The refined 1.9‐Å X‐ray crystal structure of d‐Phe‐Pro‐Arg chloromethylketone‐inhibited human α‐thrombin: Structure analysis, overall structure, electrostatic properties, detailed active‐site geometry, and structure‐function relationships

Abstract

Thrombin is a multifunctional serine proteinase that plays a key role in coagulation while exhibiting several other key cellular bioregulatory functions. The X‐ray crystal structure of human α‐thrombin was determined in its complex with the specific thrombin inhibitor d‐Phe‐Pro‐Arg chloromethylketone (PPACK) using Patterson search methods and a search model derived from trypsinlike proteinases of known spatial structure (Bode, W., Mayr, I., Baumann, U., Huber, R., Stone, S.R., & Hofsteenge, J., 1989, EMBO J. 8, 3467–3475). The crystallographic refinement of the PPACK‐thrombin model has now been completed at an R value of 0.156 (8 to 1.92 Å); in particular, the amino‐ and the carboxy‐termini of the thrombin A‐chain are now defined and all side‐chain atoms localized; only proline 37 was found to be in a cis‐peptidyl conformation.The thrombin B‐chain exhibits the characteristic polypeptide fold of trypsinlike serine proteinases; 195 residues occupy topologically equivalent positions with residues in bovine trypsin and 190 with those in bovine chymo‐trypsin with a root‐mean‐square (r.m.s.) deviation of 0.8 Å for their α‐carbon atoms. Most of the inserted residues constitute novel surface loops. A chymotrypsinogen numbering is suggested for thrombin based on the topological equivalences. The thrombin A‐chain is arranged in a boomeranglike shape against the B‐chain globule opposite to the active site; it resembles somewhat the propeptide of chymotrypsin(ogen) and is similarly not involved in substrate and inhibitor binding.Thrombin possesses an exceptionally large proportion of charged residues. The negatively and positively charged residues are not distributed uniformly over the whole molecule, but are clustered to form a sandwichlike electrostatic potential; in particular, two extended patches of mainly positively charged residues occur close to the car‐boxy‐terminal B‐chain helix (forming the presumed heparin‐binding site) and on the surface of loop segment 70–80 (the fibrin[ogen] secondary binding exosite), respectively; the negatively charged residues are more clustered in the ringlike region between both poles, particularly around the active site. Several of the charged residues are involved in salt bridges; most are on the surface, but 10 charged protein groups form completely buried salt bridges and clusters. These electrostatic interactions play a particularly important role in the intrachain stabilization of the A‐chain, in the coherence between the A‐ and the B‐chain, and in the surface structure of the fibrin(ogen) secondary binding exosite (loop segment 67–80).The most remarkable feature at the thrombin surface is the prominent canyonlike active‐site cleft mainly shaped by two characteristic insertion loops around Trp 60D and Trp 148. The deep and narrow active‐site cleft in general explains the narrow specificity of thrombin for distinct macromolecular substrates and inhibitors. Comparisons with other crystal structures of human and bovine thrombin recently determined using this PPACK‐thrombin model indicate that the first loop around Trp 60D is relatively rigid, whereas the opposite loop around Trp 148 can attain different conformations depending on complexation state and crystalline environment.The active‐site residues and the entrance to the specificity pocket are partially occluded in thrombin (much more than in the other serine proteinases) by this distinctive Trp 60D loop. The specificity pocket of thrombin resembles that of bovine trypsin but is designed to prefer arginine over lysine residues at PI. D‐Phe II and Pro 21 of the bound PPACK inhibitor fit neatly to a novel hydrophobic cleft (the aryl‐binding site) and to the cavitylike hydrophobic S2 subsite; the d‐configuration of Phe II is beneficial for binding as it allows the PPACK amino‐terminus to form hydrogen bonds to Gly 216 in addition. Some small arginine and benzamidine‐derived synthetic inhibitors owe their particularly high thrombin specificity and affinity to their exceptional steric fit to these novel hydrophobic cavities close to the thrombin active site (Bode, W., Turk, D., & Stürzebecher, J., 1990, Eur. J. Biochem. 193, 175–182).The active‐site cleft levels off in the primed direction and continues over the molecular surface of the thrombin loop Lys 70‐Glu 80 (itself structurally similar to the calcium loop in trypsin, with the distal nitrogen of Lys 70 replacing the calcium of trypsin). This site of a strong positive electrostatic surface potential probably represents the secondary site for interaction with the α‐chain of fibrinogen and fibrin (the fibrin[ogen] secondary binding exosite) and accommodates the carboxy‐terminal acidic tail part of hirudin (Rydel, T.J., Ravichandran, K.G., Tulinsky, A., Bode, W., Huber, R., Roitsch, C., & Fenton, J.W., II, 1990, Science 249, 277–280). Segment Arg 187‐Gly 188‐Asp 189, which could represent a thrombin adhesion site for cellular interactions with platelets, fibroblasts, and endothelial cells, is mainly buried in α‐thrombin; its adhesive role would thus appear to require some prior unfolding.Most of the well‐characterized sites of proteolytic cleavage leading to the degradation products β‐, γ‐, and ϵ‐thrombin of diminished or lost clotting activity are situated in exposed mobile loops of α‐thrombin. None of these segments is in a canonical conformation that would allow association with the substrate‐binding site of a cleaving serine proteinase without large conformational changes. The cleavage of the Arg 77A‐Asn 78 scissile peptide bond (leading to β‐thrombin) presumably results in the unfolding of the 70–80 loop, exposing the salt bridge‐connected residues buried in α‐thrombin to the solvent; the concomitant disruption of the surface of the fibrin(ogen) secondary binding exosite would explain the loss of binding capacity (and thus catalytic activity) toward fibrinogen. The Arg 67‐Ile 68 peptide bond is completely buried beneath this 70–80 loop surface and thus only susceptible to an attacking proteinase after prior cleavage and exposure of this loop, as observed.