolated by heparin Sepharose affinity chromatography. It had normal mobility on SDS polyacrylamide gel electrophoresis. However, the second order rate constant of inhibition of thrombin was about half that of normal, and this was compatible with a heterozygous abnormality involving the reactive site. The antithrombin was further purified by chromatography on thrombin-Sepharose (to remove the normal component), reduced, S-carboxymethylated and fragmented with cyanogen bromide. A pool containing the reactive site region was digested with trypsin and the molecular size of peptides generated determined by fast atom bombardment mass spectrometry. The two peptides adjacent to the Arg393-Ser394 bond of mass 2290 and 700 were almost absent from the mass spectrum, but an additional peptide of mass 2952 was present. Subdigestion with V8 protease reduced the mass of this peptide to 1748. These peptides generated by trypsin and V8 protease were almost identical to those obtained when another variant, antithrombin Glasgow, was treated in the same way (Erdjument et al, 1988). It is concluded that the molecular abnormality of antithrombin Sheffield is identical to that of antithrombin Glasgow, Arg393 to His.

A complete quantitative N-terminal analysis method

A complete quantitative N-terminal analysis (QNA) technique based on the application of dimethylaminoazobenzene isothiocyanate is described. The method allows recovery of all free N-terminal amino acids, including Asn, Gln, Trp, Ser, and Thr in quantitative yield. N-Termini of polypeptides as little as 5 pmol can be reliably and reproducibly determined by this method. This QNA method is useful in many aspects of protein structure analysis. (a) QNA is useful in assessing the purity, identity, and quantity of a polypeptide preparation. It has also been applied in our lab as a routine guarding step to prevent impure or ill-characterized samples from occupying the space of the gas-phase sequenator. (b) QNA of a trypsinized protein generates a miniaturized amino acid composition which is useful both in characterizing the identity of a protein and in comparing the homology of structurally related proteins. (c) QNA can be used to follow the pathway and preferential cleavage sites of limited proteolysis. (d) QNA is useful in characterizing selectively modified Lys and Arg residues. The details of this QNA method and the results of its applications are presented here.

Characterization of recombinant human alpha 2-antiplasmin and of mutants obtained by site-directed mutagenesis of the reactive site

Human alpha 2-antiplasmin (alpha 2AP) has been expressed in Chinese hamster ovary cells and purified from conditioned media. The recombinant protein (r alpha 2AP) is immunologically identical with natural alpha 2AP and indistinguishable with respect to plasmin(ogen) binding properties. Second-order rate constants (k1) for the interaction of alpha 2AP and r alpha 2AP with plasmin are both (1-2) X 10(7) M-1 s-1. In order to examine the effects of alterations within the reactive site of alpha 2AP, deletions of the P1 residue Arg-364 (r alpha 2AP-delta Arg364) or the P'1 residue Met-365 (r alpha 2AP-delta Met365) were introduced by in vitro site-directed mutagenesis. r alpha 2AP-delta Met365 completely retains its ability to inhibit both plasmin and trypsin, indicating that alpha 2AP has no absolute requirement for Met in the P'1 position. Unexpectedly, no increase in antithrombin activity was observed. r alpha 2AP-delta Arg364 has lost the ability to inhibit plasmin, trypsin, and thrombin, but unlike the wild-type protein, this variant is an effective elastase inhibitor (k1 = 1.5 X 10(5) M-1 s-1).

Hormone-like activity of human thrombin

Recently, we have shown that thrombin is a chemotaxin and growth-promoting agent for cells of the mononuclear phagocytic lineage. These activities are independent of thrombin's enzymatic activity. Unlike other chemotactic factors, thrombin is specific for monocytes and does not attract granulocytes. To further explore the cellular specificity we have used a human leukemia cell line HL-60 that is capable of in vitro differentiation toward either monocytes (HL-60/mono) following incubation with 1,25(OH)2D3, or granulocytes (HL-60/gran) following incubation with DMSO. In contrast to undifferentiated HL-60 cells or HL-60/gran, we find that HL-60/mono respond chemotactically to intact human alpha-thrombin, esterolytically inactive iPR2P-alpha-thrombin, and the thrombin-derived peptide CB67-129, previously shown to contain the thrombin chemotactic exosite. In addition, thrombin induces in HL-60/mono association of actin with the cytoskeleton and causes an increase in levels of free cytosolic Ca2+. These phenomena are well characterized as early events occurring concomitant with directed cell movement associated with exposure to chemotactic agents such as FMLP. Furthermore, in contrast to fibroblasts, both iPR2P-alpha-thrombin and the thrombin chemotactic peptide CB67-129 evoke dose-dependent [3H]TdR incorporation, protein synthesis, and cell replication in growth-arrested J-744 cells, a murine macrophage-like cell line. Limited tryptic digests of CB67-129 lose chemotactic activity but retain full mitogenic activity, demonstrating that as with PDGF, the sites on CB67-129 required for chemotaxis and mitogenesis are clearly dissociable. The mitogenic effects of the CB67-129 digest can be mimicked by a synthetic tetradecapeptide analogue of CB67-129 (residues 367-380) that includes the loop B insertion sequence, previously shown to be critical for thrombin's chemotactic effects. From these data, it is apparent that the loop B insertion is critical for thrombin's nonenzymic biological effects on cells, but additional sites are required for stimulation of cell movement.

Pasteurisation induced changes in an antithrombin III concentrate

The changes that take place when a therapeutic antithrombin III (AT III) concentrate is heated in the presence of citrate ion have been assessed. There is some loss of heparin cofactor antithrombin activity and of heparin binding ability. Protein aggregates are also formed during heating. These aggregates are not AT III but impurities in the concentrate.

Denaturation behavior of antithrombin in guanidinium chloride. Irreversibility of unfolding caused by aggregation

The structural stability of the protease inhibitor antithrombin from bovine plasma was examined as a function of the concentration of guanidinium chloride (GdmCl). A biphasic unfolding curve at pH 7.4, with midpoints for the two phases at 0.8 and 2.8 M GdmCl, was measured by far-ultraviolet circular dichroism. Spectroscopic and hydrodynamic analyses suggest that the intermediate state which exists at 1.5 M GdmCl involves a partial unfolding of the antithrombin molecule that exposes regions of the polypeptide chain through which slow, intermolecular association subsequently takes place. The partially unfolded molecule can be reversed to its fully functional state only before the aggregation occurs. Upon return of the aggregated state to dilute buffer, the partially unfolded antithrombin remains aggregated and does not regain the spectroscopic properties, thrombin-inhibitory activity, or heparin affinity of the native inhibitor. This behavior indicates that the loss of the functional properties of the proteins is caused by the macromolecular association. Comparative experiments gave similar results for the human inhibitor. Analyses of bovine antithrombin in 6 M GdmCl indicated that the second transition reflects the total unfolding of the protein to a disulfide-cross-linked random coil. This transition is spectroscopically reversible; however, on further reversal to dilute buffer, the molecules apparently are trapped in the partially unfolded, aggregated, intermediate state. The results are consistent with the existence of two separate domains in antithrombin which unfold at different concentrations of GdmCl but do not support the contention that the thrombin-binding and heparin-binding regions of the protein are located in different domains [Villanueva, G. B., & Allen, N. (1983) J. Biol. Chem. 258, 14048-14053].

Studies of the heterogeneity of antithrombin III concentrates

The biological activities of antithrombin III (At III) concentrates, prepared by several manufacturers for clinical use, have been compared by three assay methods, and their heparin-binding properties studied by crossed immunoelectrophoresis and heparin affinity chromatography. Concentrates from two of the four manufacturers showed discrepancies between assay methods, with concentrations by heparin co-factor assays significantly lower than those by immunological and progressive antithrombin methods. Heterogeneity was also found by heparin binding studies, with about half the total At III antigen in these concentrates being unable to bind to heparin. These results confirm previous findings of heterogeneity in At III concentrates and show that some concentrates contain substantial amounts of altered At III molecules in which the heparin-binding site has been denatured but the thrombin-neutralizing site left largely intact.

Localization of the disulfide bond in human antithrombin III required for heparin-accelerated thrombin inactivation

Heparin accelerates the rate of inhibition of thrombin by antithrombin III. Reduction of one of the three antithrombin disulfide bonds with dithiothreitol under mild conditions abolishes this rate-enhancing effect without affecting the rate of reaction in the absence of heparin. Alkylation of mildly reduced antithrombin III with [3H]iodacetic acid followed by digestion with cyanogen bromide yielded two major labeled peptides. The smaller peptide, containing Cys-422, was identified as extending from Gly-414 to the C-terminus, Lys-424. Our data are consistent with the larger labeled peptide being the one extending from Glu-104 to Met-243 and containing Cys-239. Cys-422 has been shown by others to be linked to Cys-239. These data indicate that the sensitive disulfide bond in antithrombin III extends between Cys-239 and Cys-422; the site at which thrombin cleaves the antithrombin III is between these two half-cystines.

The action of thrombin on peptide p-nitroanilide substrates. Substrate selectivity and examination of hydrolysis under different reaction conditions

Kinetic parameters for the action of bovine alpha-thrombin on 24 commercially available peptide p-nitroanilides have been determined. The selectivity constant, kcat/Km, ranges from 3.3 X 10(1) to 1.1 X 10(8) M-1 X S-1 for the poorest and the best substrates, respectively. The best substrates for thrombin were identified as those with arginine in the P1 position, proline or a proline homolog in the P2 position, and an apolar amino acid in the P3 position. Quantitative distinction between lysine and arginine in the P1 position and other amino acids in the P2-P4 positions of the substrate is reported from the changes in the kinetic parameters for substrates differing in only a single amino acid in these positions. Effects of NaCl, CaCl2 and poly(ethylene glycol) concentrations, pH and temperature on the action of thrombin on selected substrates have been assessed. A source of large systematic error in thrombin concentration estimates was identified as resulting from adsorption losses. These losses were eliminated by inclusion of poly(ethylene glycol) in dilution and reaction buffers.