Crystal structure of the human alpha-thrombin-haemadin complex: an exosite II-binding inhibitor

Structural requirements for the activation of human factor VIII by thrombin

The coagulation factors V (FV) and VIII (FVIII) are important at sites of vascular injury for the amplification of the clotting cascade. Natural variants of these factors frequently lead to severe bleeding disorders. To understand the mechanisms of activation of FVIII by thrombin, we used a bank of mutant thrombins to define residues important for its activation. From the initial screening of 53 mutant thrombins for the activation of human recombinant FVIII, we mapped thrombin mutants with 50% or less activity to anion-binding exosite-I (Lys21Ala, His66Ala, Lys65Ala, Arg68Ala, Arg70Ala, and Tyr71Ala) and anion-binding exosite-II (Arg98Ala), the Na(+)-binding site (Glu229Ala, Arg233Ala, Asp234Ala, and Asp193Ala/Lys196Ala), and the 50-insertion loop (Trp50Ala), which were similar to our results for the activation of FV. The role of these residues for cleavage at Arg372 and Arg1689 was investigated using plasma FVIII. Anion-binding exosite-I appears to be important for cleavage at both sites, whereas the anion-binding exosite-II residue Arg98Ala is important for cleavage at Arg372 alone. The Glu229Ala mutant, which contributes to the Na(+)-binding site, and the 50-insertion loop mutant W50A have severely impaired cleavage at Arg372 and Arg1689. This suggests that the integrity of the active site and the Na(+)-bound form of thrombin are important for its procoagulant activity against FVIII. Detailed mutagenic analysis of thrombin can assist in understanding the pathogenesis of bleeding disorders and may lead to the rational design of selective thrombin inhibitors.

Interaction of Kazal-type inhibitor domains with serine proteinases: biochemical and structural studies

The interaction of domains of the Kazal-type inhibitor protein dipetalin with the serine proteinases thrombin and trypsin is studied. The functional studies of the recombinantly expressed domains (Dip-I+II, Dip-I and Dip-II) allow the dissection of the thrombin inhibitory properties and the identification of Dip-I as a key contributor to thrombin/dipetalin complex stability and its inhibitory potency. Furthermore, Dip-I, but not Dip-II, forms a complex with trypsin resulting in an inhibition of the trypsin activity directed towards protein substrates. The high resolution NMR structure of the Dip-I domain is determined using multi-dimensional heteronuclear NMR spectroscopy. Dip-I exhibits the canonical Kazal-type fold with a central alpha-helix and a short two-stranded antiparallel beta-sheet. Molecular regions essential for inhibitor complex formation with thrombin and trypsin are identified. A comparison with molecular complexes of other Kazal-type thrombin and trypsin inhibitors by molecular modeling shows that the N-terminal segment of Dip-I fulfills the structural prerequisites for inhibitory interactions with either proteinase and explains the capacity of this single Kazal-type domain to interact with different proteinases.

Binding of exosite ligands to human thrombin. Re-evaluation of allosteric linkage between thrombin exosites I and II

The substrate specificity of thrombin is regulated by binding of macromolecular substrates and effectors to exosites I and II. Exosites I and II have been reported to be extremely linked allosterically, such that binding of a ligand to one exosite results in near-total loss of affinity for ligands at the alternative exosite, whereas other studies support the independence of the interactions. An array of fluorescent thrombin derivatives and fluorescein-labeled hirudin(54-65) ([5F]Hir(54-65)(SO(3)(-))) were used as probes in quantitative equilibrium binding studies to resolve whether the affinities of the exosite I-specific ligands, Hir(54-65)(SO(3)(-)) and fibrinogen, and of the exosite II-specific ligands, prothrombin fragment 2 and a monoclonal antibody, were affected by alternate exosite occupation. Hir(54-65)(SO(3)(-)) and fibrinogen bound to exosite I with dissociation constants of 16-28 nm and 5-7 microm, respectively, which were changed < or =2-fold by fragment 2 binding. Native thrombin and four thrombin derivatives labeled with different probes bound fragment 2 and the antibody with dissociation constants of 3-12 microm and 1.8 nm, respectively, unaffected by Hir(54-65)(SO(3)(-)). The results support a ternary complex binding model in which exosites I and II can be occupied simultaneously. The thrombin catalytic site senses individual and simultaneous binding of exosite I and II ligands differently, resulting in unique active site environments for each thrombin complex. The results indicate significant, ligand-specific allosteric coupling between thrombin exosites I and II and catalytic site perturbations but insignificant inter-exosite thermodynamic linkage.

Characterization of the residues involved in the human alpha-thrombin-haemadin complex: an exosite II-binding inhibitor

Haemadin is a 57-amino acid thrombin inhibitor from the land-living leech Haemadipsa sylvestris, whose structure has recently been determined in complex with human alpha-thrombin. Here we communicate the effect of ionic strength on the kinetics of the inhibition of human alpha-thrombin by haemadin, by using thrombin mutants modified in exosite II. Data analysis has allowed both the ionic and nonionic binding contributions to be ascertained, with the nonionic component being virtually the same for all of the thrombins that have been examined, while the ionic binding energy contributions varied from molecule to molecule. In the case of the native human alpha-thrombin-haemadin complex, ionic interactions contribute -17 kJ/mol to the Gibbs free energy of binding, this being the equivalent of up to six salt bridges. These salt bridges make up 20% of the total binding energy at zero ionic strength, and this has been attributed to the C-terminal tail alone. In addition, the contributions of the N-terminal and C-terminal regions of haemadin to its tight binding have been ascertained by using derivatives of both haemadin and thrombin. Limited proteolysis using formic acid produced haemadin cleaved between residues 40 and 41, removing the majority of the C-terminal tail. This truncated haemadin displayed a 20000-fold reduced affinity for thrombin, and was no longer a tight binding inhibitor. A form of thrombin in which the active site serine has been blocked by diisopropyl fluorophosphate binds to haemadin, but with a 72000-fold reduced affinity, indicating that the N-terminus is more important than the C-terminus for strong binding.

Structural and functional mapping of the thrombin domain involved in the binding to the platelet glycoprotein Ib

The activation of human platelets by alpha-thrombin is mediated in part by cleavage of the protease-activated receptor (PAR) 1 and 4 and by the glycoprotein (Ib)alpha, (Gp(Ib)alpha), which binds with high affinity to alpha-thrombin. Recent studies have shown that the thrombin domain referred to as heparin binding site (HBS) is involved in the interaction with the platelet Gp(Ib)alpha. The HBS is rich in basic amino acids. To identify the key amino acid residues involved in the binding to Gp(Ib)alpha, we have performed alanine scanning mutagenesis of the basic HBS R93, R97, R101, R233, K236, K240, R233/K236/Q239, as well as of the neutral Q239 residues, located in different regions of the domain. For comparison, mutation at R67 within the fibrinogen recognition site (FRS) of thrombin was performed as well. Solid-phase binding experiments showed that the Kd of thrombin-GpIb interaction was reduced 22-fold for R93A, 8-fold for R97A, 13-fold for R101A, 29-fold for R233A, 21-fold for K236A, 5-fold for K240A, and 31-fold for the triple mutant R233A/K236A/Q239A, while the Q239A and R67A forms did not show any significant affinity change. The platelet activating capacity of these mutants was evaluated as well. Using gel-filtered platelets, the EC50 value of thrombin-induced aggregation was from 5- to 13-fold higher in the HBS mutants than in the WT form, and was linearly and positively correlated with the corresponding Kd values pertaining to thrombin binding to GpIb. Measurements of PAR-1 hydrolysis on the platelet membrane showed that the HBS mutants R233A, R101A, R93A, K236A, and R233/K236/Q239 forms had a reduction of the apparent kcat/Km value. These results are a consequence of a defective binding to GpIb, which is known to optimize the interaction with PAR-1 in situ. A confirm of this hypothesis came from the demonstration that the kcat/Km value pertaining to the hydrolysis by the HBS-mutated thrombins of the synthetic PAR-1 38-60 peptide in solution was similar to that one obtained with the WT form. In conclusion, these experiments provide a structural and functional mapping of the thrombin HBS subregions involved in the binding to the platelet Gp(Ib)alpha and in the cell activation.