Title: The structure of thrombin, a chameleon-like proteinase
Abstract: α-Thrombin, the ultimate proteinase in the blood coagulation system, acts on diverse substrates and regulates a number of processes related to hemostasis and thrombosis. The first crystal structure of α-thrombin determined 15 years ago helped to explain a number of thrombin properties on a structural basis, but also stimulated the elucidation of many other thrombin-related structures, which reflect nicely the many-sidedness of this multifunctional enzyme. Up to now, 180 thrombin-related crystal structures have been deposited in the PDB [1Berman H.M. Westbrook J. Feng Z. Gilliland G. Bhat T.N. Weissig H. Shindyalov I.N. Bourne P.E. The Protein Data Bank.Nucleic Acids Res. 2000; 28: 235-42Crossref PubMed Scopus (29029) Google Scholar], and several hundred more must be stored in pharmaceutical company archives. In this historical review, I will try to show, how crystal structures contributed to the current understanding of hemostasis and thrombosis, and will give a quite personal view of our contribution to this remarkable development. The blood clotting system consists essentially of trypsin-like serine proteinase zymogens, which activate each other in succession, eventually leading to the burst-like release of active α-thrombin [2Davie E.W. A brief historical review of the waterfall/cascade of blood coagulation.J Biol Chem. 2003; 278: 50819-32Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar, 3Hougie C. The waterfall-cascade and autoprothrombin hypotheses of blood coagulation: personal reflections from an observer.J Thromb Haemost. 2004; 2: 1225-33Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar]. Because of their important physiological roles and their structural and functional versatility, I always aimed at doing some crystallographic work on clotting proteinases. This dream came true in the late 1980s and 1990s, when I, together with talented co-workers and excellent colleagues from the Martinsried lab but also from all over the world, had the privilege to determine the first structures of thrombin and some other coagulation factors [4Stubbs M.T. Bode W. A player of many parts: the spotlight falls on thrombin's structure.Thromb. Res. 1993; 69: 1-58Abstract Full Text PDF PubMed Scopus (452) Google Scholar, 5Stubbs M.T. Bode W. Coagulation factors and their inhibitors.Curr Opin Struct Biol. 1994; 4: 823-32Crossref PubMed Scopus (63) Google Scholar, 6Stubbs M.T. Bode W. The clot thickens: clues provided by thrombin structure.Trends Biochem Sci. 1995; 20: 23-8Abstract Full Text PDF PubMed Scopus (180) Google Scholar, 7Bode W. Brandstetter H. Mather T. Stubbs M.T. Comparative analysis of haemostatic proteinases: structural aspects of thrombin, factor Xa, factor IXa and protein C.Thromb Haemost. 1997; 78: 501-11Crossref PubMed Scopus (99) Google Scholar], in this way contributing to a better understanding of hemostasis and thrombosis. In 1972, after having done my PhD work on biophysical/biochemical aspects of bacterial flagella, I joined Robert Huber at the newly founded Max-Planck-Institute of Biochemistry in Martinsried to learn to apply protein crystallography. The first structure elucidation I became involved in was that of the trypsin-basic pancreatic trypsin inhibitor complex [8Huber R. Kukla D. Bode W. Schwager P. Bartels K. Deisenhofer J. Steigemann W. Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor.J Mol Biol. 1974; 89: 73-101Crossref PubMed Scopus (473) Google Scholar]. This analysis revealed how proteinases interact not only with inhibitors but also with protein substrates, and provided the structural data allowing Michael Laskowski/Purdue to define his ‘standard-mechanism’ concept [9Laskowski Jr, M. Kato I. Protein inhibitors of proteinases.Annu Rev Biochem. 1980; 49: 593-626Crossref PubMed Scopus (2015) Google Scholar] about the interaction of ‘canonical’ proteinase inhibitors with trypsin-like serine proteinases (see [10Bode W. Huber R. Natural protein proteinase inhibitors and their interaction with proteinases.Eur J Biochem. 1992; 204: 433-51Crossref PubMed Scopus (1029) Google Scholar]). Immediately after, I solved the free and inhibited bovine trypsin structures, providing the first refined structure of any proteinase, and allowing accurate definition of the active-site geometry and identification of the stabilizing calcium binding site that trypsin shares with other pancreatic and coagulation enzymes [11Bode W. Schwager P. The refined crystal structure of bovine beta-trypsin at 1.8 A resolution.J Mol Biol. 1975; 98: 693-717Crossref PubMed Scopus (427) Google Scholar]. Shortly after, we also solved and refined the structure of the zymogen trypsinogen, and formulated the concept of the transformable ‘activation domain’ [12Bode W. Fehlhammer H. Huber R. Crystal structure of bovine trypsinogen at 1.8 Å resolution.J Mol Biol. 1976; 106: 325-35Crossref PubMed Scopus (91) Google Scholar, 13Bode W. Schwager P. Huber R. The transition of bovine trypsinogen to a trypsin-like state upon strong ligand binding. The refined crystal structures of the bovine trypsinogen-pancreatic trypsin inhibitor complex and of its ternary complex with Ile-Val at 1.9 Å resolution.J Mol Biol. 1978; 118: 99-112Crossref PubMed Scopus (256) Google Scholar, 14Huber R. Bode W. Structural basis of the activation and action of trypsin.Acc Chem Res. 1978; 11: 114-22Crossref Scopus (635) Google Scholar, 15Bode W. The transition of bovine trypsinogen to a trypsin-like state upon strong ligand binding. II. The binding of the pancreatic trypsin inhibitor and of isoleucine-valine and of sequentially related peptides to trypsinogen and to p-guanidinobenzoate-trypsinogen.J Mol Biol. 1979; 127: 357-74Crossref PubMed Scopus (171) Google Scholar]. This highly cooperative subdomain of trypsinogen (and related serine proteinases) can alternatively exist in a more flexible, inactive, as well as in a structured, active state. Upon activation cleavage, a new Ile16Try-Val-Gly-Gly-like amino-terminus (chymotrypsinogen nomenclature) is created, which inserts into the preformed Ile16-pocket forming an internal salt-bridge with Asp194. Rotation of the Asp194 side chain triggers the generation of a functional active site and a correctly shaped substrate binding region. In 1976 I showed that under special conditions Ile-Val-like dipeptides can stabilize the active enzyme form of trypsinogen without activation cleavage [16Bode W. Huber R. Induction of the bovine trypsinogen-trypsin transition by peptides sequentially similar to the N-terminus of trypsin.FEBS-Let. 1976; 68: 231-6Crossref PubMed Scopus (124) Google Scholar]. These experiments indicated that the zymogen activation mechanism was fundamentally conformational and that zymogens are in, typically highly unfavorable, equilibrium with the corresponding active forms. When Agnes Henschen determined the first three Ile1SK-Ala-Gly amino-terminal residues of streptokinase (SK), I speculated that this non-enzymatic activator might activate plasminogen through induction of a conformational change by inserting its Ile1SK-Ala amino-terminus into the corresponding Val16-pocket of plasminogen, a mechanism that we coined ‘molecular sexuality’ [16Bode W. Huber R. Induction of the bovine trypsinogen-trypsin transition by peptides sequentially similar to the N-terminus of trypsin.FEBS-Let. 1976; 68: 231-6Crossref PubMed Scopus (124) Google Scholar]. I speculated further that a similar triggering mechanism might be utilized by some (at that time unknown) proteinase-activatable receptors [15Bode W. The transition of bovine trypsinogen to a trypsin-like state upon strong ligand binding. II. The binding of the pancreatic trypsin inhibitor and of isoleucine-valine and of sequentially related peptides to trypsinogen and to p-guanidinobenzoate-trypsinogen.J Mol Biol. 1979; 127: 357-74Crossref PubMed Scopus (171) Google Scholar, 16Bode W. Huber R. Induction of the bovine trypsinogen-trypsin transition by peptides sequentially similar to the N-terminus of trypsin.FEBS-Let. 1976; 68: 231-6Crossref PubMed Scopus (124) Google Scholar], an idea more recently verified by Shaun Coughlin by his ‘tethered-ligand’ mechanism of the thrombin receptor PAR-1 [17Vu T.K. Wheaton V.I. Hung D.T. Charo I. Coughlin S.R. Domains specifying thrombin-receptor interaction.Nature. 1991; 353: 674-7Crossref PubMed Scopus (519) Google Scholar]. After having solved some more structures of serine proteinases, thrombin came into my scientific focus. At that time, three-dimensional models of the thrombin B-chain had been proposed based on trypsin and chymotrypsin structures [18Magnusson S. Peterson T.E. Sottrup-Jensen L. Claeys H. Complete primary structure of prothrombin: Isolation, structure and reactivity of ten carboxylated glutamic acid residues and regulation of prothrombin activation by thrombin.in: Reich E Rifkin DB Shaw E Proteases and Biological Control. Cold Spring Harbour, Cold Spring Harbour Lab., 1975: 123-49Google Scholar]. These models could not satisfactorily explain the particular specificity features of thrombin, however, asking strongly for an authentic thrombin structure. Although single crystals of bovine [19Tsernoglou D. Walz D.A. McCoy L.E. Seegers W.H. An x-ray crystallographic study of thrombin.J Biol Chem. 1974; 249: 999Abstract Full Text PDF PubMed Google Scholar] and human thrombin [20McKay D.B. Kay L.M. Stroud R.M. Lundblad RL Fenton JW Mann KG Chemistry and Biology of Thrombin. Ann Arbor Science Publ., 1977: 113-21Google Scholar] had been described, they had not lead to structures. In 1985, I was invited to the third Joint Meeting of the Biochemical Societies of France, Germany and Switzerland in Basel, to talk about my findings of ‘Non-enzymatic (Ile-Val-induced) Activation of Trypsin-like Serine Proteinases’ [21Bode W. Nonenzymatic activation of trypsin-like serine proteinases.Biol Chem Hoppe-Seyler. 1985; 366: 767Google Scholar]. At that occasion, Elliot Shaw acquainted me with two talented young biochemists, Stuart Stone and Jan Hofsteenge, who worked on the kinetics of α-, β- and γ-thrombin and their differential interactions with substrates, cofactors, and inhibitors. Encouraged by my recent results obtained with human chloromethylketone-blocked leukocyte elastase [22Wei A. Mayr I. Bode W. The refined 2.3 Å crystal structure of human leukocyte elastase in a complex with a valine chloromethyl ketone inhibitor.FEBS Lett. 1988; 234: 367-73Crossref PubMed Scopus (81) Google Scholar], we attempted crystallization with Stone's human D-Phe-Pro-Arg-chloromethylketone (PPACK)-α-thrombin [23Bode W. Mayr I. Baumann U. Huber R. Stone S.R. Hofsteenge J. The refined 1.9 Å crystal structure of human alpha-thrombin: interaction with D-Phe-Pro-Arg chloromethylketone and significance of the Tyr-Pro-Pro-Trp insertion segment.EMBO J. 1989; 8: 3467-75Crossref PubMed Scopus (888) Google Scholar]. After a while, we obtained a macroscopically twinned PPACK-thrombin crystal, whose separated fragments (Fig. 1) diffracted to beyond 1.9 Å resolution. By merging X-ray data obtained with rotation camera films and a television area detector, I arrived at a reasonable 1.9 Å X-ray data set [24Bode W. Turk D. Karshikov A. The refined 1.9-Å X-ray crystal structure of D-Phe-Pro-Arg chloromethylketone-inhibited human alpha-thrombin: structure analysis, overall structure, electrostatic properties, detailed active-site geometry, and structure-function relationships.Protein Sci. 1992; 1: 426-7127Crossref PubMed Scopus (674) Google Scholar]. Because others had failed to solve the thrombin phase problem by replacement methods before, I anxiously explored various search models to determine initial orientation and translation parameters using house-made search functions. These Patterson searches were surprisingly unequivocal, immediately yielding the correct solution. I still remember the night, when I continued to stay at the institute to have a look at my first 3 Å Sim-weighted 2Fobs-Fcalc electron density map of PPACK-thrombin calculated by virtue of a correctly placed trypsin/chymotrypsin chimera, using Alwyn Jones’ Martinsried-originating PSFRODO version on an Evans-and-Sutherland interactive display system. This density immediately showed most of the characteristic features of thrombin [23Bode W. Mayr I. Baumann U. Huber R. Stone S.R. Hofsteenge J. The refined 1.9 Å crystal structure of human alpha-thrombin: interaction with D-Phe-Pro-Arg chloromethylketone and significance of the Tyr-Pro-Pro-Trp insertion segment.EMBO J. 1989; 8: 3467-75Crossref PubMed Scopus (888) Google Scholar] including the canyon-like active-site cleft narrowed by thrombin's uniquely juxtaposed 60- and 148-insertion loops, the deeply buried active-site residues, and the D-Phe-P3, L-Pro-P2 and L-Arg-P1 side chains of PPACK slotting into the characteristic S4-groove, S2-cavity, and the deep polar S1-specificity pocket, respectively (Fig. 2). In the course of further refinement [24Bode W. Turk D. Karshikov A. The refined 1.9-Å X-ray crystal structure of D-Phe-Pro-Arg chloromethylketone-inhibited human alpha-thrombin: structure analysis, overall structure, electrostatic properties, detailed active-site geometry, and structure-function relationships.Protein Sci. 1992; 1: 426-7127Crossref PubMed Scopus (674) Google Scholar] other fascinating details of the thrombin molecule became obvious, in particular two large positively charged surface patches, one extending to the right-hand side where the active-site cleft levels off to the surface, called the fibrinogen recognition exosite or anion binding exosite-I, and the other one towards the top, called anion binding exosite-II (see Fig. 2).Figure 2Solid Connolly surface representation of α-thrombin covalently substituted at the active-site residues Ser195 and His57 (chymotrypsinogen nomenclature) by the D-Phe-Pro-Arg-chloromethyl ketone inhibitor (stick model, colored according to the standard atom colors [24Bode W. Turk D. Karshikov A. The refined 1.9-Å X-ray crystal structure of D-Phe-Pro-Arg chloromethylketone-inhibited human alpha-thrombin: structure analysis, overall structure, electrostatic properties, detailed active-site geometry, and structure-function relationships.Protein Sci. 1992; 1: 426-7127Crossref PubMed Scopus (674) Google Scholar]). The thrombin molecule is shown in the proteinase standard orientation, i.e. with the active-site cleft horizontally extending across the molecular surface such that peptidic chains to be cleaved usually run from left (amino-terminus) to right (carboxy-terminus). The coloration of the thrombin surface is made according to the electrostatic surface potential, extending from −20 e/kT (intense red) to +20 e/kT (intense blue).View Large Image Figure ViewerDownload Hi-res image Download (PPT) It was immediately clear that our thrombin structure provided an excellent tool for the rational design of small molecule antithrombotics. In 1989, only a few compounds with high affinity and specificity for thrombin had been developed, such as the Arg-chloromethyl ketones [25Kettner C. Shaw E. D-Phe-Pro-ArgCH2C1-A selective affinity label for thrombin.Thromb Res. 1979; 14: 969-73Abstract Full Text PDF PubMed Scopus (296) Google Scholar], mechanism-based isocoumarin inhibitors [26Kam C.M. Fujikawa K. Powers J.C. Mechanism-based isocoumarin inhibitors for trypsin and blood coagulation serine proteases: new anticoagulants.Biochemistry. 1988; 27: 2547-57Crossref PubMed Scopus (84) Google Scholar], tripeptidyl aldehyde analogs of PPACK [27Bajusz S. Barabas E. Tolnay P. Szell E. Bagdy D. Inhibition of thrombin and trypsin by tripeptide aldehydes.Int J Pept Protein Res. 1978; 12: 217-21Crossref PubMed Scopus (122) Google Scholar], stereoisomeric arginine derivatives including the antithrombotic argatroban [28Kikumoto R. Tamao Y. Tezuka T. Tonomura S. Hara H. Ninomiya K. Hijikata A. Okamoto S. Selective inhibition of thrombin by (2R,4R)-4-methyl-1-[N2-[(3-methyl-1,2,3,4-tetrahydro-8-quinolinyl) sulfonyl]-l-arginyl)]-2-piperidinecarboxylic acid.Biochemistry. 1984; 23: 85-90Crossref PubMed Scopus (304) Google Scholar], and a number of benzamidine derivatives developed by the Marquardt/Stürzebecher/Wagner group [29Stürzebecher J. Markwardt F. Voigt B. Wagner G. Walsmann P. Cyclic amides of N-alpha-arylsulfonylaminoacylated 4-amidinophenylalanine – tight binding inhibitors of thrombin.Thromb Res. 1983; 29: 635-42Abstract Full Text PDF PubMed Scopus (122) Google Scholar] in Erfurt/Leipzig, Eastern Germany, in those days infinitely away for West Germans. Early in 1989, Jörg Stürzebecher, on his officially allowed trip to the Upper Bavaria/Penzberg site of Boehringer-Mannheim, stopped over unofficially in Martinsried. To determine the binding modes and the correct chirality of his stereo-isomeric inhibitors, we first soaked his NAPAP and related benzamidine compounds into bovine trypsin crystals, due to lack of thrombin crystals with an accessible substrate binding region [30Bode W. Turk D. Stürzebecher J. Geometry of binding of the benzamidine- and arginine-based inhibitors NAPAP and MQPA to human alpha-thrombin. X-ray crystallographic determination of the NAPAP-trypsin complex and modeling of NAPAP-thrombin and MQPA-thrombin.Eur J Biochem. 1990; 193: 175-82Crossref PubMed Scopus (131) Google Scholar]. Superposition on the thrombin structure revealed directly the binding mode of NAPAP and argatroban [31Matsuzaki T. Sasaki C. Okumura C. Umeyama H. X-ray analysis of a thrombin inhibitor-trypsin complex.J Biochem (Tokyo). 1989; 105: 949-52Crossref PubMed Scopus (30) Google Scholar] in the thrombin cleft, and explained nicely the high affinity and preference of both compounds for thrombin (Fig. 3). These initial thrombin publications spawned and intensified enormous efforts in the pharmaceutical industry worldwide, aimed at rationally designing and elaborating specific direct thrombin inhibitors, which could help to prevent venous thromboembolic events (see e.g. [32Steinmetzer T. Stürzebecher J. Progress in the development of synthetic thrombin inhibitors as new orally active anticoagulants.Curr Med Chem. 2004; 11: 2297-321Crossref PubMed Scopus (66) Google Scholar]). Considering the amount of effort, it is surprising, however, that only very recently has the first orally available direct thrombin inhibitor, the prodrug Ximelagatran, been approved for (restricted) clinical use [33Gustafsson D. Oral direct thrombin inhibitors in clinical development.J Intern Med. 2003; 254: 322-34Crossref PubMed Scopus (86) Google Scholar]. At that time, Phil Martin from Wayne State/Detroit came along with well-diffracting orthorhombic crystals of benzamidine-inhibited bovine ɛ-thrombin, which allowed us to determine thrombin interactions with synthetic inhibitors directly. Our studies essentially confirmed the thrombin binding geometries already derived from our trypsin work, and allowed (as also [34Banner D.W. Hadvary P. Crystallographic analysis at 3.0-Å resolution of the binding to human thrombin of four active site-directed inhibitors.J Biol Chem. 1991; 266: 20085-93Abstract Full Text PDF PubMed Google Scholar]) formulation of some general rules governing the inhibitor interactions with thrombin [35Brandstetter H. Turk D. Hoeffken H.W. Grosse D. Stürzebecher J. Martin P.D. Edwards B.F. Bode W. Refined 2.3 Å X-ray crystal structure of bovine thrombin complexes formed with the benzamidine and arginine-based thrombin inhibitors NAPAP, 4-TAPAP and MQPA. A starting point for improving antithrombotics.J Mol Biol. 1992; 226: 1085-99Crossref PubMed Scopus (192) Google Scholar]. This crystal form served for binding studies of lead compounds, until Alex Tulinsky/East Lansing detected superior monoclinic crystals consisting of human α-thrombin in a complex with hirugen [derived from the C-terminal tail of hirudin (Fig. 4A), see also Fig. 5], which likewise were suitable for soaking-in larger synthetic inhibitors [36Skrzypczak-Jankun E. Carperos V.E. Ravichandran K.G. Tulinsky A. Westbrook M. Maraganore J.M. Structure of the hirugen and hirulog 1 complexes of alpha-thrombin.J Mol Biol. 1991; 221: 1379-93Crossref PubMed Scopus (269) Google Scholar].Figure 5Solid surface representation of the ternary complex [47Stubbs M.T. Oschkinat H. Mayr I. Huber R. Angliker H. Stone S.R. Bode W. The interaction of thrombin with fibrinogen. A structural basis for its specificity.Eur J Biochem. 1992; 206: 187-95Crossref PubMed Scopus (210) Google Scholar] of human α-thrombin in complex with an FPA-derived decamere (yellow stick model, with blue and red N and O atoms) and hirugen (white stick model, with blue and red N and O atoms). The solid surface calculated for the thrombin moiety alone is colored according to the electrostatic surface potential.View Large Image Figure ViewerDownload Hi-res image Download (PPT) The late 1980s showed a strong scientific and industrial interest in protein inhibitors of serine proteinases. In Martinsried, we laid with the structure of cleaved α1-proteinase inhibitor the structural fundament for serpin inhibitors [37Löbermann H. Lottspeich F. Bode W. Huber R. Interaction of human alpha 1-proteinase inhibitor with chymotrypsinogen A and crystallization of a proteolytically modified alpha 1-proteinase inhibitor.Hoppe Seylers Z Physiol Chem. 1982; 363: 1377-88Crossref PubMed Scopus (26) Google Scholar, 38Loebermann H. Tokuoka R. Deisenhofer J. Huber R. Human alpha 1-proteinase inhibitor. Crystal structure analysis of two crystal modifications, molecular model and preliminary analysis of the implications for function.J Mol Biol. 1984; 177: 531-57Crossref PubMed Scopus (668) Google Scholar], and studied the crystal structures of a number of ‘canonical’ inhibitors and their complexes [10Bode W. Huber R. Natural protein proteinase inhibitors and their interaction with proteinases.Eur J Biochem. 1992; 204: 433-51Crossref PubMed Scopus (1029) Google Scholar]. In agreement with Michael Laskowski's pioneering kinetic, thermodynamic, and biochemical data, these structures showed that these canonical protein inhibitors interact with their target enzymes like tight-binding, slow-turnover substrates. In those days, hirudin, a very potent and selective thrombin inhibitor derived from the European medicinal leech Hirudo medicinalis, was in the focus of pharmaceutical companies as a potentially useful anticoagulant. Biochemical results suggested that the hirudin-thrombin complex should differ grossly from previous inhibitor complexes. Therefore many scientists were quite eager to know the structure of the complex and the underlying mechanism of inhibition. In collaboration with Alex Tulinsky and Markus Grütter/Ciba-Geigy we were able to determine the structures of two hirudin variants in complex with human thrombin, using our thrombin model for phasing [39Rydel T.J. Ravichandran K.G. Tulinsky A. Bode W. Huber R. Roitsch C. Fenton II, J.W. The structure of a complex of recombinant hirudin and human alpha-thrombin.Science. 1990; 249: 277-80Crossref PubMed Scopus (690) Google Scholar, 40Grutter M.G. Priestle J.P. Rahuel J. Grossenbacher H. Bode W. Hofsteenge J. Stone S.R. Crystal structure of the thrombin-hirudin complex: a novel mode of serine protease inhibition.EMBO J. 1990; 9: 2361-5Crossref PubMed Scopus (317) Google Scholar]. Both structures revealed that the amino-terminus of hirudin is aligned close to the thrombin active-site in reverse direction compared with PPACK, that the second residue (Thr) spans across the entrance to the specificity pocket without utilizing it, and that the central compact domain slots into the apolar ‘left-side’-region of thrombin (Fig. 4A). Particularly fascinating was to see the extended acidic tail of hirudin interacting with thrombin's strongly positively charged exosite-I (Fig. 5). Surprisingly, most exosite interactions are not direct salt bridges, but are made via hydrophobic and ‘through-space’ electrostatic interactions. Fascinated by the hirudin results, we became interested in the design strategies used by other hematophagous animals to prevent clot formation during blood sucking from their prey [41Ascenzi P. Amiconi G. Bode W. Bolognesi M. Coletta M. Menegatti E. Proteinase inhibitors from the European medicinal leech Hirudo medicinalis: structural, functional and biomedical aspects.Mol Aspects Med. 1995; 16: 215-313Crossref PubMed Scopus (34) Google Scholar]. At that time I found a paper describing the properties of a femtomolar binding, thrombin-specific inhibitor called rhodniin, which had been cloned from the assassin bug Rhodnius prolixus, the transmitter of Chagas disease in South America. We expressed, purified and co-crystallized this inhibitor with bovine thrombin, and solved its structure [42van de Locht A. Lamba D. Bauer M. Huber R. Friedrich T. Kröger B. Höffken W. Bode W. Two heads are better than one: crystal structure of the insect derived double domain Kazal inhibitor rhodniin in complex with thrombin.EMBO J. 1995; 14: 5149-57Crossref PubMed Scopus (170) Google Scholar]. This structure (Fig. 4B) revealed a quite different approach to prevent blood clotting: a double-headed inhibitor consisting of two covalently connected Kazal-type domains, with the amino- and carboxy-terminal domains binding canonically into the active-site cleft and interacting electrostatically with exosite-I of thrombin, respectively. Rhodniin particularly nicely exemplifies the advantage of bivalent inhibitors: Each separate domain binds with only submicromolar affinity, while the covalent tandem domains act cooperatively, giving rise to an extremely tight binding and selective thrombin inhibitor. A similar thrombin complex with the specific thrombin inhibitor ornithodorin, derived from the soft tick Ornithodorus moubata, revealed another double-headed inhibitor docking to similar thrombin sites [43van de Locht A. Stubbs M.T. Bode W. Friedrich T. Bollschweiler C. Höffken W. Huber R. The ornithodorin-thrombin crystal structure, a key to the TAP enigma?.EMBO J. 1996; 15: 6011-7Crossref PubMed Scopus (179) Google Scholar]. Ornithodorin, in contrast to rhodniin, consists of two Kunitz domains, with the amino-terminal domain not (like BPTI) binding in a substrate-like manner, but (similar to hirudin) through the amino-terminus (Fig. 4C). As proposed, this amino-terminal domain also exemplified the previously enigmatic binding mode of the tick anticoagulant peptide, TAP, from the same organism to factor-Xa [44Wei A. Alexander R.S. Duke J. Ross H. Rosenfeld S.A. Chang C.H. Unexpected binding mode of tick anticoagulant peptide complexed to bovine factor Xa.J Mol Biol. 1998; 283: 147-54Crossref PubMed Scopus (87) Google Scholar]. The structure of a thrombin complex with the triatomine bug-derived inhibitor triabin (Fig. 4D) revealed a completely different inhibition mechanism [45Fuentes-Prior P. Noeske-Jungblut C. Donner P. Schleuning W.D. Huber R. Bode W. Structure of the thrombin complex with triabin, a lipocalin-like exosite-binding inhibitor derived from a triatomine bug.Proc Natl Acad Sci U S A. 1997; 94: 11845-50Crossref PubMed Scopus (125) Google Scholar]: a single lipocalin-like domain docked to the distal part of exosite-I, leaving the active site and most of the cleft fully accessible to small substrates, in excellent agreement with the fully maintained hydrolytic activity of the thrombin component toward tripeptide chromogenic substrates (Fig. 4D). The structures of exosite-directed inhibitors gave a rough idea about the enormous extension of the fibrinogen interface into exosite-I of thrombin, what has been shown only very recently in more detail in a thrombin complex with the central fibrinogen fragment E [46Pechik I. Madrazo J. Mosesson M.W. Hernandez I. Gilliland G.L. Medved L. Crystal structure of the complex between thrombin and the central ‘‘E“ region of fibrin.Proc Natl Acad Sci U S A. 2004; 101: 2718-23Crossref PubMed Scopus (100) Google Scholar]. In the early 1990s, we became interested in the interaction of fibrinogen with the thrombin apolar binding site and specificity pocket. We thus solved two different crystal structures of thrombin complexes formed with fibrinopeptide A (FPA): one of human thrombin in a complex with a covalently bound FPA-mimicking decapeptide and hirugen (Fig. 5 [47Stubbs M.T. Oschkinat H. Mayr I. Huber R. Angliker H. Stone S.R. Bode W. The interaction of thrombin with fibrinogen. A structural basis for its specificity.Eur J Biochem. 1992; 206: 187-95Crossref PubMed Scopus (210) Google Scholar]), and the other of bovine thrombin in a non-covalent complex with this decapeptide [48Martin P.D. Robertson W. Turk D. Huber R. Bode W. Edwards B.F. The structure of residues 7–16 of the A alpha-chain of human fibrinogen bound to bovine thrombin at 2.3-Å resolution.J Biol Chem. 1992; 267: 7911-20Abstract Full Text PDF PubMed Google Scholar]. Both structures showed exactly the same picture: the FPA fragment of the intact Aα-chain of fibrinogen does not bind thrombin in an extended but compact conformation, with the benzyl Phe8 side chain, placed nine residues before the Arg16FibAα-Gly17 scissile peptide bond, slotting into the hydrophobic S4-cleft, and with the Gly12FibAα-Gly13-Gly14 segment approaching the active site of the cognate thrombin molecule (Fig. 5). As shown in numerous mutagenesis studies [49Malkowski M.G. Martin P.D. Lord S.T. Edwards B.F. Crystal structure of fibrinogen-Aalpha peptide 1–23 (F8Y) bound to bovine thrombin explains why the mutation of Phe-8 to tyrosine st